Neuroimaging Clinics of North America

Register or Login:

Password:

SIGN IN

Auto-Login [Reminder]

Search This Clinic

for

GO

Advanced Search - MEDLINE - My Recent Searches - My Saved Searches - Search Tips

CLINIC HOME CURRENT ISSUE PREVIOUS ISSUES SEARCH THIS CLINIC FORTHCOMING ISSUES CLINIC INFORMATION

Issue
Alert me when new journal issues are available. Add TOC Alert May 2002 (Vol. 12, Issue 2) View Selected Abstracts Preface Display: without preview

Consulting Editor Author Information Abstracting/Indexing Media Information Contact Information Permissions Buy Back Issues
RELATED SITES

More periodicals:
FIND A PERIODICAL FIND A PORTAL GO TO PRODUCT CATALOG

Imaging of head trauma by Go JL, Zee CS page xi Full Text | PDF (40 KB) Clinical evaluation of patients with head trauma by Ko DY pages 165-174 Full Text | PDF (100 KB) Pathology of head trauma by Hardman JM, Manoukian A pages 175-187 Full Text | PDF (263 KB) Imaging of traumatic intracranial hemorrhage by Young RJ, Destian S pages 189-204 Full Text | PDF (1922 KB) Diffuse axonal injuries: pathophysiology and imaging by Hammoud DA, Wasserman BA pages 205-216 Full Text | PDF (614 KB) Imaging of acquired cerebral herniations by Johnson PL, Eckard DA, Chason DP, Brecheisen MA, Batnitzky S pages 217-228 Full Text | PDF (1007 KB) Radiographic assessment of cranial gunshot wounds by Kim PE, Go JL, Zee CS pages 229-248 Full Text | PDF (860 KB) Traumatic vascular injuries and their management by Larsen DW pages 249-269 Full Text | PDF (1026 KB) Imaging of pediatric head trauma by Poussaint TY, Moeller KK pages 271-294 Full Text | PDF (1043 KB) Imaging of facial trauma by Sun JK, LeMay DR pages 295-309 Full Text | PDF (1076 KB)

file:///C|/temporal/Neuroimaging%20Clinics%20Mayo%202002.html (1 de 2)07/03/2004 9:57:47

Neuroimaging Clinics of North America

Orbital trauma by Go JL, Vu VN, Lee KJ, Becker TS pages 311-324 Full Text | PDF (722 KB) Imaging of sequelae of head trauma by Zee CS, Hovanessian A, Go JL, Kim PE pages 325-338 Full Text | PDF (820 KB) Surgical management of head trauma by Gruen P pages 339-343 Full Text | PDF (55 KB) Index pages 345-349 PDF (44 KB) View Selected Abstracts

Display: without preview

2004 Elsevier, Inc. | Privacy Policy | Terms & Conditions | Feedback | About Us | Help | Contact Us

file:///C|/temporal/Neuroimaging%20Clinics%20Mayo%202002.html (2 de 2)07/03/2004 9:57:47

Neuroimag Clin N Am 12 (2002) xi

Preface

Imaging of head trauma

John L. Go, MD Chi-Shing Zee, MD Guest Editors

The role the neuroradiologist plays in the diagnosis and assessment of patients who have sustained head trauma is a vital one. Prior to the advent of CT, plain film radiography and angiography were the mainstay in the radiologic assessment in these groups of patients. The sticks and stones of a previous era have now been replaced with state-of-the-art imaging with we are now accustomed to. CT and MR imaging provide our clinicians with the tools they need in making an immediate diagnosis, deciding on management, and determining the future prognosis of these patients. In this issue of Neuroimaging Clinics of North America, we have attempted to provide the reader with some of the tools they can apply in providing a thoughtful assessment and perspective in assessing patients with head trauma. By laying a foundation of both the pathology and early neurologic assessment of head trauma in the early articles, we can then build on some of the more germane topics of what is seen during an everyday practice in the emergency room. Though CT remains the main modality in assessing the patient with head trauma, MR imaging is beginning to play a larger role in the early diagnosis and determination of prognosis for these patients. Topics of interest were also included that we felt would be of

interest to you as a reader. Finally, a neurosurgic perspective on head trauma is needed so that, as you sit and interpret the neuroradiologic images, the important information that neurosurgeons need can be incorporated into your interpretation in the surgical management of this subset of patients. We would like to thank Dr. Burton Drayer for asking us to begin this worthwhile endeavor. In addition, we would like to thank Mr. Barton Dudlick at W.B. Saunders Company for his patient perseverance in the planning and implementation of this project. The valuable contributions made by all of our authors goes without saying. The administrative assistance of Yvonne Klausmeier has been invaluable. Finally, we would like to thank our families and friends for their patience and support. John L. Go, MD Chi-Shing Zee, MD Division of Neuroradiology Department of Radiology Keck School of Medicine University of Southern California 1200 North State Street Los Angeles, CA 90033, USA

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 1 4 - X

Neuroimag Clin N Am 12 (2002) 165 174

Clinical evaluation of patients with head trauma

David Y. Ko, MD*
Keck School of Medicine, University of Southern California, Los Angeles, CA 90089, USA Los Angeles County and University of Southern California Medical Center, 1200 North State Street, General Hospital Room 5641, Los Angeles, CA 90033, USA

Head trauma is unfortunately a very common condition; it affects every age group, with a wide range of outcomes from no sequelae to death. There are many causes of head trauma. Young children fall, and even suffer nonaccidental trauma such as shaken baby syndrome, which has a high incidence [1]. Older children and young adults have outdoor-activity related injuries from bicycling, rollerblading, snowboarding, or skiing, as well as sport injuries from soccer or football. Other common injuries occur from falls, assault, and gunshot wounds. Even the medical community contributes to head injury, with chiropractic neck manipulation resulting in carotid or vertebral dissection. The most common cause of serious head trauma is from motor vehicle accidents. Head injury has been increasing at an alarming rate and is a public health problem. The human brain is our most important organ; it is of incredible complexity in design and function, but vulnerable to injury. Head trauma can have immediate transient symptoms, or lead to chronic signs and symptoms. The medical management of acute head trauma can be quite complicated and expensive. Serious head trauma can result in chronic brain injury. This can be very debilitating to the individual and family, as well as being a huge financial burden on society. The apparent seriousness of head injury can be misleading. Minor trauma can result in serious problems, or patients who sustain serious head trauma may survive, as in the case of Phineas

Gage, who in 1848 had an iron rod go through his head and suffered only personality changes [2]. As more medical and surgical therapies for head trauma are available, the role of neuroimaging has become more important.

Medical evaluation The medical evaluation of head trauma needs to be done quickly but thoroughly. History is important, because a patient may have a seizure, stroke, or subarachnoid hemorrhage and then fall and sustain secondary head traumathe primary event still needs to be medically addressed. A history of prior head trauma is particularly important, because head trauma effects can be cumulative. The details of the head trauma are crucial, with particular emphasis on the location and force of impact. One important distinction is penetrating injury versus closed head injury, with the former having a breach in the cranium and thus a higher incidence of infection. Symptoms of nausea and vomiting associated with head trauma can be signs of increased intracranial pressure (ICP) and are worrisome. A complete neurologic examination is mandatory in head trauma; simply stating that a patient has a nonfocal examination is inadequatea dead person has a nonfocal examination. In assessing the severity of head injury, the presence or absence of loss of consciousness and the duration of the loss of consciousness are of paramount importance. The patients recollection of loss of consciousness may be hazy, so anothers account may be more accurate. Head trauma without loss of consciousness can be as serious as that with loss of consciousness, and this is being recognized more

* Los Angeles County and University of Southern California Medical Center, 1200 N. State Street, General Hospital Room 5641, Los Angeles, CA 90033. E-mail address: dko@hsc.usc.edu (D.Y. Ko).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 1 0 - 2

166

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

in sports concussion injuries. The presence of a seizure with acute head trauma indicates a more serious head insult, although the prognosis for developing epilepsy is still small, and a seizure occurring more remotely is more worrisome for the development of epilepsy. The medical evaluation of a head trauma patient begins with vital signs. The classic vital sign change of increased ICPthe Kocher-Cushing reflex of hypertension and bradycardiamay be seen. If patients have fever, it may be due to subarachnoid blood producing hyperpyrexia, or hypothalamic dysfunction, which can also cause low temperature. Patients with increased ICP may have many different types of cardiac abnormalities on EKG, although a primary cardiac event always needs to be evaluated for [3]. Patients with increased ICP can also have altered breathing patterns, which has localizing value [4]. If the respiratory status is labored and airway protection is needed, intubate electively. A general examination is important, looking specifically for signs of head trauma such as Battles sign and raccoon sign, or depressed skull fracture. Fluid from the nose (rhinorrhea) or ear (otorrhea) should be checked for possible evidence of cerebrospinal fluid (CSF) leak. Normal nasal or ear discharge is without glucose, but CSF contains glucose that can be quickly checked by dipstick. The neck examination is important in head trauma, and the neck needs to be cleared for cervical spine injury or instability. After the neck has been cleared, it needs to be assessed for stiffness or meningismus of meningitis or subarachnoid hemorrhage. Carotid auscultation is important to pick up a bruit or lack of pulse from carotid dissection or occlusion, which then needs to be evaluated with appropriate diagnostic testing.

Neurologic examination The neurologic examination in head trauma needs to cover certain essential features. When head trauma patients are first brought to the emergency room, a rudimentary neurologic status is obtained by getting a Glasgow coma scale (GCS) score. This is a commonly used standardized scale, and it is useful because many different health care professionals can score it as well as physicians. The GCS covers three spheres: best verbal response, best motor response, and best eye opening response; and ranges from 3 to 15. The advantage of the GCS is that it is simple and allows for serial examinations, and it is written in as part of emergency room and intensive care unit nursing sheets in many institutions. The

GCS is by itself not sufficient, however; a complete neurologic evaluation is needed. Every neurologic evaluation requires a mental status examination. Mental status can be assessed as hyperalert (occurs with stimulant use), alert, clouding of consciousness, delirious (agitated, with irritability and misperceptions of stimuli, and possibly with delusions and hallucinations), obtunded, stuporous, and comatose [4]. These are the most commonly used terms, although actual descriptions of the individual response may be more informative. Orientation to person, place, and date are fairly standard neurologic examination items. Testing of both short-term and long-term memory should also be performed. Evidence of amnesia may suggest a more serious head injury. Memory of past events may be impaired (retrograde), or there may be problems with making new memory (antegrade). Assessment of language for expressive and receptive dysfunction is important in head injury. Aphasia needs to be differentiated from slurred speech or dysarthria. A patient that is confused may have aphasia, which is an important localizing sign. Tests of judgment and executive function are informative. Personality changes may be difficult to assess acutely, but are seen chronically. Higher cognitive dysfunction may be best evaluated by formal neuropsychologic testing. A rapidly fluctuating mental status in an acute setting is worrisome and needs a thorough evaluation, including the appropriate neuroimaging and electroencephalogram. The cranial nerve (CN) examination is important in head trauma, with particular emphasis on the eye examination [5]. Although head trauma can lead to anosmia, it is not routinely tested for. Pupillary examination is very important to detect anisocoria, although a small percentage of the population has unequally sized pupils naturally. A fixed dilated pupil, which may signify uncal herniation, is worrisome, although I have been called to the ER for a fixed dilated pupil only to discover the patient had a prosthetic glass eye. Another situation that has caused misinterpretation is when mydriatics are used to dilate the eyes and the pupils will be less than fully reactive, although the medication effect is temporary. A procedural note can help by documenting that medication was administered. Studies of comatose patients revealed that the absence of pupillary reflexes was a poor prognostic sign [6]. A Horners syndrome that has the clinical triad of pupillary constriction, ptosis, and anhydrosis may be seen with carotid dissection. Visual field needs to be tested to pick up temporal and occipital lesions, or even cortical blindness. There is the rare situation with Antons syndrome of patients being cortically blind but denying their

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

167

visual deficits. The fundoscopic examination is important to detect signs of increased intracranial pressure, such as the presence of papilledema and the loss of venous pulsation (the presence of venous pulsation is a reassuring sign of not having significantly increased ICP). Retinal hemorrhage, sometimes called subhyloid hemorrhage, can be seen with subarachnoid hemorrhage. Extraocular movement not only assesses CN III, IV, and VI but other diffuse areas such as brainstem structures which may include the parapontine reticular formation, medical longitudinal fasiculus, and cortical areas frontal eye fields parietal visual association areas which are necessary for conjugate eye movements. A CN VI palsy may be an early sign of increased ICP. Corneal reflex assesses the sensory component of CN V and the motor component of CN VII. The auditory and vestibular system as part of cranial nerve VIII can be affected and readily tested. In comatose patients the oculo-vestibular system is checked with the classical oculo-cephalic test (Dolls eye test) and oculo-vestibular reflex test (ice water calorics) after the neck has been cleared. Other cranial nerve abnormalities can be seen with head trauma, although if patients are intubated the lower cranial nerves may be difficult to assess. Gag reflex needs to be assessed carefully in patients with altered mental status, to prevent aspiration. If a patient is thought to be brain dead, all the cranial nerves that can be tested need to be documented. The motor examination is important in head trauma for assessing functional status as well as having localizing value. Motor examination begins with observation of spontaneous movements such as hemiballismus, often associated with a subthalamic lesion; or myoclonus, which has multiple etiologies, including cortical or even spinal cord generators. Muscle mass asymmetry signifies a prior dysfunction. Muscle tone is important, because with acute central lesion there may be flaccid paralysis, while marked hypertonia suggests a prior central nervous system dysfunction. Motor examination includes voluntary testing of muscle strength, which is highly effort dependent, so someone in pain may not be able to exert full effort. Useful screens for weakness include the pronator drift and forearm roll test [7]. For the lower extremities the gait examination can be informative, especially the heel-and-toe walk. In stuporous or comatose patients the motor response to noxious stimuli needs to be performed. A patients localizing pain is more favorable than posturing. The two classic motor signs described include decorticate posturing (arms flexed and legs extended) and decerebrate posturing (arms extended

and internally rotated and legs extended), with the latter being a more ominous sign. Patients may have decerebrate posturing on one side and decorticate sign on the other side. The noxious stimuli can be administered by pinching the skin of the chest, but in patients with trauma and possible cervical spinal cord damage, one should test noxious stimuli above the neck, on the face. Minor motor signs such as facial twitching or finger twitching sometimes may be the only overt manifestation of a seizure, and this needs to be explored further with electroencephalogram. A promising technique for assessing the motor pathways is the use of transcranial motor stimulation to get motor-evoked potentials, which is more tolerable than electrical stimulation of motor-evoked potentials. The sensory examination is important in localizing the lesion, particularly in localizing spinal cord lesions that may occur with head trauma, but it is highly dependent on patient cooperation. The sensory level would direct imaging at the appropriate region. Because high-dose steroids have been shown to improve outcome in traumatic spinal cord lesions, these lesions need to be detected as quickly as possible. Another spinal cord condition that can be associated with trauma is the development of syringomyelia. The advent of magnetic resonance imaging (MRI) has vastly improved the diagnostic workup of spinal cord lesions. Computed tomography (CT) scans, while superior for bone, are not particularly informative for the spinal cord proper, and CT myelogram is more invasive. Somatosensory evoked potentials is an extension of the sensory examination that is an objective measure of a relatively subjective clinical examination, although its sensitivity is less than its specificity. Reflex examination yields very important localizing information, with emphasis on asymmetric deep tendon reflexes (DTR) or pathologic reflexes such as the Babinski response. The jaw jerk reflex is helpful in a patient with diffuse hyper-reflexia, for it can be helpful in localizing the hyper-reflexia between spinal cord localization (normal jaw jerk) and cerebral localization (brisk jaw jerk). One of the easiest signs to elicit is the Babinski sign and its many variations. Other pathologic signs include clonus and Hoffmans sign. Other reflexes that are helpful are the frontal release signs, such as palmomental, rooting, sucking, and glabellar/Myerson sign. Cerebellar examination is important to highlight cerebellar dysfunction, such as axial ataxia for more midline localization, or appendicular ataxia for more lateral cerebellar dysfunction. A head tilt may signify brainstem or cerebellar dysfunction. The cerebellar signs are important to pick up because cerebellar

168

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

lesions such as hemorrhage need to be assessed quickly. The posterior fossa is a small space and herniation can be rapid; an upward herniation needs to be treated immediately with surgical evacuation. Gait examination is important because it an important functional status that can be impaired. Anterior cerebral artery aneurysms can predominately affect gait. Effects of hydrocephalus produce gait difficulties in the form of gait apraxia, along with cognitive dysfunction (dementia) and urinary incontinence. Hydrocephalus can be treated with shunting, and with normal pressure hydrocephalus (NPH) the sooner the shunting, the better the outcome. A Romberg test will also be informative. A thorough neurologic examination is important, and examining skills will always be in demand and cannot be replaced by neuroimaging. It is important to get serial follow-up neurologic examinations, such as neruochecks for inpatients, to detect changes that may prompt further workup and therapeutic implementation. Neuroimaging is an important and informative extension of the neurologic examination and its role is ever increasing. The sensitivity of MRI for picking up silent ischemic events belies the classical neurologists adage, there are no noneloquent areas of the brain, just noneloquent examiners.

Laboratory and diagnostic testing Head trauma patients need to have a basic laboratory examination, including complete blood chemistry, electrolytes, serum osmolality, and arterial blood gas, among other elements. Toxic screening needs to be conducted and should include alcohol level, and levels of medication that are taken by the patient, if known, which may have played a role in the head trauma. Creatinine phosphokinase MB fraction may be elevated with cerebral injury and can be measured, although questions about its sensitivity remain. Another marker for neuronal injury is neuron specific enolase [8]. These surrogate markers of head injury may be important in future therapeutic trials. The electroencephalogram (EEG) was used as the screening examination for cerebral dysfunction before modern neuroimaging began with the CT scan. The EEG used to be the examination for stroke or brain tumor because it may show focal slowing. Compared with CT, the EEG was more effective in detecting a cortical stroke early, although with MRI and its many sequences the acute stroke can be picked up quickly. EEG examination should be performed in patients who may be having subclinical seizures, which can be the cause of altered mental

status. An offshoot of EEG that became popular for a time among non-neurologists was brain mapping, which is computerized power spectral analysis of an EEG. Brain mapping is, however, technically difficult, so the American Academy of Neurologys Therapeutics and Technology Committee does not recommend its use with head injury [9]. The EEG gives functional information about the central nervous system and its use in head trauma may be helpful in the evaluation of chronic symptoms. Somatosensory evoked potentials (SSEP) of the upper and lower extremities have been helpful in assessing sensory pathways [10]. The absence of cortical potentials in the SSEP in nontraumatic coma indicates a poor prognosis [11]. The use of intraoperative monitoring has become the standard of practice in a number of neurosurgical procedures, to minimize surgical risk in preserving neurologic function. The relatively new technique of transcranial magnetic stimulation may be a useful test to assess motor function neurophysiologically in an objective manner. PET and SPECT scans may show perfusion defects that may correlate with cognitive dysfunction. Rapid advances in functional MRI may also assess cerebral blood flow, thus making MRI useful for both structural and functional assessment of the brain. Functional MRI is still in its infancy, but its clinical utility is very promising, because it is safer because it involves no radiation. Another promising advance in neruroimaging is the emerging field of magnetic resonance spectroscopy, which will be useful in assessing cerebral neurochemical status noninvasively.

Pathophysiology of head trauma The pathophysiology of head trauma can be complex. A simple classification is closed head injury, sometimes called blunt head trauma or penetrating head injury. Head trauma can produce coupe and contra-coupe injury, and pathologic or neuroradiologic studies may reveal diffuse axonal injury far removed from the area of impact. Certain forms of trauma can cause focal cerebral contusion, diffuse cerebral edema, diffuse axonal injury, subdural hematoma, epidural hematoma, subarachnoid hemorrhage, and intraparenchymal hemorrhage [12]. Depending on degree of increased ICP and location of pathology, there can be different forms of herniation, including subfalcine, uncal, rostrocaudal, and even upward herniation from a posterior fossa lesion. A neuroimaging study is critical in determining the change in intracranial pressure in a patient who

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

169

may be herniating. Subfalcine herniation usually is not symptomatic, but uncal herniation is a critical sign that needs to be acted on quickly. Cerebellar herniation can also occur and my cause upward (caudal-rostral) herniation. Cerebral herniation needs to be identified quickly so medical and surgical measures to decrease ICP can be instituted promptly. In large, severe strokes with significant edema, there has been some limited success with craniotomy as a last-resort measure to accommodate the massively increased ICP. When a head trauma patients clinical status changes, a follow-up neuroimaging study is warranted and highlights the value of serial studies. An important sequela of head trauma may be the development of obstructive hydrocephalus, and this needs to be picked up early, because these patients can benefit from shunting to prevent herniation. Cerebral edema is a common finding in head trauma and may change over time. Delayed intracerebral hemorrhage may also occur. Another consequence of head trauma may be the development of syndrome of inappropriate antidiuretic hormone (SIADH) secretion with hyponatremia, which may effect brain swelling. The overly rapid correction of the hyponatremia may lead to central pontine myelinosis. Hypopituitarism has also been seen with head trauma. It is not uncommon for a minor head injury to prompt evaluation with neuroimaging, with the discovery of probably unrelated structural lesions such as vascular lesions (cavernous angiomas and arteriovenous malformations), cortical migration defects, prior ischemic events, and even brain tumors. An important factor in evaluating an ischemic lesion on CT and MRI is classifying the lesion as acute or chronic. The newer sequences on MRI of perfusion and diffusion have been very helpful in determining the age of the cerebral insult with more precision than in the past. A stroke in a head injury patient always raises the possibility of a carotid vessel dissection and requires the appropriate neurovascular imaging studies. It is standard practice to get neuroimaging before performing a lumbar puncture to decrease the risk of increased ICP causing herniation. Only in the emergency situation of an infection such as bacterial meningitis can one even consider bypassing the neuroimaging study. In patients with clinically suspected subarachnoid hemorrhage, the absence of blood does not preclude a lumbar puncture being performed. It is crucial to measure the opening pressure during a lumbar puncture. Head trauma can cause a disruption of the bloodbrain barrier. This is often picked up as CSF leaks from

the nose or ears, but leaks may not be present and neuroimaging may be necessary to determine if disruption is present. Skull fracture is important to identify because it can increase the risk of CNS infection either as meningitis, encephalitis, abscess, or subdural empyema, which can occur soon after injury or be delayed. Head trauma can lead to the development of hydrocephalus, which needs to be differentiated from post-traumatic ventriculomegaly resulting from atrophy (hydrocephalus ex-vacuo). Head trauma has been associated with development of traumatic aneurysms, which are rare, and even brain tumors, although this association is controversial. Neuroimaging has allowed clinicians to get a sense of the location and type of pathology noninvasively. The neuroimaging studies lead to higher yields from brain biopsies, using neuroimaging guided stereotactic methods that are minimally invasive. MRI neuroimaging is now becoming available in real time in the operating room. Neuroimaging studies can result in quick medical or surgical treatment, which in the past had to wait until neuropathology conferences were used to definitively diagnose, and which often came too late to be of benefit to the patient.

Classification of head trauma The classification of head trauma has been evolving. The Glasgow coma scale (GCS) is the most common general scale, but the Hunt and Hess scale is useful with subarachnoid hemorrhage, and the NIH stroke scale is being used with stroke [13,14]. The GCS has three components: eye opening, verbal response, and motor response, which are graded as follows: Eye opening Never: 1 To pain: 2 To verbal stimuli: 3 Spontaneously: 4 Verbal response No response: 1 Incomprehensible sounds: 2 Inappropriate words: 3 Disoriented and converses: 4 Oriented and converses: 5 Motor response No response: 1 Extensor posturing: 2 Flexor posturing: 3 Flexion withdrawal: 4

170

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

Localizes pain: 5 Obeys: 6 The GCS score has been subdivided as follows: someone having a score of 13 to 15 is considered as having minor injuries; a score of 9 to 13 indicates moderate injury; 5 to 8 indicates severe injury, and 3 to 4 indicates very severe injury [15]. There is also a variation of the GCSthe Glasgow outcome scale. The use of standardized scales helps in evaluating different studies and trials, which has been difficult in the past. The seriousness of a head injury is often a difficult clinical question, and labeling head injury as mild or a concussion as minor can be misleading [16,17]. Concussion is defined as a trauma-induced alteration in mental status, often with confusion and amnesia, that may or may not involve loss of consciousness [18]. The concussion history of athletes has shown that the accumulation of minor concussion can lead to serious and chronic signs and symptoms, which has led to the early retirement of athletes such as Steve Young, Troy Aikman, and Pat LaFontaine. Repetitive head injury is not limited to athletes and occurs widely in the form of motor vehicle accidents, sports injury, assault, and other accidents in various combinations. Alcoholics are prone to repetitive head injuries. The American Academy of Neurologys Quality Standards Subcommittee has issued a practice parameter summary statement of the management of concussion in sports, which is increasing at an alarming rate [19]. It is difficult to decide which patients with mild to moderate head injury should be admitted to the hospital, which is not an issue with serious head trauma. Patients should be admitted if they have had loss of consciousness for longer than 5 or 10 minutes, focal neurological deficit, seizure, skull fracture, or if they are difficult to evaluate fully, as with young children or patients with alcohol or drug abuse [20]. Patients with severe brain injury may be slow to wake up or may show no improvement over time. Prognosis of patients who are comatose is always difficult, and neuroimaging can give some guidance. One condition that should not be overlooked is a patient who is locked-in due to a pontine lesion resulting in almost total paralysisthese patients are fully awake but cant interact with the environment except with eye blinks. Prognosis depends on the clinical examination, but can be aided by the use of clinical neurophysiologic studies such as electroencephalogram (EEG) and SSEP. Serial EEG can either show improvement in brain activity or show it become slower with lower voltage, which indicates poor prognosis.

Some patients with severe head injury will be brain dead, and neruoimaging plays a role in certain cases. Brain death is a clinical diagnosis in adults [21]. Brain death in adults does not require a confirmatory test, but the American Academy of Neurology subcommittee on Brain Death recommended confirmatory testing in difficult cases. The recommendations are different in children, especially in neonates and infants, where confirmatory tests are required. The most sensitive confirmatory test is a cerebral angiogram, in which a patient with massive ICP will not have the intravenous dye enter the brain. The most commonly used confirmatory test is the EEG, but another neuroimaging test, the nuclear brain flow scan, can also be used and may be easier to interpret. Other confirmatory tests include transcranial Doppler and evoked potentials. Brain death needs to be determined quickly, so that a death can be declared and the family and loved ones can move on and not drag out the process of prolonged suffering and grieving, and also so that the patient does not continue to use precious intensive care resources. Brain death also comes into play in organ donation. With the list of patients waiting for major organs growing, the fast use of organs from a brain-dead head trauma patient is critically important. As head trauma patients can have a quick course, the organs tend to be more viable and suitable than in patients who have had chronic illness.

Therapy Therapy for head trauma has been a combination of medical and surgical management [22,23]. Optimal management starts with maintaining vital signs by providing adequate oxygenation and assuring adequate blood pressure to maintain cerebral perfusion pressure. The standard therapy for increased ICP is to keep the head elevated about 30 degrees, keep the patient fluid status even or slightly negative, and use normal saline or lactated ringers (avoid free water or diluted saline) to keep the fluids intravascular instead of intracellular, which can worsen cerebral edema [24]. Long term monitoring of ICP is also being used more frequently to manage these complex patients [25]. Some patients with increased ICP may need ventriculostomy to drain CSF. Cases of traumatic subarachnoid hemorrhage can be complicated by vasospasm and further ischemia. These patients are optimally managed by serial transcranial Doppler studies to assess for velocity change of vasospasm, and the use of volume loading and calcium channel blockers. If

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

171

signs of herniation appear, patients should be hyperventilated to a pCO2 of about 25 mm Hg, avoiding too vigorous hyperventilation, which can cause oligemia and may extend the area of damage. Hypothermia has potential to minimize neuronal damage and is being studied for neural protection [26]. Phenobarbital can be used in massive ICP, as it is reported to decrease cerebral metabolic activity, which may help allow areas that have decreased cerebral perfusion to survive on lowered essential substrates such as oxygen and glucose. Be aware, though, that phenobarbital can cause hypotension, which would be harmful in patients with increased ICP who need adequate cerebral perfusion pressure. Steroids, whose optimal role is reducing vasogenic edema from brain tumor, are used for cytotoxic edema from severe head trauma. There are some reports of benefit, but steroid use in this situation is still controversial [27 31]. Steroid use will increase the risk of infection and gastrointestinal bleeding, and elevate blood sugars. Keeping the blood sugar regulated, as opposed to uncontrolled hyperglycemia, has been associated with better outcome due to osmotic lactic acid levels and pH changes. If a patient has severe head trauma, surgery such as lobectomy may be helpful [32]. The last resort measure is to use acute diuretics such as mannitol to decrease ICP, but their effects are short lived and overly vigorous diuresis can cause excessive increase in serum osmolality, impairing neuronal function. A long-standing controversy in head trauma is prophylactic use of seizure medications to decrease the incidence of post-traumatic seizures [33]. The trials with pheyntoin, phenobarbital, and valproic acid have shown no anti-epileptogenic effect, and seizure medication prophylaxis is not encouraged with the agents currently available [34]. The duration of treatment with an anticonvulsant after early onset post-traumatic seizure is also uncertain, although a year may be sufficient [35]. Even in patients with subarachnoid hemorrhage, where there is a high incidence of seizures (18%), the benefits of prophylactic antiepiletpic drugs are still unclear [36]. The focus of the development of the next generation of seizure medicationsto have antiepileptogenic medications that prevent the development of seizuresmay change this practice. Some of the newer anticonvulsants, such as topiramate, have shown neuroprotective properties in animal studies, but clinical neuroprotective effects in humans remain to be demonstrated [37]. While the course of acute head trauma can present a rapidly changing clinical picture, the long term sequelae can be very slow to change or not change

significantly. Head trauma may require many therapies, such as physical therapy, occupational therapy, speech therapy, cognitive therapy, and neuropsychologic and psychiatric support [38].

Chronic symptoms Symptoms after head trauma include physical, cognitive, and psychologic signs and symptoms [39]. Head trauma can produce physical changes, including various cranial nerve abnormalities. Head trauma can lead to loss of smellin fact head trauma is one of the most common causes of anosmia. There can be disturbances of oculomotor cranial nerves that should be evaluated by CT scan of the orbits. Head trauma can produce vestibular and auditory dysfunction [40]. The vestibular symptoms can be dizziness, lightheadedness, and vertigo, which can be difficult to treat. Vestibular dysfunction may be best assessed by electronystagnogram. Auditory changes can include tinnitus, hyperacusis, and hearing loss. Hearing tests can be useful for hearing difficulties: brainstem auditory evoked potentials may be helpful to distinguish between a peripheral and a central dysfunction of the auditory pathways. There are many physical conditions that result from head trauma, with headache being very common. The headache after head trauma can be a tension type headache as well as a migraine, and there have even been reports of cluster headache [41]. Post-traumatic headache may not only be episodic, but can cause chronic daily headache. Occipital neuralgia may also cause headache. Normal pressure hydrocephalus may also present as headache and may need to be evaluated with a cisternogram. Head trauma headache is also rarely associated with increased intracranial pressure such as benign intracranial hypertension A post-traumatic seizure is not uncommon, but only a small number of patients will develop posttraumatic epilepsy, a primary tendency to have recurrent seizures. The seizure that occurs acutely with the trauma is generally less likely to lead to epilepsy than the seizure that occurs remotely from the trauma. The seizures of head trauma are partial seizures and can be intractable. Sleep disturbances are not uncommon after head trauma, sometimes secondary to pain and often compounded by psychologic components such as anxiety and even post-traumatic stress disorder. Movement disorder such as Parkinsonism can occur from head trauma, and trauma can produce other movement disorders such as tremor and dystonia. A chronic pain condition may arise from head trauma.

172

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

This has had many different names in the past. It was called causalgia by Weir Mitchell during the Civil War period, later called reflex sympathetic dystrophy, and is now called complex regional pain syndrome. This pain syndrome can be from minor trauma but lead to debilitating pain and disability [42]. Head traumas cognitive and psychologic changes can be the most debilitating. Usually the severity of head trauma correlates with the cognitive and psychologic changes, but this is a complex relationship. Patients may develop post-concussion syndrome, post-traumatic stress disorder, depression, impaired attention and concentration, memory disturbance, and personality changes. Another role of neuroimaging far removed from the acute injury is to detect the presence of atrophy and degree of encephalomalcia after injury. There is some epidemiologic evidence that head trauma can lead to dementia (eg, dementia pugilistica, often associated with cavum septum pellucidi), and there is even a risk for Alzheimers dementia [43]. The neuropsychologic changes of head trauma can include both cognitive and behavioral changes that fall under the rubric of post-concussion syndrome or organic brain syndrome [44]. There are many possible signs and symptoms, including decreased concentration and attention, memory disturbance, reduced information processing speed, altered language capabilities, irritability, anxiety, asthenia, low energy and increased or more quickly reached fatigue [45]. These symptoms need to be addressed, and one important condition that may bring these symptoms out or worsen them may be depression. Careful psychiatric and neuropsychologic assessment needs to be undertaken and appropriate therapy given to maximize recovery. Due to the many signs and symptoms of head trauma, some patients are treated with numerous medications, which can lead to polypharmacy. It is not uncommon for the head trauma patients to be on antidepressants, anticonvulsants, pain medication, anti-anxiety medications, medications for spasticity, muscle relaxants, and sleep aids among many others, and often more than one medication from each class. This polypharmacy can lead to many drug interactions and contribute many side-effects, compounding the patients signs and symptoms.

treatment may be altered, by legal proceedings [46]. The high incidence of shaken baby syndrome requires high suspicion, and this diagnosis should not be missed in helpless infants. Neuroimaging studies often play a critical role in the courtroom, since they provide objective measurement in obvious cases, but in other cases there may be no obvious changes, although patients may have clinical signs and symptoms. Some claims associated with head trauma, such as the development of multiple sclerosis, can not be substantiated, but controversy persists, particularly in medical-legal cases. The long-term symptoms of head trauma are often subjective complaints and are neither easily tested nor quantifiable. Functional MRI may be potentially useful for cerebral functional assessment, although this needs further study. A common injury with varying outcomes is the acceleration-deceleration injury, whiplash, often associated with motor vehicle accidents, which can result in neurologic symptoms and is also often associated with legal action. The clinical interpretation of neuroimaging studies always requires clinical correlation. Clinicians who order the examinations by giving detailed information of the clinical circumstances help the neuroradiologist answer specific questions and aid in providing the best care to the patient. Because of the medical-legal climate, there may be more unnecessary neuroimaging studies ordered as defensive medicine, although good clinical judgement still should be the rule of thumb.

Prevention Prevention of head trauma still is the most effective way to decrease its devastating effects. Head trauma has been decreased and the sequelae lessened with the wider use of protective headgear and better safety equipment, especially in motor vehicles [47]. It often takes legislation requiring protective headgear for the bicyclists and motorcyclist to get good compliance. With recent high-profile skiing accident deaths of a Kennedy and Congressman Bono, it may not be unreasonable to require helmets to ski. The number of diving accidents has decreased with rules and ordinances prohibiting diving in pools except in certain safe areas. The use of firearms is one of the more common causes of head trauma, and gun regulation is always a heated topic that fluctuates with the political climate. Accidental discharge of firearms hurts children particularly, and better safety

Medical-legal Patients with head trauma are often involved in medical-legal issues. The signs and symptoms of head trauma may be influenced, and the response to

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174

173

measures are needed. Firearm use in illegal activities is unfortunately too common and is growing. Motor vehicle accidents are by far the largest cause of serious head trauma. Better car design to minimize head trauma is helping reduce head trauma; one major safety measure is the inflatable safety bag. Federal and state legislation of lowered speed limits, which were employed to save energy, also decreased serious head trauma. Lower legal blood alcohol limits and tighter screening of drivers have been somewhat effective. For professional truckers, limiting driving time can reduce fatigue and sleep deprivation-related accidents. As technology is rapidly advancing, it needs to be applied to motor vehicle safety. Natures protection of the brain with the solid cranium was evolutionarily helpful, but as we expand the potential means of inducing head injury, the cranium cannot expand to accommodate any significant swelling or mass effect.

radiologists; as well as allied health care workers such as physical, occupational, and speech therapists, clinical psychologists, neuropsychologists, and many others. Head trauma needs to be approached by a mutlidisciplinary team because it is complex. Specialized trauma centers incorporate all of these specialists and the best medical technology for optimal management of head trauma. The following chapters cover the use of different neuroimaging techniques, including CT scan and MRI, that greatly aid clinicians in evaluation and management of head trauma patients. These advances have truly revolutionized medicine and it has happened rapidlypneumoencephalography was the neuroimaging study of choice less than half a century ago. The future of neuroimaging in head trauma will undoubtedly include advances we can not yet foresee but that will allow clinicians to continue to improve patient care.

Future treatment Intensive research and development of a number of pharmacologic therapies to minimize neuronal injury from head trauma and improve clinical outcome have produced promising animal studies; however, human trials have been unimpressive. Due to the complex cascade involved in neuronal injury and cell death, neuroprotection is difficult. Nevertheless, the quest for the Holy Grail, an effective and safe neuroprotective agent for head trauma treatment, continues, and neuroimaging will play an important role in these trials. Besides minimizing neuronal damage in head trauma, another therapeutic goal would be treatments to restore function in damaged areas. Restorative therapy is another promising research avenue that is in its infancy, and may involve pharmacologic therapies such as nerve growth factors and transplantation therapy, among others. A pilot trial with transplantation for stroke has already taken place. The combination of neuroprotection and restorative therapy may significantly improve outcome in devastating conditions such as head trauma.

References
[1] Geddes JF, Hackshaw AK, Vowles GH, et al. Neuropathology of inflicted head injury in children. Brain 2001;124(part 7):1290 306. [2] Damasio H, et al. The return of Phineas Gage: clues about the brain from the skull of a famous patient. Science 1994;264(5162):1102 5. [3] Valerino J, Elson J. Electrocaridographic changes in central nervous system disease. Neurol Clin N Am 1993;11(2):257 72. [4] Plum F, Posner J. The diagnosis of stupor and coma. 3rd edition. Philadelphia: F.A. Davis Company; 1982. p. 1 368. [5] Keane J, Baloh R. Posttraumatic cranial neuropathies. Neurology of trauma. Neurol Clin 1992;10(4):849 64. [6] Levy DE, Knill-Jones RP, Plum F. The vegetative state and its prognosis following non-traumatic coma. Ann N Y Acad Sci 1978;315:293 306. [7] Sawyer RN, Hanna JP, Ruff RL, Leigh JL. Asymmetry of foream rolling as a sign of unilateral cerebral dysfunction. Neurology 1993;43(8):1596 8. [8] Marangos PJ. Clinical studies with neuron specific enolase. Prog Clin Biol Res 1985;175:285 94. [9] Nuwer M. Assessment of digital EEG, quantitative EEG, and EEG brain mapping: report of the American Academy of Neurology and the American Clinical Neurophysiology Society. Neurology 1997;49(1):277 92. [10] Hume AL, Cant BR. Central somatosensory conduction after head injury. Ann Neurol 1981;10:411 9. [11] Madl C, Kramer L, Yeagnehfar W, et al. Detection of nontraumatic comatose patients with no benefit of intensive care treatment by recording of sensory evoked potentials. Arch Neurol 1996;53:512 6. [12] Adams JH, Graham JI, Murray LS, et al. Diffuse axonal injury due to non missile head injury in

Summary Head trauma is a very common and sometimes life-threatening medical condition that involves sports medicine physicians, emergency room physicians, neurologists, neurosurgeons, orthopedists, anesthesiologists, rehabilitation physicians, psychiatrists, and

174

D.Y. Ko / Neuroimag Clin N Am 12 (2002) 165174 humans: analysis of 45 cases. Ann Neurol 1982;12: 557 63. Hunt WE, Hess RM. Surgical risks as related to time of intervention in the repair of intracranial aneursyms. J Neurosurg 1968;28:14 20. Teasdale G, Jennett B. Assessment and prognosis of coma after head injury. Acta Neurochir 1976;34:45. Rimel RW, Girodani B, Barth JT, et al. Disability caused by minor head injury. Neurosurgery 1981;9:221 9. Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM. Indications for computed tomography in patients with minor head injury. N Engl J Med 2000;343(2):100 5. Mandel S. Minor head injury may not be minor. Postgrad Med 1989;85(6):213 25. Kelly JP, Rosenberg JH. Diagnosis and management of concussion in sports. Neurology 1997;48:575 80. Report of the Quality Standards Subcommittee. Practice parameter: the management of concussion in sports. Neurology 1997;48:581 5. Eisenberg HM, Weiner RL, Tabaddor K. Emergency care: initial evaluation. In: Cooper P, editor. Head injury. 2nd edition. Baltimore: Williams and Wilkins; 1987. p. 20 33. Report of the Quality Standards Subcommittee of the American Academy of Neurology. Practice parameters for determining brain death in adults. Neurology 1995; 45:1012 4. Maas AI, Dearden M, Teasdale GM, Braakman R, Cohadon F, Iannotti F, et al. EBIC-guidelines for management of severe head injury in adults. Acta Neurochir 1997;139(4):286 94. White RJ, Likavec MJ, The diagnosis and initial management of head injury. N Eng J Med 1992;327(21): 1507 11. Simma B, Burger R, Falk M, Sacher P, Fanconi S. A prospective, randomized, and controlled study of fluid management in children with severe head injury: lactated Ringers solution versus hypertonic saline. Crit Care Med 1998;26(7):1265 70. Saul TG, Ducker TB. Effect of intracranial pressure monitoring and aggressive treatment on mortality in severe head injury. J Neurosurg 1982;56(4):498 503. Shiozaki T, Sugimoto H, Taneda M, Yoshida H, Iwai A, Yoshioka T, et al. Effect of mild hypothermia on uncontrollable intracranial hypertension after severe head injury. J Neurosurg 1993;79(3):363 8. Dearden NM, Gibson JS, McDowall DG, et al. Effect of high-dose dexamethasone on outcome from severe head injury. J Neurosurg 1986;64(1):81 8. Eisenberg HM, Frankowski RF, Contant CF, et al. High-dose barbiturate control of elevated intracranial pressure in patients with severe head injury. J Neurosurg 1988;69(1):15 23. Giannotta SL, Weiss MH, Apuzzo ML, Martin E. High dose glucocorticoids in the management of severe head injury. Neurosurgery 1984;15(4):497 501. [30] Newell DW, Temkin NR, Bullock R, Choi S. Corticosteroids in acute brain injury [letter]. BMJ 1998; 316:396. [31] Saul TG, Ducker TB, Salcman M, Carro E. Steroids in severe head injury: a prospective randomized clinical trial. J Neurosurg 1981;54(5):596 600. [32] Litofsky NS, Chin LS, Tang G, Baker S, Giannotta SL, Apuzzo ML. The use of lobectomy in the management of severe closed-head trauma. Neurosurgery 1994; 34(4):628 33. [33] Young B, Rapp RP, Norton JA, et al. Failure of prophylactically administered phenytoin to prevent late postraumatic seizures. J Neurosurg 1983;58:236 41. [34] Schierhout G, Roberts I. Prophylactic antiepileptic agents after head injury: a systemic review. J Neurol Neurosurg Psychiatry 1998;64:108 12. [35] Epstein FM, Ward JD, Becker. Medical complications of head injury. In: Cooper P, editor. Head injury. 2nd edition. Baltimore: Williams and Wilkins;1987. p. 390 421. [36] Rhoney DH, Tipps LB, Murry KR, Basham MC, Michael DB, Coplin WM. Anticonvulsant prophylaxsis and timing of seizures after aneurysmal subarachnoid hemorrhage. Neurology 2000;55(2):258 65. [37] Koh S, Jensen FE. Topiramate blocks perinatal hypoxiainduced seizures in rat pups. Ann Neurol 2001;50(3): 366 72. [38] Boake C, Bobetic KM, Bontke CF. Rehabilitation of the patient with mild traumatic brain injury. Neurorehabil 1991;1:70 8. [39] Denny-Brown D. Disability arising from closed head injury. JAMA 1945;127:429 36. [40] Tuohimaa P. Vestibular disturbances after acute mild head injury. Acta Otolaryngol 1978;359(suppl):1 59. [41] Weiss HD, Stern BJ, Goldberg J. Post-traumatic migraine: chronic migraine precipitated by minor head or neck trauma. Headache 1991;31:451 6. [42] Schwartzman R. Reflex sympathetic dystrophy and causalgia, neurology of trauma. Neurol Clin 1992; 10(4):953 73. [43] Roberts GW, Allsop D, Bruton C. The occult aftermath of boxing. J Neurol Neurosurg Psych 1990;53:373 8. [44] Binder LM. Persisting symptoms after mild head injury: a review of the postconcussive syndrome. J Clin Exp Neuropysch 1986;8:323 46. [45] Barth JT, Macciocchi SN, Giordan B, et al. Neuropsychologic sequelae of minor head injury. Neurosurgery 1983;13:529 33. [46] Binder LM, Rohling ML. Money matters: a metaanalytic review of the effects of financial incentives on recovery after closed-head injury. Am J Psychiatry 1996;153:7 10. [47] Thomas S, Acton C, Nixon J, Battistutta D, Pitt WR, Clark R. Effectiveness of bicycle helmets in preventing head injury in children: case-control study. BMJ 1994; 308(6922):173 6.

[13]

[14] [15] [16]

[17] [18] [19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

Neuroimag Clin N Am 12 (2002) 175 187

Pathology of head trauma

John M. Hardman, MD*, Anthony Manoukian, MD
Department of Pathology, John A. Burns School of Medicine, 1960 East-West Road, Honolulu, HI 96822, USA

Traumatic brain injury (TBI) causes one third of all injury deaths in the United States [1]. Motor vehicle crashes, falls, assaults, guns, sports, and recreational activities are the major causes of TBI [2]. TBI is highest in young adults aged 15 to 24 years and higher in men than women in all age groups. Secondary peak incidences of TBI occur in infants and children and the elderly. Conditions that increase risk for accidents include alcoholism, prior head injury, prior meningitis, seizure disorders, mental retardation, and psychiatric disorders [3,4]. Collision within or ejection from a vehicle produces TBI [5]. Crash investigations have identified and led to correction of design flaws that made vehicles hazardous. The use of shoulder and lap belts, air bags, and rollover bars have all improved crash survivability. Properly fitting crash helmets significantly reduce the incidence of severe TBI and death of motorcyclists and bicyclists [6 9]. Helmet use is protective in any activity (eg, contact sports, bike riding, skate boarding) [7]. However, gunshot wounds to the head are steadily increasing and since 1990 have caused more deaths each year than motor vehicle accidents [1]. The incidence, severity, etiology, and specific types of injuries have been assessed in clinicopathologic studies of head injuries [10 15]. Gennarelli and his coworkers have replicated the fundamental pathophysiologic and neuropathologic features of human head injuries in primates [16 19]. Patients with TBI have abnormalities that are attributable to primary and secondary brain damage. Primary injuries occur at the moment of injury and secondary brain damage evolves and complicates the primary injuries. Primary injuries include fractures, contusions and lacerations,

diffuse axonal injury, and intracranial hemorrhage. Secondary injuries consist of complications such as brain swelling, hypoxia/ischemia, increased intracranial pressure, and infection. The pathologic features of both the primary and secondary lesions attributed to TBI should be understood by anyone caring for head-injured patients. The computed tomography (CT) and magnetic resonance (MR) images mirror the pathologic abnormalities found in head trauma [20 22]. Radiologists must accurately interpret the CT and MR images of injured patients. Forensic pathologists have long appreciated the characteristic focal lesions, such as coup and contracoup contusions, that occur in falls or vehicle accidents, but the understanding of diffuse injuries has been more elusive. In the past 30 years, the characterization and significance of diffuse brain injury has been clarified. Adams, Gennarelli, Graham and coinvestigators have clearly demonstrated that diffuse lesions occur commonly in fatal head injuries after vehicle accidents [3,11,12,17 19,23 33]. Such injuries more fully explain the pathophysiologic abnormalities that result from TBI. Understanding the nature of the focal and diffuse injuries is critical to understanding the morbidity and mortality of brain injury. A review of the essential primary and secondary injuries attributable to TBI follows.

Mechanisms of traumatic brain injury Traumatic brain injuries are produced by the transmission of physical energy to the head by contact and acceleration/deceleration. When a blunt object contacts the head, shock waves emanate from the point of impact and travel rapidly through the bone and brain, creating additional tissue strains. When the head is set in motion, or when the moving

* Corresponding author. E-mail address: hardman@hawaii.edu (J.M. Hardman).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 9 - 6

176

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187 Table 2 Data from a consecutive series of 635 fatal nonmissile head injuries over a 25-year period (1968 1982) Condition Sex distribution Male Female Types of injury Road traffic accident Falls Assaults Other Incidence of Fracture of skull Surface contusions Diffuse axonal injury Intracranial hematoma Elevated intracranial pressure Ischemic brain damage Brain swelling Intracranial infection % 78 22 53 35 5 7 75 94a 29 60b 75 55 51c 4

head is stopped without striking an object, or when the moving head is stopped by impact, the resulting TBI is caused by the forces transmitted by translation, rotation, and angulation [33]. Strain causes tissue injury by either contact or inertia. Tissues will withstand injury better if they are deformed slowly. Contact forces cause local deformations of the skull and produce local or remote compressive, shear, and tensile strains in the underlying skull and brain. The summation of these forces produces the injury: skull fracture, contusion, and some of the intra- or extraaxial hematomas. Head acceleration or deceleration of appropriate magnitude, rate of onset, and direction are the determinants of the type and severity of head and brain injuries. Impact to the skull is not essential to cause injury by shear, tensile, and compressive strains within the head. Concussive injuries, diffuse axonal injury, and most acute subdural hematomas and contrecoup contusions may be produced solely by acceleration/deceleration of the head. Even so, a combination of contact and acceleration/deceleration forces is the likely cause of most TBI.

Classification of brain injury A classification of the principal types of injuries in nonmissile and missile injuries of the CNS is given in Table 1. A summary of the key features of a large series of nonmissile TBI autopsy cases is presented in Table 2.

Abbreviations: EDH, extradural (epidural) hematoma; SDH, subdural hematoma. From Graham DI. Neuropathology head injury. In: Narayan RK, Wilberger JE Jr, Povlishock JT, editors. In: Neurotrauma. New York: McGraw-Hill Companies Inc.; 1996. p. 44; with permission. a Mild in 6%, moderate in 78% severe in 10%. b EDH in 10%, SDH in 18%, intracerebral in 16%, burst lobe in 23%. c Unilateral in 34%, bilateral in 17%.

Table 1 Classification of primary head injuries Skull fractures Linear Depressed Basilar Focal injuries Contusions Coup Contrecoup Intermediate Fracture Herniation Lacerations Hematomas Extradural Subdural Subarachnoid Intracerebral Diffuse injuries Concussion Diffuse axonal injury Mild Moderate Severe Gliding contusions Brain swelling

CT and MR images accurately capture focal abnormalities as seen in life, but the images of diffuse brain injury are not so easily distinguished or identified. The pathologic features of the focal and diffuse injuries of the brain are presented in the following sections. Skull fractures Skull fractures are found in only 25% of fatal head injuries at autopsy [34]. They can be best visualized with nonenhanced CT images [34]. Fractures of the base of the skull are not reliably depicted by routine radiographic examination [35,36]. Hence, the routine removal of the dura mater from the skull is necessary to identify basilar fractures at autopsy. The pattern of a skull fracture may show the direction, location, and force of the impact producing the injury. Fracture lines usually radiate from the impact point. They often extend into the base of the skull and connect with cranial nerve foramina. Interlacing fracture lines suggest that multiple blows produced the fractures. The appearance and the

Modified from Hardman JM. Cerebrospinal trauma. In: Davis RL, Robertson DM, editors. Textbook of neuropathology. Baltimore: Williams & Wilkins; 1997. p. 1180; with permission.

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

177

location of the impact site may identity the cause of the injury and determine the pattern of fractures. Fractures may be closed or open (compound). Skull fractures are classified by the patterns they display [35 39]: A. A linear fracture shows a straight crack or break and is produced by a blow to the skull. Linear fractures frequently result from falls and commonly occur in children. B. A depressed skull fracture is identified by having the bone depressed by a blow. Depressed skull fractures commonly result from low-velocity impacts to a limited area of the calvarium. Such fractures commonly form in the frontal and parietal regions. C. If the fracture shows bending outward distal to the impact site, it is termed a bursting fracture. D. Diastatic fractures show separation of the cranial sutures. Diastatic fractures usually form between the petrous portion of each temporal bone, the greater wing of each sphenoid bone, and the petrosquamosal fissures. E. Hinge fractures may develop when the head is run over and crushed by a heavy object, such as a truck wheel [40]. These fractures commonly cross the dorsum sellae to allow the base of the skull to move like a hinge. F. Ring fractures circumscribing the foramen magnum may result from impacts to the base of the spine, such as in falls impacting on the buttocks. G. Remote fractures occur at a distance from the point of cranial impact. Remote fractures of the orbital roofs and ethmoid plates occur in crushing injuries, falls, and gunshot wounds [41,42]. Periorbital ecchymoses (black eyes) herald the presence of these fractures. H. Compound basal skull fractures may produce cerebrospinal fluid fistulae [43]. When the fractures traverse the paranasal sinuses, rhinorrhea results. Fractures of the petrous portion of the temporal bone may cause otorrhea. Rarely, pneumocephalus may also be demonstrable by plain skull films or CT scans [20,44]. The cerebrospinal fistulae may be complicated by CNS infection. I. A growing fracture results from an enlarging traumatically induced leptomeningeal cyst [45 48]. Growing fractures nearly always occur in children less than three years old and are not associated with any significant injury of the brain. The cyst is lined by fibrous connective tissue and is filled by cerebrospinal

fluid. It protrudes through a tear in the dura mater, enlarges a linear fracture, and erodes the bone along the fracture margin. The enlargement and erosion of bone are believed to be caused by pulsation of the brain.

Contusions A cortical contusion is a bruise of the brains surface. The overlying dura mater nearly always remains intact. A cerebral contusion is a distinctive lesion of mechanical injury. Lindenberg and Freytag identified different types of contusions and considered each one to be caused by different biomechanical forces [49,50]. Cortical contusions involve the crowns of gyri, are wedge-shaped, and cause full-thickness necrosis and hemorrhage of the cortex and leptomeninges (Fig. 1). Bleeding into the Virchow-Robin spaces produces linear hemorrhages oriented perpendicular to the pial surface of the gyri. Blood extends to the cortical surface and spreads in the cerebrospinal fluid. Types of contusions: A. Coup contusions form at the site of cranial impact B. Contrecoup contusions develop opposite the cranial impact. These contusions have a predilection for the inferior frontal and anterior temporal lobes. C. Herniation contusions develop at the margins of brain hernias (eg, falx cerebri, tentorium, or foramen magnum). D. Fracture contusions form along the edges of skull fractures. E. Gliding contusions, as described by Lindenberg, are hemorrhagic lesions found in the parasagittal white matter and are a type of diffuse brain damage that is associated with diffuse axonal injury and deep hemispheric traumatic hematomas [50,51]. The types of contusions formed depend on the position and motion of the body at the time of impact. Experimental studies using the monkey as a model show the underlying pathophysiologic events in the formation of cortical contusions [17,19]. The injury patterns differ depending on whether the victims head is stationary or in motion at the moment of impact. When the free mobile head is motionless at the moment of impact, contusions appear most prominently beneath the point of cranial impact. When a free mobile head is accelerating, as in a fall, and

178

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

Fig. 1. Contracoup contusions of each gyrus rectus and inferolateral frontal cortex. Note the linear petechial hemorrhages (arrow). Injuries are the result of a left frontal impact in an auto accident. Coronal section of frontal lobes at level of genu of corpus callosum.

impacts against a firm or unyielding surface, contusions form most prominently on cortical surfaces opposite the cranial impact site [52]. The inferior frontal lobes and the temporal poles are particularly vulnerable to such injuries. The contusions are usually more prominent on one side and rarely affect the superior and posterior aspects of the cerebral hemispheres. Patients with unilateral intracranial hematomas or unilateral brain swelling have significantly more severe ipsilateral cerebral contusions [53]. Lateral cranial impacts cause prominent contusions of the contralateral frontotemporal regions (Fig. 2). Contusions of the inferior temporal and frontal regions likely result from movement of the brain over the irregular bony prominences found in the anterior and middle cranial fossae. A contusion index may be used to determine the depth and extent of contusions in various parts of the brain [54]. Contusions are more severe in patients with a skull fracture than those without. They are less severe in patients with diffuse axonal injury than those without. They are more severe in patients who do not have a lucid interval after injury than those who do. Frontal and temporal regions are the more susceptible to injury at low accelerations than are occipital impacts. The greater the acceleration, the greater the extent of injury to the brain from occipital impacts.

Laceration A laceration is a mechanical tear or rent in normal tissues. Such tears appear in the same sites as contusions, but they result from greater force. Lacerations are most frequently found along fracture lines and are invariably found around penetrating or perforating wounds. Lacerations may form with or without associated skull fractures [55]. They may produce microscopic hemorrhages or large hematomas. Cerebral lacerations preferentially involve the inferior frontal lobes and temporal tips. With diffuse axonal injury, lacerations commonly occur in the corpus callosum and the rostral brainstem. Tears of the pontomedullary junction and cerebral peduncles are presumably caused by hyperextension [56]. Such injuries are reported to occur in up to 3.6% of fatal road crashes [57]. They may occur without other apparent gross brain injury and are often associated with atlantooccipital dislocations or high cervical fracture dislocations. They may also occur with other brainstem lacerations or other brain injuries. In such cases transverse middle fossa fractures are usually found. This lesion is found in diffuse axonal injury (DAI). Studies by others have further characterized these tears [10,13,15,57 64]. The dislocations or fractures allow excessive movement of the brain with hyperextension and tears of the pontomedullary junction or cerebral

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

179

Fig. 2. Gliding contusions of the frontal lobes (arrows) and focal hemorrhages in left side of corpus callosum (open arrow) following severe impact to crown of skull incurrred in fall from a four story building. Note there is diffuse subarachnoid hemorrhage and contracoup contusions of both inferolateral temporal lobes (arrowheads).

peduncles. This injury is nearly always fatal. Rare survivors may exhibit the locked in syndrome [58]. Such injuries may also become visible with CT and MR scans [60,65].

Diffuse brain injury Diffuse brain injury indicates that widespread damage to gray and white matter has occurred. The principal types of diffuse brain injury are diffuse axonal injury, brain swelling, ischemic/hypoxic damage, and increased intracranial pressure. DAI is a distinct clinicopathologic entity that may develop after mild concussion or after TBI, causing persistent posttraumatic coma [24,62,66 68]. Theoretical, experimental, and morphologic studies show that diffuse axonal injury likely results from sudden angular rotation of the head [15,44,67,69 72]. Strich believed that shear forces caused the damage [73]. Using angular acceleration, Gennarelli et al produced injury patterns in monkeys that are indistinguishable from DAI seen in humans [17,19]. Such injuries result from shear or tensile strains of axons in the cerebral white matter, corpus callosum, and brainstem. The severity of axonal injury is proportional to rapidity of angular acceleration of the head. Recent experimental studies also show that minor trauma will cause significant damage of axons [53,68]. These studies suggest that concussion may not be the reversible

physiologic disorder we have always thought. For example, we know that boxers suffer permanent brain damage after repeated concussive insults [70,71,74]. The finding of occasional clusters of microglia in patients dying soon after a minor head injury indicates that axonal injury had occurred [69]. More recently, using an antibody against b-amyloid precursor protein, axonal damage has been demonstrated in patients who died from unrelated causes after sustaining mild head injury [75 78]. Patients dying with DAI have a statistically significant lower incidence of a lucid interval after TBI, skull fractures, cerebral contusions, intracerebral hematomas, and increased intracranial pressure compared with patients without this type of brain injury [24]. In 1956, Strich recognized DAI as an essential component of post-traumatic dementia [73]. DAI develops most frequently in victims of vehicle accidents, though it may appear after falls [79,80], after assaults [31,81], and with child abuse [44]. DAI may occur with contusions, increased intracranial pressure, brain swelling, intracranial hematomas, and hypoxic/ischemic brain damage. The diagnosis of DAI depends upon finding axonal bulbs microscopically. Early injury of axons is best demonstrated immunocytochemically [82 84]. The most sensitive indicator of injured axons is finding b-amyloid precursor protein (bAPP) in the injured axons [63,85]. Damaged axons may be seen within two to three hours of injury or as long as 99 days after

180

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

injury [75]. Patients that die within one to two hours will not have sufficient survival time for axonal bulbs to develop, even though DAI likely occurred [85]. In experimental models, damage of the axolemma leads to focal axonal swelling that is followed by axotomy four to six hours later. Physiologic data indicate that the axons are damaged but not usually sheared at the time of injury [33,86,87]. In addition, hypoxia, brain swelling, and increased intracranial pressure are not initiators of axonal injury [33]. DAI produces distinctive gross and microscopic abnormalities. The grossly visible lesions include multiple small (0.5 1.5 cm) oval or elliptical lesions that appear swollen and hemorrhagic. The lesions involve the lobar white matter, internal capsule, corpus callosum, and the dorsolateral aspect of the brain stem. The basal ganglia, thalamus, and cerebellum can also be affected. These grossly visible lesions are best shown by CT and MR imaging [34,88]. How-ever, such studies do not permit quantification of neural injury after TBI. Magnetization transfer ration and proton MR spectroscopy can quantify the damage after TBI [21]. Delayed imaging may show abnormalities not apparent initially. CT scans show tiny petechial hemorrhages at the gray-white matter junction and within the corpus callosum. Hemorrhagic lesions are hyperintense on T1-weighted MR scans. The more common nonhemorrhagic lesions are difficult to identify on T1-weighted scans but appear as multiple hyperintense lesions on proton-density weighted, T2weighted, and fluid-attenuated inversion-recovery images. Hypointense lesions may persist for years on T2-weighted images because of hemosiderin deposits left as remnants of hemor-rhage. In mild head trauma, hemorrhagic foci tend to occur in the lobar white matter of the frontal and temporal lobes, usually at the gray-white matter junction. These focal hemorrhagic lesions are gliding contusions as described by Lindenberg, but are now believed to result from the same forces that cause DAI [89,90]. Small hemorrhages may be found in the frontoparietal white matter [50]. They also may develop in the walls of the third ventricle, the hippocampus, within each striatum, and in the thalamus. With survival, these lesions become soft and granular and form brown cystic scars containing hemosiderin. With more severe injury, the posterior half of the corpus callosum (see Fig. 2) and the dorsolateral aspect of the midbrain and rostral pons appear hemorrhagic. These hemorrhagic lesions are usually at sites of axonal damage that can only be seen microscopically. Early injury of the corpus callosum appears hemorrhagic. Lesions of the corpus callosum usually lie lateral to the midline, affect the inferior and posterior

part of the corpus callosum, and may extend several centimeters. When they extend to the midline, the interventricular septum may be disrupted and cause hemorrhage into the lateral ventricles. The callosal injury may be focal, segmental, or extend from the genu to the splenium. Such lesions are often hemorrhagic and may disrupt or tear the corpus callosum. A grossly apparent hemorrhagic lesion may involve one or both dorsolateral quadrants of the brainstem. The histologic appearance of DAI depends upon the survival time of the victim after TBI. Diffuse injury of axons can only be seen microscopically. With survival of a few days, many axonal bulbs form in the white matter of the cerebral hemispheres, corpus callosum, cerebellum, and rostral brainstem. Irregular swellings of axons appear as oval or rounded bulbs at the end of axons. They are eosinophilic in hematoxylin and eosin-stained sections or argyrophilic in silverstained preparations. With survival of a few weeks, small clusters of microglia replace the axonal bulbs. Finally, with prolonged survival and severe disability, symmetrical enlargement of the ventricles and Wallerian degeneration of long tracts of the cerebral hemispheres, the brainstem, and the spinal cord occur. Small brownish-pigmented lesions of the corpus callosum and the rostral brainstem, and perhaps some small healed superficial cortical contusions will be all that remain. Occasionally patients die with DAI that have no apparent macroscopic abnormalities. These cases of microscopic DAI were found in 1.8% of 434 cases of fatal non-missile head injury [25]. Adams et al and Blumbergs et al have proposed grading systems to measure the severity of injury for DAI [26,66]. Using these systems, many patients with the lesser degrees of DAI may have complete or partial lucid intervals after TBI. Blumbergs et al found 50% of the less severely damaged cases had a partial lucid intervals before death [91]. Adams et al found a lower frequency [26].

Hemorrhage Hemorrhages develop in the epidural, subdural, and subarachnoidal spaces of the brain in victims of TBI. The resultant mass lesions often produce devastating neurophysiologic effects and even death. Studies of about 11,000 head injury victims admitted to hospitals in Brisbane, Australia, documented 1.5% extradural, 5% subdural, and 0.6% intracerebral hematomas among these patients [92 97]. The comprehensive review by Bullock and Teasdale summarizes the experimental and clinical features of intracranial hematomas [92].

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

181

Epidural hematoma Epidural hematomas form when bleeding occurs between the calvarium and the dura mater (endosteum). Blows to the head that fracture the temporal bone with laceration of the middle meningeal artery typically cause such hemorrhages [97 106]. Hematomas of variable size may be localized or diffusely layered between the calvarium and dura. The hematomas frequently form in the temporoparietal region, less commonly in the frontal and occipital regions, and rarely in the posterior fossa. Nearly onethird of patients have other significant brain injuries, such as contusions and subdural hematomas [97,100]. The blood in an epidural hematoma seldom liquefies. The inflammatory and reparative responses are comparable to those observed in subdural hematomas [100,107]. A thin layer of granulation tissue develops on the surface of the dura adjacent to the blood clot. A well-formed neomembrane of vascular fibrous tissue will usually encapsulate the loculated blood a month or more after injury. Ossification develops infrequently in the interface between the granulation tissue and the dura. False aneurysms of the middle meningeal artery may rarely occur. Epidural hematomas usually have an acute, rapidly progressing clinical course. CT and MR aid the physician in identifying epidural hematomas and associated conditions. With these evaluations, the surgical mortality rates have improved significantly [106,108]. Untreated hematomas usually enlarge and may rapidly progress to cause coma and death of the victim. Epidural hematomas are typically hyperattenuating extra-axial fluid collections as seen on unenhanced CT scans. Active bleeding may produce small areas of low attenuation that have a swirl configuration [109]. Lacerations of the middle meningeal artery most frequently cause bleeding in acute epidural hemorrhages. Torn veins or dural sinuses and bone fragments account for the remainder. Occasionally epidural hematomas do not become symptomatic for a protracted period after injury and sometimes they are chronic [106]. Such hematomas are less likely to be caused by bleeding arteries [110]. In severe burns of the head, a thin diffuse layer of coagulated blood forms in the epidural space [14]. The intense heat rather than trauma is the likely cause of this change. Subdural hematoma Subdural hematomas are accumulations of blood in the subdural space that form by bleeding from torn bridging parasagittal veins after TBI. Subdural hematomas may be caused by open (penetrating) or

closed head injuries. Such hemorrhages occur commonly in vehicular accidents, after falls and assaults, in victims of child abuse, and in sporting accidents. Important predisposing factors for their development are the older age of victims and alcoholism. Rarely, ventricular decompression for hydrocephalus and pneumoencephalography have caused such hemorrhages [83,111 113]. The absence of a history of head injury occurs surprisingly often and was reported in 23.6% of 212 surgically diagnosed subdural hematomas [83]. Finally, about 4% of subdural hematomas arise in the margins of old traumatic scars [114]. Subdural hematomas remain an important cause of death of severely injured patients. CT scans may be used to determine the internal architecture of chronic subdural hematomas. This assessment may be useful in predicting the success of operative treatment for removal of subdural hematomas. In a recent study, the laminar stage hematomas had a high recurrence rate [115]. These hematomas were identified by finding a high-density laminar structure running along the inner membrane. The high density likely represents fresh bleeding from the vascularized inner membrane. Acute, subacute, and chronic subdural hematomas are usually distinguished by the time interval between injury and the onset of symptoms. Hematomas that become symptomatic within 3 days are considered acute. If they become symptomatic between 3 and 21 days, they are considered subacute and if they become symptomatic after 21 days they are considered chronic [94]. Some physicians prefer to identify the age of a lesion by its pathologic appearance [116]. For example, if well-formed neomembranes encase a hematoma, they consider such a hematoma chronic regardless of its age. Intracerebral hematoma Hematomas of the brain develop after closed or open head injury. Intracerebral hematomas after TBI measuring 0.5 cm or more in diameter developed in 0.4 to 9% of all intracranial hematomas treated neurosurgically [117]. They occur concomitantly in about 20% of acute subdural hematomas, but they occurred less often with an epidural hematoma alone, or combined with a subdural hematoma. Intracerebral hematomas are more likely to be solitary in nonlethal injuries. At autopsy, intracerebral hematomas may be found in up to 40% of cases [10,32]. Hematomas develop most frequently in the temporal and frontal lobes in association with cortical contusions and lacerations. A burst lobe applies to an intraparenchymal hematoma arising in the margin of a cortical laceration or contusion [118] (Fig. 3). A

182

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

subdural hematoma nearly always accompanies such lesions. Frontal and temporal hematomas extend into the lateral ventricles more often than hematomas of the caudate nucleus or thalamus [119]. Laceration of the septum pellucidum and fornices may produce a ventricular hematoma. Hematomas confined to the parasagittal white matter of the frontal lobes (so called gliding contusions) typically occur in a patient whose head is in motion at the time of impact [50]. Hematomas of the basal ganglia and thalamus and DAI often accompany these gliding contusions [89,120]. In the past, intracerebral hematomas that become symptomatic after a delay of a week or more were classified as delayed posttraumatic hematomata (of Bollinger) [121]. These hematomas may become symptomatic within 24 hours of injury and be identified by serial CT scans [122]. The development of intracerebral hematomas remains a complex pathophysiologic process. Alcohol intoxication directly contributes to the development of intracerbral hematomas. Alcohol may alter platelet function, coagulation, or membrane-bound enzymes [123,124]. Thrombocytopenia also appears to enhance bleeding [125,126]. Open or closed head injuries that damage the brain release tissue thromboplastins into the circulation that initiates disseminated intravascular coagulation [127]. Brain swelling Brain swelling develops with other injuries of the brain. One or both cerebral hemispheres may swell independently of the focal injuries. Two closely related patterns of diffuse brain swelling were identified by CT scanning [128 130]. These patterns occur almost exclusively in children and adolescents [11]. In one pattern, the diffuse brain swelling is attributed to brain edema. In the other, the swelling is believed to be caused by severe hyperemia and increased blood flow [129 132]. Diffuse brain swelling often follows a relentless course to death; whereas cerebral vascular congestion is usually followed by complete recovery. In young children, diffuse brain swelling may evolve without other apparent injury. Rapid neurologic deterioration may begin within minutes to hours after head injury. Cerebral blood flow and CT density studies suggest that this swelling is not due to edema. Such children with Glascow coma scores of > 8 make a complete recovery and are normal on follow-up CT studies. Children with coma scores of < 8 usually do well, though a few die. The few autopsies done on such cases seem to confirm the findings shown by CT and blood flow studies [133]. By comparison, adults may undergo similar clinical neurologic deteriFig. 3. Laceration and intracerebral and subdural hemorrhages of a burst left temporal lobe following severe impact to the right side of head in an auto accident. Basal view of the brain after formalin fixation [55].

oration but usually have a mass lesion, often an expanding hematoma. Children who are unconscious immediately after an injury often have diffuse brain edema and other injuries comparable with those found in adults. This response is especially difficult to assess in cases of child abuse. In a review of non-accidental head injury in 53 children, 82% died from brain swelling with raised intracranial pressure, 77% had severe hypoxic brain damage, 72% had acute subdural bleeding, and retinal hemorrhage occurred in 71%. Surprisingly, only 6% had DAI. However, 21% had localized axonal injury of the craniocervical junction [134,135]. Such patients may remain comatose for weeks to months. In follow-up CT examinations of these patients, other injuries often become visible. In one study, 17 of 34 children developed increased intracranial pressure, and 5 of 34 patients died [129]. Follow-up CT examinations demonstrated subarachnoidal collections of presumed cerebrospinal fluid in 27% of patients and ventricular dilation and cortical atrophy in 35%. Such

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187

183

changes suggest that these patients suffered severe diffuse injury of the white matter. In a series of 589 severe head injury cases of adults, 11% (66 cases) presented with unilateral enlargement of a cerebral hemisphere [136]. The cerebral swelling was associated with an ipsilateral subdural hematoma in 88% of cases, a large epidural hematoma in 7%, and isolated cerebral swelling in 4%. These cases had the most uncontrollable intracranial hypertension, highest mortality rate, and shortest survival among patients in this series. Diffuse swelling causes flattening of gyri, narrowing of sulci, and symmetrical collapse of the ventricular system. With localized swelling, distortion and herniation may appear. Minimal swelling is not so easily recognized. Cerebral edema produces pallor of the myelin, distention of the perivascular and pericellular spaces, rarefaction of subpial spaces, a vacuolar appearance of the neuropil, and pools of protein-rich fluid in the spongy-appearing areas [137,138]. Ischemic/hypoxic damage Ischemic brain damage is common in patients dying from TBI [28,30]. Such damage has a predilection for the hippocampus, basal ganglia, and cerebellum and occurs soon after injury [139,140]. Such injury is more common in patients who have had known hypoxia and in patients with increased intracranial pressure [28,29]. In a retrospective study of 16 patients with posttraumatic cerebral infarctions, infarcts were found in the distribution of all the major cerebral arteries [141]. Infarcts in the posterior cerebral artery distribution were most frequent and occurred in nine patients. Acquired intracranial herniation or focal mass effect explained 81.2% of cases. Vascular injury with dissection of the carotid artery was documented angiographically and may be a cause of a cerebral infarction [141]. Increased intracranial pressure Pressure necrosis of one or both parahippocampal gyri indicates that intracranial pressure has been high during life [55]. Graham et al found that 75% of 434 patients dying with TBI had such lesions [29]. Increased intracranial pressure may occur whenever there is a mass lesion like an intracranial hematoma or brain swelling. Subdural hygroma A subdural hygroma (hydrome) is the accumulation of watery (serous) fluid resembling cerebrospinal

fluid in the subdural space [142]. In addition, meningitis serosa traumatica and traumatic subdural effusion are terms used to describe this condition [143]. The fluid may be clear, pink, or xanthochromic and may be under variable pressure. Hygromas may develop immediately or after a delay following head injury. They are caused by motor vehicle accidents, domestic accidents, fights, and child abuse [39,44,144]. Subdural hygromas develop in males and females of all ages. Most hygromas form above the tentorium [136]. They may form unilaterally or bilaterally over the cerebral hemispheres. Hygromas occur alone or concomitantly with other head and brain injuries [44,136,143]. Chronic subdural hematoma, epidural hematoma, skull fractures, or ventricular enlargement may accompany them. Traumatic subdural hygromas make up 13% of operated subdural lesions [143]. Symptoms may vary and depend on the amount of fluid and how rapidly the fluid collects. Some patients remain asymptomatic. In others, the fluid accumulation forms a progressively expanding mass lesion that may lead to coma and death. CT scanning now makes preoperative recognition of hygromas possible [44,143,145,146]. Hygromas differ from hypodense subdural hematomas in their lower attenuation coefficients on CT scans. The low protein content of cerebrospinal fluid compared with that of blood produces these low values [147]. Delayed magnetic resonance imaging with Gadolinium-diethylenetriaminepenta-acetic acid (Gd-DTPA) shows significantly greater enhancement that is attributed to the increased capillary density in the outer membrane of the effusion (hypodense subdural hematoma) [148]. Nevertheless, some authors contend that differentiation of a hypodense subdural hematoma from a subdural hygroma is not always possible [107,149]. Tears of the arachnoid or impaired cerebrospinal fluid absorption likely produce hygromas [114,144,150].

References
[1] Sosin DM, Sniezek JE, Waxweller RJ. Trends in death associated with traumatic brain injury, 1979 through 1992. Success and failure. JAMA 1995; 273:1778 80. [2] Kraus JF, McArthur DL. Epidemiology of brain injury. In: Cooper PR, Golfinos JG, editors. Head injury. New York: McGraw-Hill Medical Publishing Division; 2000. p. 1 26. [3] Adams JH, Graham DI, Scott G, et al. Brain damage in fatal non-missile head injury. J Clin Pathol 1980; 33:1132 45. [4] Kraus JF, McArthur DL. Epidemiology of head injury.

184

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187 In: Cooper PR, editor. Head injury. Baltimore: Williams & Wilkins; 1993. p. 1 26. Gikas P. Forensic aspects of the highway crash. Pathology Annual 1983;18:147 63. Baker SP, Baker ON, Karpf RS. The injury fact book. Lexington: D.C. Heath and Co.; 1984. Wasserman R, Waller J, Mounty M, et al. Bicycles, helmets, and head injuries. Am J Public Health 1988; 78:1220 1. Watson GS, Zador PL, Wilks A. The repeal of helmet laws and increased motorcycle mortality in the United States 1975 78. Am J Public Health 1980;70:579 85. Watson GS, Zador PL, Wilks A. Helmet use, helmet use laws and motorcycle fatalities. Am J Public Health 1981;71:297 300. Clifton GL, McCormick WF, Grossman RG, et al. Neuropathology of early and late deaths after head injury. Neurosurgery 1981;8:309 14. Graham DI, Ford I, Adams JH, et al. Fatal head injury in children. J Clin Pathol 1989;42:18 22. Graham D, Adams JH, Nicoll JAR, et al. The nature, distribution and causes of traumatic brain injury. Brain Pathol 1995;5:397 400. Hardman JM. The pathology of traumatic brain injuries. Adv Neurol 1979;22:15 50. Hardman JM. Cerebrospinal trauma. In: Davis RL, Robertson DM, editors. Textbook of neuropathology. Baltimore: Williams & Wilkins; 1997. p. 1179 232. Lindenberg R. Trauma of meninges and brain. In: Minckler J, editor. Pathology of the nervous system. New York: McGraw-Hill; 1971. p. 1705. Adams JH, Graham DI, Gennarelli TA. Acceleration induced head injury in the monkey. Acta Neuropathol (Berl) 1981;7(Suppl):26 8. Gennarelli TA, Adams JH, Graham DI. Acceleration induced head injury in the monkey: the model, its mechanical and physiologic correlates. Acta Neuropathol (Berl) 1981;7(Suppl):23 5. Gennarelli TA, Spielman GM, Langfit TW. Influence of the type of intracrannial lesion on outcome from severe head injury: a multicenter study using a new classification system. J Neurosurg 1982;56:26 32. Gennarelli TA, Thibault LE, Adams JH, et al. Diffuse axonal injury and traumatic coma in the primate. Ann Neurol 1982;12:564 74. Osborn AG, Daines JH, Wing SD, et al. Intracranial air on computerized tomography. J Neurosurg 1978; 48:355 9. Sinson G, Bagley LJ, Cecil KM, et al. Magnetization transfer imaging and proton MR spectroscopy in the evaluation of axonal injury: correleation with clinical outcome after traumatic brain injury. AJNR Am J Neuroradiol 2001;22:143 51. Udstuen GJ, Claar JM. Imaging of acute head injury in the adult. Sem Ultrasound, CT, and MRI 2001; 22:135 47. Adams JH, Mitchell DE, Graham DI, et al. Diffuse brain damage of immediate impact type. Brain 1977; 100:489 502. [24] Adams JH, Graham DI, Murray LS. Diffuse axonal injury due to non-missile head injury in humans: an analysis of 45 cases. Ann Neurol 1982;12:557 63. [25] Adams JH, Doyle D, Graham DI, et al. Microscopic diffuse axonal injury in cases of head injury. Med Sci Law 1985;25:265 9. [26] Adams JH, Doyle D, Ford I, et al. Diffuse axonal injury in head injury: definition, diagnosis and grading. Histopathology 1989;15:49 59. [27] Gennarelli TA. The spectrum of traumatic axonal injury. Neuropathol Appl Neurobiol 1996;22:509 15. [28] Graham DI, Adams JH, Doyle D. Ischaemic brain damage in fatal non-missile head injuries. J Neurol Sci 1978;39:213 34. [29] Graham DI, Lawrence AE, Adams JH, et al. Brain damage in fatal non-missile head injury secondary to high intracranial pressure. Neuropathol Appl Neurobiol 1987;13:209 17. [30] Graham DI, Ford I, Adams JH, et al. Ischaemic brain damage is still common in fatal non-missile head injury. J Neurol Neurosurg Psychiatry 1989;52:346 50. [31] Graham DI, Clark JC, Adams JH, et al. Diffuse axonal injury caused by assault. J Clin Pathol 1992; 45:840 1. [32] Graham DI. Neuropathology head injury. In: Narayan RK, Wilberger JE Jr, Povlishock JT, editors. Neurotrauma. New York: McGraw-Hill Companies Inc.; 1996. p. 43. [33] Graham DI, McIntosh TK, Maxwell WL, et al. Recent advances in neurotrauma. J Neuropathol Exp Neurol 2000;59:641 51. [34] Osborn AG, Blaser SI, Salzman KL. Pocket radiologist-brain top 100 diagnoses. Salt Lake City: Amirsys Inc./W.B. Saunders Co.; 2002. p. 1 3. [35] De Grood MPAM. Skull fractures. In: Vinken PJ, Bruyn GW, editors. Handbook of clinical neurology, vol. 23. Amsterdam: North-Holland; 1975. p. 387 402. [36] Ehler E, Ivankievicz D, Schomacher GH. Diagnosis of skull fractures by autopsy and radiology. Acta Morphol Acad Sci Hung 1980;28:291 6. [37] Braakaman R, Jennett B. Depressed skull fracture (non-missile). In: Vinken PB, Bruyn GW, editors. Handbook of clinical neurology, vol. 23. Amsterdam: North-Holland; 1975. p. 403 15. [38] LeCount ER, Apfelbach CW. Pathologic anatomy of traumatic fractures of cranial bones and concomitant injuries. JAMA 1920;74:501 11. [39] Vance BM. Fractures of the skull. Complications and causes of death: a review of 512 necropsies and 61 cases studied clinically. Arch Surg 1927;14:1023 92. [40] Russell R, Schiller F. Crushing injury of the skull: clinical and experimental observation. J Neurol Neurosurg Psychiatry 1949;12:52 60. [41] Hirsch CS, Haufman B. Contrecoup skull fractures. J Neurosurg 1975;42:530 4. [42] Sato O, Kamitani H, Kamitani T. Blow-in fractures of both orbital roofs caused by shear strain of the skull. J Neurosurg 1978;49:734 8. [43] Laun A. Traumatic cerebrospinal fluid fistulas in the

[5] [6] [7]

[8]

[9]

[10]

[11] [12]

[13] [14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187 anterior and middle cranial fossae. Acta Neurochir 1982;60:215 22. Zimmerman RA, Bilaniuk LT, Bruce D, et al. Computed tomography of craniocerebral injury in the abused child. Radiology 1979;130:687 90. Kingsley D, Till K, Hoare R. Growing fractures of the skull. J Neurol Neurosurg Psychiatry 1978;41: 312 8. Lende RA, Erickson TC. Growing fractures of childhood. J Neurosurg 1961;18:479 89. Matson DD. Leptomeningeal cyst. Neurosurgery of infancy and childhood. Springfield: Charles C. Thomas; 1969. p. 304. Taveras J, Ransokoff J. Leptomeningeal cysts of the brain following trauma with erosion of the skull: a study of seven cases treated by surgery. J Neurosurg 1953;10:233 41. Lindenberg R, Freytag E. Morphology of cortical contusions. Arch Pathol Lab Med 1957;63:23 42. Lindenberg R, Freytag E. A mechanism of cerebral contusions. A pathologic-anatomic study. Arch Pathol Lab Med 1960;69:440 69. Adams JH, Doyle D, Graham DI, et al. Gliding contusions in non-missile head injuries in humans. Arch Pathol Lab Med 1986;110:485 8. Dawson SL, Hirsch CS, Lucas F, et al. The contrecoup phenomenon. Hum Pathol 1980;11:155 66. Rodda RA, Adams JH, Graham DI, et al. Fatal nonmissle head injury: a cohort study of patients with and without an intracranial hematoma. Clin Neuropathol 1987;6:179 84. Adams JH, Doyle D, Graham DI, et al. The contusion index: a reappraisal in man and experimental nonmissile head injury. Neuropathol Appl Neurobiol 1985;11:299 308. Adams JH, Graham DI. The relationship between ventricular fluid pressure and neuropathology of raised intracranial pressure. Neuropathol Appl Neurobiol 1976;2:323 32. Lindenberg R, Freytag E. Brainstem lesions characteristic of traumatic hyperextension of the head. Arch Pathol 1970;90:509 15. Simpson DA, Blumbergs PC, Cooter RD, et al. Pontomedullary tears and other gross brainstem injuries after vehicular accidents. J Trauma 1989;29:1519 25. Britt RH, Herrick MK, Mason RT, et al. Traumatic lesions of the ponto-medullary junction. Neurosurgery 1980;6:623 31. Klingele TG, Schulz R, Murphy MG. Pontine gaze paresis due to traumatic craniocerebral hyperextension. Report of two cases. J Neurosurg 1980;53:249 51. Pilz P, Strohecker J, Grobovschek M. Survival after traumatic ponto-medullary tear. J Neurol Neurosurg Psychiat 1982;45:422 7. Pilz P. Axonal injury in head injury. Acta Neurochir 1983;32Suppl(Wein):119 23. Pilz P. Survival after trauma: the pontomedullary junction. Acta Neurochir 1983;32Suppl(Wein):75 8. Sheriff FE, Bridges LR, Sivaloganathan S. Early de-

185

[44]

[64] [65]

[45]

[46] [47]

[66]

[67]

[48]

[68]

[49] [50]

[69]

[70]

[51]

[52] [53]

[71]

[72] [73]

[54]

[74] [75]

[55]

[76]

[56]

[57]

[77]

[58]

[78]

[59]

[79]

[60]

[80] [81]

[61] [62] [63]

[82]

tection of axonal injury after human head trauma using immunocytochemistry for b-amyloid precursor protein. Acta Neuropathol 1994;87:55 62. von Wuermeling HB, Struck G. Brain stem ruptures in traffic accidents. Beitr Gerichtl Med 1965;23:297. Zampella EJ, Duvall ER, Langford KH. Computed tomography and magnetic resonance imaging in traumatic locked-in syndrome. Neurosurgery 1988;22: 591 3. Blumbergs PC, Jones NR, North JB. Diffuse axonal injury in head trauma. J Neurol Neurosurg Psychiatry 1989;52:838 41. Clark JM. Distribution of microglial clusters in the brain after head injury. J Neurol Neurosurg Psychiatry 1974;37:463 74. Povlishock JT, Becker DP, Cheng CLY, et al. Axonal change in minor head injury. J Neuropathol Exp Neurol 1983;42:225 42. Oppenheimer DR. Microscopic lesions in the brain following head injury. J Neurol Neurosurg Psychiatry 1968;31:299 306. Roberts AH. Brain damage in boxers. A study of the prevalence of traumatic encephalopathy among exprofessional boxers. London: Pitman Medical & Scientific Publishing; 1969. Roberts GW, Allsop D, Bruton C. The occult aftermath of boxing. J Neurol Neurosurg Psychiatry 1990; 53:373 8. Tomlinson BE. Brain stem lesions after head injury. J Clin Pathol 1970;23(Suppl 4):154 65. Strich SJ. Diffuse degeneration of the cerebral white matter in severe dementia following head injury. J Neurol Neurosurg Psychiatry 1956;19:163 85. Corsellis JAN. Boxing and the brain. BMJ 1989; 298:105 9. Blumbergs PC, Scott G, Manavis J, et al. Staining of amyloid precursor protein to study axonal damage in mild head injury. Lancet 1994;344:1055 6. Gentleman SM, Nash MJ, Sweeting CJ, et al. b-amyloid precursor protein (b-APP) as a marker of axonal injury after head injury. Neurosci Lett 1993;160: 139 44. Grady MS, McLaighlin MR, Christman CW, et al. The use of antibodies targeted against the neurofilament subunits for the detection of diffuse axonal injury. J Neuropathol Exp Neurol 1993;52:143 52. Gultekin SH, Smith TW. Diffuse axonal injury in craniocerebral trauma. A comparative histologic and immunohistochemical study. Arch Pathol Lab Med 1994;118:168 71. Adams JH, Graham DI, Doyle D, et al. Diffuse axonal injury in head injuries caused by a fall. Lancet 1984; 2:1420 2. Imajo T, Kazee AM. Diffuse axonal injury in a simple fall. Am J Forensic Med Pathol 1992;13:169 72. Imajo T, Challener RC, Roessman U. Diffuse axonal injury by assault. Am J Forensic Med Pathol 1987; 8:217 9. Sherriff FE, Bridges LR, Gentleman SM, et al.

186

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187 Markers of axonal injury in post mortem human brain. Acta Neuropathol 1994;88:433 9. Wintzen AR. The clinical cause of subdural hematoma. Brain 1980;103:855 67. Yaghmai A, Povlishock J. Traumatically induced reactive changes as visualized through the use of monoclonal antibodies to neurofilament subunits. J Neuropathol Exp Neurol 1992;51:158 76. McKenzie KJ, McLellan DR, Gentleman SM, et al. Is b-APP a marker of axonal damage in short surviving head injury? Acta Neuropathol (Berl) 1996; 92:608 13. Maxwell WL, Povlishock JT, Graham DL. A mechanistic analysis of non-disruptive axonal injury; a review. J Neurotrauma 1997;14:419 40. Povilshock JT, Jenkins LW. Are the pathobiological changes evoked by traumatic brain injury immediate and irreversible? Brain Pathol 1995;5:415 26. Bradley Jr WG, Brant-Zawadzki M, Cambry-Forker J. Case 10 shear injury. In: Lufkin RB, Bradley WG Jr, Brandt-Zawadzki MB, editors. MRI of the brain I. 2nd edition. Philadelphia: Lippincott-Williams & Wilkins; 2001. p. 20 2. Adams JH, Doyle D, Graham DI, et al. Deep intracerebral (basal ganglia) hematomas in fatal non-missile head injury in man. J Neurol Neurosurg Psychiatry 1986;49:1039 43. MacPherson P, Teasdale E, Dhaker S, et al. The significance of traumatic haematoma in the region of the basal ganglia. J Neurol Neurosurg Psychiatry 1986; 49:29 34. Blumbergs PC, Scott G, Manavis J, et al. Topography of axonal injury as defined by amyloid precursor protein and the sector scoring method in mild and severe closed head injury. J Neurotrauma 1995;12:565 72. Bullock R, Teasdale G. Surgical management of traumatic intracranial hematomas. In: Braakman R, editor. Handbook of clinical neurology: head injury, vol. 57 (Revised series 13). Amsterdam: North-Holland; 1990. p. 249. Jamieson KG, Yelland JDN. Extradural hematoma: report of 167 cases. J Neurosurg 1968;29:13 23. Jamieson KG, Yelland JDN. Surgically treated traumatic subdural hematoma. J Neurosurg 1972;37: 137 49. Jamieson KG, Yelland JDN. Traumatic intracerebral hematoma: report of 63 surgically treated cases. J Neurosurg 1972;37:528 32. Jamieson KG. Posterior fossa hematoma. In: Vinken PJ, Bruyn GW, editors. Handbook of clinical neurology: injuries of the brain and skull, Part II. Amsterdam: North-Holland Publishing Co.; 1976. p. 343 50. Jamieson KG, Yelland JDN. Epidural hematoma. In: Vinken PJ, Bruyn GW, editors. Handbook of clinical neurology: injuries of the brain and skull, part II, vol. 24. Amsterdam: North-Holland; 1976. p. 261 74. Cordobes F, Lobato RD, Rivas JJ, et al. Observations on 82 patients with extradural hematoma. J Neurosurg 1981;54:179 86. [99] Hernesniemi J. Outcome following subdural hematoma. Acta Neurochir (Wien) 1979;49:191 8. [100] Hirsh LF. Chronic epidural hematomas. Neurosurgery 1980;6:508 12. [101] Iwakuma T, Brunngraber CV. Chronic extradural hematomas. A study of 21 cases. J Neurosurg 1973; 38:488 93. [102] Mathur PPS, Dharker SR, Agarwall SK, et al. Fluid chronic extradural hematoma. Surg Neurol 1980; 14:81 2. [103] McKissock W, Taylor JC, Bloom WH, et al. Extradural hematoma. Observation on 125 cases. Lancet 1960;2:167 72. [104] Parkinson D, Reddy V, Taylor J. Ossified epidural hematoma: case report. Neurosurgery 1980;7:171 3. [105] Phonprasert C, Suwanwela C, Hongsaprabhas C, et al. Extradural hematoma: analysis of 138 cases. J Trauma 1980;20:679 83. [106] Rivas JJ, Lobato RD, Sarabia R, et al. Extradural hematoma: analysis of factors influencing the courses of 161 patients. Neurosurgery 1988;23:44 51. [107] Munro D, Merritt HH. Surgical pathology of subdural hematoma. Based on a study of one hundred and five cases. Arch Neurol Psychiatry 1936;35:64 78. [108] Baykaner K, Alp H, Ceviker N, et al. Observation of 95 patients with extradural hematoma and review of the literature. Surg Neurol 1988;30:339 41. [109] Al-Nakshabandi NA. The swirl sign. Radiology 2001; 218:433. [110] Aronson S, Okazaki H. A study of some factors modifying response of cerebral tissue to subdural hematomata. J Neurosurg 1963;20:89 93. [111] Miyaishi S, Moriya F, Yamamoto Y, et al. Massive pulmonary embolization with cerebral tissue due to gunshot wound to the head. Brain Inj 1994;8:559 64. [112] Stehbens W. Pathology of the cerebral blood vessels. St Louis: C.V. Mosby; 1972. p. 207. [113] Yashon D, Jane JA, White RJ, et al. Traumatic subdural hematoma in Infancy. Arch Neurol 1968; 18:370 7. [114] Stone JL, Rifai MHS, Sugar O, et al. Subdural hematomas. I. Acute subdural hematoma: progress in definition, clinical pathology, and therapy. Surg Neurol 1983;19:216 31. [115] Nakaguchi H, Tanishima T, Yoshimasu N. Factors in the natural history of chronic subdural hematomas that influence their postoperative recurrence. J Neurosurg 2001;95:256 62. [116] Fogelholm R, Heiskanen O, Waltimo O. Chronic subdural hematoma in adults. Influence of patients age on symptoms, signs, and thickness of hematoma. J Neurosurg 1975;42:43 6. [117] Rivano C, Borzone M, Carta F, et al. Traumatic intracerebral hematomas. Seventy-two cases surgically treated. J Neurosurg Sci 1980;24:77 84. [118] McCormick WF. Trauma. In: Schochet SS Jr. Neuropathology, vol. 3. In: Rosenberg RN, editor. The clinical neurosciences. New York: Churchill Livingstone; 1983. p. 111.

[83] [84]

[85]

[86]

[87]

[88]

[89]

[90]

[91]

[92]

[93] [94]

[95]

[96]

[97]

[98]

J.M. Hardman, A. Manoukian / Neuroimag Clin N Am 12 (2002) 175187 [119] Fujitsu K, Kuwabara T, Murainoto M, et al. Traumatic intraventricular hemorrhage: report of twenty-six cases and consideration of the pathogenic mechanism. Neurosurgery 1988;23:423 30. [120] Boto GR, Lobato RD, Rivas JJ. Basal ganglia hematomas in severely head injured patients; clinicoradiological analysis of 37 cases. J Neurosurg 2001; 94:224 32. [121] Brown FD, Mullan S, Duda EE. Delayed traumatic intracerebral hematomas. J Neurosurg 1978;48: 1019 22. [122] Young HA, Gleave JRW, Schmidek HH. Delayed traumatic intracerebral hematoma: report of 15 cases operatively treated. Neurosurgery 1984;14:22 5. [123] Cowan DH. The platelet defect in alcoholism. Ann NY Acad Sci 1975;252:328 41. [124] Flamm ES, Demopoulous HB, Seligman ML, et al. Ethanol potentiation of central nervous system trauma. J Neurosurg 1977;46:328 39. [125] Auer L. Disturbances of the coagulatory system in patients with severe cerebral trauma. I. Acta Neurochir (Wein) 1978;43(1 2):51 9. [126] Auer L. Disturbances of the coagulatory system in patients with severe cerebral trauma. II. Acta Neurochir 1979;49:219 26. [127] Kumura E, Sato M, Fukuda A, et al. Coagulation disorders following acute head injury. Acta Neurochir (Wien) 1987;85:23 8. [128] Bruce D, Obrist W, Zimmerman R, et al. The pathophysiology of acute severe brain swelling following pediatric head trauma. Cerebral blood flow and metabolism. Acta Neurol Scand 1979;60(Suppl 72):372 3. [129] Bruce DA, Alavi A, Bilaniu L, et al. Diffuse cerebral swelling following head injuries in children: the syndrome of malignant brain edema. J Neurosurg 1981;54:170 8. [130] Bruce DA. Special considerations of the pediatric age group. In: Cooper PR, editor. Head injury. Baltimore (MD): Williams and Wilkins; 1982. p. 315. [131] Bell WE, McCormick WF. Increased intracranial pressure in children. In: Schaefer A, editor. Major problems in clinical pediatrics. Philadelphia: W.B. Saunders; 1971. p. 123. [132] Obrist WD, Thompson Jr HK, Wang HS, et al. Regional cerebral blood flow estimated by Xenon inhalation. Stroke 1975;6:245 56. [133] Snoek J, Jennett B, Adams JH, et al. Computerized tomography after recent severe head injury in patients without acute intracranial hematoma. J Neurol Neurosurg Psychiatry 1979;42:215 25. [134] Geddes JF, Hackshaw AK, Vowels GH, et al. Neuropathology of inflicted head injury in children I. Patterns of brain damage. Brain 2001;124:1290 8. [135] Geddes JE, Vowles GH, Hackshaw AK, et al. Neuropathology of inflicted head injury in children II. Microscopic brain injury in infants. Brain 2001;124: 1299 306.

187

[136] Sarabia R, Lobato RD, Rivas JJ, et al. Cerebral hemisphere swelling in severe head injury patients. Acta Neurchir Suppl (Wein) 1988;42:40 6. [137] Bakay L, Lee JC. Cerebral edema. Springfield: Charles C. Thomas; 1965. [138] Feigin I. Sequence of pathological changes in brain edema. In: Vinken PJ, Bruyn GW, editors. Brain edema. New York: Springer-Verlag; 1967. p. 129. [139] Bouma GJ, Muizelaar JP, Stringer WA, et al. Ultraearly evaluation of regional cerebral blood flow in severely head-injured patients using xenon-enhanced computerized tomography. J Neurosurg 1992;77:360 8. [140] Kotapka MJ, Graham DI, Adams JH, et al. Hippocampal damage in fatal non-missile human head injury: frequency and distribution. Acta Neuropathol (Berl) 1992;83:530 40. [141] Server A, Dullerud R, Haakonsen M, et al. Post-traumatic cerebral infarction. Neuroimaging findings, etiology and outcome. Acta Radiol 2001;42:254 60. [142] Gurdjian ES, Brihaye J, Christenson JC, et al. Glossary of neurotraumatology. Acta Neurochir 1978; 25(Suppl):36. [143] Stone JL, Lang RGR, Sugar O, et al. Traumatic subdural hygroma. Neurosurgery 1981;8:542 50. [144] St John JN, Dila C. Traumatic subdural hygroma in adults. Neurosurgery 1981;9:621 6. [145] French BN, Dublin AB. The value of computerized tomography in the management of 1000 consecutive head injuries. Surg Neurol 1977;7:171 83. [146] Sanders MD, Hoyt WF. Hypoxic ocular sequelae of carotid-cavernous fistulae. Br J Ophthalmol 1969;53: 82 97. [147] Norman D, Price D, Boyd D, et al. Quantitative aspects of computed tomography of blood and cerebrospinal fluid. Radiology 1977;123:335 8. [148] Mori K, Maeda M. Delayed magnetic resonance with Gd-DTPA differentiates subdural hygroma and subdural effusion. Surg Neurol 2000;53:303 10. [149] Scotti G, Terbrugge K, Melancon D, et al. Evaluation of the age of subdural hematomas by computerized tomography. J Neurosurg 1977;47:311 5. [150] Naffziger HC. Subdural fluid accumulations following head injury. JAMA 1924;82:1751.

Further Readings
Patscheider H. Zur Entstehung von Ringbruchen des Schadelgrundes. Dtsch Z Gestamte Gerichtl Med. 1962;52:13. Ryan GA, McLean AJ, Vilenius ATS, et al. Brain injury patterns in fatally injured pedestrians. J Trauma 1994;36:469. Yaghamai A, Povlishock JT. Traumatically induced reactive axonal change as visualised through the use of monoclonal antibodies targeted neurofilament subunits. J Neuropathol Exp Neurol 1992;51:151.

Neuroimag Clin N Am 12 (2002) 189 204

Imaging of traumatic intracranial hemorrhage

Robert J. Young, MDa,*, Sylvie Destian, MDa,b
a

Department of Radiology, Saint Lukes Roosevelt Hospital Center, 1000 Tenth Avenue, New York, NY 10019, USA b Department of Radiology, Saint Vincent Catholic Medical Center, Manhattan Division, 153 West Eleventh Street, New York, NY 10011, USA

Traumatic brain injury is a leading cause of morbidity and mortality in the United States, with an incidence of 95 per 100,000 population [1]. The most frequent causes are motor vehicle accidents, falls, assault, and firearm-related incidents [2]. The Centers for Disease Control and Prevention estimate that 22% of people suffering traumatic brain injury die from their injuries. Deaths attributable to intracerebral and intracranial hemorrhage range from 5.0 to 5.6 per 100,000 population [3]. Head trauma is particularly common in adolescents, young adults and the elderly. Increased age is an independent risk factor for acute traumatic intracranial hemorrhage, perhaps due to cerebral atrophy and an increased propensity for bridging veins to tear, or increased friability of vessels secondary to atherosclerotic disease or amyloid angiopathy [4]. The risk in males is twice the risk in females for all age groups [2]. Motor vehicle accidents are responsible for up to 35% of all cases of head trauma [2]. Polytrauma victims with signs of circulatory compromise have a relatively low incidence of traumatic intracranial hematomas, and should undergo prompt evaluation for serious injuries to the chest, abdomen and pelvis [5]. Falls cause up to 33% of all head trauma [6], and are the most common cause of hemorrhage in the elderly, particularly patients anticoagulated with sodium warfarin (Coumadin), who are at increased risk for hemorrhage [7]. Traumatic intracranial hematomas are common in head trauma patients with known or suspected seizure disorders, who appear postictal, or who are found convulsing after a fall [6].

Many patients may be managed medically [8]. Ten percent of medically managed patients may eventually require neurosurgical intervention. In a series reported by Patel et al [8], surgical evacuation was eventually required in 17% of epidural hematomas, 14% of subdural hematomas, and 10% of intraparenchymal hematomas, with frontal intraparenchymal hematomas particularly prone to early medical failure.

Imaging techniques Computed tomography Computed tomography (CT) is the imaging modality of choice in the evaluation of acute head trauma, due to its widespread availability, speed, and compatibility with life support and monitoring devices. Motion due to uncooperative patients is less important with the increasing availability of fast multidetector CT scanners [9]. If images are degraded by motion artifact, those particular slices can be selectively rescanned without repeating the entire scan. At our institution, we routinely view images in brain (window 80, level 40), subdural (window 200, level 70), and bone (window 3000/ level 400) formats to evaluate for parenchymal, extra axial, and osseous injuries, respectively. Subdural windows are particularly useful in detecting superficial hemorrhage, shallow contusions, and small extra axial collections, where the high attenuation of blood may blend into the adjacent high attenuation bone (Fig. 1). Limitations of CT include beam hardening effects, which may partially obscure blood in the posterior fossa, subtemporal, and subfrontal regions; and volume averaging of small

* Corresponding author. E-mail address: youngrobert@yahoo.com (R.J. Young).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 3 - 5

190

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

Fig. 1. Subtle acute subdural hematoma in a 19-year-old male. (A) Noncontrast CT image viewed with brain window demonstrates asymmetry of the sylvian fissures and calvarium. (B) Noncontrast CT image viewed with subdural window allows separation of the small high density left subdural hematoma from the higher density calvarium.

amounts of blood that occupy less than the acquired slice thickness with normal brain. In those instances where CT is equivocal, lesions such as small subdural hematoma will be more readily apparent on magnetic resonance (Fig. 2). Magnetic resonance Magnetic resonance (MR) is an alternative initial modality with greater sensitivity for detecting abnormalities and predicting prognosis [10], particularly in assessing injury to the brainstem. T2-weighted images are most useful for lesion detection, and T1weighted images most useful for anatomic localization [11]. The appearance of blood on MR imaging depends on the predominant type of hemoglobin present within the hematoma. In uncooperative, unstable, or claustrophobic patients, ultrafast sequences such as the gradient recalled echo single shot echo planar imaging sequence may be used, although characterization of hemorrhagic foci and sensitivity at the skull base are inferior to traditional gradient echo sequences [12]. Abbreviated MR studies using fast pulse sequences on ultra-low, low, or intermediate field strength systems, with less stringent ferromagnetic equipment restrictions, may offer answers to the crucial questions in the shortest time possible [13].

MR is inferior to CT in evaluation of injuries to the skull vault. Contraindications to MR imaging include pacemakers, noncompatible vascular clips, metallic implants, and ocular foreign bodies. Other imaging modalities Ultrasound is portable, inexpensive and does not subject patients to ionizing radiation. Neonates are the most suitable patients, where the open anterior and posterior fontanelles are suitable imaging windows. Compared to CT and MR, however, ultrasound has relatively lower sensitivity and specificity for intracranial hemorrhage and ischemia [14]. The spatial resolution of positron emission tomography (PET) and 99m technetium hexamethylpropyleneamineoxime (HMPAO) brain perfusion with single photon emission computed tomography (SPECT) is inferior to that of CT and MR, but the former may be more sensitive in detecting brain abnormalities in chronic traumatic brain injury patients [15,16].

Mass effect and herniation Regardless of the mechanism of injury, intra-axial and extra-axial hemorrhage may develop enough mass

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

191

Fig. 2. Acute subdural hematoma in a 37-year-old male. Contrast enhanced CT images viewed with (A) brain and (B) subdural windows reveal only mild left sulcal effacement. (C) Fluid attenuated inversion recovery MR image at similar level easily demonstrates the small left subdural hematoma.

effect to compress adjacent vascular structures and cause ischemia or infarct. Alternatively, mass effect may herniate part of the brain from one compartment into another. In subfalcine herniation, increased pressure directed medially causes the cingulate gyrus to shift beneath and across the falx cerebri. This may

cause compression of the anterior cerebral artery or internal cerebral veins. In descending transtentorial herniation, increased supratentorial pressure directed medially and inferiorly causes the temporal lobe to herniate over the tentorium, potentially compressing the oculomotor nerve, posterior cerebral, and anterior

192

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

choroidal arteries, or midbrain. In ascending transtentorial herniation, mass effect from the posterior fossa directed superiorly causes the cerebellum to herniate upward through the tentorial incisura. In uncal or medial transtentorial herniation, the uncus herniates through the tentorial incisura. Uncal herniation causes enlargement of ipsilateral basal cisterns and the contralateral temporal horn. In tonsillar herniation, increased pressure in the posterior fossa causes the cerebellar tonsils and cerebellum to herniate inferiorly through the foramen magnum, compressing the fourth ventricle, and thus producing acute obstructive hydrocephalus. The upper cervical cord may also be compressed.

Traumatic intra-axial injury Contusion Contusions are bruises of the brain. They occur in up to 43% of patients with blunt or nonpenetrating head injuries [11]. Neurosurgical intervention is more likely to improve outcome for hemorrhagic contusions in patients after a fall, patients who have a low Glasgow Coma score, patients with anisocoria, or patients older than 60 [4]. Contusions have traditionally been described as coup or contrecoup injuries due to deceleration/acceleration trauma. Coup contusions occur when the moving brain strikes the stationary calvarium. The brain decelerates as it hits the calvarium, and the calvarium is deformed temporarily, physically distorting the underlying brain parenchyma [17]. There is mechanical damage to neurons, with or without petechial hemorrhage or torn capillaries. Contrecoup contusions occur distant or opposite from the site of initial impact, when the brain is set in motion relative to the stationary calvarium [17]. With frontal impact trauma, the brain moves over the roughened edges of the inner table of the skull, particularly the floor of the anterior cranial fossa, and slams into the sphenoid wings and petrous ridges, explaining why contusions occur most commonly in the inferior frontal, anterior temporal, and lateral temporal regions (Fig. 3). Paramedian bony irregularities may cause superior frontal and parasagittal contusions. The acute contusion less than 12 hours old is composed mostly of intracellular oxyhemoglobin with the edematous brain undergoing necrosis. Noncontrast CT will show low attenuation if hemorrhage is absent and mixed, or high attenuation if hemorrhage is present. At this point, the high attenuation (50 70 Hounsfield units) is from high protein concentration within intact red blood cells and not iron content [18].

Fig. 3. Hemorrhagic contusions in a 38-year-old female. Noncontrast CT image shows bilateral inferior frontal and right anterior temporal hemorrhagic contusions.

On T2-weighted MR images, acute hemorrhagic contusions will exhibit inhomogeneous signal due to hypointense deoxyhemoglobin and hyperintense, edematous cortical tissue. MR is less sensitive than CT in the hyperacute stage because diamagnetic intracellular oxyhemoglobin lacks unpaired electrons and thus clot signal is close to normal brain parenchyma normal to slightly lower signal on T1-weighted images and slightly higher signal on T2-weighted images [19,20]. Repeat imaging is indicated to monitor the size of the hemorrhage and the development of delayed hemorrhage and vasogenic edema. Large hemorrhagic contusions commonly increase in size within the first 48 hours (Fig. 4). Acute hemorrhagic contusions one to three days old are composed mostly of paramagnetic intracellular deoxyhemoglobin. The deoxyhemoglobin is formed by the dissociation of oxygen from hemoglobin, a process that begins within several hours. Because the deoxyhemoglobin within intact, clotted hypoxic red blood cells does not cause T1 shortening, the hemorrhagic contusion will have normal to slightly lower signal on T1-weighted MR images. The concentration of red blood cells with clot and the concentration of fibrin cause T2 shortening, with areas of very low signal on T2-weighted spin echo and T2 * -weighted gradient echo images [20].

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

193

Fig. 4. Contusions in a 66-year-old male. (A) Initial noncontrast CT image demonstrates small hemorrhagic contusions in the left frontal lobe. (B) Noncontrast CT image 5 days later demonstrates a hemorrhagic contusion in the left temporal lobe and an increase in the size of the left frontal hemorrhagic contusions with surrounding edema.

Within a few days, the subacute contusion will start to undergo liquefaction with development of vasogenic edema. As the edema increases over the first week, it may be great enough to cause herniation. The edema has fluid or water characteristics: iso- to hypointense on T1-weighted images, and hyperintense on T2-weighted images (Fig. 5). With oxidation of deoxyhemoglobin to strongly paramagnetic intracellular methemoglobin, proton-electron dipole-dipole interactions between hydrogen atoms and the paramagnetic centers of methemoglobin will cause marked T1 shortening and very high signal intensity on T1-weighted images [20] within the periphery of the hematoma (Fig. 6). The intracellular methemoglobin will cause T2 shortening and very low signal on T2-weighted images. After erythrocyte membrane breakdown and extracellular migration of methemoglobin, there is neovascularization with removal of blood components and debris by macrophages. The new blood vessels at the periphery of the lesion lack the tight endothelial junctions of an intact blood brain barrier, and so there is intense enhancement of the margins on both contrast CT and MR [13]. The fragile granulation tissue vessels predispose the patient to additional episodes of acute hemorrhage. CT will show a decrease in the density of the contusion and decrease in the mass effect, the latter due to a decrease in

Fig. 5. Hemorrhagic contusion in a 58-year-old female. Axial T2-weighted image shows a left temporal contusion with hypointense acute blood anteriorly (*) and hyperintense subacute blood posteriorly (x). Intermediate increased signal around the contusion represents edema (arrowheads).

194

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

onstrate a homogenous high attenuation consolidation with well-defined margins (Fig. 7). Surrounding edema increases and peaks at one week. Subacute hematomas three to seven days old may give the appearance on CT of layering fluid-blood levels within the hematoma or clot retraction. Late subacute hematomas 7 to 14 days old will decrease in attenuation from the periphery inward, of approximately one to two Hounsfield units each day. Chronic cerebral hematomas more than two weeks old are composed primarily of intracellular ferritin and lysosomal hemosiderin. On CT, the hematoma will continue to decrease in attenuation. Within 3 to 10 weeks, chronic hematomas will become isodense with normal brain parenchyma and very difficult to detect. Continued proteolysis, phagocytosis and adjacent atrophy will eventually replace the hematoma with an area of encephalomalacia. Diffuse axonal injury Diffuse axonal injuries occur in up to 48% of patients with closed head injuries [11] when the shearing forces of rapid rotational acceleration or deceleration cause axonal disruption. Axonal injury ranges from incomplete disruption detected at only the

Fig. 6. Early subacute hemorrhagic contusion in a 78-yearold male. Sagittal T1-weighted image demonstrates high signal intensity at the periphery of the contusion, consistent with extracellular methemoglobin.

edema. MR will exhibit the persistent high signal of extracellular methemoglobin on T1- and T2-weighted images [20] for up to a year. The peripheral rim of hemosiderin and ferritin has slightly low signal on T1- and marked low signal on T2-weighted images [20] from the susceptibility effect of hemosiderin within macrophage lysosomes. Clot resorption begins from the periphery inward, and depending on the size of the hematoma, may vary from one to six weeks in duration. Necrotic tissue is sloughed and cystic cavities are formed over the next 6 to 12 months. Focal atrophy is characterized by a decrease in the size of cortical gyri, with compensatory enlargement of cerebrospinal fluid spaces and dilatation of the adjacent ventricle. Cystic cavities are surrounded by gliosis and hemosiderin scarring. Fibroglial scars may adhere dura to adjacent brain, and cause seizures in post-traumatic patients. Intraparenchymal hematoma Intraparenchymal hematomas unrelated to contusions are generally the result of penetrating trauma such as gunshot or stab wounds. Missile trauma with metallic objects may preclude MR imaging as a diagnostic modality. Noncontrast CT will demFig. 7. Intraparenchymal hemorrhage in a 54-year-old male. Axial CT image demonstrates acute hemorrhage in the left frontal lobe causing midline shift. There is a small left subdural hematoma. A subarachnoid hemorrhage outlines the basal cisterns.

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

195

microscopic level during postmortem analysis, to complete disruption associated with shearing hemorrhage from adjacent capillary lacerations. Severe diffuse axonal injury is responsible for coma and poor outcome in nearly half of patients with significant closed head trauma from motor vehicle accidents. Poor prognostic indicators include low Glasgow coma scale scores, concomitant shearing hemorrhages in the brainstem and corpus callosum, and hemorrhagic and nonhemorrhagic space-occupying lesions such as subdural hematomas, and hemorrhagic and nonhemorrhagic contusions [10,21]. Characteristically, diffuse axonal injuries are distributed at points of maximal shearing stress along white matter tracts. Nearly two-thirds of the lesions occur in lobar white matter at the corticomedullary junction, including frontal parasagittal region, temporal periventricular region, and less commonly the parietal and occipital lobes. The corpus callosum is commonly involved, especially the splenium and posterior body (Fig. 8). Intraventricular hemorrhage suggests injury to the corpus callosum with concomitant injury to the subependymal capillary and venous plexus along the ventricular surface of the corpus callosum, fornix, or septum pellucidum [22,23]. With

Fig. 9. Diffuse axonal injury in a 7-year-old. Noncontrast CT image demonstrates punctate hemorrhagic foci at the frontal gray-white junctions (arrows).

Fig. 8. Hemorrhagic shearing injury in a 77-year-old male. Sagittal T1-weighted image demonstrates high signal intensity representing subacute hemorrhage at the junction of the posterior body and splenium of the corpus callosum. In addition, a subacute subdural hematoma is seen posteriorly (arrows).

more severe head trauma, there is involvement of the basal ganglia, including the internal and external capsules and corona radiata. Lesions may also involve the rostral dorsolateral brainstem adjacent to the superior cerebellar peduncles and medial lemnisci, within the midbrain and upper pons. The initial CT is often normal, particularly when nonhemorrhagic shearing injury has occurred. Since more than 80% of shearing injuries are nonhemorrhagic [24], noncontrast CT underestimates diffuse axonal injury. Acute tissue tear hemorrhages, or shearing hemorrhages, occur in more severe instances or in up to 13% of diffuse axonal injuries and up to 8% of all head injuries [21]. With complete axonal disruptions and associated tissue tears, CT will demonstrate multiple 0.5 to 1.5 cm foci of high attenuation, representing hemorrhagic axonal injuries surrounded by rims of low attenuation edema (Fig. 9). Hemorrhagic foci are more easily detected on delayed than initial CT scans. Almost one-third of patients with normal CT findings after head trauma will have evidence of diffuse axonal injury on MR [25]. T2-weighted images are especially sensitive for detecting axonal injuries, which are hyperintense on T2-weighted images and iso- to hypointense on T1-weighted images. They are round or ovoid, with the long axis parallel to fiber bundle direction. Because of their greater sensitivity

196

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

Fig. 10. Diffuse axonal injury in a 20-year-old male. (A) T2-weighted spin-echo MR image demonstrates no abnormal signal. (B) T2*-weighted gradient-echo image demonstrates hypointensity at the right occipital gray-white junction, compatible with diffuse hemorrhagic axonal injury (arrows). (From Mittl RL Jr, Grossman RI, Hiehle JF, et al. Prevalence of MR evidence of diffuse axonal injury in patients with mild head injury and normal head CT findings. Am J Neuroradiol 1994;15:1583 9; with permission. D by the American Society of Neuroradiology [www.ajnr.org].)

Fig. 11. Diffuse axonal injury in a 52-year-old male. (A) Isotropic diffusion weighted image demonstrates hyperintense signal within the splenium, consistent with cellular edema. (B) Diffusion trace image demonstrates corresponding decreased ADC values. (From Liu AY, Maaldjian JA, Bagley LJ, et al. Traumatic brain injury: diffusion-weighted MR imaging findings. Am J Neuroradiol 1999;20:1636 41; with permission. D by the American Society of Neuroradiology [www.ajnr.org].)

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

197

for susceptibility effects, gradient echo sequences such as 2D fast low angle shot (FLASH) are more sensitive for detecting hemorrhagic shear injuries [25,26] (Fig. 10). Acute to subacute nonhemorrhagic axonal injuries have increased signal on diffusion-weighted images and corresponding areas of decreased signal on trace apparent diffusion coefficient maps, consistent with cellular swelling or cytotoxic edema [27] (Fig. 11). Secondary Wallerian degeneration will occur in areas of axonal disruption and lead to diffuse degeneration and atrophy over months or years. Deep cerebral gray matter and brainstem injury Traumatic injury to the thalamus, basal ganglia, and brainstem is relatively uncommon. Severe shearing forces associated with diffuse axonal injury are usually responsible for the disruption of small perforating vessels. CT may be normal or may demonstrate multiple hemorrhagic foci near the lentiform nucleus and external capsule. The basal cisterns are often obliterated after brainstem injury, due to diffuse cerebral edema. MR provides better evaluation for brainstem contusion and hemorrhage, since it is more sensitive for nonhemorrhagic lesions and less susceptible to posterior fossa artifacts. Frequently coexisting lesions include subarachnoid hemorrhage,

contusion, extra-axial hematoma, and intraventricular hemorrhage. Poor prognosis is reliably predicted by age greater than 60, low Glasgow coma scale score, abnormal pupil response, impaired oculocephalic response, and abnormal motor response to painful stimuli [13,28].

Traumatic extra-axial hermorrhage Subdural hematoma Subdural hematomas are seen in 10% to 20% of patients with head trauma. The mortality is between 50% and 85%. Acute subdural hematomas, with or without associated cerebral contusions, are the most commonly encountered operable intracranial hematoma [4]. Subdural collections, blood or cerebrospinal fluid, dissect into the potential space between the dura and arachnoid membranes. Subdural hematomas may cross suture lines, but not dural reflections such as the falx and tentorium. Subdural hematomas are usually caused by traumatic tears of cortical bridging veins between the temporal lobes and sphenoparietal or petrosal sinuses. Unlike epidural hematomas, subdural hematomas are not commonly associated with skull fractures in adults. Subdu-

Fig. 12. Subdural hematoma in a 15-month-old male with multiple healing long bone metaphyseal fractures after child abuse. (A) Axial CT image shows faint high density subdural blood along the right tentorium. (B) Coronal T1-weighted MR image shows bright signal in the right suboccipital region (arrows) consistent with a subacute subdural hematoma. (Courtesy of C. Hilfer, MD, New York, NY.)

198

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

Fig. 13. Acute subdural hematoma in an 81-year-old male. (A) Initial CT image demonstrates a left subdural hematoma causing midline shift. There is compression of the ipsilateral ventricle and dilatation of the contralateral ventricle and midline shift. (B) Follow-up CT image 15 days later reveals the left subdural collection to have significantly decreased in density, now similar to CSF attenuation.

Fig. 14. Bilateral subacute subdural hematomas in a 52-year-old male. (A) Noncontrast CT image demonstrates bilateral isodense subdural hematomas displacing the gray-white junction medially (arrows). There is sulcal effacement, but the two subdural hematomas balance each other and result in no midline shift. (B) Contrast enhanced CT image demonstrates enhancing cortical veins along the surface of the brain (arrows), medial to the bilateral subdural hematomas.

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

199

ral hematomas are common in cases of child abuse (Fig. 12). The typical appearance of an acute subdural hematoma on CT is a crescentic hyperdense collections between the cerebral hemisphere and inner table of the skull, extending from front to back around one hemisphere. Mass effect from the subdural hematoma will displace the gray-white junction away from the inner table and may cause midline shift, with compression of the ipsilateral lateral ventricle and dilatation of the contralateral ventricle (Fig. 13A). Bilateral subdural hematomas may have a balanced effect without midline shift despite significant mass effect (Fig. 14A). Small subdural hematomas may be missed because of high convexity locations, beam hardening artifacts, or narrow window settings. As noted above, wider subdural windows help to differentiate pixels of similar brightness, thus differentiating acute blood from bone (see Fig. 1). Acute subdural hematomas may uncommonly present as isodense or hypodense collections because of marked anemia and paucity of hemoglobin within the hematoma, clotting abnormalities, or cerebrospinal fluid dilution from associated arachnoid tears, similar in appearance to subacute hematomas. The right and left interhemispheric subdural spaces are separated by the attachment of the falx to the inferior sagittal sinus. Disruption of bridging veins feeding into the superior sagittal sinus causes blood to collect in the subdural space along one side of the falx cerebri. The medial border will be straight as it is bounded by the rigid falx, and the lateral border will be convex as it displaces brain away from the midline. The interhemispheric subdural hematoma is particularly common in abused children. Hematomas may collect in the potential subdural space along the tentorium. Supratentorial extra-axial collections are more commonly subdural hematomas than epidural hematomas, and infratentorial extraaxial collections are more commonly epidural hematomas than subdural hematomas. The subdural hematoma will appear as hazy high attenuation along the tentorium. Volume averaging and beam hardening artifacts may partially obscure the subtemporal hematoma. The predictable evolution of blood within unevacuated subdural hematomas produces typical changes on noncontrast CT scan. The subdural collection will gradually decrease in size and attenuation as the hemoglobin, platelet, or fibrin clot, and cellular components are resorbed. The attenuation of subacute subdural hematomas 4 to 20 days old will approach the attenuation of normal brain parenchyma. Isodense subdural hematomas are difficult to detect with non-

contrast CT, but should be suspected when the graywhite matter interface is displaced away from the ipsilateral inner table of skull and the sulci are effaced or difficult to trace to the brain surface (Fig. 15). Contrast CT will demonstrate enhancement of the inner membrane or cortical veins defining the surface of the brain (see Fig. 14B). Subacute subdural hematomas are more easily detected with MR imaging because of their high signal intensity on T1-weighted images [29] (see Fig. 12B). Chronic subdural hematomas more than two to three weeks old are typically lower in attenuation than brain on CT (see Fig. 13B). Acute on chronic subdural hematomas develop from rehemorrhage into previous subdural collections. Fluid-blood levels are seen when sedimented fresh blood gravitates dependently and the proteinaceous fluid layers on top. Subdural hygromas are cerebrospinal fluid collections that develop 6 to 30 days after traumatic arachnoid tears. On CT, subdural hygromas will have the same density as cerebrospinal fluid. Blood vessels crossing through the extra-axial collection will help differentiate subdural hygromas from chronic subdural hematomas. On T1-weighted MR images, subdural hygromas will have the same signal as cerebrospinal fluid, although slightly higher signal from increased

Fig. 15. Isodense subdural hematoma in a 70-year-old male. Left subdural hematoma (arrows) has the same density as normal brain parenchyma. Associated mass effect causes left-to-right midline shift.

200

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

Fig. 16. Acute epidural hematoma in a 16-year-old male. (A) Noncontrast CT image reveals a convex left frontal extra-axial hematoma. (B) The same CT image viewed with bone window demonstrates the associated frontal bone fracture (arrows), typical for an injury to the anterior division of the middle meningeal artery.

protein content may be observed on occasion. Since subdural hygromas lack the methemoglobin present in chronic subdural hematomas, they are hypointense, similar to cerebrospinal fluid (CSF), on T1-weighted images [30]. Epidural hematoma Only 1% to 4% of patients with head trauma have epidural hematomas. The overall mortality in these patients is 5%. More than 90% of epidural hematomas in adults are associated with fractures [31]. Fractures of the inner table may cause lacerations of the middle meningeal artery because the dural blood supply lies within the epidural space. Less commonly, venous bleeding can occur from disruption of a meningeal vein or a dural venous sinusthe transverse or sigmoid sinus in the posterior fossa and superior sagittal sinus in the parasagittal region. Blood dissects into the potential space between the inner table of the skull and periosteal dura. The periosteal dura has its strongest attachment at the sutures. Therefore, in contrast to subdural hematomas, epidural hematomas do not cross sutures and have a characteristic convex shape. They do, however, cross dural reflections such as the falx. Epidural hematomas classically occur unilaterally in the temporoparietal region. Ninety-five percent are supratentorial.

Epidural hematomas in the posterior fossa are rare, and have a higher morbidity and mortality. Acute epidural hematomas are typically seen on noncontrast CT as hyperdense lenticular (Fig. 16A) or biconvex (Fig. 17) extra-axial collections adjacent to a fracture. (see Fig. 16B). Areas of low attenuation within the hematoma represent mixing of unclotted blood and serum separating from clotted blood [31,32], and are suggestive of active arterial bleeding. (see Fig. 17). Hyperacute and chronic epidural hematomas demonstrate exhibit signal intensity similar to CSF on T1- and T2-weighted images. The acute epidural hematoma is isointense on T1-weighted images and iso- to hypointense on T2-weighted images. Subacute and early chronic epidural hematomas are hyperintense on T1- and T2-weighted images. Large acute epidural hematomas with significant mass effect may displace and compress adjacent brain, and lead to brainstem herniation and death. These cases are neurosurgical emergencies that require prompt evacuation. Small epidural hematomas that are asymptomatic, located along the convexities, are less than 1.5 cm in maximum width, and have minimal or no midline shift require no intervention, provided they do not increase in size [33,34]. Slow bleeds may increase pressure within the hematoma before blood has time to dissect further into the epidural space, leading to tamponade of the bleeding vessel. Depend-

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

201

Fig. 17. Acute epidural hematoma in a 41-year-old male. Noncontrast CT image demonstrates a biconvex left frontotemporal epidural hematoma. Epidural blood does not cross the coronal suture. Small areas of lower attenuation represent mixing of clotted and unclotted blood. No fracture was present, but the location is typical for an injury to the main branch of the middle meningeal artery.

the sylvian fissures, superior cerebellar cisterns, and sulci over the cerebral convexities (Fig. 18). Nonvisualization of the interpeduncular cistern may be a clue that a small amount of isodense subarachnoid blood is present. Small amounts of subarachnoid hemorrhage may not be visualized, since blood that occupies less than the full thickness of a single CT slice will be volume averaged with brain. The low hematocrit and low deoxyhemoglobin of acute subarachnoid blood gives signal similar to brain parenchyma on T1- and T2-weighted spin echo images [30]. Therefore, MR may underestimate or fail to detect acute subarachnoid hemorrhage. Fluid attenuated inversion recovery imaging (FLAIR) may detect small areas of acute or subacute subarachnoid hemorrhage, however, that are not detected by conventional MR images or CT scans, as hyperintense signal within the sylvian fissures or cerebral sulci [40] (Fig. 19). Hydrocephalus is the most common complication after subarachnoid hemorrhage. Acute obstructive hydrocephalus may develop within the first week from ependymitis or intraventricular blood obstructing the aqueduct of Sylvius or outlet of the fourth ventricle. Communicating hydrocephalus may develop within hours or after the first week, if the arachnoidal villi are blocked by the corpuscular elements of the blood, leading to fibroblastic proliferation in the subarachnoid space and blockage of the

ing upon the series, enlargement of epidural hematomas has been reported in 10% to 64.9% [35 37] of patients, usually within 48 hours after injury [35,37]. Therefore, serial neurological exams with low thresholds for repeat scanning are indicated to detect enlarging epidural hematomas [38]. Subarachnoid hemorrhage Subarachnoid hemorrhage occurs in up to 11% of traumatic brain injuries [39]. It is the result of injury to small bridging cortical vessels on the pia or arachnoidal leptomeninges crossing the subarachnoid space. The very young and very old are especially vulnerable to subarachnoid hemorrhage because of their relatively larger subarachnoid spaces. Alternatively, blood from an intracerebral hematoma may decompress directly into the subarachnoid space or dissect into the ventricular system. When occurring in conjunction with other forms of traumatic brain injury, subarachnoid hemorrhage is often focal, next to a contusion. CT will demonstrate high attenuation blood within the basal cisterns and subarachnoid spaces such as

Fig. 18. Acute subarachnoid hemorrhage in a 31-year-old female. Noncontrast CT image demonstrates high density blood filling the interpeduncular cistern (arrows).

202

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

Fig. 19. Subarachnoid hemorrhage in a 62-year-old male. (A) Noncontrast CT image demonstrates sulcal effacement on the left and questionable high attenuation in a few right parietal sulci. (B) FLAIR MR image clearly demonstrates subarachnoid hemorrhage within posterior sulci bilaterally (arrowheads).

pacchionian granulations. Blood is a direct irritant to vessels. Subarachnoid hemorrhage may induce cerebral vasospasm. The risk of severe ischemia and subsequent infarction peaks at 5 to 15 days. Vasospasm is more likely when subarachnoid hemorrhage is accompanied by subdural hematoma, intraventricular hemorrhage, cerebral contusion, or intracerebral hemorrhage [41]. Intraventricular hemorrhage Intraventricular hemorrhage occurs in 2.8% of all patients with blunt head trauma [42]. The incidence is higher, ranging from 9.5 22% [22,43], in those patients with Glasgow coma scores less than or equal to eight. Isolated intraventricular hemorrhage is relatively uncommon. Superficial contusions and subarachnoid hemorrhage are common associated injuries. Intraparenchymal hematomas may dissect along white matter tracts into the ventricular system. Hemorrhagic injuries to the corpus callosum and brain stem, often due to diffuse axonal injury, are associated with intraventricular hemorrhage and poor outcome [42]. LeRoux et al [42] report a series where intracranial pressure monitors were placed in 39 patients with intraventricular hemorrhage. Intracranial pressure rose in 46%, acute hydrocephalus developed in 7%, and ventricular drainage was required in 10% of the

patients [42]. Mortality in patients with intraventricular hemorrhage has been reported to range from 21% to 77% [43,44], although outcome is likely related more to severity of the primary brain injury than directly to intraventricular hemorrhage [43]. The intrinsic antithrombotic properties of fibrinolytic activators within CSF often cause intraventricular hemorrhage to layer within the ventricular system rather than clot [45]. Noncontrast CT will demonstrate a fluid-fluid layer with the high attenuation blood layering dependently within the ventricle and the lower attenuation, lower density supernatant, or CSF on top. On T2-weighted MR imaging, the layering of acute blood will appear isointense to hypointense inferiorly and CSF intensity superiorly. Acute blood will appear hyperintense on FLAIR sequences sensitive to T2 prolongation, while nulling the normal cerebrospinal fluid background. There is evidence that FLAIR and fast spin echo FLAIR (fast FLAIR) sequences more conspicuously show acute intraventricular hemorrhage during the first 48 hours than noncontrast CT [46]. T1- and T2-weighted images will show acute intraventricular hemorrhage as isointense or hypointense. Cerebrospinal fluid pulsation artifacts within the ventricles, particularly within the posterior fossa, may be confused for intraventricular hemorrhage [46].

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204

203

Summary Traumatic intracranial hemorrhage is a leading cause of morbidity and mortality in the United States. CT remains the primary imaging modality for initial evaluation of patients who have sustained head trauma. MR imaging, which has always been important for the evaluation of subacute and chronic head trauma, has been gaining popularity and recognition as an alternative primary imaging modality.

References
[1] Center for Disease Control and Prevention, National Center for Health Statistics. Epidemiology of traumatic brain injury in the United States. Atlanta 2000. p. 1 8. [2] Thurman DJ, Alverson C, Browne D, et al. Traumatic brain injury in the United States: a report to Congress. Atlanta: Centers for Disease Control and Prevention; 1999 p. 1 22. [3] U.S. Census Bureau. Statistical abstract of the United States: 2000. Vital Statistics 63, Section 2. [4] Gutman MB, Moulton RJ, Sullivan I, et al. Risk factors predicting operable intracranial hematomas in head injury. J Neurosurg 1992;77:9 14. [5] Gutman MB, Moulton RJ, Sullivan I, et al. Relative incidence of intracranial mass lesions and severe torso injury after accidental injury: implications for triage and management. J Trauma 1991;31:974 7. [6] Zwimpfer TJ, Brown J, Sullivan I, et al. Head injuries due to falls caused by seizures: a group at high risk for traumatic intracranial hematomas. J Neurosurg 1997; 86:433 7. [7] Li J, Brown J, Levine M. Mild head injury, anticoagulants, and risk of intracranial injury. Lancet 2001;357: 771 2. [8] Patel NY, Hoyt DB, Nakaji P, et al. Traumatic brain injury: patterns of failure of nonoperative management. J Trauma 2000;48(3):367 74. [9] Jones TR, Kaplan RT, Lane B, et al. Single- versus multi-detector row CT of the brain: quality assessment. Radiology 2001;219:750 5. [10] Paterakis K, Karantanas AH, Komnos A, et al. Outcome of patients with diffuse axonal injury: the significance and prognostic value of MRI in the acute phase. J Trauma 2000;49(6):1071 5. [11] Gentry LR, Godersky JC, Thompson B. MR imaging of head trauma: review of the distribution and radiopathologic features of traumatic lesions. Am J Roentgenol 1988;153(3):663 72. [12] Liang L, Korogi Y, Sugahara T, et al. Detection of intracranial hemorrhage with susceptibility-weighted MR sequences. Am J Neuroradiol 1999;20(8):1527 34. [13] Gentry LR. Imaging of closed head injury. Radiology 1994;191:1 17.

[14] Blankenberg FG, Norbash AM, Lane B, et al. Neonatal intracranial ischemia and hemorrhage: diagnosis with US, CT, and MR imaging. Radiology 1996;199: 253 9. [15] Hofman PA, Stapert SZ, van Kroonenburgh MJ, et al. MR imaging, single-photon emission CT, and neurocognitive performance after mild traumatic brain injury. Am J Neuroradiol 2001;22:441 9. [16] Ichise M, Chung DG, Wang P, et al. Technetium-99mHMPAO SPECT, CT and MRI in the evaluation of patients with chronic traumatic brain injury: a correlation with neuropsychological performance. J Nucl Med 1994;35:217 26. [17] Gurdjian ES, Gurdjian ES. Cerebral contusions: reevaluation of the mechanism of their development. J Trauma 1976;16:35 51. [18] New PF, Aronow S. Attenuation measurements of whole blood and blood fractions in computed tomography. Radiology 1976;121:635 40. [19] Atlas SW, Thulborn KR. MR detection of hyperacute parenchymal hemorrhage of the brain. Am J Neuroradiol 1998;19:1471 507. [20] Gomori JM, Grossman RI, Goldberg HI, et al. Intracranial hematomas: imaging by high-field MR. Radiology 1985;157:87 93. [21] Wilberger JE, Rothfus WE, Tabas J, et al. Acute tissue tear hemorrhages of the brain: computed tomography and clinicopathological correlations. Neurosurgery 1990;27:208 13. [22] Abraszko RA, Zurynski YA, Dorsch NW. The significance of traumatic intraventricular haemorrhage in severe head injury. Br J Neurosurg 1995;9(6): 769 73. [23] Gentry LR, Thompson B, Godersky JC. Trauma to the corpus callosum: MR features. Am J Neuroradiol 1988;9:1129 38. [24] Gentry LR, Godersky JC, Thompson B, et al. Prospective comparative study of intermediate-field MR and CT in the evaluation of closed head trauma. Am J Roentgenol 1988;150(3):673 82. [25] Mittl RL, Grossman RI, Hiele JF, et al. Prevalence of MR evidence of diffuse axonal injury in patients with mild head injury and normal head CT findings. Am J Neuroradiol 1994;15:1583 9. [26] Zimmerman RA. Craniocerebral Trauma. In: Lee SH, Rao KCVG, Zimmerman RA, editors. Cranial MRI and CT. 4th edition. New York: McGraw-Hill; 1999. p. 435 8. [27] Liu AY, Maldjian JA, Bagley LJ, et al. Traumatic brain injury: diffusion weighted MR imaging findings. Am J Neuroradiol 1999;20:1636 41. [28] Lee JP, Wang AD. Post-traumatic basal ganglia hemorrhage: analysis of 52 patients with emphasis on the final outcome. J Trauma 1991;31(3):376 80. [29] Ebisu T, Naruse S, Horikawa Y, et al. Nonacute subdural hematoma: fundamental interpretation of MR images based on biochemical and in vitro MR analysis. Radiology 1989;171:449 53. [30] Barkovich AJ, Atlas SW. Magnetic resonance imaging

204

R.J. Young, S. Destian / Neuroimag Clin N Am 12 (2002) 189204 of intracranial hemorrhage. Radiol Clin North Am 1988;26(4):801 20. Zimmerman RA, Bilaniuk LT. Computed tomographic staging of traumatic epidural bleeding. Radiology 1982; 144:809 12. Peterson OF, Espersen JO. How to distinguish between bleeding and coagulated extradural hematomas on CT scanning. Neuroradiology 1984;26:285 92. Hamilton M, Wallace C. Nonoperative management of acute epidural hematoma diagnosed by CT: the neuroradiologists role. Am J Neuroradiol 1992;13:853 9. Pozzati E, Tognetti F. Spontaneous healing of acute extradural hematomas: study of twenty-two cases. Neurosurgery 1986;18:696 700. Osborn AG, Davis WL, Jacobs J. Craniocerebral trauma. In: Osborn AG, editor. Diagnostic neuroradiology. St. Louis: Mosby Year-Book; 1994. p. 199 247. Sakai H, Takagi H, Ohtaka H, et al. Serial changes in acute extradural hematoma size and associated changes in level of consciousness and intracranial pressure. J Neurosurg 1988;68:566 70. Sullivan TP, Jarvik JG, Cohen WA. Follow-up of conservatively managed epidural hematomas: implications for timing of repeat CT. Am J Neuroradiol 1999;20: 107 13. Smith HK, Miller JD. The danger of an ultra-early computed tomographic scan in a patient with an evolving acute epidural hematoma. Neurosurgery 1991;29: 258 60. Greene KA, Marciano FF, Johnson BA, et al. Impact of traumatic subarachnoid hemorrhage on outcome in nonpenetrating head injury. Part I: a proposed computerized tomography grading scale. J Neurosurg 1995; 83:445 52. Singer MB, Atlas SW, Drayer BP. Subarachnoid space disease: diagnosis with fluid-attenuated inversionrecovery MR imaging and comparison with gadolinium-enhanced spin-echo MR imagingblinded reader study. Radiology 1998;208:417 22. Martin NA, Doberstein C, Alexander M, et al. Posttraumatic cerebral arterial spasm. J Neurotrauma 1995; 12(5):897 901. LeRoux PD, Haglund MM, Newell DW, et al. Intraventricular hemorrhage in blunt head trauma: an analysis of 43 cases. Neurosurgery 1992;31(4): 678 84. Fujitsu K, Kuwabara T, Muramoto M, et al. Traumatic intraventricular hemorrhage: report of twenty-six cases and consideration of the pathogenic mechanism. Neurosurgery 1988;23(4):423 30. Cordobes F, de la Fuenta M, Lobato RD, et al. Intraventricular hemorrhage in severe head injury. J Neurosurg 1983;58:217 22. Hayman LA, Pagani JJ, Kirkpatrick JB, et al. Pathophysiology of acute intracerebral and subarachnoid hemorrhage: applications to MR imaging. Am J Roentgenol 1989;153:135 9. Bakshi R, Kamran S, Kinkel PR, et al. Fluid-attenuated inversion-recovery MR imaging in acute and subacute cerebral intraventricular hemorrhage. Am J Neuroradiol 1999;20:629 36.

[31]

[40]

[32]

[33]

[41]

[34]

[42]

[35]

[43]

[36]

[44]

[37]

[45]

[38]

[46]

[39]

Neuroimag Clin N Am 12 (2002) 205 216

Diffuse axonal injuries: pathophysiology and imaging

Dima A. Hammoud, MD, Bruce A. Wasserman, MD *
Division of Neuroradiology, Johns Hopkins University School of Medicine, 600 North Wolfe Street, Phipps B-100, Baltimore, MD 21287, USA

Head trauma is a major public health problem in the United States, with a significant social and economic impact. The yearly incidence has been reported as high as 130 to 150 per 100,000 persons [1]. Head trauma can result in various types of primary injury occurring at the moment of impact, including lacerations, skull fractures, contusions, and diffuse axonal shear injury [2]. While direct kinetic energy transmission or penetrating injury result in local contusions that are regions of damage to the surface of the brain, rotational shear-strain deformation results in widespread bilateral damage involving both the deep and superficial structures, close to or remote from the site of impact [3]. This is referred to as diffuse axonal injury (DAI), or more accurately, axonal shear injury. DAI is a frequent imaging and pathologic finding of severe head injury, accounting for about 48% of primary traumatic brain injuries [4].

Pathophysiology of diffuse axonal injury Mechanism of injury in DAI The mechanism of DAI is thought to be shearstrain injury. The deceleration occurring in a linear or rotational manner and associated with vehicle accidents is usually responsible. As early as 1945, Holbourn [3] reasoned that as the skull rotates, the brain has to depend on its rigidity to avoid being left behind. However, some brain structures are left behind to a considerable degree, and anatomic dis-

* Corresponding author. E-mail address: bwasser@rad.jhu.edu (B.A. Wasserman).

tortion can occur [3]. Because of the relative fixation of certain parts of the brain to the rigid skull, the deep and superficial portions may not move at the same rate, and can even move in different directions. This will result in shear strain that manifests across the axons and results in axonal injury and rupture [5]. Since different brain parts have different consistencies depending on cell morphology and cell concentration, strains usually result at their junction [6]. In summary, shear-strain deformation develops upon exposure to rotational acceleration forces because of differential movements of one portion of the brain with respect to another [6]. Different tissue consistencies and different degrees of fixation of brain parts account for the mechanism and location of DAI injury. These shear-strain forces thus develop in special locations such as the white-gray matter junction, brain/pia-arachnoid interface, cerebrospinal fluid (CSF)/brain interface, and skull/dura interface. [6] The most common location of shear injuries is at the white-gray matter junction (Fig.1A) and in the deep centroaxial white matter such as the corpus callosum (see Fig.1B), periventricular and hippocampal areas, cerebral peduncles, brachium conjunctivum, superior colliculi, and deep reticular formation [7]. Gentry et al [8] studied 63 cases of acute head injury and 15 patients with chronic head injury. They detected corpus callosum injuries in 47% of the patients. The corpus callosum is prone to injury because of its rigid attachment to the falx and its relationship to the more mobile cerebral hemispheres. The bulky and independently mobile cerebral hemispheres are connected by the less mobile corpus callosum, and with lateral or oblique lateral movements of the head, the rigid falx prevents the cerebral hemispheres from moving across the midline. Shear strains develop across the connecting point of the two hemispheres,

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 1 1 - 4

206

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

207

at the corpus callosum. Less strain develops anteriorly, where the falx is shorter and can allow transient displacement of portions of the brain across the midline. Posteriorly, the falx is broader and effectively prevents this displacement, allowing greater shear and tensile strains to develop within the fibers of the corpus callosum [8]. Gentry et al [8] also found a significantly greater number of lesions in the septum pellucidum and fornix in patients with callosal injury, as compared with those without such injury. This is secondary to shear strains that develop at the junction of the corpus callosum and septum pellucidum and corpus callosum and fornix. The fornix and septum pellucidum are very delicate, tenuous structures that connect the relatively mobile corpus callosum to the relatively immobile diencephalon. When the massive cerebral hemispheres and attached corpus are displaced, shearing stresses are produced at the junction of the corpus callosum with the septum pellucidum and fornix, resulting in tears at the junction site. Such an injury is of clinical concern because this structure comprises projection and commissural fibers of the limbic system and hippocampal formation, and is related to memory and many other complex functions. The authors observed that the diffuse axonal lesions of the lobar white matter and brainstem are usually small to microscopic in size, while those of the corpus callosum typically are larger and more easily identified on computed tomography (CT) and magnetic resonance imaging (MRI). The importance of observing diffuse axonal lesions of the corpus callosum thus lies in the fact that they serve as easily visible markers of more widespread, but often less visible, diffuse axonal injury [8]. Intraventricular hemorrhage is common in injuries of the corpus callosum, and in shear strain injuries in general [8]. The hemorrhage is attributed to a shearing rupture of subependymal veins [5]. Brainstem injury is also commonly seen with shear-stress types of trauma. Brain stem injuries are usually divided into primary and secondary types. The primary type of brainstem injury is the shear type of injury, which tends to occur in the dorsolateral aspect of the upper brainstem and spare the ventral aspect. The lesions are usually edematous but with no associated hemorrhage, and are therefore difficult to visualize on CT scan. Hem-

orrhagic petechial lesions can also be seen in the periaqueductal region [4,5]. The secondary type of brain stem injury is not related to the basic mechanism of trauma but rather to the end result. In this case, brain swelling and brainstem compression associated with hippocampal herniation result in edematous or hemorrhagic changes in the anterolateral aspect of the midbrain [4,5]. The thalamus and hypothalamus are also involved in axonal shear injury, but with less frequency [5]. Such injuries are usually hemorrhagic because of the high vascularity of this region. Histopathology of DAI DAI is thought to result from mechanical forces that damage the integrity of the axon at the node of Ranvier, with subsequent alterations in axoplasmic flow [7]. Animal studies have shown that cytoskeletal perturbation at injured sites results in focal axonal swelling and accumulation of neurofilaments and organelles, with subsequent formation of what are referred to as retraction balls (Fig. 2) [9]. Later, increased numbers of microglia are seen in the areas of injury in the cerebral cortex and subsequently, degeneration of the involved fiber tracts occurs [7]. An alternative theory presented by Povlishock and Jenkins [10] suggests that the axonal damage is a delayed consequence of complex axolemmal or cytoskeletal collapse and impairment of axoplasmic transport, leading to axonal swelling and disconnection. The histopathology of DAI is characterized by the wide but often asymmetric distribution of axonal swellings that appear within hours of the injury and may persist for much longer. These are best demonstrated with silver impregnation techniques or with immunoperoxidase methods for axons [7]. Another method of identifying traumatic damage to axons is by detecting b APP-immunoreactive axons or bulbs, scattered or in groups in hemispheric white matter, corpus callosum, and internal capsule [11]. Clinical presentation The clinical presentation of DAI depends on the severity of injury. Both severe (Glascow coma scale GCS 3 to 7) and moderate (GCS 8 to 12) head injuries have been associated with DAI, with white matter

Fig. 1. (A,B) Gross pathology specimens of DAI showing the typical hemorrhagic pattern of shear injury at the splenium of the corpus callosum, the gray-white matter junction, and the deep centroaxial white matter. (Courtesy of J. Troncoso, MD, Baltimore, MD.) Fig. 2. Photomicrograph of DAI showing the typical axonal retraction balls of axonal shearing injury (Courtesy of J. Troncoso, MD, Baltimore, MD.)

208

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

shearing injury demonstrated neuropathologically in these patients [1,2]. Another subset of patients is the group with mild head injury. These patients usually present with a history of brief loss of consciousness and good neurologic status without subsequent deterioration [1]. However, one-third to one-half of mild head injury patients develop a postconcussion syndrome. They usually have subjective complaints including difficulty concentrating, memory problems, headache, or disequilibrium; and neuropsychologic testing reveals deficits in information processing. Mittl et al [1] assessed twenty subjects with closed head trauma, brief loss of consciousness, and normal CT scan of the brain on presentation, and demonstrated MR abnormalities compatible with DAI in 30% of them. The lesions in mildly injured patients were not as widely distributed as in moderate to severe head injury patients, and occurred solely in subcortical white matter. No lesions were located in the corpus callosum or brainstem, common locations in more severely injured patients [1]. The shear-strain pattern of injury seen in infants less than 12 months old is different from the pattern seen in children and adults. Geddes et al [12] found that the predominant neurohistologic abnormality in infants is due to hypoxia and not diffuse axonal injury. This is partly attributed to the smooth skull of infants and soft consistency of the incompletely myelinated brain. Axonal damage was found to be mainly limited to the craniocervical junction. This is usually associated with hyperextension injury and the secondary stretch to the neuraxis. Infants have been shown to be susceptible to high cervical cord injury without radiologic evidence of bony injury, as a result of which they may suffer apnea and cardiorespiratory arrest, or severe hypotension [12,13]. The hypoxic damage leads to severe brain swelling, which leads to death [12,13]. Prognostic factors Prediction of the course of recovery and outcome in patients with DAI has been thoroughly investigated in relation to specific severity variables, including the Glascow coma scale on admission, duration of post-traumatic amnesia (PTA), age of the patient, and CT findings. In a study by Katz and Alexander [14], outcome of patients with DAI was found to be directly proportional to severity variables, the most sensitive being the duration of PTA. Patients with PTA of less than 12 weeks had a much more favorable recovery after rehabilitation than patients with PTA for more than 12 weeks. Age was found to

have a complex effect on recovery. Age effect was not found to contribute to the duration of coma but rather to the duration of PTA for any given depth or duration of coma. The age effect was apparent in patients older than 40 and very significant in patients older than 60 [14]. Imaging may play an important role in assessing prognosis, though this may rely on advanced techniques. CT has been proven of low prognostic value in patients with DAI [1,4,15]. MRI, although more sensitive than CT, was not found to identify a constant relationship between the pattern of neuropsychologic impairment and the site of injury [16]. The role of functional imaging, quantitative imaging, and PET scanning is presently being investigated in the subacute and chronic stages of DAI to better assess the posttraumatic cognitive outcome. These methods will be discussed in further detail in the section on imaging of DAI.

Imaging of DAI MRI and CT scan In the setting of acute closed head injury, an adequate diagnostic imaging study should fulfill several criteria. It should be available in or near the emergency room and should be safely and easily

Fig. 3. Axial nonenhanced CT (NECT) scan in a young adult involved in a motor vehicle accident showing multiple hemorrhagic foci of DAI at the gray-white matter junction (arrows).

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

209

performed on a critical patient. Its sensitivity must be sufficient to determine the severity, type, and anatomic location of injury. It should be specific enough to determine the category of injury (eg, surgical versus nonsurgical) in order to guide therapy accordingly. The introduction of CT revolutionized the assessment of patients suffering from acute head injury. It is fast, easy to perform, usually adequate to determine the need for surgery such as for massive extra axial or intraparenchymal hematomas, and is the diagnostic modality of choice for assessment of bony injuries. CT findings in diffuse axonal injury or axonal shear injury include discrete foci of hypodensity in the brain. They are typically less than 1 cm in size and spare the adjacent cortical surface of the brain [15]. Lesions are usually bilateral and located entirely within the white matter or at the gray-white matter interface (Fig. 3) [15]. The vast majority are nonhemorrhagic and can involve the corpus callosum (Fig. 4), internal capsule, posterolateral aspect of the brainstem, cerebellar hemispheres (Fig. 5), and sometimes the basal ganglia and thalami. Bilateral cerebral swelling with compression of the ventricular system can be an early finding. Within two to three weeks,

Fig. 5. Axial NECT scan of the brain in a 10-year-old patient with closed head trauma and coma, showing a small hemorrhagic focus in the anterior aspect of the body of the corpus callosum (arrow). Note mild surrounding hypodensity compatible with edema.

Fig. 4. Axial NECT scan of patient with accelerationdeceleration injury and coma showing a hemorrhagic focus with surrounding edema in the right cerebellar hemisphere, at the gray-white matter junction, compatible with axonal shear injury (arrow).

atrophic enlargement of the ventricles occurs, with well defined foci of hypodensity seen in the white matter (Fig. 6) [17]. Hemorrhagic shear injury lesions are more easily identified on CT in the acute stage, except in the brain stem and posterior fossa, due to beam hardening artifacts from bone. However, CT remains limited. Many patients were found to have discrepancy between a normal CT scan of the brain and bad neurologic status [17]. As early as 1978, Zimmerman et al [17] noted a marked discrepancy between the CT findings and the clinical status of patients presenting after head trauma. This has led to thorough comparison of CT and MRI in the setting of DAI. In 1987, Gentry et al [15] demonstrated marked difference in the sensitivity of CT and MRI in the detection of nonhemorrhagic axonal shear injury lesions. CT was very insensitive, detecting only 19% of lesions, while T1-weighted images and T2-weighted images detected 72.3% and 92.4% of the lesions respectively. In 1994, Mittl et al [1] demonstrated that abnormalities compatible with diffuse axonal injury were seen on MRI (spin-echo T2-weighted and T2*-weighted gradient echo images) in 30% of patients with mild head injury and normal head CT findings. It is now recommended that all patients with moderate to severe head injury get an

210

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

Fig. 6. (A) An 8-year-old female who sustained a severe acceleration-deceleration injury with multiple foci of diffuse axonal injury in the frontal lobes bilaterally, seen in the right frontal white matter on this section (arrows). (B) Same patient after 16 months, showing encephalomalacic changes and marked loss of volume for age, most prominent in the frontal lobes.

MRI examination within the first two weeks of injury for better evaluation of the full extent of brain involvement, which is not feasible by CT only [6]. MRI findings in DAI The conspicuity of a lesion on MRI depends on many factors, including its location, size, presence of hemorrhage, and most important, the pulse sequence

used [6]. Long TR/long TE scans are among the most sensitive for detection of DAI lesions (Fig. 7). T2-weighted hyperintensities are usually secondary to infiltration by inflammatory cells, demyelination, serum extravasation, or post-traumatic secondary ischemic damage, depending on the time after injury of the study [18]. However, this sequence is limited whenever the lesions are in paraventricular or cortical locations.

Fig. 7. (A) Sagittal T1-weighted image in a young patient who sustained a severe closed head injury, showing a faint low signal intensity in the splenium of the corpus callosum with slight expansion (black arrow). (B) Axial T2-weighted and (C) FLAIR images of the same patient, obtained at the same time, showing enlargement and increased signal intensity of the splenium (white arrowheads). Long TR/long TE sequences are better than short TR/short TE sequences in detection of diffuse axonal injury.

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

211

Fluid attenuated inversion recovery (FLAIR) With the advent of FLAIR imaging, DAI lesion detection improved significantly [19]. FLAIR is an inversion recovery sequence designed to null or greatly reduce signal from CSF, thus enabling T2 weighting without the interference of the high signal of CSF. Ashikaga et al demonstrated that FLAIR is superior to T2-weighted spin echo images in 50% of cases of traumatic closed head injury (Fig. 8) [19]. T2*-weighted gradient recall echo (GRE) Another helpful sequence is the T2*-weighted gradient recall echo (GRE), which is highly sensitive to small hemorrhagic lesions. This is due to the shortening of T2* by the heterogeneous local magnetic field arising from paramagnetic blood breakdown products. These changes remain conspicuous months to years after the injury, unlike FLAIR and FSE T2W abnormalities, which may become indiscernible (Fig. 9) [18]. T2*-weighted images are thus essential in the evaluation of patients with remote head injury. In a study conducted by Yanagawa et al [18], the number of lesions per patient detected by GRE imaging was significantly higher than that found by T2-FSE ( p < 0.001). They found that lesions on GRE correlated positively with both the duration of unconsciousness and with GCS, whereas those on T2-FSE did not show any significant correlation. Limitations of this sequence include artifacts from the frontal or temporal bones. Furthermore, small (less than 3 mm) nonhemorrhagic lesions are not as well seen as on T2-weighted FSE images [18].

Fig. 9. (A) Axial gradient recall echo images of the brain in a patient with a remote history of closed head trauma showing an area of hemosiderin deposition in the posterior aspect of the pons on the left side (small white arrow), corresponding to an area of axonal shear injury. Other hypointense foci of hemosiderin from axonal injury are identified in the right cerebellar hemisphere (white arrowhead) and (B) in the temporal (small black arrow) and occipitotemporal (black arrowhead) lobes on the left side.

Fig. 8. (A) Axial T2-weighted image of the brain of the same patient as in Fig. 7, showing a faint focus of increased signal intensity (black arrow) at the gray-white matter junction in the right frontal lobe, compatible with shear injury associated with severe closed head trauma. (B) FLAIR image at the same level shows the same focus as well as an additional focus of hyperintensity at the gray-white matter junction in the right parietal region (white arrows). This latter is barely visualized on the T2-weighted image due to high signal intensity of adjacent CSF spaces.

Diffusion weighted imaging (DWI) Animal studies of apparent diffusion coefficient (ADC) changes due to DAI have shown conflicting results [20 23]. While some investigators observed an increase in ADC values in experimental models of traumatic head injury [21], others observed a decrease [20], or an original increase followed by a progressive decline [22]. In the first human study, Liu et al [24] performed DWI of DAI in the acute and subacute stages, using echo-planar imaging in three orthogonal directions. They showed that significant decreases in ADC can be seen in regions of DAI up to 18 days after injury (Fig. 10). This probably reflects cellular swelling or cytotoxic edema in the acute setting. The

212

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

Fig. 10. (A) Axial T2-weighted image acquired in the setting of an acute head trauma showing a faint hyperintensity in the splenium of the corpus callosum (small white arrow). (B) DWI at the same level showing corresponding high signal intensity (small white arrowhead). Restricted diffusion is commonly seen in diffuse axonal injury.

authors hypothesize that very low ADC values, beyond those typically seen in acute ischemia, might be due to the presence of blood products and ruptured

axons with membrane fragmentation, thus increasing the restriction to free movement of water molecules and producing low ADC values for longer periods of

Fig. 11. (A,B) Axial diffusion weighted image of the same patient as in Fig. 7, showing high signal intensity in the splenium of the corpus callosum at the same location of T2 hyperintensity, with corresponding low signal intensity on ADC map (white arrowheads) compatible with restricted diffusion.

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

213

time (Fig. 11) [24]. The physiology remains unclear and more studies are needed. Magnetization transfer imaging Magnetization transfer (MT) imaging exploits the principle that protons bound in macromolecular structures exhibit T1 relaxation coupling with protons in the aqueous phase. An off-resonance saturation pulse can be applied to selectively saturate bound protons. Subsequent exchange of longitudinal magnetization with the free water protons leads to a reduction in signal intensity detected from these free protons. The MT ratio (defined by the equation MTR = [MMs]/ M0 100%, where M0 is the average signal intensity in the absence of the saturation pulse and Ms is the average signal intensity in its presence) provides a quantitative index of this MT effect, and may be viewed as a quantitative measure of the structural integrity of tissues [25]. MT ratio (MTR) imaging in experimental models of DAI in pigs showed reduced MTR values in regions of histologically proven axonal injury [26]. When applied to 28 patients with closed head injury and loss of consciousness, Bagley et al [26] demonstrated that average MTR values were higher in all locations (areas of abnormal T2-weighted hyperintensity as well as normal appearing white matter) in patients without persistent neurologic deficit than in patients with such deficits. They concluded that detection of abnormal MTR in normal appearing white matter may predict a poor patient outcome. The presence of normal MTR in these areas does not necessarily confer a good outcome. MR spectroscopy Although MRI and CT are valuable tools for assessment of brain injury after closed head trauma, they remain insensitive to the detection of microscopic injury seen at histology [27]. MR Spectroscopy (MRS) offers a noninvasive, nonradioactive method of assessing microscopic injury and predicting outcome. The balance of brain chemicals such as NAA (N-acetyl aspartase), lactate, choline, and high energy phosphates is perturbed in DAI. Clinical studies have shown that diffusely elevated lactate levels, even in normal appearing tissues on MRI, correlate with poor clinical outcome [28]. Increased lactate confined to areas of injury does not, however, correlate with a poor outcome [27,28]. Reduced NAA levels, reflective of neuronal injury or metabolic depression, were observed in patients with elevated lactate up to 24 hours after injury. In the subacute phase, there is marked loss of NAA or NAA/cre (NAA/creatine ratio) within the white matter. However, in some patients

normal levels of NAA were demonstrated six months after the injury, suggesting a partially reversible process [27]. In patients with poor clinical outcome, elevated choline levels were observed at three and six months after injury, suggestive of continuing inflammation [29]. Choline levels tend to normalize at six months in patients with more favorable outcome [27]. Combining MRI information with MRS mapping of cellular integrity and metabolism may improve the prediction of the clinical outcome of patients and suggest rehabilitation strategies. More studies are needed to determine whether MRS is useful for monitoring of the cellular response to therapeutic interventions [27]. Quantitative MRI The association between the neurologic outcome of patients with DAI in the subacute and chronic stages and morphometric measurements of the brain has been investigated. Earlier studies correlated the volumes of different intracranial structures to postinjury neurologic outcome. Traumatic brain injury was found to result in significant hippocampal atrophy and temporal horn enlargement (Fig. 12) [30]. In children, the growth of the corpus callosum is reduced after severe head injury [31]. Gale et al [32] were able to demonstrate a significantly larger ventricle-to-brain ratio and temporal horn volumes in patients with DAI compared to normal controls, as well as significantly smaller fornix-tobrain ratios and corpus callosum area measurements. In a more recent study, the authors further divided DAI patients into two groups according to their performance on post injury IQ testing (patients with IQ < 90 and patients with IQ > 90) [33]. Using complex volumetric measurements of the MRI scans acquired at least 90 days after the insult, they were able to demonstrate that the third ventricular size and temporal horn size are the most sensitive determinants of neurologic outcome after closed head injury. The hippocampal volume was found to be less sensitive [33]. Functional imaging in DAI Functional neuroimaging techniques directly assess regional cerebral flow with techniques such as SPECT (Single photon emission computed tomography), or regional cerebral metabolism with the use of PET (positron emission tomography) and functional MRI. This allows insight into cerebral dysfunctions that occur after traumatic brain injury, even in regions that appear intact on MR [34].

214 D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216 Fig. 12. (A) CT scan of the brain in an 8-year-old boy who was involved in a motor vehicle accident, showing a small hemorrhagic focus in the left parietal lobe at the gray-white matter junction. (B) Proton-density MR image of the same patient showing multiple foci of hyperintensity at the gray-white matter junction of the left frontal lobe, typical of DAI. (C) Same patient, 5 years later, with prominent diffuse volume loss.

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216

215

SPECT imaging has been investigated by Prayer et al [35] as a tool for assessment of clinical outcome of patients after traumatic brain injury, as defined by the Glascow outcome scale (GOS). MRI provided detailed anatomic information and was more sensitive in demonstrating WM lesions, while SPECT offered an insight into the functional status of different brain structures and was clearly superior in the assessment of cortical and deep gray matter alterations [35]. They concluded that the synthesis of both morphologic and functional impairment available through MRI and SPECT is important for prognosis [35]. SPECT was also used for follow-up of patients in rehabilitation programs [36]. Those patients showed improvements in neuropsychologic test scores and in relative blood flow at the end of rehabilitation when compared with pretreatment values [36]. PET imaging can assess regional cerebral metabolism with higher resolution than SPECT. PET was used to explore the relationship between regional cerebral glucose metabolism and neurobehavioral status after traumatic brain injury [37]. Fontaine et al discovered a close link between decreased metabolism in the prefrontal and cingulate cortex and cognitive and behavioral disorders in patients with DAI [37]. Brain activation in response to different tasks such as memory and planning after traumatic brain injury has been studied by functional MRI in both animals and humans [38 40]. It was shown that in response to motor or sensory tasks, traumatic brain injury patients showed activation of areas adjacent to the damaged brain as well as areas in the contralateral sensorimotor cortex, when compared with control subjects (healthy monozygotic twins of patients) [39,40]. The studies suggested that behavioral recovery after traumatic brain injury can be accounted for by a process of reorganization and neuromodulation of the brain [40]. However, further research is needed to determine the clinical usefulness of these modalities and their potential for prediction of outcome.

such as MRS show preliminary evidence of some utility in determining outcome. References
[1] Mittl RL, Grossman RI, Hiehle JF, et al. Prevalence of MR evidence of diffuse axonal injury in patients with mild head injury and normal CT findings. AJNR 1994;15:1583 9. [2] Goetz CG, Pappert EJ, Schmitt B, editors. Acute head injury. Textbook of clinical neurology. 1st edition. Philadelphia: WB Saunders Co.; 1999. p. 1035 6. [3] Holbourn AHS. The mechanics of brain injuries. Br Med Bull 1945;3:147 9. [4] Gentry LR, Godersky JC, Thompson B. MR imaging of head trauma: review of the distribution and radiopathologic features of traumatic injury. AJNR 1988;9: 101 2. [5] Taveras J, Pile-Spellman J. Shear-strain or axonal injury. In: Taveras J, editor. Neuroradiology. 3rd edition. Lippincott, Williams & Wilkins; 1996. p. 353 4. [6] Gentry LR. Head trauma. In: Scott Altas, editor. Magnetic resonance imaging of the brain and spine. 3rd edition. Lippincott, Williams and Wilkins; 2001. p. 1069 57. [7] Cotran R, Kumar V, Collins T, editors. Trauma to the brain. In: Robbins pathologic basis of disease. 6th edition. WB Saunders, 1999. p. 1301 4. [8] Gentry LR, Thompson B, Godersky JC. Trauma to the corpus callosum: MR features. AJNR 1988;9: 1129 38. [9] Nakayama Y, Aoki Y. Mechanism responsible for the formation of focal swellings on injured neuronal processes using a novel in vitro model of axonal injury. Forensic Sci Int 2000;113:245 9. [10] Povlishock JT, Jenkins LW. Are the pathological changes evoked by traumatic brain injury immediate and irreversible? Brain Pathol 1995;5(4):415 26. [11] Geddes JF, Hackshaw AK, Vowles GH, et al. Neuropathology of inflicted head injury in children. I. Patterns of brain damage. Brain 2001;124(7):1290 8. [12] Geddes JF, Vowles GH, Hackshaw AK, et al. Neuropathology of inflicted head injury in children. II. Microscopic brain injury in infants. Brain 2001;124(7): 1299 306. [13] Graham DI. Pediatric head injury. Brain 2001;124(7): 1261 2. [14] Katz DI, Alexander MP. Traumatic brain injury; predicting course of recovery and outcome for patients admitted to rehabilitation. Arch Neurol 1994;51:661 70. [15] Gentry LR, Godersky JC, Thompson B, et al. Prospective comparative study of intermediate field MR and CT in the evaluation of closed head trauma. AJNR 1988;9:91 100. [16] Levin HS, Williams DH, Eisenberg HM, et al. Serial MRI and neurobehavioural findings after mild to moderate closed head injury. J Neurol Neurosurg Psychiatry 1992;55:255 62.

Summary Diffuse axonal shear injury is a common traumatic brain injury, with significant neurologic and behavioral impact on patients. Radiologic recognition of this entity and understanding of its sequelae can be of utmost importance in the prediction of outcome and planning for rehabilitation. MRI has proven to be the optimal means of detection and characterization of DAI lesions, with GRE and FLAIR sequences being particularly helpful, and more advanced techniques

216

D.A. Hammoud, B.A. Wasserman / Neuroimag Clin N Am 12 (2002) 205216 injury: a quantitative proton magnetic resonance study. J Neurotrauma 2000;17:629 40. Bigler ED, Blatter DD, Anderson CV, et al. Hippocampal volume in normal aging and traumatic brain injury. AJNR 1997;18:11 23. Levin HS, Benavidez DA, Verger-Maestre K, et al. Reduction of corpus callosum growth after severe traumatic brain injury in children. Neurology 2000;544(3): 647 53. Gale SD, Johnson SC, Bigler ED, et al. Nonspecific white matter degeneration following traumatic brain injury. J Int Neuropsychol Soc 1995;1(1):17 28. Bigler ED, Johnson SC, Blatter DD. Head trauma and intellectual status: relation to quantitative magnetic resonance imaging findings. Appl Neuropsychol 1999; 6(4):217 25. Azouvi P. Neuroimaging correlates of cognitive and functional outcome after traumatic brain injury. Curr Opin Neurol 2001;3:665 9. Prayer L, Winberger D, Oder W, et al. Cranial MR imaging and cerebral 99m Tc HM-PAO-SPECT in patients with subacute or chronic severe closed head injury and normal CT examinations. Acta Radiol 1993; 34:593 9. Laatsch L, Pavel D, Jobe T, et al. Incorporation of SPECT in a longitudinal cognitive rehabilitation therapy programme. Brain Inj 1999;13:555 70. Fontaine A, Azouvi P, Remy P, et al. Functional anatomy of neuropsychological deficits after severe traumatic brain injury. Neurology 1999;53(9):1963 8. McAllister TW, Saykin AJ, Flashman LA, et al. Brain activation during working memory 1 month after mild traumatic brain injury: a functional MRI study. Neurology 1999;53(6):1300 8. Cioni G, Montanaro D, Tosetti M, et al. Reorganisation of the sensorimotor cortex after early focal brain lesion: a functional MRI study in monozygotic twins. Neuroreport 2001;12(7):1335 40. Abo M, Chen Z, Lai LJ, et al. Functional recovery after brain lesioncontralateral neuromodulation: an fMRI study. Neuroreport 2001;12(7):1543 7.

[17] Zimmerman RA, Bilaniuk LT, Genneralli T. Computed tomography of shearing injuries of the cerebral white matter. Radiology 1978;127:393 6. [18] Yanagawa Y, Tsushima Y, Tokumaro A, et al. A quantitative analysis of head injury using T2-weighted gradient-echo imaging. J Trauma 2000;49:272 7. [19] Ashikaga R, Araki Y, Ishida O. MRI of head injury using FLAIR. Neuroradiology 1997;39:239 42. [20] Alsop D, Murai H, Detre J, et al. Detection of acute pathologic changes following experimental traumatic brain injury using diffusion-weighted magnetic resonance imaging. J Neurotrauma 1996;13:515 21. [21] Hanstock C, Faden A, Bendall M, et al. diffusion weighted imaging differentiates ischemic tissue from traumatized tissue. Stroke 1994;25:843 8. [22] Barzo P, Marmarou A, Fatouros P, et al. Contribution of vasogenic and cellular edema to traumatic brain swelling measured by diffusion-weighted imaging. J Neurosurg 1997;87(6):900 7. [23] Ito J, Marmarou A, Barzo P, et al. Charactetrization of edema by diffusion-weighted imaging in experimental traumatic brain injury. J Neurosurg 1996;84:97 103. [24] Liu AY, Maldjian JA, Bagley LJ, et al. Traumatic brain injury: diffusion weighted MR imaging findings. AJNR 1999;20:1636 41. [25] Bagley LJ, Mc Gowan JC, Grossman RI, et al. Magnetization transfer imaging of traumatic brain injury. J Magn Reson Imaging 2000;11:1 8. [26] McGowan JC, McCormack TM, Grossman RI, et al. Diffuse axonal pathology detected with magnetization transfer imaging following brain injury in the pig. Magn Reson Med 1999;41:727 33. [27] Brooks WM, Friedman SD, Gasparovic C. Magnetic resonance spectroscopy in traumatic brain injury. J Head Trauma Rehabil 2001;16(2):149 64. [28] Condon B, Oluoch-Olunya D, Hadley D, Teasdale F, et al. Early 1H magnetic resonance spectroscopy of acute head injury: four cases. J Neurotrauma 1998;15: 563 71. [29] Brooks WM, Stidley CA, Petropoulos H, et al. Metabolic and cognitive response to human traumatic brain

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

Neuroimag Clin N Am 12 (2002) 217 228

Imaging of acquired cerebral herniations

Philip L. Johnson, MD a,*, Donald A. Eckard, MD a, David P. Chason, MD b, Marci A. Brecheisen, MD a, Solomon Batnitzky, MD a
a

Department of Radiology, Kansas University Medical Center, 3901 Rainbow Boulevard, Kansas City, KS 66160, USA b Department of Radiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390, USA

The cranium provides protection for the brain against external forces. The brain is suspended in the cerebrospinal fluid (CSF) and supported by the skull and dural septa, which divide the cranial cavity into compartments and prevent excessive movement of the brain [1]. The unyielding cranium is a closed system and hence there is a fixed volume for the brain, CSF, and blood vessels and very little room for expansion in the face of increased intracranial volume and pressure. The brain is essentially incompressible (its shape can be altered, but its volume is basically constant); therefore, the vascular and CSF spaces must compensate for alterations in intracranial volume because of mass lesions or cerebral edema [2,3]. Unfortunately, this capacity is very limited, and when it is exceeded, cerebral displacement and herniation result. Cerebral herniation is the displacement of brain tissue from one compartment to the other. Of course, there are numerous variables that influence the degree of herniation and its untoward neurologic effects, such as the location of the mass, the volume of the mass, the intracranial pressure, the rate at which alterations in pressure and volume develop, brain atrophy, and individual anatomical variations. There are several different types of cerebral herniation and these are defined by the anatomic boundary through which the herniating tissue traverses (Fig. 1) [1,4]. Subfalcine, transtentorial, and tonsillar herniation are well known. Transtentorial herniation can be further classified as central (ascending and descending) and lateral (anterior and posterior).

Transphenoidal and external hernias are less wellknown patterns of herniation. A thorough understanding of the anatomy of the brain, skull, and dural septa is imperative to understanding cerebral displacement and herniation [5]. Recognition of cerebral displacement and herniation is critical, because the clinical consequences are devastating. Cerebral herniation causes compression of the brain, cranial nerves, and blood vessels and results in serious neurologic morbidity and even death.

Anatomy The cranial cavity is incompletely divided into fossae by bony ridges and further divided by dural septa (Fig. 2). The base of the skull provides support for the brain and consists of three fossae: anterior, middle and posterior, which are separated by bony ridges. The foramen magnum is a large orifice in the floor of the posterior fossa that allows passage of the spinal cord, nerves and blood vessels. Although there are numerous skull base foramina, the foramen magnum is the only opening of any sig-nificant size. The dural septa represent reflections of the dura mater that further divide the intracranial cavity. Dura is the Latin word for hard, an accurate description of this tough, fibrous membrane. The two main dural reflections are the falx cerebrei, which separates the cerebral hemispheres, and the tentorium cerebelli, which separates the posterior fossa from the cerebral hemispheres defining the supratentorial and infratentorial compartments. The falx cerebelli is a smaller dural reflection, which partially separates the cerebellar hemispheres.

* Corresponding author. E-mail address: pjohnson@kumc.edu (P.L. Johnson).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 8 - 4

218

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

Fig. 1. Patterns of cerebral herniation. (A) Coronal, (B) axial, and (C) sagittal diagrams demonstrate the different types of cerebral herniation: (1) subfalcine, (2) uncal, (3) descending transtentorial, (4) tonsillar, (5) ascending transtentorial, and (6) external herniation.

The falx cerebrei is a sickle-shaped septum that attaches anteriorly to the crista galli, and posteriorly to the internal occipital protuberance and tentorium cerebelli. The anterior part of the falx cerebrei has a relatively narrow sagittal width, whereas the posterior part of the falx cerebrei has a broad sagittal width where it attaches to the tentorium cerebelli and contains the straight sinus [6,7]. Superiorly, the falx cerebrei attaches to the inner

table of the skull in the midline and contains the superior sagittal sinus, whereas the inferior free edge of the falx cerebrei contains the inferior sagittal sinus [6]. The corpus callosum is located immediately inferior to the free edge of the falx and the distance between the two decreases posteriorly. The cingulate gyrus is just lateral to the inferior aspect of the falx. The pericallosal artery, a branch of the anterior cerebral artery, runs in the

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

219

Subfalcine herniation Subfalcine herniation, also known as midline shift or cingulate herniation, is the most common type of cerebral herniation and is easily recognized on CT and MR imaging. It is caused by unilateral frontal, parietal, or temporal lobe mass effect or edema [1]. When subfalcine herniation occurs, the anterior falx, although rigid, will tilt away from the mass effect. The posterior falx is wider, more rigid, and more resistant to displacement. This explains why subfalcine herniations occur anteriorly. As the mass effect increases, the ipsilateral cingulate gyrus will be dis-

Fig. 2. Dural reflections. Diagram of the cranial dural reflections and tentorial incisura.

pericallosal sulcus between the corpus callosum and cingulate gyrus [6]. The tentorium cerebelli is a transverse dural septum that separates the posterior fossa from the cerebral hemispheres. The tentorium, as its name suggests, is shaped like a tent, with its apex suspended in the midline by the posterior inferior falx cerebrei, and the base attached to the inner surface of the occipital bone and upper border of the petrous temporal bone [7]. The transverse sinuses and superior petrosal sinuses are contained within these basal dural attachments. The opening of the tentorium is the tentorial incisura, also known as the hiatus or notch (see Fig. 2). The incisura has a semioval shape and is attached at its base to the clinoid processes and at its apex to the falx cerebrei [6,7]. Otherwise, the margins of the incisura are free, though quite rigid. The midbrain and cerebral peduncles pass through the incisura. The anatomy of the tentorial incisura and adjacent structures is complex. The uncus and hippocampus of the medial temporal lobe slightly overhang the incisura. The posterior cerebral arteries, anterior choroidal arteries, and internal cerebral veins pass around the midbrain in the ambient cistern in close proximity to the free edge of the tentorium. The oculomotor nerve exits the midbrain and crosses the interpeduncular cistern between the posterior cerebral and superior cerebellar arteries, and courses just medial to the uncus on its way to the cavernous sinus. Both computed tomography (CT) and magnetic resonance (MR) imaging demonstrate the anatomy of the cranial cavity and enable diagnosis of cerebral herniation, although MR imaging provides superior anatomic detail (Fig. 3).

Fig. 3. Normal anatomy. (A) Coronal T2-weighted MR image demonstrates the falx cerebrei (large black arrow), the cingulate gyrus (outlined arrow), corpus callosum (circle) and tentorium cerbelli (small black arrows). (B) Axial T1-weighted MR image demonstrates the uncus (outlined arrow) and hippocampal gyrus (black arrow).

220

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

Transtentorial herniation Transtentorial herniation occurs when brain tissue is displaced into the tentorial notch, and is perhaps the most important herniation phenomenon in terms of frequency and clinical consequences. Transtentorial herniation is a complex process that can be loosely classified into overlapping patterns. Lateral transtentorial herniation can be anterior (uncal) or posterior (parahippocampal); whereas central transtentorial herniation can be either ascending or descending. Only a few millimeters of space are present between the midbrain and rigid tentorial edge, and the oculomotor nerve and posterior cerebral artery are in close proximity [5,13]. Consequently, very little displacement can be tolerated, and a shift of a few millimeters will result in compromise of these vital structures, with serious neurologic sequelae.

Fig. 4. Subfalcine herniation. Coronal enhanced T1weighted MR image demonstrates a large enhancing mass (large arrow) within the left frontal lobe causing subfalcine herniation. The cingulate gyrus (outline arrow) is displaced beneath the falx cerebrei (small black arrow) and there is downward displacement of the corpus callosum (circle).

Uncal herniation Uncal herniation represents anterior lateral transtentorial herniation. It is perhaps the best-known type of transtentorial herniation and is readily identified on both CT and MR. Uncal herniation typically occurs from a unilateral, expanding supratentorial lesion, especially in the middle cranial fossa. In this pattern

placed beneath the free edge of the falx cerebrei, pushing down the ipsilateral corpus callosum. With further progression, the contralateral cingulate gyrus and contralateral corpus callosum are compressed [2]. These findings are best demonstrated on coronal MR imaging (Fig. 4). When subfalcine herniation becomes severe, both foramina of Monro become compressed and obstructed. This results in dilatation of the contralateral lateral ventricle, but the ipsilateral ventricle usually remains compressed because of mass effect (Fig. 5). A wedge of pressure necrosis may occur where the cingulate gyrus is squeezed against the falx [2]. Also, the pericallosal artery may be pressed against the unyielding falx, resulting in occlusion and brain infarction [8,9]. Quantification of the degree of midline shift is most easily done by measuring the degree of displacement of the septum pellucidum from the midline, on axial images with either CT or MRI. The midline is usually determined by halving the distance between the inner tables at the level of the septum pellucidum. The degree of septal shift, not surprisingly, is predictive of patient prognosis [10 12]. Ross et al [12] reported that greater midline shift on CT scans correlated with a significantly lower likelihood of recovery in patients with acute intracranial hematomas. All patients in their series with a septal shift over 15 mm had a poor outcome. Conversely, no patient with septal shift less than 5 mm shift had a poor outcome. Pineal shift and aqueductal shift appear to have less predictive value [12].

Fig. 5. Subfalcine herniation. Axial CT scan reveals a large acute subdural hematoma (large arrow) causing subfalcine herniaton. Note the displacement of the anterior falx cerebrei (small arrow) and extensive midline shift. The ipsilateral ventricle is compressed due to mass effect, but the atrium of the contralateral lateral ventricle (outlined arrow) is dilated due to obstruction of the foramen of monro.

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

221

anterior choroidal and superior cerebellar arteries may be compromised, resulting in infarction [5]. Uncal herniation is well demonstrated on both CT and MRI (Fig. 6) [18 20]. Early on there is displacement of the uncus, resulting in effacement of the lateral suprasellar cistern. As herniation progresses, there is widening of the ipsilateral ambient and lateral pontine cistern, with displacement and rotation of the brainstem. With more advanced herniation, there is obliteration of the cisternal spaces, and the midbrain becomes compressed and elongated in the anteroposterior dimension.

Posterior (parahippocampal) tentorial herniation In patients with occipital and posterior temporal masses, the herniation of the medial temporal lobe occurs more posteriorly (Fig. 7). Consequently, the herniating hippocampus impinges less upon the cerebral peduncle, but more on the tectum at the level of the superior colliculus [21]. This may result in Parinauds syndrome, with paralysis of upward conjugate gaze. There is also relatively less compression of the oculomotor nerve and posterior cerebral artery. However, as herniation becomes more severe, these structures will be compromised. Early on, MR and CT will demonstrate effacement of the ipsilateral

Fig. 6. Uncal herniation. (A) Axial and (B) coronal T1weighted MR images after the administration of gadolinium demonstrate a large enhancing temporal lobe mass causing herniation of the uncus (arrow) into the suprasellar cistern.

of herniation, the uncus is displaced over the free edge of the tentorium [14]. Consequently, the midbrain is displaced and the opposite cerebral peduncle is squeezed against the contalateral tentorial edge. This damages the descending corticospinal and corticobulbar tracts, resulting in paralysis on the same side of the supratentorial mass. This false-localizing sign occasionally resulted in surgery on the wrong side of the brain in the days before CT [15]. The groove where the midbrain is compressed against the contalateral tentorial edge is known as Kernohans notch after the author who first described it [16]. Mass effect on the ipsilateral cerebral peduncle can also lead to a contralateral hemiparesis. As herniation progresses, the midbrain is further compressed, narrowing it in the transverse dimension, which can affect the ascending reticular activating system and result in progressive loss of consciousness. Compression of the oculomotor nerve results in ipsilateral third nerve palsy (blown pupil). If the posterior cerebral artery is compressed, infarction of the occipital lobe ensues and produces homonymous hemianopsia, though by this time the patient is usually unconscious [2,17]. Rarely, the

Fig. 7. Uncal and hippocampal herniation. Axial T2weighted MR image of the brain demonstrates uncal and hippocampal herniation due to multiple abscesses in the right temporal and occipital lobes. The largest abscess is in the occipital lobe resulting in parahippocampal herniation (black arrow) and compression of the right midbrain tegmentum (outlined arrow).

222

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

Fig. 8. Complications of transtentorial herniation. (A) Post mortem specimen in a patient with the brainstem removed demonstrates bilateral wedging and hemorrhage of the uncus. (B) In another post mortem specimen, the transtentorial herniation is more severe, demonstrating bilateral necrosis of the uncus and hippocampus due to pressure against the tentorium. (Courtesy of K. SantaCruz, MD, Kansas City, KS)

simultaneously with uncal and hippocampal herniation. Furthermore, with unilateral lesions, subfalcine herniation commonly precedes transtentorial herniation; whereas midline masses, bilateral supratentorial masses, or supratentorial hydrocephalus cause transtentorial descent without midline shift [1]. As descending transtentorial herniation progresses, there is further downward displacement of brain tissue, with disastrous consequences (Fig. 8). Damage is caused by local pressure from crowding of tissue within the incisura, which can lead to compromise of the third cranial nerve, posterior cerebral artery, and midbrain, and there are also traction forces wreaking havoc. The lower brainstem is less mobile because of the upper cervical dentate ligaments. The longitudinal compression of the upper brain stem against the unyielding medulla leads to buckling of the brainstem and further compression of the midbrain tegmentum [22]. Downward displacement of the brainstem also produces traction on the ouclomotor nerve, causing third nerve palsy. Stretching and shearing of the perforating branches of the basilar artery occur, resulting in ischemia and hemorrhage within the brainstem, called Durets hemorrhage (Fig. 9) [23,24]. This occurrence is a late finding and portends a poor prognosis, usually death [22,25]. Hydrocephalus develops because of distortion and obstruction of the cerebral aqueduct, further increasing the volume and pressure of supratentorial contents. Stovring pointed out that preferential dilatation of the contralateral

quadrigeminal plate and ambient cisterns, while there is usually no effacement of the lateral suprasellar cistern. Compression of the tectum and rotation of the midbrain can be seen as herniation progresses, and finally the cisternal spaces will be obliterated and the midbrain displaced and deformed. Compared with uncal herniation, this pattern of herniation appears to be much less commonly recognized as an isolated clinical presentation. In many cases there is overlap of uncal and hippocampal herniation.

Descending transtentorial herniation Descending transtentorial herniation consists of caudal descent of brain tissue through the tentorial incisura, and occurs mainly in response to mass effect in the frontal, parietal, and occipital lobes [1]. Since intracranial masses are rarely in the midline, there is considerable overlap of horizontal and vertical shifts during transtentorial herniation; such that descent of the diencephalon, midbrain, and pons tends to occur

Fig. 9. Durets hemorrhage. Axial CT scan of the head in a patient who had a large traumatic acute subdural hematoma (not shown) and unilateral descending transtentorial herniation reveals a hemorrhage (arrowhead) within the dorsal brainstem consistent with a Durets hemorrhage.

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

223

temporal horn tends to occur, because it is relatively shielded within the middle cranial fossa from compression by the supratentorial mass [26]. There also is a propensity for dilatation of the contralateral occipital horn because of relative shielding by the rigid posterior falx cerebrei. For the same reason, preferential dilatation of the contralateral temporal horn and occipital horn may also occur with subfalcine herniation, although the cause of the hydrocephalus is the result of obstruction of the foramen of Monro, rather than the aqueduct. Further descent can result in tonsillar herniation, with compression and distortion of the medulla oblongata causing cardiorespiratory compromise. It has been well established clinically that patients who decompensate from descending transtentorial herniation do so in an orderly rostral to caudal fashion of diencephalic and brainstem dysfunction [27]. The clinical consequences of progressive central herniation are oculomotor paresis, progressive alteration of consciousness, decerebrate rigidity, coma, and death. Effacement of the cisternal spaces, caudal displacement of the basilar artery and pineal gland, deformity and displacement of the midbrain, hydrocephalus, and infarction of the posterior cerebral artery have all been described on CT as findings associated with descending transtentorial herniation [17,28 30]. Most of these findings are indirect, secondary, and late manifestations of descending transtentorial herniation. Effacement of the perimesencephalic cisterns, though qualitative, is probably the most useful and consistent finding on CT (Fig. 10). With its better soft tissue definition and multiplanar imaging, MR provides superior anatomical detail of the tentorial incisura and adjacent structures. MRI findings of descending transtentorial herniation include downward descent of the brainstem, effacement of the basal cisterns, flattening of the pons against the clivus, and inferoposterior displacement of the quadrigeminal plate. As herniation progresses, deformity and buckling of the brainstem occur. Hydrocephalus develops because of aqueductal obstruction. Numerous authors have attempted to quantitate the degree of downward shift on MR images [19,31,32]. Ropper et al measured horizontal and vertical components of brain displacement on coronal MR images in 10 patients with supratentorial masses [33]. Feldman et al compared vertical brainstem position on 50 images from normal patients and 21 images from 15 patients with large supratentorial masses [31]. The methods of both Ropper and Feldman are cumbersome to employ in the clinical setting, and neither study demonstrated significant correlation between measured vertical descent and neurologic signs. Reich

Fig. 10. Descending transtentorial herniation. Axial CT scan of the head demonstrates obliteration of the basal cisterns (arrows) in a patient who sustained massive head trauma. There is downward displacement of the diencephalon and medial temporal lobes through tentorial incisura, resulting in effacement of the perimesencephalic cisterns and compression and displacement of the brainstem.

et al measured the relationship of the iter (proximal opening of the aqueduct of sylvius) to the incisural line on mid-sagittal MR images [19]. The incisural line was defined as a straight line from the anterior tuberculum sellae to the inferior point of the confluens of the straight sinus, the vein of Galen, and the inferior sagittal sinus. In 123 normal patients, the iter was 0.2 .8 mm below the incisural line. Displacement of the iter by more than two standard deviations (1.8 mm) below the incisural line was considered as MRI criterion for descending transtentorial herniation. Eighteen patients with supratentorial masses demonstrated downward displacement of the iter ranging from 2.0 mm to 11.0 mm, which correlated well with clinical signs of acute brain herniation, but less so with chronic cases [19]; however, in clinical practice, qualitative assessment of downward shift on MRI scans is probably more widely applied, though measurement of displacement of the iter relative to the incisural line can be useful.

Ascending transtentorial herniation Just as a supratentorial masses may cause descent of brain tissue through the hiatus, an infratentorial

224

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

mass may cause brain tissue to ascend through the hiatus [30,34,35]. Ascending transtentorial herniation has been most frequently reported in association with a cerebellar mass or a hydrocephalic (trapped) fourth ventricle. Cerebellopontine angle masses and intrinsic masses of the pons have also been reported, albeit much less commonly [34,36,37]. The propensity for ascending tentorial herniation to occur is more likely when the mass is near the incisura (ie, superior cerebellar vermis), when shunting of the lateral ventricle reduces hydrocephalus and pressure above, and when the opening of the incisura is large [34]. MR and CT findings of ascending transtentorial herniation include effacement of the superior cerebellar cistern, superior displacement of the superior vermis through the incisura, compression of the midbrain, and forward displacement of the pons against the clivus (Fig. 11) [31,34]. Ascending tentorial herniation can compress the posterior cerebral artery or superior cerebellar arteries against the tentorium, resulting in infarctions, or it can compress the aqueduct of sylvius, resulting in hydrocephalus. Obstruction of venous outflow by compression of the vein of Galen and basal vein of Rosenthal may occur and further increase intracranial pressure [38,39].

Tonsillar (foramen magnum) herniation Tonsillar herniation is defined as inferior displacement of the cerebellar tonsils through the foramen magnum into the cervical spinal canal. This is most commonly caused by a posterior fossa mass, but may also be caused by a supratentorial mass that causes downward transtentorial herniation [40]. In fact, tonsillar herniation occurs in association with descending transtentorial herniation anywhere from 20 to 50% of the time [14,19,41,42]. The propensity of tonsillar herniation to follow descending tentorial herniation is related to the size and shape of the incisura. If the incisura is small, the patient will be less likely to have tonsillar herniation. When tonsillar herniation progresses, pressure necrosis may develop where the tonsil is squeezed against the unyielding foramen magnum. Compression of the medulla oblongata by the displaced cerebellar tonsils may lead to loss of consciousness secondary to involvement of the ascending reticular activating system. Further progression of tonsillar herniation may damage cardiac and respiratory centers of the brainstem and even result in death. Additionally, occlusion of the posterior inferior cerebellar artery between the displaced tonsils and foramen magnum may result in cerebellar infarction.

Fig. 11. Ascending transtentorial hemorrhage. (A) Axial T1weighted MR image and (B) sagittal enhanced T1-weighted MR image demonstrate ascending transtentorial herniation in a patient with a cerebellar lung metastases. Note the upward displacement of the superior cerebellar vermis (black arrows) through the incisura, compression of the fourth ventricle, and anterior displacement of the pons (arrowhead) against the clivus.

Obstruction of the outlet foramina of the fourth ventricle causes hydrocephalus, further increasing intracranial pressure within the posterior fossa and worsening the herniation. There are large variations in the anatomy of the foramen magnum and the inferior cerebellum [2]. In a group of 200 normal patients, Barkovich et al demonstrated that the position of the tonsils ranged from 8 mm above the foramen magnum to 5 mm below the

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

225

Fig. 12. Tonsillar herniation. Sagittal enhanced T1-weighted MR image demonstrates a large enhancing cerebellar mass causing both tonsillar herniation and ascending transtentorial herniation. Note the inferior displacement of the tonsils (arrow) below the foramen magnum and the effacement of the surrounding csf spaces. The brainstem is compressed and displaced against the clivus and there is upward displacement of the superior cerebellar vermis (arrowhead) through the incisura.

concluded that extension of the tonsils below the foramen magnum is considered normal up to 3 mm, borderline between 3 mm and 5 mm, and clearly pathological when it exceeds 5 mm. Perhaps more important than strict quantitative criteria of tonsillar displacement are the morphologic appearance of the brainstem and the surrounding CSF spaces. When the tonsils extend below the foramen magnum, anterior displacement of the lower brain stem and loss of CSF space surrounding the brainstem are important in distinguishing tonsillar herniation from normal variations [45]. Tonsillar herniation can be difficult to assess on axial CT because of beam hardening artifacts at the skull base and partial volume average effects. MR imaging is the study of choice and sagittal images best delineate the relationship of the tonsils to the foramen magnum (Fig. 12). The foramen magnum can be identified on sagittal MR images by drawing a line from the inferior tip of the clivus (basion) to

foramen magnum [43]. Also, Aboulez et al reported the average distance of the tonsillar tips from the foramen magnum was 2.9 3.4 mm above the foramen magnum in 82 normal patients [44]. They

Fig. 13. Transphenoidal herniation. Sagittal enhanced T1weighted MR image demonstrates a large ring enhancing mass (arrowhead) in the right frontal lobe causing descending transphenoidal herniation (arrow).

Fig. 14. (A) Axial CT scan and (B) axial T1 weighted MR image of the head demonstrate a large right frontal external hernia in a patient who underwent a wide scalp excision, craniectomy, and free flap for an infiltrative squamous cell carcinoma. The flap became infected and failed, and the patient developed encephalitis resulting in external herniation. (Courtesy of R. Strang, MD, Kansas City, KS)

226

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

the posterior lip of the foramen magnum (opisthion). Identification of the tonsils below the foramen magnum, anterior brainstem displacement, and loss of CSF surrounding the brainstem indicate tonsillar herniation [46]. Of course, chronic tonsillar herniation, such as occurs with Chiari malformations, must be distinguished from acquired tonsillar herniation. Fortunately, the clinical history and imaging are characteristic and rarely cause diagnostic difficulty. The exception may be the presence of a posterior fossa mass with a pre-existing Chiari I malformation.

Transphenoidal (transalar) herniation Transphenoidal herniation is less well known, but occurs when brain tissue is displaced across the superior sphenoid alae, which incompletely separate the anterior and middle cranial fossae [2]. Areas of necrosis and hemorrhage may develop where the brain is displaced against the bony sphenoid ridge. There are two types of transphenoidal herniation: descending and ascending. Descending transphenoidal herniaton occurs when anterior cranial fossa mass effect causes

Fig. 15. Intracranial hypotension. (A) Coronal enhanced T1-weighted MR image demonstrates diffuse pachymeningeal enhancement (arrows) in a patient with severe postural headaches. (B) Sagittal enhanced T1-weighted MR image reveals tonsillar herniation (arrow), descending transtentorial herniation (note downward descent of the brainstem, loss of surronding csf spaces and flattening of the pons against the clivus) and pachymeningeal enhancement (arrowheads). (C) Postmyelogram CT scan of the thoracic spine shows an extradural accumulation of contrast (arrow) within the spinal canal consistent with a CSF leak. The extradural and intradural contrast medium outlines the dura (arrowhead) at the T10 level. A blood patch performed at this level resulted in resolution of the patients symptoms.

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228

227

displacement of the posterior frontal lobe over the sphenoid wing into the middle cranial fossa (Fig. 13). Ascending transphenoidal herniation is produced by middle cranial fossa mass effect, which causes displacement of the anterior temporal lobe over the sphenoid ridge into the anterior cranial fossa. The middle cerebral artery can become compressed between the displaced brain and the sphenoid ridge, resulting in middle cerebral artery infarction [4]. Although vascular compromise may occur with transphenoidal herniation, it is rare. Moreover, the clinical features of this type of herniation are poorly defined.

External herniation External herniations, also known as fungus cerebrei, are the rarest of acquired herniations [45]. They are most frequently caused by post surgical and posttraumatic defects that allow swollen or displaced brain to pass through. This type of herniation may be beneficial by relieving intracranial pressure and thereby preventing more devastating complications due to internal herniation; however, external exposure of the brain can result in infectious complications, and many of these patients have a poor prognosis because of the severity of their underlying illness. Both CT and MR imaging are effective at demonstrating these herniations (Fig. 14). MR imaging, however, is clearly superior at defining soft tissue, which can be critical in certain clinical scenarios. Distinguishing brain tissue from other tissues such as granulation tissue, cholesteatoma, tumor, or other lesions involving defects of the skull is essential, and MR imaging accomplishes this well [4,47]. MR imaging is also superior for evaluation of congenital encephaloceles, which are essentially congenital external hernias.

brain substance, CSF, and blood [3]. Since brain volume is constant, the volumes of CSF and blood fluctuate reciprocally. This buffering capacity is very limited, but does explain the imaging findings in patients with intracranial hypotension. With a loss in CSF volume, there is a greater increase in blood volume. This results in dural venous hyperemia and pachymeningeal venous engorgement and edema, which can be identified on MR imaging as diffuse pachymeningeal enhancement (Fig. 15). Furthermore, the decrease in volume of the suspending CSF results in downward descent of the brain and can cause descending central transtentorial herniation and tonsillar herniation. A blood patch or surgical repair of the dural defect is usually required.

Summary The consequences of cerebral herniation are compression of the brain, cranial nerves, and blood vessels that may result in serious neurologic morbidity, coma, and even death. A thorough understanding of the various patterns of cerebral herniation is essential, and it is important to remember that many of these patterns of herniation overlap. CT and MR imaging are effective at establishing the diagnosis of cerebral herniation, which will guide important decisions regarding therapeutic options and prognosis.

References
[1] Plum F, Posner JB. The diagnosis of stupor and coma. 3rd edition. Philadelphia: F.A. Davis; 1980. [2] McComb JG, Davis RL. Choroid plexus, cerebrospinal fluid, hydrocephalus, cerebral edema, and herniation phenomena. In: Davis RL, Robertson DM, editors. Textbook of neuropathology. Baltimore: Williams & Wilkins; 1985. p. 164 71. [3] Mokri B. The Monro-Kellie hypothesis, applications in CSF volume depletion. Neurology 2001;56:1746 8. [4] Laine FJ, Sheden AI, Dunn MM, Ghatak NR. Acquired intracranial herniations: MR imaging findings. AJR 1995;165:967 73. [5] Coburn MW, Rodriquez FJ. Cerebral herniations. Appl Radiol 1998;27:10 16. [6] Gray H. Grays anatomy. In: Warwick R, Williams PL. Philadelphia: Running Press; 1974. p. 512 23. [7] Strong OS, Elwyn A. Strong and Elwyns neuroanatomy. 5th edition. In: Truex RC, Carpenter MB, editors. Baltimore (MD): 1987. [8] Hassler O. Arterial pattern of human brainstem: normal appearance and deformation in expanding supratentorial conditions. Neurology 1967;17:1 6. [9] Sohn D, Levine S. Frontal lobe infarcts caused by brain

Intracranial hypotension Intracranial hypotension is a cause of cerebral herniation that should be considered in patients who do not harbor an intracranial mass or edema, or in patients in whom the degree of herniation is out of proportion to the degree of mass effect. Intracranial hypotension is caused by a CSF leak, either iatrogenic or spontaneous. As previously stated, the intracranial cavity is essentially a closed system and the volume is constant. These principles were described as early as the 1780s by Scottish anatomists Alexander Monro and George Kellie in what is known as the Monro Kellie doctrine, which states that intracranial volume is constant and is represented by the summed volumes of

228

P.L. Johnson et al / Neuroimag Clin N Am 12 (2002) 217228 herniation: compression of anterior cerebral artery branches. Arch Pathol Labe Med 1967;84:509 12. Marshall LF, Toole BM, Bowers SA. The National Traumatic Coma Data Bank. Part 2: patients who talk and deteriorate: implications for treatment. J Neurosurg 1983;59:285 8. Miller JD, Becker DP, Ward JD, et al. Significance of intracranial hypertension in severe head injury. J Neurosurg 1977;47:503 16. Ross DA, Walter OL, Ross AM, et al. Brain shift, level of consciousness and restoration of consciousness in patients with acute intracranial hematoma. J Neurosurg 1989;71:498 502. Finney LA, Walker AE. Transtentorial herniation. Springfield (IL): Thomas; 1962. Meyer A. Herniation of the brain. Arch Neurol 1920;4: 387 400. Collier J. The false localizing signs of intracranial tumors. Brain 1904;27:490 508. Kernohan JW, Woltman HW. Incisura of the curs due to contralateral brain tumor. Arch Neurol Psychiatry 1929;21:274 87. Sato M, Tanaka S, Kohama A, Fujii C. Occipital lobe infarction caused by tentorial herniation. Neurosurgery 1986;18:300 5. Osborn AG. Diagnosing of descending transtentorial herniation by cranial computed tomography. Radiology 1976;123:93 6. Reich JB, Sierra J, Camp W, et al. Magnetic resonance imaging: measurements and clinical changes accompanying transtentorial and foramen magnum brain herniation. Ann Neurol 1993;33:159 70. Stovring J. Descending tentorial herniation: findings on computed tomography. Neuroradiology 1977;14: 101 5. Johnson RT, Yates PO. Clinico-pathological aspects of pressure changes at the tentorium. Acta Radiol 1956b; 46:250 6. Howell DA. Longitudinal brain stem compression with buckling. Arch Neurol 1961;4:572 9. Fisher CM. Acute brainstem herniation: a revised concept. Semin Neurol 1984;4:417 21. Thompson RK, Scalman M. Brainstem hemorrhages: historical perspective. Neurosurgery 1988;22:623 8. Scheinker IM. Transtentorial herniation of the brainstem; a characteristic clinicopathologic syndrdome: pathogenesis of hemorrhages in the brainstem. Arch Neurol Psychiatr 1945;53:289 98. Stovring J. Contralateral temporal horn widening in unilateral supratentorial mass lesions: a diagnostic sign indicating tentorial herniation. J Comput Assist Tomogr 1977;1:319 23. McNealy DE, Plum F. Brainstem dysfunction with supratentorial mass lesions. Arch Neurol 1962;7:10 32. Hahn F, Gurney J. CT signs of central descending transtentorial herniation. AJNR 1985;6:844 5. Hahn F, Rim K, Schapiro RL. The normal range and position of the pineal gland on computed tomography. Radiology 1976;119:599 600. [30] Speigelman R, Hadani M, Ram Z, et al. Upward transtenorial herniation: a complication of postoperative edema at the cervicomedullary junction. Neurosurg 1989;24:284 8. [31] Feldman E, Gandy SE, Becker R, et al. MRI demonstrates descending transtenorial herniation. Neurology 1988;38:697 701. [32] Ropper AH. A preliminary MRI study of the geometry of brain displacement and level of consciousness with acute intracranial masses. Neurology 1989;39:622 7. [33] Ropper AH. Lateral displacement of the brain and level of consciousness in patients with an acute hemispheral mass. N Engl J Med 1986;314:953 8. [34] Cuneo RA, Caronna JJ, Pitts L, et al. Upward transtentorial herniation. Seven cases and a literature review. Arch Neurol 1979;36:618 23. [35] Ecker A. Upward transtentorial herniation of the brainstem and cerebellum due to tumor of the posterior fossa: with a special note on tumors of acoustic nerve. J Neurosurg 1948;5:51 61. [36] Binkow SM, Smirnov NA, Onischenko GT, translator. Lindsley DB, editor. Brain displacements and deformations. New York: Plenum Press; 1971. [37] Dinsdale HB. Spontaneous hemorrhage in the posterior fossa. Arch Neurol 1964;10:98 115. [38] Azambuja N, Lindren E, Sjorgen SE. Tentorial herniations. I,II,III. Acta Radiol 1956;46:215 41. [39] Sunderland S. The tentorial notch and complications produced by herniations of the brain through that aperture. Br J Surg 1958;455:422 38. [40] Cushing H. Some experimental and clinical observations concerning states of increased intracranial tension. Am J Med Sci 1902;124:375 400. [41] Ryder JW, Kleinschmidt-Demasters BK, Keller TS. Sudden deterioration and death in patients with benign tumors of the third ventricle area. J Neurosurg 1986; 64:216 23. [42] Schwarz GA, Rosner AA. Displacement and herniation of the hippocampal gyrus through the incisura tentorii. Arch of Neurol Psychiatry 1941;46:297 321. [43] Barkovich AJ, Wippold FJ, Sherman JL, Citrin CM. Significance of cerebellar tonsil position on MRI. Am J Neuroradiol 1986;6:795 9. [44] Aboulezz AO, Sartor K, Geyer CA, et al. Position of cerebellar tonsils in the normal population and in patients with Chiari Malformation: a quanitative approach with MR imaging. J Comput Assist Tomogr 1985;9: 1033 6. [45] Andrews BT, Chiles BW, Olsen WL, et al. The effect of intracerebral hematoma location on the risk of brainstem compression and on clinical outcome. J Neurosurg 1988;69:518 22. [46] Ishikawa M, Kikuch H, Fujisawa I, Yonekawa Y. Tonsillar herniation on magnetic resonance imaging. Neurosurgery 1988;22:77 81. [47] Kaseff LG, Seidenwurm DJ, Nieberding PH, et al. Magnetic resonance imaging of brain herniation into the middle ear. Am J Otol 1992;13:74 7.

[10]

[11]

[12]

[13] [14] [15] [16]

[17]

[18]

[19]

[20]

[21]

[22] [23] [24] [25]

[26]

[27] [28] [29]

Neuroimag Clin N Am 12 (2002) 229 248

Radiographic assessment of cranial gunshot wounds

Paul E. Kim, MD*, John L. Go, MD, Chi-Shing Zee, MD
Division of Neuroradiology, Department of Radiology, Keck School of Medicine, University of Southern California, 1200 North State Street, Room 3740, Los Angeles, CA 90033, USA

Although the number of crimes committed with firearms in the United States declined from the peak of nearly 600,000 in 1993 to approximately 350,000 in 2000 [1], firearm violence continues to plague the public health infrastructure of urban areas, and, more recently, suburban areas also. Gun homicide is the fourth leading cause of death for people aged 10 to 14 years and the second leading cause of death for people aged 15 to 24 [2]. The majority of investigative work to date worldwide in the diagnostic and therapeutic management of penetrating craniocerebral injuries has come from military experience [3 6]. In the United States, however, firearms have long been prevalent in civilian culture, so forensic investigations in this country have yielded significant contributions. Small shrapnel injuries, which are significantly lower in energy and less lethal than direct missile wounding of the brain, are more prevalent in the combat setting [7]. Civilian forensic work has an inherent bias toward fatal forms of gunshot injury, since the majority of victims do not survive to receive treatment. [6,8 11]. Four types of cranial gunshot injuries are classically described, based on the degree of penetration of the cranial vault [4,12,13]. Superficial injuries are defined as bullets trapped within the scalp or skull at impact without penetrating the cranial vault (Fig. 1). Tangential injuries are produced by bullets that graze the head, again without penetration of the cranial vault (Fig. 2). Despite the absence of direct penetration of the calvarium, both superficial and tangential gunshot wounds can result in significant intracranial lesions, either through direct energy

* Corresponding author. E-mail address: pek@hsc.usc.edu (P.E. Kim).

transfer or through the production of secondary penetrating missiles, as indriven bone fragments from calvarial fractures. Penetrating injuries result from bullets that enter into and lodge within the cranial vault (Fig. 3). The term spent bullet then applies to both superficial and penetrating injuries, in which the missile has come to rest within the victim. Perforating injuries describe through-andthrough lesions having both entrance and exit wounds to the cranial vault (Figs. 4 6). It should be noted that the distinction between penetrating and perforating injuries rests on whether the skull as a whole is traversed, not only the brain, because in the majority of cases defined as penetrating, that is, not through-and-through injuries, the bullet actually enters and exits the brain itself within the cranial vault before coming to rest. Brain tissue is relatively soft and gelatinous, offering little resistance to the advancing bullet. In fact, Kirkpatrick and Di Maio [14] reported 100% perforation of the brain once penetration of the cranial vault had occurred, indicating that bullets lodged within the brain parenchyma typically have ricocheted before coming to rest (see Fig. 3). Overall, more than 80% of gunshot wounds to the head penetrate the calvarium [15]. Unlike radiographic assessment of closed head injury, the evaluation of gunshot injuries is made more complex by the involvement of multiple anatomic levels simultaneouslyscalp, skull, and the multiple levels within the cranial vault itself. Despite this greater complexity, the spectrum of potential traumatic lesions and pathophysiologic sequelae are essentially the same as those encountered in closed head injury: scalp laceration, skull fracture, brain laceration, brain contusion, virtually all forms of intracranial hemorrhage, infarction, brain swelling,

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 7 - 2

230

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

herniation, and later complications of infection and traumatic aneurysm formation.

Plain skull films Compared to blunt trauma, skull films are more likely to provide at least some potentially useful information in the setting of cranial gunshot trauma. Since nearly all gunshot injuries involve comminuted, depressed skull fractures or foreign bodies, plain films may provide a reasonable overview of fractures, as well as a general idea of missile trajectory and location. They may be a helpful adjunct to computed tomography (CT) when metallic streak artifact limits visualization of areas of the scan, or perhaps when linear skull fractures are hidden on CT because they lie in the same plane as the scan acquisition. Nevertheless, the utility of plain radiographs is limited, so they should be obtained only if it can be done so without delaying CT. Some of the plain film information can even be obtained by examination of the CT scout image obtained before the actual scan [15]. Routine views include lateral, Towne, and Caldwell projections, but antero-posterior (AP) and lateral views are typically sufficient to give orthogonal information necessary for localization purposes.

Computed topography More than a quarter of a century since its advent, computed tomography continues to be the primary means of radiographic evaluation of the head in all acute trauma settings. All patients with cranial gunshot injuries should be imaged on an immediate emergent basis by CT, whether or not there is clinical evidence of penetration, except in the most extreme situations, in which immediate direct surgical intervention may be clinically mandated. Recent technological advances, including helical and multi-detector row CT scanners, have so drasti-

Fig. 1. Superficial injury. (A) CT scout image is useful substitute for plain films, showing spent bullet in occipital region. (B) CT bone window shows a portion of the bullet embedded within the calvarium and slightly depressed inner table fracturing (spallation). (C) Softtissue windowing reveals small hemorrhagic contusion and subarachnoid hemorrhage posteriorly beneath impact site. Subgaleal hematoma in right frontal region is presumably from secondary blunt trauma. Note that the subgaleal collection lies over the temporalis muscle within the subgaleal space (arrow).

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

231

Fig. 2. Tangential injury. (A) Bullet path was through the left orbit with deflection off the left sphenoid greater wing (black arrows). Indriven bone fragments are visible on the CT scout (white arrow). (B) Note comminuted fractures of the left sphenoid greater wing (arrow) and left orbital emphysema. (C) Small area of spallation of the inner table along the squamosal portion of the temporal bone, at some distance from the impact point. (D) Soft-tissue window shows indriven bone fragments, surrounding edema, and extra-axial pneumocephaly.

cally reduced total scanning time that the entire head can now be imaged in seconds, not including routine data entry and preliminary scout generation. Logistically, patient transport and scanner setup is the timelimiting factor for an available scanner. Intravenous contrast in the acute setting is generally not indicated and should be deferred. Older generation CT scanners acquired contiguous individual slices, typically at

5 mm intervals from skull vertex to the skull base. In contrast, new helical (spiral) CT scanners obtain a continuous three-dimensional data set, and photographed images may be formatted to whatever slice thickness is desired [16,17]. Both soft tissue and bone windows should be photographed. Three potential artifacts may be encountered: metallic streak artifact, partial volume averaging,

232

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

and motion. Bullets are typically made of high atomic weight metals [18], so there is an extreme discrepancy of radiographic density between projectile and brain, resulting in severe star-shaped streak artifact often obscuring the entire image within the same image slice (Fig. 7). If time permits, repeat scanning with the gantry angled to avoid the metal projectile may be performed to clearly image an area of the

head that is of particular concern. Partial volume averaging is encountered in the head adjacent to the calvarium, resulting in an artifactual thickening of the skull that can obscure thin extracerebral blood collections when photographed using standard soft tissue window levels. A modified, wider soft tissue window setting will better evaluate extra-axial hemorrhage. Motion artifact is encountered less fre-

Fig. 3. Penetrating bihemispheric injury with ricochet. (A) Right frontal hemorrhagic entry wound (black arrow) and ricochet track (white arrow). (B) Note the inner table beveling and indriven bone fragments at the entry wound. (C) Ricochet point near the contralateral left frontal bone vertex (arrow). (D) Spent bullet in the left occipital lobe and subarachnoid hemorrhage in the right sylvian fissure. (E) CT scout demonstrates initial missile track (black arrow) and ricochet track (white arrow).

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

233

Fig. 3 (continued ).

quently, because of the rapidity of current scanners, but when present, it is no less devastating to image quality. Motion can even create phantom lesions, such as artifactual hematomas [19].

Magnetic resonance imaging Magnetic resonance imaging (MRI) has been studied to a limited extent in gunshot wounds to other areas of the body [20 22], but except for investigation of MRI safety issues, studies on its evaluation of cranial gunshot injuries have been exceedingly sparse. The vast majority of commercial sporting bullets manufactured in the United States are non-ferromagnetic, but military and paramilitary ammunition (used in assault weapons such as the AK-47) commonly contain ferromagnetic materials, typically in the jacket covering a lead or antimony core, and exhibit varying degrees of rotation and displacement in a strong magnetic field. Therefore, although the vast majority of criminal and police shootings encountered in the United States involve non-ferromagnetic bullets that are safe for MRI, a cautious approach should be used and each case should be considered individually. Plain films may help determine bullet type, although the most accurate way to identify bullet type is by examining spent casings, or perhaps other fragments surgically removed from the patient [23 25]. Non-ferromagnetic bullets generate only minor image-degrading artifacts, and show good spatial resolution for bullet localization and proximity to vital structures, often better than CT, which is limited by streak artifact. On the other hand, ferromagnetic intracranial foreign

bodies, aside from being hazardous in strong magnetic fields, generate extreme artifacts, even in minute quantities [26]. Magnetic resonance imaging, like CT, continues to undergo technological advances that will likely expand its role in head trauma. For instance, the relatively recent development of a clinically practical fluid attenuated inversion recovery (FLAIR) technique allows MRI detection of acute extra-axial hemorrhage, including subarachnoid hemorrhage, with a greater degree of sensitivity than CT [27,28]. Diffusion-weighted imaging has facilitated earlier imaging diagnosis of early acute infarction; however, the primary limitations of MRI in the acute setting of trauma continue to be logistical and related to scan acquisition times, scanner availability, and difficulty in monitoring critically ill patients. With gunshot injuries, the consistent additional presence of potentially hazardous ferromagnetic metal bullet fragments and skull fractures, which MRI plays little role in assessing, leads to the obvious choice of CT as the primary imaging modality of cranial gunshot injuries in the acute setting, where the added information provided by MRI is very unlikely to affect surgical management [29]. In the subacute setting, however, MRI is superior to CT for assessment of delayed complications, such as acute infarction or infection, in which the earlier and more specific diagnosis offered by MRI impacts clinical management.

Angiography Although some investigators have advocated the widespread use of cerebral angiography in the acute setting [30], most advocate its use selectively, such as in the presence of significant intracerebral hematomas or in proximity injuries, such as missile injuries involving the base of the brain or structures near major vessels [5,31]. Recent technological advancements in noninvasive vascular imaging have been realized in both magnetic resonance angiography (MRA) and computed tomography angiography (CTA). There are technical and logistical differences among conventional angiography, MRA, and CTA that weigh on the specific clinical utility of these approaches. Metal fragments are typically devastating to MRA and CTA image quality, so their proximity would likely preclude the use of these modalities. Delayed subarachnoid hemorrhage or clinical signs of an arteriovenous fistula are presumptive evidence of a significant vascular abnormality, and should be evaluated directly by conventional angiography; however, in the ab-

234

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

sence of metal fragments that would affect image quality, MRA or CTA may be used as a screening test in the subacute period for proximity injuries or large intracerebral hematomas. MRA and CTA are superior to conventional angiography for detailed assessment of intravascular lesions such as intimal flap dissections. Conventional angiography can only demonstrate a narrowed lumen and does not give information about the state of the arterial wall itself.

CTA is theoretically superior to MRA because of its higher resolution and absence of flow-related artifacts, but further studies are needed to compare the specific advantages of these techniques [32 34]. Imaging features Penetrating missiles generate tissue injury through three mechanisms, the predominant mechanism

Fig. 4. Fatal bihemispheric perforating injury. (A,B) Right frontal entrance wound shows inner table beveling and indriven bone fragments. Note the paradoxical progressive increase in diameter of hemorrhagic wound track. (C) Note also the left frontal exit wound, pneumocephaly, large parafalcial subdural hematoma, evidence of global ischemic injury and extensive skull fractures. (D) Severe swelling/edema with intracranial hypertension evidenced by effacement of basilar cisterns. (E) Typical exit wound outer table beveling in another patient.

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

235

pressure gradient that may suction debris into the wound is also created within this cavity. With the highest velocities ( > 1000 m/sec), cavitational motion may be so outwardly explosive that it literally bursts the skull [36,37]. Scalp There are five anatomic layers of the scalp, from superficial to deep. These consist of: skin, subcutaneous fibrofatty tissue, galea aponeurotica, loose areolar connective tissue, and periosteum [38]. In all but the most superficial tangential gunshot injuries, all five layers are typically lacerated. Hemorrhage may accumulate in three potential spaces within the scalp, depending on the age of the patient: subcutaneous, subperiosteal, and subgaleal. Subcutaneous hematomas (caput succedaneum) (Fig. 8) and subperiosteal locations are much more common in infants, seen most commonly in relation to minor birth trauma. Subgaleal hematomas are the most common type seen in adults in all types of trauma, whether penetrating or closed. On CT, these hematomas are seen as high density collections overlying the temporalis muscle, which lies deep to the galea aponeurotica within the subgaleal space (see Fig. 1C). Like subdural hematomas, hemorrhagic collections in the subgaleal space tend to spread more diffusely over the skull, because of a relative lack of impedance within this space. In contrast, subperiosteal hematomas, much like epidural hematomas intracranially, tend to be more confined and lenticular in shape, and do not cross suture lines, because of tight skull attachments of the periosteum. MRI findings parallel those of CT, with signal intensities of blood collections also varying primarily as a function of time. MRI delineates the anatomic layers of the scalp better than CT, though this is usually not of clinical significance. A subgaleal hygroma is a collection of cerebrospinal fluid within the subgaleal space. On CT, a hygroma may be distinguished from hemorrhage by its low attenuation, unless it is contaminated by blood, which increases its attenuation. Its presence is inevitable evidence of an underlying skull fracture and dural laceration, whether due directly to missile penetration or to secondary fracturing. MRI will show signal intensity paralleling cerebrospinal fluid (CSF), again unless there is contamination by blood. Skull Evaluation of the skull is accomplished primarily by CT, with plain films sometimes providing supplementary information. MRI does not provide the degree of bony detail necessary for evaluation of skull frac-

Fig. 4 (continued ).

being dependent on the velocity of the missile: (1) direct laceration, (2) shock wave transmission, and (3) cavitation. Direct laceration is the primary mode of injury with low impact velocities, meaning velocities at impact of less than 100 m/sec. (Muzzle velocities do not necessarily represent missile energies at impact, as a bullets velocity is affected by distance, altitude, pitch, yaw, and caliber. Thus, a bullet fired at high velocity from a distance may impact the target at a significantly lower energy state than a low-velocity gunshot from close range.) This is the case with all stabbing injuries and many handgun injuries, particularly if fired from a distance. Shock wave transmission and cavitation become increasingly prominent mechanisms of damage with increasing projectile velocity, and become the predominant mechanisms at greater than 320 m/sec, the speed of sound. Thus, a muzzle velocity of 320 m/sec is somewhat arbitrarily designated as the point of division between high- and low-velocity firearms. Shock waves are violent pulsations of high-pressure waves traveling at the speed of sound and emanating from the front of the advancing projectile. They may reflect off the skull and summate with increasing tissue damage. Tangential injuries can produce extensive damage through this mechanism, with or without secondary injury resulting from indriven fracture fragments [12,35]. Cavitation is the predominant mechanism of tissue injury at the highest velocities. The explosive centrifugal motion generated by the advancing missiles volume creates a temporary conical cavity that pulsates several times before collapsing. A negative

236

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

Fig. 5. Unilobe perforating injury. (A) Bullet track is anterior to posterior, with bullet fragments and (B) deeply indriven bone fragment adjacent to the path. (C) Two days later, after removal of deep bone fragment, acute bilateral anterior cerebral artery infarcts have developed, as well as infarcts of right posterior frontal and right posterior thalamic regions.

tures. Irregular and stellate fractures typically radiate from the point of impact on the skull, but remote linear

fractures are usually seen at a distance from the impact point (see Fig. 4), the result of a complex set of forces

Fig. 6. Fatal unilateral hemispheric perforating injury. Missile path is cephalo-caudad and posterior to anterior. (A) Inner table beveling demarcates entrance wound just above. Proximal wound canal (B) diameter is larger than distal portion near exit through skull base (C). (D) Image above (C) shows ricochet of a portion of the bullet. (E) CT scout shows missile path (black arrows) and partial ricochet fragments (white arrow).

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

237

238

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

Fig. 7. Severe streak artifact.

that can be imparted by the missile. At the point of impact, focal inbending of the skull first occurs, resulting in a tensile stress to the inner table. Forces insufficient to overcome the tensile resilience of the inner table result in rebounding of the skull without a fracture. Simultaneous outbending of the skull in a circumferential manner also occurs at a distance from the point of impact, with tensile stress applied instead to the outer table [39]. Isolated inner or outer table spallation, usually not visible on plain radiograph, may occur when forces are insufficient to cause complete fracture (see Figs. 1B and 2C). Comminution results from a sufficiently large amount of energy dissipated over a short period of time in a relatively localized area. The inner table is typically more comminuted than the outer table at the impact site. In high-velocity penetrating/perforating injuries (impact velocities of greater than 320 m/sec), the bullet is a rapidly advancing, space-occupying volume that can impart an explosive outward displacement of brain tissue against calvarial bone, with enough energy to generate extensive skull fracturing at a distance from the impact site [36]. Depressed skull fractures require special consideration because they are associated with a higher risk of complications, including hemorrhage, infection, CSF leak, and epilepsy. On plain films, these may appear as linear or irregular densities rather than lucencies, because of the overlap of edges of fracture components. Complete assessment of all but the simplest of skull fractures, including delineation of

overlap, displacement, comminution, and indriven fragments, requires additional imaging with computed tomography. Depending on missile velocity, skull fractures can appear as simple beveled perforations or, with higher velocity ballistics, may be quite complex, with multiple fractures arising at remote distances from the point of impact. The site of penetration can be identified by the pattern of calvarial beveling on CT. Entrance sites typically show beveling and comminution of the inner table (Figs. 3B and 4A), while the exit site in perforating (through-and-through) injuries will show the reversebeveling and greater comminution of the outer table (see Fig. 4E). The size of an entrance wound is also typically smaller than that of the exit wound [37,40]. In most cases, entrance wounds will also show adjacent indriven bone fragments, which arise primarily from the inner table [15]. Some interpretative caution should be exercised when an intracranial bullet has ricocheted and possibly fractured the inner table at multiple sites. As noted previously, intracranial ricocheting is an extremely common phenomenon, and the presence of a bullet lodged within brain parenchyma usually indicates that ricocheting has occurred [14]. The type of skull fracture seen in tangential injuries depends on the depth and angle of the bullet path as well as the amount of energy deposited, which can be considerable even in superficial injuries. If a

Fig. 8. Transventricular missile path with extensive intraventricular hemorrhage. Note the conical shape of hemorrhagic wound canal that tapers longitudinally along the missile track.

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

239

tangential injury is the result of a high-velocity missile, extensive skull fracturing can also be encountered remote from the impact site [12,35]. Depressed, comminuted fractures are the most common type seen with superficial/tangential injuries, and CT is best used to delineate these fractures, as well as the size, number, and location of indriven fragments. Superficial/tangential injuries also may result in incomplete fractures of the skull that are not visible on plain radiographs, and that require CT for their detection. Spallation is a type of incomplete fracturing of the inner table with sparing of the outer table [12], caused by a combination of inbending of the skull and shock wave transmission at impact. Guttering is a specific type of skull fracture seen only with tangential injuries, consisting of a groove created by the bullet in its path along the outer portions of the calvarium [12,35]. Fractures of the skull base require CT for adequate delineation. The incidence of fractures of the sphenoid bone has probably been underestimated in the past [15,41], when plain films would frequently be negative or show only secondary signs of a skull base fracture, such as clouding of air cells of the sinuses or middle ear. Sphenoid fractures have a high association with complications such as cerebrospinal fluid rhinorrhea, cranial nerve and optic nerve injury, and injury to the internal carotid artery and cavernous sinus. Many of these fractures are in the horizontal plane and sometimes poorly seen on axial CT scans. Either direct coronal imaging or high-resolution, three-dimensional (3D) helical scanning with multiplanar reconstructions formatted in multiple planes may be required to visualize these fractures. Temporal bone fractures are even more easily missed on routine axial CT acquisitions because of the thickness of the slices obtained. The presence of fluid within the middle ear or mastoids, CSF otorrhea, hearing loss, vertigo, or facial nerve injury should prompt targeted CT imaging of the temporal bones in 1.5 mm thick sections. Frequently, clinical factors preclude imaging in both axial and direct coronal planes, but most often axial images suffice. Again, high resolution reformatting in the coronal plane from 3Dacquisitions in the axial plane with current helical CT scanners may be performed for further evaluation when needed. Cranial vault CT should be utilized in the acute setting to evaluate most of the intracranial consequences of gunshot injury, whether penetrating or not. These include: pneumocephaly, extra-axial hemorrhage (epidural, subdural, subarachnoid), parenchymal laceration, parenchymal hemorrhage, contusion, intra-

ventricular hemorrhage, infarction, cerebral swelling and edema, and herniation. Depending on the clinical scenario, MRI is typically best suited for the subacute setting, and later for the evaluation of delayed complications. There are few references in the literature to pneumocephaly associated with gunshot injury, but one investigator reported a 79% incidence of extracerebral and 43% incidence of intracerebral pneumocephalus detected by CT [41]. In the case of non-penetrating superficial or tangential injuries, its presence always indicates the presence of a dural tear. Air within a dural venous sinus may indicate a potentially serious venous laceration [42]. Parenchymal air at a distance from the missile track has been observed and carries a negative prognostic significance [43]. Previous investigations have suggested that epidural hematomas are surprisingly uncommon in gunshot injuries. This may in part be attributable to misdiagnosis of some epidural hematomas as subdural hematomas, since approximately 15% of epidural hematomas do not appear as classic lenticular or biconvex collections, but instead mimic the crescentic configuration of subdural hematomas [41]. Many small epidural hematomas are also partially or completely hidden by volume averaging effects of the adjacent calvarium. The reported range of clinically significant epidural hematomas is 0.6% to 3% [8,44,45]. Investigations using MRI might be expected to yield a higher level of detection of epidural hematomas in gunshot injuries because of its inherently superior anatomic detail, although any hematomas that are not demonstrable by CT are unlikely to be of clinical significance. Subdural hematomas in cranial gunshot injuries are a relatively common finding (see Fig. 4). A report in the pre-CT era indicated an incidence of 21% for subdural hematomas of 20 mL or greater in volume [44]. Interestingly, the reported incidence with CT has not deviated significantly from this number [8,46 48]. Classically, subdural hematomas are crescentic in shape and tend to conform to the shape of the brain surface. There is a relative lack of impedance within the subdural space, so unlike epidural hematomas, subdural hematomas tend to spread more diffusely over the convexities and are not confined by dural attachments at calvarial suture lines. The MRI or CT appearance of hematomas (whether extra- or intra-parenchymal) is determined primarily by time, and altered if composition changes such as CSF mixing occur, as they frequently do in the setting of penetrating injury. Delayed subdural hemorrhage is an important event, because it suggests the presence of a significant delayed complication such as trau-

240

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

matic aneurysm formation or consumptive coagulopathy [5,31,49,50]. Subarachnoid hemorrhage in cranial gunshot injury has largely been ignored in the past, perhaps because it seems to have little clinical significance as an acute finding. In fact, Levy et al found no association of subarachnoid hemorrhage at presentation with delayed development of traumatic aneurysms [51]. It is likely present to some degree in most cases, except for the most superficial injuries. Subarachnoid hemorrhage large enough to be identifiable on CT is quite common, with a reported rate between 31% and 59% [43,51]. Subarachnoid hemorrhage carries greater clinical significance if it occurs as a delayed event, when it may indicate the development of a traumatic aneurysm. On CT, acute subarachnoid hemorrhage appears as increased density within the cisterns, fissures, or sulci. MRI had not been useful for the detection of subarachnoid hemorrhage until the relatively recent development of a CSF-suppression, T2-weighted technique fast enough to be clinically useful. The development in the mid-1990s of FLAIR, a fast protocol for CSF suppression, has led to its routine use in standard brain imaging protocols. A significant advantage of this technique is its ability to image acute extracerebral hemorrhage, including acute subarachnoid hemorrhage, previously invisible using routine spinecho techniques [27,28]. Despite this capability, CT continues to be the preferred method for demonstration of subarachnoid hemorrhage, because of the logistical limitations of scanner availability, cost, time, and the small risk associated with ferromagnetic bullet materials. The parenchymal laceration produced by the missile is generally conical in shape, with the base of the cone situated at the entrance site, and a gradually decreasing diameter along the missile track. Hemorrhage along this track typically clearly delineates this high attenuation conical laceration (see Fig. 8). A great variety of wound shapes are encountered, attributable to missile yaw or deformation occurring at impact with calvarial bone, which slows the missile and increases energy deposition and damage to tissues [14,52]. It should be noted that a significant number of bullet tracks are not hemorrhagic, so one cannot rely on visualization of a hemorrhagic wound canal to trace its path. The length and location of bullet tracks, including ricochets, has prognostic significance, so CT or MRI imaging may be used to identify these tracks [26]. Ricocheting, observed in 18% to 26% of cases [14,53], obviously increases the amount of laceration produced and worsens prognosis (see Fig. 3).

Intracerebral hematomas arise in one of three ways: (1) direct vascular laceration, (2) vascular injury from transmitted shear and tensile stresses, and (3) parenchymal contusion. The latter two are produced by either shock wave transmission or mechanical impact loading (contrecoup injury) as in closed head injury, except that in gunshot trauma the sudden forceful displacement of brain tissue is largely imparted by a missile, particularly if there is significant cavitational force [36,37,40]. Thus, parenchymal hematomas and contusions may be seen at a distance from the missile track (Fig. 9) [45,54,55]. Gunshot wounds of the face and neck may even transmit shock wave energy to the formation of cerebral contusions [12]. The reported incidence of significant intracerebral hematomas, defined as greater than 20 mL (or roughly a 3.3 cm diameter sphere), is 23% to 27% [8,44,56]. The CT appearance evolves with time in the same manner as extra-axial collections. The intraparenchymal clot undergoes lysis and resorption with time, beginning at the periphery and proceeding to the center. Delayed intracerebral hematomas typically occur at sites of nonhemorrhagic contusion [9,57,58]. Less commonly, they may be a secondary complication of either cerebral infarction or a consumptive coagulopathy syndrome that may accompany severe head injury [49,50]. Cerebral contusions occur through several mechanisms in gunshot injury. In non-penetrating injuries they may occur either from direct energy transmission through the calvarium to the underlying brain, or by focal inbending of the skull, with or without fracture (see Fig. 1C). In penetrating injuries, shock wave transmission and cavitational forces can form contusions remote from the missile track [55,59]. Of course, blunt head trauma from a fall subsequent to gunshot injury is possible as well. Cerebral contusions are parenchymal bruises that show perivascular extravasation and edema, which determine the gross appearance. Non-hemorrhagic contusions are the result of these changes confined to a microscopic level. With increasing vascular breakdown and extravasation, the contusion eventually becomes grossly hemorrhagic [58,60]. Nonhemorrhagic contusions may be missed by CT but are seen well and early by MRI [29,58], particularly if diffusionweighted imaging is used, as it has high sensitivity for the restricted proton diffusion of cytotoxic edema, whether because of contusion or acute infarction [61]. More diffuse parenchymal injury may result in diffuse axonal shear injury [59,62]. Most shear injury foci are not visible on CT, but occasionally they can be hemorrhagic and appear as punctate petechial or

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

241

Fig. 9. The right posterior frontal intracerebral hematoma was remote from the missile path (not shown, extending from high left frontal to right frontal vertex).

linear hemorrhages in the white matter of the cerebral hemispheres, brainstem, and cerebellum. MRI more readily detects these small lesions, particularly if they are non-hemorrhagic [61,63]. Intraventricular hemorrhage has poor prognostic significance (see Fig. 8). It results either from direct penetration, from transmitted shear stresses on the ventricular walls that result in vessel injury, or from trans-ependymal dissection of intracerebral hemorrhage. On CT, it may be seen in the acute stages as a focal clot, sometimes forming a virtual cast of the ventricle, or unclotted, as high-attenuation fluid layering dependently within the occipital horns. Intraventricular hemorrhage is a poor prognostic sign, even if ventricular penetration has not occurred, because its presence in any trauma setting signifies major brain injury [64]. Ischemic brain injury is extremely common in penetrating head injury. Apnea resulting from direct brainstem traversal is obviously acutely fatal. However, even when the missile track is several centimeters clear of the vital centers, missile-induced apnea is the most common acutely fatal physiologic circumstance, and results from energy transfer by pressure-wave phenomena. This can occur even with relatively low energy bullets [7]. Thus, global anoxic/

ischemic injuries are common in those who survive the initial respiratory insult (Figs. 4 and 6). They may result from direct vascular injury by the penetrating missile or bone fragments, or through traumatic vasospasm or thrombosis from shock wave transmission. Intracranial hypertension, swelling, and herniation are frequently present as well. CT is typically normal in the hyperacute stage of cerebral infarction, but within 12 to 24 hours areas of hypoattenuation and swelling develop, because of increasing edema resulting from conversion of predominantly cytotoxic intracellular edema to extracellular edema, as an accompaniment of cell death. MRI T2-weighted and FLAIR sequences are significantly more sensitive than CT to these cellular and extra-cellular changes, and may be abnormal within hours of the initial insult. Diffusion-weighted imaging is even more sensitive and shows signs of cytotoxic edema and restricted diffusion within one hour of the insult [65,66]. The edema usually peaks four to six days after the insult, and gradually diminishes over the ensuing weeks. The instantaneous damage inflicted by the missile is soon compounded over the next minutes to hours by a series of deleterious secondary events, many of which have been discussed previously. Intracranial hypertension is one of the most significant of these effects, and begins within several minutes [13]. Intracranial hypertension results from some combination of post-traumatic cerebral hyperemia, early edema, intracranial hemorrhagic mass effects, and obstructive hydrocephalus. Notably, when the effects of post-traumatic cerebral hyperemia predominate, brain attenuation is actually normal or slightly increased, whereas edema results in a decrease in parenchymal density on CT [67 69]. In experimental missile woundings, edema was shown to be predominantly vasogenic in craniocerebral gunshot injuries and was primarily confined to the region around the missile track. This edema is typically mild, peaks in 48 to 72 hours, and nearly always resolves within one week [70]. Signs of increased intracranial pressure on CT or MRI include effacement of sulci and basal cisterns, and compression of ventricles. Of these, basal cisternal effacement is the most reliable imaging sign of intracranial hypertension and an ominous portent of herniation or fatal transmission of pressure to the brainstem (see Fig. 4D) [71,72]. Delayed complications These include infection in all of its intracranial forms, cerebrospinal fluid fistula, hydrocephalus, and

242

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

delayed hemorrhage. Post-traumatic epilepsy following penetrating cranial gunshot injuries occurs at a rate of approximately 30% [73]. Although imaging plays little role in predicting its onset, a seizure may be the earliest sign of developing complication, such as infection, hemorrhage, or infarction, and should prompt expeditious imaging, preferably MRI with and without contrast, barring the presence of significant ferromagnetic hazards. Intracranial bullet migration is uncommon, but usually significant. Migration most commonly indicates development of an intracerebral abscess [74,75]. Bullets may also move in an evolving hematoma [76] or within the ventricular system. Migrating intracranial bullets, whether intraventricular or parenchymal, may precipitate a number of other delayed complications, including seizures, delayed hematomas, obstructive hydrocephalus, and endocrine dysfunction [74,77,78]. Virtually all forms of intra- and extra-cranial infection may occur: scalp wound infection, osteomyelitis of the skull, epidural and subdural empyema, meningitis, ventriculitis, cerebritis, and cerebral abscess. Although markedly decreased with antibiotics and modern neurosurgical techniques, infection continues to be an important complication and is seen in approximately 5% to 7% of cases [79,80]. Infectious complications may even be delayed for years in the presence of an intracranial foreign body [81]. Risk factors include external wound infection, cerebrospinal fluid fistula, paranasal sinus injury, and retained external organic material such as hair or clothing. Notably, retained bullets do not in themselves increase the risk of infectious complications. Infection risk because of retained intracranial bone fragments has been the source of controversy in the surgical literature, but relatively recent reports indicate that retained bone fragments do not by themselves increase infection risk unless there is also wound dehiscence or a CSF fistula. This may mitigate the need for additional debridement and its accompanying morbidity [56,80]. Scalp wound infections may be difficult to differentiate on CT or MRI from postsurgical changes, so a search for secondary clues, such as development of air or fluid collections not previously present, should be made, but clinical evidence is typically paramount. Osteomyelitis of the skull is clinically somewhat more problematic, and imaging may be needed for confirmation of a clinical suspicion. Plain radiographs are the least sensitive, requiring sufficient lytic bony destruction to allow for grossly visible density changes. CT is more sensitive to subtle lytic bone changes that are quite specific for osteomyelitis.

MRI is more sensitive than CT, because the signal intensity changes that characterize infection occur before the bony destructive changes that are required for CT visualization. MRI findings are somewhat less specific, however, because edema resulting from bony contusions, fractures, and surgery can mimic infection. Nuclear medicine bone scintigraphy may be of help when the other tests are equivocal. Epidural and subdural empyemas are typically identical in configuration to hemorrhagic collections in these locations. Empyemas tend to be low attenuation on CT, only slightly higher density than CSF, unless complicated by hemorrhage. MRI signal characteristics will vary depending on protein and blood content [82,83], but are typically mildly hyperintense to CSF on T1 and equal to CSF on T2. Both entities will exhibit restricted diffusion on diffusion-weighted imaging (DWI). Contrast enhancement at the periphery may be seen in both hematomas and empyemas, but tends to be more intense when infection is present [84 86]. Evidence for underlying leptomeningitis should be sought as a potential secondary finding. In practice, it is not usually possible to distinguish between extracerebral empyemas and delayed hematomas by imaging alone in the subacute post-traumatic period and later, and clinical suspicion is paramount in this circumstance. Hemorrhagic collections may themselves be complicated by infection, and signal variability arising from varying mixtures of cerebrospinal fluid, blood, and serous fluid can further complicate the imaging appearance on both CT and MRI, so caution is advised in interpretation. The diagnosis of meningitis is best established by examination of cerebrospinal fluid, but imaging evidence includes abnormal contrast enhancement in the sulci and cisterns on CT and MRI, and sulcal hyperintensity on FLAIR acquisitions [58,87]. Ventriculitis is seen as intense enhancement of ventricular walls and choroids plexus, and is often associated with obstructive hydrocephalus. Cerebritis refers to an area of infection and inflammatory response without necrosis. CT demonstrates a nonspecific area of hypoattenuation that commonly enhances, reflecting the presence of edema fluid and inflammation. MRI will show nonspecific hypointensity on T1- and hyperintensity on T2-weighted images, and enhancement and restricted diffusion on DWI. With progression, central necrosis and cavitation walled off by a thin reticular-collagen capsule develops, forming a cerebral abscess [88,89]. Cerebral abscess has a reported incidence of 2% to 3% in gunshot injuries [80,90]. This usually occurs from two to four weeks following injury [80,89,90]. CT typically demonstrates a circumscribed, rounded, low density cystic mass with a

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

243

thin enhancing capsule and surrounding edema, while MRI also shows a cystic mass with proteinaceous fluid, a thin enhancing capsule that is hypointense on T2-weighted sequences and characteristically thinner on the ventricular side, and restricted diffusion of the cyst fluid on DWI [91]. Cerebrospinal fluid fistulae are common following gunshot wounds to the head, seen in approximately 9% of cases [92], and have a high association with the development of meningitis. These are most commonly the result of fractures of the skull base that involve a paranasal sinus (CSF rhinorrhea) or temporal bone (otorrhea) and arise most frequently in orbitofacial penetrating gunshot wounds [9,19]. Radiologic detection of the site of leak is difficult. Nuclear medicine cisternography using Tc-99m diethylene triamine-pentaacetic acid (DTPA) and cotton pledgets in the ear or nose can be used to confirm a leak if the diagnosis is clinically equivocal, but CT cisternography must be used to demonstrate the exact site of the defect. Hydrocephalus may have an acute or delayed onset and may be caused by subarachnoid hemorrhage, intraventricular hemorrhage, ventriculitis, mass effect from intracerebral hematomas, edema or swelling, or an intraventricular missile or fragment. Unlike berry aneurysms of the circle of Willis, traumatic aneurysms tend to occur more distally

and do not have a predilection for branching points (Fig. 10). If a traumatic aneurysm in penetrating trauma is the result of direct laceration, a false aneurysm would be the result; however, strong pressure waves and energy transfer in gunshot injuries may give rise to traumatic aneurysms just as in closed head injury, and the proportion of true versus false aneurysms of traumatic origin is not clear [93]. Traumatic aneurysms complicating cranial gunshot injuries occur at a reported rate of approximately 3% to 4% [3,31,51], when angiograms are obtained in the subacute period. One study suggested that the incidence of traumatic aneurysms might be significantly higher if angiography were performed earlier [30]. Early diagnosis is important, because there is 50% mortality associated with rupture. As previously noted, magnetic resonance angiography and CT angiography are frequently limited by the presence of metal fragments in the setting of gunshot injury, but if such interference is limited, these techniques can be used as screening tests for traumatic aneurysms.

Radiographic analysis of outcome Numerous investigations have employed an array of pathophysiologic etiologic, epidemiological, and radiographic criteria in an attempt to arrive at mean-

Fig. 10. Delayed traumatic aneurysm from a nonpenetrating shotgun blast injury, presumably the result of energy transfer via shock wave phenomena. (A) CT with contrast shows an aneurysm collection in the anterior interhemispheric fissure (arrow). Extracranial shotgun pellets were present on the left. (B) Conventional angiogram shows the aneurysm at the A2 segment of the right anterior cerebral artery.

244

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248

ingful prognostic parameters for cranial gunshot injury. The ethical, medicolegal, and public health ramifications of these prognostications and their impact on resource utilization and treatment are beyond the scope of this writing, but are difficult to ignore when the overall mortality rate is so high (91.7%) [11]. The majority of patients, in fact, expire before receiving treatment. Reported mortality rates in patients who do reach the hospital vary considerably but are still high, from 35% to 84% [43,94]. Outcome analysis is a difficult endeavor because of the number and complexity of interactions between parameters, none of which can be legitimately studied independently. Variables that have been considered for their prognostic significance include patient age, Glasgow coma scale (GCS) score, intracranial pressure monitoring, number of lobes involved, traversing of midsagittal or midcoronal planes by the missile, intraventricular hemorrhage, basal cistern effacement, intracerebral pneumocephaly, intracerebral hemorrhage, extra-axial hemorrhage, infection, and bullet or bone fragment retention. Of these, the presenting neurologic status (GCS) is the single best predictor of outcome, with mortality approximating 95% when the GCS is less than 5 [94 97]. The remaining parameters do not appear to be significant factors independent of the admission GCS, and might impact prognostic accuracy only when the GCS is relatively high. Narayan et al reported that CT data used independently might be an ineffective prognosticator unless combined with clinical data [98]. Careful analysis of missile path with CT to examine the planes traversed or number of lobes involved can be difficult because of the relatively high frequency of ricochet injuries and the occasional presence of non-hemorrhagic missile tracks. Missile path anatomy correlates with outcome in the most obvious instance, when a vital structure such as the brainstem is directly involved. These injuries are typically fatal [99]. In patients with GCS of 5 to 13, Kennedy et al found 88.9% survival when a single lobe was involved, compared with 54.2% survival with multilobe involvement [94]. When multi-lobe injury is bihemispheric (ie, the missile crossed the midline), a 15% to 17% survival has been reported, whereas 65% survival was seen with unilateral multi-lobe injury [96,97]. If both midsagittal and midcoronal planes are traversed, survival decreases to less than 8% [11,43,47,95,96]. Of the various forms of intracranial hemorrhage, intraventricular hemorrhage is the most significant for prognosis [11]. Direct transventricular injury is particularly associated with a dismal outcome, with most reports indicating less than 14% survival, and most

survivors neurologically devastated [8,11,95,96]. Clark et al reported a poor Glasgow outcome scale score in 89.7% of patients with ventricular injury against 24.1% of patients without such injury [8]. As noted previously, intraventricular hemorrhage even without transventricular injury may have negative prognostic significance, because it usually indicates severe brain injury [64]. Basal cisternal effacement on imaging is directly correlated with intracranial hypertension, impending herniation, and brainstem injury [71,72,100]. Intracerebral hematomas are associated with a lower survival rate (36%) than extracerebral collections (50%) [96], though obviously the size of the hemorrhagic collection plays a significant role in individual cases. Infection also negatively impacts outcome. Analysis of deaths in the Vietnam War among patients who survived the acute period through the first week showed that delayed development of intracranial infection resulted in a decrease from 93% to 69% in subsequent survival. Development of a cerebral abscess is the most severe complication and carries a mortality rate of over 50% [90,101]. Additional CT findings reported to be of some negative predictive value include small air bubbles dissecting into the white matter at a distance from the missile track [43], midline shift of greater than 10 mm [96,102], and presence of bone and metal fragments away from the missile track [97].

Summary Though advances in MRI will undoubtedly increase its use, particularly in the subacute period, CT will likely continue its primary role in the management of these injuries in the foreseeable future. The spectrum of imaging features of cranial gunshot injuries is vast, because they encompass all of the findings encountered in closed head injury in addition to the wide variety of problems associated with penetration. Thus, only a brief summary of the many varied aspects of this complex problem is presented here as a review of the more salient issues.

References
[1] FBI. The Uniform Crime Reports (UCR) 2000. [2] National Center for Health Statistics. 1997. [3] Achram M, Rizk G, Haddad FS. Angiographic aspects of traumatic intracranial aneurysms following war injuries. Br J Radiol 1980;53:1144 9. [4] Haddad FS. Nature and management of penetrating

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248 head injuries during the civil war in Lebanon. Can J Surg 1978;21:233 40. Haddad FS, Haddad GF, Taha J. Traumatic intracranial aneurysms caused by missiles: their presentation and management. Neurosurgery 1991;28:1 7. Hammon WM. Analysis of 2187 consecutive penetrating wounds of the brain from Vietnam. J Neurosurg 1971;34:127 31. Carey ME. Experimental missile wounding of the brain. Neurosurg Clin N Am 1995;6:629 642. Clark WC, Muhlbauer MS, Watridge CB, et al. Analysis of 76 civilian craniocerebral gunshot wounds. J Neurosurg 1986;65:9 14. Kaufman HH. Treatment of civilian gunshot wounds to the head. Neurosurg Clin North Am 1991;2: 387 97. Rosenwasser RH, Andrews DW, Jimenez DF. Penetrating craniocerebral trauma. Surg Clin North Am 1991;71:305 16. Siccardi D, Cavaliere R, Pau A, et al. Penetrating craniocerebral missile injuries in civilians: a retrospective analysis of 314 cases. Surg Neurol 1991; 35:455 60. Copley IB. Cranial tangential gunshot wounds. Br J Neurosurg 1991;5:43 53. Crockard HA, Brown FD, Calica AB, et al. Physiological consequences of experimental cerebral missile injury and use of data analysis to predict survival. J Neurosurg 1977;46:784 94. Kirkpatrick JB, Di Maio V. Civilian gunshot wounds of the brain. J Neurosurg 1978;49:185 98. Gean AD. Imaging of head trauma. New York: Raven Press; 1994. Alberico RA, Loud P, Pollina J, et al. Thick-section reformatting of thinly collimated helical CT for reduction of skull base-related artifacts. Am J Roentgen 2000;175:1361 6. Schwartz RB. Helical (Spiral) CT in neuroradiologic diagnosis. Radiol Clin North Am 1995;33:981 95. Ravereby M. Analysis of long-range bullet entrance holes by atomic absorption spectrophotometry and scanning electron microscopy. J Foren Sci 1982;27: 92 112. Becker DF, Gade GF, Young HF, et al. Diagnosis and treatment of head injury in adults. In: Youmans JR, editor. Neurological surgery. 3rd edition. Philadelphia: WB Saunders; 1990. p. 2017 48. Bashir EF, Cybulski GR, Chaudhri K, et al. Magnetic resonance imaging and computed tomography in the evaluation of penetrating gunshot injury of the spine. Spine 1993;18:772 3. Ebraheim NA, Savolaine ER, Jackson WT, et al. Magnetic resonance imaging in the evaluation of a gunshot wound to the cervical spine. J Orthop Trauma 1989;3:19 22. Parker SJ, Jarvis LJ, Dale RF. Magnetic resonance imaging in the evaluation of a high-velocity gunshot wound to the thigh. Br J Surg 1997;84:1119. Dodd III GD, Budzik Jr RF. Identification of retained

245

[5]

[24]

[6]

[25]

[7] [8]

[26]

[27]

[9]

[28]

[10]

[11]

[29]

[12] [13]

[30]

[31]

[14] [15] [16]

[32]

[33]

[17] [18]

[34]

[35] [36]

[19]

[20]

[37] [38] [39] [40]

[21]

[22]

[41] [42]

[23]

firearm projectiles on plain radiographs. Am J Roentgen 1990;154:471 5. Gieszl R, Williams KD, Drayer BP, et al. Magnetic resonance imaging and ferromagnetic bullets. Assoc Firearm and Tool Mark Exam J 1989;21:595 604. Teitelbaum GP, Yee CA, Van Horn DD, et al. Metallic ballistic fragments: MR imaging safety and artifacts. Radiology 1990;175:855 9. Hess U, Harms J, Schneider A, et al. Assessment of gunshot bullet injuries with the use of magnetic resonance imaging. J Trauma 2000;49:704 9. Mitchell P, Wilkinson ID, Hoggard N, et al. Detection of subarachnoid haemorrhage with magnetic resonance imaging. J Neurol Neurosurg Psych 2001;70: 205 11. Noguchi K, Seto H, Kamisaki Y, et al. Comparison of fluid-attenuated inversion-recovery MR imaging with CT in a simulated model of acute subarachnoid hemorrhage. AJNR Am J Neuroradiol 2000;21:923 7. Snow RB, Zimmerman RD, Gandy SE, et al. Comparison of magnetic resonance imaging and computed tomography in the evaluation of head injury. Neurosurgery 1986;18:45 52. Jinkins JR, Dadsetan MR, Sener RN, et al. Value of acute-phase angiography in the detection of vascular injuries caused by gunshot wounds to the head: analysis of 12 cases. Am J Roentgenol 1992;159:356 68. Aarabi B. Traumatic aneurysms of brain due to high velocity missile head wounds. Neurosurgery 1988;22: 1056 63. Latchaw RE. The role of CT angiography in the long-term management of cerebrovascular dissection. [letter, comment] AJNR Am J Neuroradiol 1998;19: 992 3. Leclerc X, Lucas C, Godefroy O, et al. Helical CT for the follow-up of cervical internal carotid artery dissections. AJNR Am J Neuroradiol 1998;19: 831 7. Nakatsuka M, Mizuno S. Three-dimensional computed tomographic angiography in four patients with dissecting aneurysms of the vertebrobasilar system. Acta Neurochir 2000;142:995 1001. Dodge PR, Meirowsky AM. Tangential wounds of scalp and skull. J Neurosurg 1952;9:472 86. Butler EG, Puckett WO, Harvey EN, et al. Experiments on head wounding by high velocity missiles. J Neurosurg 1945;2:358 63. Hopkinson DAW, Marshall TK. Firearm injuries. Br J Surg 1967;54:344 53. Woodburne RT. Essentials of human anatomy. 6th edition. New York: Oxford University Press; 1978. Gurdjian ES, Webster JE, Lissner HR. The mechanisms of skull fracture. J Neurosurg 1950;7:106 14. Harsh III GR, Harsh IV GR. Penetrating wounds of the head. In: Wilkins RH, Rengachary SS, editors. Neurosurgery. New York: McGraw-Hill; 1985. p. 1670 8. Tsai FY, Teal JS, Hieshima GB. Neuroradiology of head trauma. Baltimore: University Park Press; 1984. Crone KR, Lee KS, Moody DM, et al. Superior sagit-

246

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248 tal sinus air after penetrating craniocerebral trauma. Surg Neurol 1986;25:276 8. Shoung HM, Sichez JP, Pertuiset B. The early prognosis of craniocerebral gunshot wounds in civilian practice as an aid to the choice of treatment. A series of 56 cases studied by computerized tomography. Acta Neurochir 1985;74:27 30. Barnett JC, Meirowsky AM. Intracranial hematomas associated with penetrating wounds of the brain. J Neurosurg 1954;12:34 8. Stone JL, Lichtor T, Fitzgerald LF, et al. Civilian cases of tangential gunshot wounds to the head. J Trauma 1996;40:57 60. Miner ME, Cabrera JA, Ford E, et al. Intracranial penetration due to BB air rifle injuries. Neurosurgery 1986;19:952 4. Selden BS, Goodman JM, Cordell W, et al. Outcome of self-inflicted gunshot wounds of the brain. Ann Emerg Med 1988;17:247 53. Sherman WD, Apuzzo MLJ, Heiden JS, et al. Gunshot wounds to the braina civilian experience. West J Med 1980;132:99 105. Goodnight SH, Kenoyer G, Rapaport SI, et al. Defibrination after brain tissue destruction. A serious complication of head injury. N Engl J Med 1974; 290:1043. Kaufman HH, Moake JL, Olson JD, et al. Delayed and recurrent intracranial hematomas related to disseminated intravascular clotting and fibrinolysis in head injury. Neurosurgery 1980;7:445 9. Levy ML, Rezai A, Masri LS, et al. The significance of SAH after penetrating craniocerebral injury: correlations with angiography and outcome in a civilian population. Neurosurgery 1993;32:532 40. Amato JJ, Rich NM. Temporary cavity effects in blood vessel injury by high velocity missiles. J Cardiovasc Surg 1972;13:147 55. Freytag E. Autopsy findings in head injuries from firearms. Statistical evaluation of 254 cases. Arch Pathol 1963;76:215 25. Freytag E, Vogel FS. The anatomy of head trauma. Pathol Annu 1971;6:321 36. Matson DD, Wokin J. Hematomas associated with penetrating wounds of the brain. J Neurosurg 1946; 3:45 53. Brandvold B, Levi L, Feinsod M, et al. Penetrating craniocerebral injuries in the Israeli involvement in the Lebanese conflict, 1982 1985. Analysis of a less aggressive surgical approach. J Neurosurg 1990;72: 15 21. Ninchoji T, Uemura K, Shimoyama I, et al. Traumatic intracerebral haematomas of delayed onset. Acta Neurochir 1984;71:69 90. Tanaka T, Sakai T, Uemura K, et al. MR imaging as predictor of delayed posttraumatic cerebral hemorrhage. J Neurosurg 1988;69:203 9. Strich SJ. Shearing of nerve fibers as a cause of brain damage due to head injury. A pathological study of 20 cases. Lancet 1961;2:443 8. [60] Smith DR, Ducker TB, Kempe LG. Experimental in vivo microcirculatory dynamics in brain trauma. J Neurosurg 1969;30:664 72. [61] Alsop DC, Murai H, Detre JA, et al. Detection of acute pathologic changes following experimental traumatic brain injury using diffusion-weighted magnetic resonance imaging. J Neurotrauma 1996;13: 515 21. [62] Zimmerman RA, Bilaniak LT, Genarelli TA. Computed tomography of shearing injuries of the cerebral white matter. Radiology 1978;174:393 6. [63] Liu AY, Maldjian JA, Bagley LJ, et al. Traumatic brain injury: diffusion-weighted MR imaging findings. AJNR Am J Neuroradiol 2000;21:808 9. [64] Tsai FY, Huprich JE, Gardner FC, et al. Diagnostic and prognostic implications of computed tomography of head trauma. J Comput Assist Tomogr 1978;2: 323 31. [65] Fitzek C, Tintera J, Muller-Forell W, et al. Differentiation of recent and old cerebral infarcts by diffusionweighted MRI. Neuroradiology 1998;40:778 82. [66] Moseley ME, Cohen Y, Mintorovitch J, et al. Early detection of regional cerebral ischemia in cats: comparison of diffusion- and T2-weighted MRI and spectroscopy. Magn Reson Med 1990;14:330 46. [67] Bruce DA, Alavi A, Bilaniuk L, et al. Diffuse cerebral swelling following head injuries in children: the syndrome of malignant brain edema. J Neurosurg 1981;54:170 8. [68] Cooper PR, Maravilla K, Cone J. Computerized tomographic scan and gunshot wounds of the head: Indications and radiographic findings. Neurosurgery 1979;4:373 80. [69] Miller JD, Gudeman SK, Kishore PS, et al. Computed tomography in brain edema due to trauma. Adv Neurol 1980;28:413 22. [70] Carey ME, Sarna GS, Farrell JB. Brain edema following an experimental missile wound to the brain. J Neurotrauma 1990;7:13. [71] Toutant SM, Klauber MR, Marshall LF, et al. Absent or compressed basal cisterns on first CT scan: ominous predictors of outcome in severe head injury. J Neurosurg 1984;61:691 4. [72] Weinstein JD, Langfitt TW, Bruno L, et al. Experimental study of patterns of brain distortion and ischemia produced by an intracranial mass. J Neurosurg 1968;28:513 21. [73] Weiss GH, Salazar AM, Vance SC, et al. Predicting posttraumatic epilepsy in penetrating head injury. Arch Neurol 1986;43:771 3. [74] DesChamps Jr GT, Morano JU. Intracranial bullet migrationa sign of brain abscess: case report. J Trauma 1991;31:293 5. [75] Wood EH. The diagnostic significance of change in position of metallic foreign bodies in brain abscess. AJR AM J Roentgen 1948;59:52 8. [76] Martin J, Campbell EH. Early complications following penetrating wounds of the skull. J Neurosurg 1946;3:58 73.

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54] [55]

[56]

[57]

[58]

[59]

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248 [77] Medina M, Melcarne A, Ettorre F, et al. Clinical and neuroradiological correlations in a patient with a wandering retained air gun pellet in the brain. Surg Neurol 1992;38:441 4. [78] Milhorat TH, Elowitz EH, Johnson RW, et al. Spontaneous movement of bullets in the brain. Neurosurgery 1993;32:140 3. [79] Hagan RE. Early complications following penetrating wounds of the brain. J Neurosurg 1971;34:132 41. [80] Taha JM, Haddad FS, Brown JA. Intracranial infection after missile injuries to the brain: report of 30 cases from the Lebanese conflict. Neurosurgery 1991; 29:864 8. [81] Drew JH, Fager CA. Delayed brain abscess in relation to retained intracranial foreign bodies. J Neurosurg 1954;11:386. [82] Ahmadi J, Destian S, Apuzzo ML, et al. Cystic fluid in craniopharyngiomas: MR imaging and quantitative analysis. Radiology 1992;182:783 5. [83] Ahmadi J, Savabi F, Apuzzo ML, et al. Magnetic resonance imaging and quantitative analysis of intracranial cystic lesions: surgical implication. Neurosurgery 1994;35:199 207. [84] Baum PA, Dillon WP. Utility of magnetic resonance imaging in the detection of subdural empyema. Ann Otol Rhinol Laryngol 1992;101:876 8. [85] Rich PM, Deasy NP, Jarosz JM. Intracranial dural empyema. Br J Radiol 2000;73:1329 36. [86] Takamura Y, Uede T, Igarashi K, et al. Magnetic resonance imaging of supratentorial and parafalcial empyema. No Shinkei Geka 1995;23:61 4. [87] Singer MB, Atlas SW, Drayer BP. Subarachnoid space disease: diagnosis with fluid-attenuated inversion-recovery MR imaging and comparison with gadolinium-enhanced spin-echo MR imaging blinded reader study. Radiology 1998;208:417 22. [88] Britt RH, Enzmann DR. Clinical stages of human brain abscesses on serial CT scans after contrast infusion. Computerized tomographic, neuropathological, and clinical correlations. J Neurosurg 1983;59: 972 89. [89] Enzmann DR, Britt RH, Placone Jr RC. Staging of human brain abscess by computed tomography. Radiology 1983;146:70. [90] Rish BL, Caveness WF, Dillon JD, et al. Analysis of brain abscess after penetrating craniocerebral injuries in Vietnam. Neurosurgery 1981;9:535 41. [91] Ramsay DW, Aslam M, Cherryman GR. Diffusionweighted imaging of cerebral abscess and subdural empyema. AJNR Am J Neuroradiol 2000;21:1172. [92] Meirowski AM, Caveness WF, Dillon JD, et al. Cerebrospinal fluid fistulas complicating missile wounds of the brain. J Neurosurg 1981;54:44. [93] Aarabi B. Management of traumatic aneurysms caused by high-velocity missile head wounds. Neurosurg Clin N Am 1995;6:775 97. [94] Kennedy F, Gonzalez P, Dang C, et al. The Glasgow Coma Scale and prognosis in gunshot wounds to the brain. J Trauma 1993;35:75 7.

247

[95] Grahm TW, Williams FC, Harrington T, et al. Civilian gunshot wounds to the head: a prospective study. Neurosurgery 1990;27:696 700. [96] Kaufman HH, Makela ME, Lee KF, et al. Gunshot wounds to the head: a perspective. Neurosurgery 1986;18:689 95. [97] Nagib MG, Rockswold GL, Sherman RS, et al. Civilian gunshot wounds to the brain: prognosis and management. Neurosurgery 1986;18:533 7. [98] Narayan RK, Greenberg RP, Miller JD, et al. Improved confidence of outcome prediction in severe head injury. A comparative analysis of the clinical examination, multimodality evoked potentials. CT scanning and intracranial pressure. J Neurosurg 1981;54: 751 62. [99] Kline DG, LeBlanc HJ. Survival following GSW of the pons: neuroanatomic considerations. Case report. J Neurosurg 1971;35:342 7. [100] Inao S, Kuchiwaki H, Kanaiwa H, et al. Magnetic resonance imaging assessment of brainstem distortion associated with a supratentorial mass. J Neurol Neurosurg Psychiatry 1993;56:280 5. [101] Rish BL, Dillon JD, Weiss GH. Mortality following penetrating craniocerebral injuries. An analysis of the deaths in the Vietnam Head Injury Registry population. J Neurosurg 1983;59:775 80. [102] Eisenberg HM, Gary Jr HE, Aldrich EF, et al. Initial CT findings in 753 patients with severe head injury. J Neurosurg 1990;73:688 98. [103] Abe T, Nakamura N, Sugishita M, et al. Partial disconnection syndrome following penetrating stab wound of the brain. Eur Neurol 1986;25:233 9. [104] Bradley Jr WG. Magnetic resonance imaging in the central nervous system: comparison with computed tomography. In: Kressel HY, editor. Magnetic resonance annual. New York: Raven Press; 1986.

Further Readings
Crow W. Aspects of neuroradiology of head injury. Neurosurg Clin North Am 1991;2:321 39. Cushing H. A study of a series of wounds involving the brain and its enveloping structures. Br J Surg 1918;5:558 684. Davis PL, Crooks L, Arakawa M, et al. Potential hazards in NMR imaging: heating effects of changing magnetic fields and RF fields on small metallic implants. Am J Roentgenol 1981;137:857 60. Grafman J, Salazar AM, Weingartner H, et al. Isolated impairment of memory following a penetrating lesion of the fornix cerebri. Arch Neurol 1985;42:1162 8. Graham DI, Adams JH, Doyle D. Ischaemic brain damage in fatal non-missile head injuries. J Neurol Sci 1978;39:213 34.

248

P.E. Kim et al / Neuroimag Clin N Am 12 (2002) 229248 Oliver C, Kabala J. Air gun pellet injuries: the safety of MR imaging. Clin Radiol 1997;52:299 300. Shellock FG, Curtis JS. MR imaging and biomedical implants, materials, and devices: an updated review. Radiology 1991;180:541 50. Taoka T, Yuh WT, White ML, et al. Sulcal hyperintensity on fluid-attenuated inversion recovery MR images in patients without apparent cerebrospinal fluid abnormality. AJR Am J Roentgenol 2001;176:519 24. Unger JM, Gentry LR, Grossman JE. Sphenoid fractures: prevalence, sites, and significance. Radiology 1990;175: 175 80. van der Sande JJ, Veltkamp JJ, Boekhout-Mussert RJ, et al. Hemostasis and computerized tomography in head injury. Their relationship to clinical features. J Neurosurg 1981; 55:718 24. Wilson J, Wiedmann KD, Hadley DM, et al. Early and late magnetic resonance imaging and neurophysiological outcome after head injury. J Neurol Neurosurg Psychiatry 1988;51:391 6. Yashon D, Jane JA, White RJ. Arteriographic observations in craniocerebral bullet wounds. J Trauma 1968;8:238 55.

Kanal E, Shellock FG, Talagala L. Safety considerations in MR imaging. Radiology 1990;176:593 606. Kaufman HH, Levin HS, High Jr WM, et al. Neurobehavioral outcome after gunshot wounds to the head in adult civilians and children. Neurosurgery 1985;16: 754 9. Leaver RC, Loeser JD. Lhermittes phenomenon after high velocity missile injuries of the brain. J Neurosurg 1971;34:159. Levy ML, Masri LS, Levy KM, et al. Penetrating craniocerebral injury resultant from gunshot wounds: Gang-related injury in children and adolescents. Neurosurgery 1993;33: 1018 25. Ludlow CL, Rosenberg J, Salazar A, et al. Site of penetrating brain lesions causing chronic acquired stuttering. Ann Neurol 1987;22:60 6. Meirowski AM, Rish BL, Mohr JP, et al. Definitive care of cerebral missile injuries crossing the midline. Milit Med 1980;145:246 50. Okuchi K, Nagata K, Fujioka M, et al. A case with traumatic internal carotid artery dissection in which the extent of the pseudolumen was defined by MR angiography. Neurol Surg 1999;27:889 94.

Neuroimag Clin N Am 12 (2002) 249 269

Traumatic vascular injuries and their management

Donald W. Larsen, MD, MBA*
Department of Neurological Surgery and Radiology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA Department of Interventional Neuroradiology, Los Angeles County and University of Southern California Medical Center, Los Angeles, CA 90033, USA

Extracranial vascular trauma Traumatic vascular injury to the extracranial circulation appears to relatively underdiagnosed, in part because of the frequent coexistence of traumatic brain injuries that can obscure the diagnosis. Management of these cases can be somewhat challenging, especially when multiple traumatic injuries are present in critically ill patients, and the use of anticoagulants may be contraindicated. Traditionally, medical management aimed at stroke prevention included antiplatelet or anticoagulant agents [1 4]. In recent years, groups have reported on the use of endovascular techniques, including stenting to correct dissections of the carotid and vertebral arteries. These techniques may be the only alternative in those critically injured patients, when more conservative medical management is contraindicated. Injuries to the extracranial circulation can be subdivided into (1) those injuries incurred as a result of blunt or penetrating trauma to the carotid and vertebral arteries, or (2) those injuries in the face involving branches of the external carotid artery. Blunt neck injuries Blunt injuries to the carotid and vertebral arteries are rare and present late with devastating strokes. The diagnosis is often delayed and usually occurs only after

* Department of Neurological Surgery and Radiology, Keck School of Medicine, University of Southern California, 1510 San Pablo Street, Suite 268, Los Angeles, CA 90033. E-mail address: dwlarsen@neurosolutions.org (D.W. Larsen).

a stroke, resulting in neurologic morbidity rates of 40% to 80% and mortality rates of 5% to 40% [5 7]. Recent literature shows an increase in the frequency of blunt carotid and vertebral artery injury; however, there does not appear to be a significant improvement in the diagnosis before a stroke has occurred. Identifying those patients who may be at risk for these injuries remains paramount in the evaluation of the trauma patient if strokes are to be prevented. Blunt carotid injuries (BCI) have been historically considered rare, yet are associated with possibly devastating sequelae [2,3,5 37]. The true incidence of traumatic injury to the extracranial cerebral vessels because of blunt trauma is unknown. Although traditionally the reported incidence of traumatic carotid injury has been less than 0.1% [6], at centers where the diagnosis is aggressively sought, the incidence ranges from 0.33% to 0.45% [2,38]. Recently, Fabian and colleagues, using an aggressive diagnostic approach, found BCI to occur in 0.67% of patients after motor vehicle accidents [2]. The true incidence, however, cannot be known, because of the inability to quantitate all asymptomatic or missed injuries. Estimated mortality rates are reported to be 20% to 40%, with permanent neurologic sequelae present in 33% to 75% of survivors [2,5,6,18,20,25,33,37,39]. Blunt extracranial vascular injuries occur most commonly from motor vehicle accidents, mostly involving young patients. The most common mechanism of injury postulated includes injury secondary to stretching of the artery because of rapid deceleration. The internal carotid artery is thought to stretch over the lateral masses of the third and fourth cervical vertebrae, producing an intimal tear, commonly in the distal internal carotid artery. This can result in a dissection in the wall of the artery for a variable

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 6 - 0

250

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

distance [21]. Some tears probably heal spontaneously, but others lead to dissection, with or without pseudoaneurysm formation, and some thrombose. Other mechanisms reported include direct blows to the head, neck, or face; strangulation injuries; basilar skull fractures resulting in arterial transaction; blunt intraoral trauma hyperflexion of the neck, causing compression of the artery between the mandible and the cervical spine; and hyperflexion of the neck, causing compression of the artery between the mandible and the cervical spine [21,29]. Various types of vascular pathologies have been identified and include (1) occlusion or thrombosis, (2) intimal flap or dissection, (3) complete transection, (4) pseudoaneurysm, (5) carotid cavernous fistula, or (6) a combination of injuries [21]. Bilateral dissections have been reported to occur in up to 45% of patients [29]. Symptoms from BCIs may develop in hours to weeks after the injury [23]. A delay in recognition is common, because the vascular injury itself is frequently asymptomatic, which delays the diagnosis until a stroke develops. Other factors that mask the neurologic deficits of stroke, such as associated head trauma and alcohol or drug intoxication, may contribute to a delay in diagnosis. On physical examination, specific signs of BCI are frequently absent. Unilateral Horners syndrome, neck hematoma, bruit, pulsatile neck mass, and palpable thrill, if present, should alert the clinician to the possibility of BCI and prompt further imaging interrogation. In one series it has been reported that up to 50% of patients with BCI had no external signs of neck trauma [37], attesting to the need for a low index of suspicion and an aggressive diagnostic approach to identify these lesions in the asymptomatic or obtunded patient. Extracranial vertebral artery dissection, like carotid dissection, can occur spontaneously [40,41] and after blunt trauma [42,43] to the head and neck. Causes of nonpenetrating trauma to the vertebral artery resulting in dissection include chiropractic manipulation, tennis, seat belt use, yoga, head banging, and kickboxing [44]. The majority of extracranial vertebral dissections present with acute or delayed vertebrobasilar ischemia or infarction. If the contralateral vertebral artery is large enough and patent, or if the patient has an adequate circle of Willis collateral circulation through the posterior communicating arteries, a vertebral occlusion may be tolerated without evidence of vertebrobasilar insuffiency; however, embolic ischemic events are possible, especially if an anticoagulant and or antiplatelet regimen have not been instituted. The natural history of vertebral artery dissection includes spontaneous healing, distal embolization

secondary to areas of stasis and intimal exposure, and formation of pseudoaneurysm [45]. Contrast cervical and cerebral angiography is the gold standard imaging modality for the diagnosis of blunt injuries to the carotid and vertebral arteries. A complete assessment of the status of the brachiocephalic arteries, including their origins from the aortic arch to the circle of Willis, is possible. Besides identifying intercranial arterial injuries, contrast angiography can asses the status of the intercranial collateral circulation through circle of Willis and pial collateral arteries. This information is vital to determine the need for various revascularization techniques. Extracranial carotid artery dissection is seen angiographically as an irregular narrowing of the artery, which can less characteristically be associated with tapered occlusions, intimal flaps, dissecting aneurysms, or extraluminal pouches. Occasionally, filling can be seen to traverse the stenotic dissection flap and create a string sign analogous to the angiographic finding of near occlusive atherosclerotic disease of the carotid artery (Fig. 1). Computed tomography (CT) is useful to assess the brain for associated injury because of direct trauma (hematomas, contusions, skull fractures) as well as identifying ischemic or embolic sequelae of more proximal vascular injury in the brachiocephalic arteries (infarction). CT angiography (CTA) of the head and neck may be the most useful diagnostic resource to the clinician in the acute setting, when there may be a delay in availability of contrast angiography [46]. One diagnostic algorithm could include CTA of the head and neck at the time of performance of a non-contrast head CT performed for assessment of head trauma. In a recent report, the authors doubled their rate of early diagnosis of BCI with the use of contrast-enhanced spiral CTA of the neck [47]. However, in another recent report, the authors were able to improve their rate of early diagnosis without the need for a special screening protocol or technique [21]. Magnetic resonance angiography (MRA) has been found by Levy et al to accurately demonstrate carotid dissection with a sensitivity of 95% and specificity of 99% [48]. Although the use of magnetic resonance imaging (MRI) [49 54] and MRA [55] may be helpful to identify these lesions, routine use of these modalities in the acute setting is limited now because of lack of equipment, increased cost, and the need to monitor and possibly resuscitate the critically ill patient. Duplex ultrasound has been found useful by some clinicians as a rapid noninvasive means for detecting carotid arterial injuries in the acute setting [56 62]; however, it is criticized for failing to visualize

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

251

fistulae, vessel transections, intimal flaps, dissections, and occlusions to the carotid and or vertebral arteries. Clinical presentation Clinically, these lesions are less occult than those due to blunt trauma. Signs of external injury are usually evident with or without frank hemorrhage or hematoma formation. Treatment The management of carotid and vertebral artery injuries depends on the type and location of injury, associated injuries, neurologic status of the patient, and the natural history of the injury. Treatment paradigms should be directed at preventing cerebral ischemic as well as potentially hemorrhagic lesions such as pseudoaneurysms. It is important to determine if the patient is at risk of future thromboembolic events from the lesion, or in the case of traumatic occlusions or stenoses, whether he is at risk of hemispheric ischemia. Signs and symptoms of ischemia may be evident, especially if neurologic symptoms are present without evidence of stroke or contusion on CT or MR. Diffusion and perfusion MR imaging may be useful to determine the presence of an ischemic cervical lesion that may require urgent revascularization. Transcranial Doppler sonography with emboli counting may be indicated to document those intimal flaps, pseudoaneurysms, or dissections [64] that are actively generating emboli. In most situations, an occluded artery can be well tolerated if there is adequate collateral circulation to the brain; however, in the event of an incomplete circle of Willis, contralateral occlusions or stenoses, or if coexistent atherosclerotic lesions in collateral circulations are present, there may be a need for surgical or endovascular revascularization. If a traumatically occluded artery is not tolerated because of an ischemia, then surgical bypass procedures should be considered. If a traumatic dissection results in a stenosis resulting an ischemic condition, then surgical bypass, surgical arterial reconstruction, or endovascular revascularization may be indicated. Endovascular stents or stent-graft, although not currently approved for this indication, can be helpful to provide urgent revascularization at the time of the diagnostic angiogram (Fig. 2). No large series or long-term data are available now regarding the safety and efficacy of the off-label use of these devices (Fig. 3). If the patient is able to tolerate the occluded artery without an ischemic threat, conservative medical management aimed at preventing embolic events, including the use of antiplatelet and/or anticoagulant medications, may be instituted if not contraindicated by the pres-

Fig. 1. A 54-year-old male who sustained blunt cervical trauma. An angiographic string sign is demonstrated (arrowheads). The reconstituted cavernous carotid artery fills through the small lumen of the partially dissected internal carotid artery. This patient was treated conservatively with aspirin therapy without evidence of further neurologic sequelae.

the distal internal carotid artery, where most injuries occur. Davis et al noted that 33% of occlusions noted on ultrasonography were actually patent (though narrowed) [6]. Kraus et al reviewed reports from several authors that overall, ultrasound demonstrated a 92% sensitivity for detection of blunt carotid injuries, and that approximately 8% of carotid injuries were missed [21]. They conclude that since tight stenoses may be misdiagnosed as occlusions, they cannot recommend ultrasonography as a sole diagnostic method. Carotid duplex ultrasound may show reduced or absent distal carotid flow at the level of the bifurcation, but is rarely able to demonstrate tapering luminal stenoses or a double lumen [63]. Penetrating neck injuries Penetrating injuries to the neck from gunshot wounds, stab wounds, and other lacerations can result in the formation of pseudoaneurysms, arteriovenous

252

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

ence of associated injuries. Fabian et al [2] and other authors [6,23] have reported favorable results with anticoagulation as the primary therapeutic modality for these patients. The rationale for systemic anticoagulation has been (1) to minimize clot formation at the site of intimal injury; (2) to decrease further propagation of clot that has formed, allowing the internal fibrinolytic system to dissolve the clot, and (3) to prevent embolization of clot from the sac of pseudoaneurysms [4]. Fabian et al [2] report that

over time, 62% of nonocclusive dissections resolve with anticoagulation. Also, 29% of initial dissections develop pseudoaneurysms on follow-up angiography. When pseudoaneurysms are present, they are unlikely to resolve. Mokri found that of 14 pseudoaneurysms identified, only 3 decreased in size or resolved [28]. It was also noted that many of these were a source of emboli, producing neurologic injury several weeks and even years after the injury. These findings suggest that patients with pseudoaneurysms

Fig. 2. A 35-year-old male, stabbed in zone three of the neck on the left, behind the angle of the mandible. Initial neurologic examination was negative. Surgical exploration revealed an active, pulsatile hemorrhage from active bleeding from small arteries of the muscle and skin. The following day the patient became lethargic, globally aphasic, and hemiparetic on the right. A CT scan of the brain demonstrated a large infarct in the distribution of the left middle cerebral artery. (A) Angiography demonstrates complete occlusion of the internal carotid artery (arrow). This is a characteristic sign of a traumatic carotid dissection. Because of the established cerebral infarction, acute revascularization of the internal carotid artery occlusion was considered contraindicated. The patient was started on daily aspirin therapy. Over the next several days, his neurologic deficits improved considerably. (B) Follow-up angiography demonstrates development of a 12-mm pseudoaneurysm (arrow) with recanalization of the previous carotid occlusion. The pseudoaneurysm was felt to pose a threat of distal embolization and local hemorrhage. For this reason, occlusion of the parent vessel was considered. Due to the patients poor collateral supply to the brain, there was impetus to preserve vessel patency. Direct surgical repair was considered challenging due to the location of this lesion near the skull base. The patient was referred for endovascular treatment two days later. (C,D) Angiogram at that time demonstrates enlargement of the pseudoaneurysm (C) with filling of multiple lobules of the pseudoaneurysm (arrows), which becomes more evident on late venous phases (D). (E) A 6 20 mm Wallgraft was placed across the origin of the pseudoaneurysm in the internal artery. The stent margins are indicated (arrowheads), and permeation through the wall of the Wallgraft material is demonstrated (arrow). (F) This permeation ceased 15 minutes later. (G) Final angiographic result, unsubtracted, demonstrates the stent graft to best advantage with stent margins outlined (arrowheads).

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

253

Fig. 2 (continued ).

254

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

access to lesions at or near the skull base, (2) avoidance of general anesthesia, and (3) preservation of the parent artery without the need for a bypass surgical procedure. The attractive aspect of endovascular therapy is that it provides an alternative treatment for patients for whom conservative medical management may not be the best option. Gomez et al [63] recently reported the following subpopulations of patients who may be good candidates for endovascular therapy: Patients who have symptomatic dissections and who cannot be anticoagulated because of concurrent injuries Patients who have symptomatic dissections despite anticoagulation, particularly if they are surgically inaccessible Patients who have asymptomatic dissection but little neurologic vascular reserve (eg, poor cerebral collateral circulation) Rarely, an occluded artery can demonstrate subsequent self-recanalization, which on follow-up angiography reveals an underlying intimal flap, stenotic dissection, or pseudoaneurysm. One reported series revealed that only one of five patients with an occluded internal carotid artery recanalized after treatment with long-term anticoagulation [2]. Another follow-up angiography study revealed no recanalization in five BCI occlusions [28]. Fabian et al also reported that five of eleven patients with unilateral, occluded internal carotid arteries had good neurologic outcomes, and two had moderate neurologic outcomes [2]. These patients had satisfactory crosscirculation through relatively intact circles of Willis. The authors suggest that those patients with bad outcomes probably had borderline cross filling, which was compromised because of blood loss or hypotension resulting in cerebral ischemic insults. Facial trauma Facial vascular injuries usually result from the penetrating trauma of stab or gunshot wounds, and less often from blunt trauma, accompanied by multiple facial fractures. Clinical presentation Facial vascular injuries consist of occlusion, fistulas, pseudoaneurysms, or gross extravasation. Arterial occlusions are usually well tolerated and are probably asymptomatic in many blunt facial injuries, most likely because of the rich collateral circulation present in the face. Occlusion of an entire external

Fig. 2 (continued ).

should receive anticoagulants until they resolve. If they fail to resolve, or enlarge on close angiographic follow-up, surgical or endovascular eradication are indicated. Mokri et al [29] reported the surgical repair of nine post-traumatic pseudoaneurysms in eight patients. Surgical intervention was selected because all were producing symptoms related to embolic ischemia. Excision with reversed saphenous vein grafting was performed in five, and four required extracranialintracranial bypass. There are multiple case reports and small series describing the endovascular treatment of pseudoaneurysms and traumatic arteriovenous fistulae of the brachiocephalic arteries [38,44,65 92]. Earlier reports of endovascular treatment consisted of balloon or coil occlusion of the pseudoaneurysm and sacrifice of the parent artery. In recent years, endovascular techniques have allowed occlusion of the offending pseudoaneurysm or post-traumatic fistula, with preservation of the parent artery. These authors have reported the use of a combination of endovascular techniques, including the use of stents alone, stent assisted coiling, and covered stent graft treatment of these post-traumatic lesions (Fig. 4). The advantages to the endovascular approach include (1)

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

255

Fig. 3. (A) The Wallgraft (Boston Scientific, Natick, MA) is a flexible Dacron-coated stent. It should be noted that its applications in Figs. 2 and 3 represent an off label use for this device. It is FDA approved for use as a tracheobroncial endoprosthesis. (B) It is deployed over a sheathed catheter and has an outer diameter of 9-French. (Courtesy of Boston Scientific, Natick, MA.)

carotid artery can occur while tissue perfusion is preserved by way of collateral branches of the contralateral internal maxillary artery, facial artery, and ipsilateral vertebral artery. Fistulas may not be immediately evident in the wake of other, more serious injuries. In fact, they may initially be too small to detect, and may require weeks to years to develop a high enough flow rate to be clinically evident. When large enough, these highflow fistulas can produce a thrill or symptomatic bruit. Rarely, they may cause a steal phenomenon, resulting in relative surrounding ischemic changes.

Pseudoaneurysms can also become apparent in a delayed fashion. Treatment Pseudoaneurysms and fistulas are not likely to resolve on their own, and can be well managed with endovascular techniques involving coil embolization, or by surgical eradication. Like pseudoaneurysms of the cervical carotid and vertebral arteries, these lesions can enlarge over time with risk of rupture. Endovascular coil embolization offers a minimally invasive alternative to surgical intervention for these

256

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

lesions. Arteriovenous fistulas should be occluded at the exact site of the fistula. An embolization or sur-

gical ligation performed proximal to the site of the fistula allows recruitment of collateral branches distal

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

257

to the occlusion and keeps the fistula patent. Further attempts to access the lesion from an endovascular approach will be met with difficulty.

Intracranial vascular trauma Though uncommon, traumatic injury to the intracranial arteries may have sudden and devastating consequences. Many of these injuries may present in a delayed fashion, making diagnosis and treatment a clinical challenge. These injuries result from a variety of vascular insults, including those due to closed head trauma, penetrating skull injury, and iatrogenic injuries secondary to surgical interventions. Endovascular therapeutic techniques are frequently the treatment of choice for these lesions, since direct surgical treatment can be limited by inaccessibility, urgency, and the presence of comorbid conditions [93]. Traumatic injuries of the intracranial arteries can be divided into three main categories: dissections, aneurysms, and fistulas. These may present individually, or in combination. Dissections A dissection occurs when an injury to one or more arterial layers allows the force of the blood to separate them along a dissection plane. Intracranial arteries differ from arteries elsewhere in the body in lacking an external elastic lamina and having a thinner adventitia [94]. Most intracranial dissections occur as a result of a defect in the internal elastic lamina [95]. Intracranial dissection that originates within the media or adventitia is less common. Although the intima may be spared, subarachnoid hemorrhage may occur from rupture through the adventitia [96]. This is the type of dissection sustained during severe head injury and implicated in the development of intracranial pseudoaneurysms. With these injuries, the distinction between dissection and pseudoaneurysm is less well defined.

Clinical presentation As with blunt extracranial carotid injuries, the actual incidence of traumatic intracranial dissection is difficult to assess, partly because angiography is no longer routinely performed for brain trauma. Also, many lesions remain asymptomatic or are not identified in those who have not survived. Contrast angiography has been the gold standard test of choice to diagnose these lesions in the past; however, some researchers have shown that MRI and MRA may be helpful as well [49,50]. MRI has the added benefit of identifying associated brain parenchymal injury and indications of intracranial ischemia, with the use of diffusion and perfusion sequences. Although any intracranial artery may be involved, the middle cerebral is most commonly implicated, followed by the basilar artery [97]. The natural history of these lesions is unknown, given their rarity and presence in asymptomatic patients. Treatment Surgical exploration and repair has been advocated by some [94,95], especially for those dissections that involve the posterior circulation. Anticoagulation has not been routinely used for fear of lesion extension and intracranial hemorrhage. Endovascular therapy with coil embolization and subsequent occlusion of the artery itself is usually sufficient to prevent subsequent rupture [73]; however, if the collateral circulation is insufficient to allow endovascular sacrifice of the parent artery, covered stents (stent-grafts) may be helpful to preserve luminal patency and prevent further rupture. Unfortunately these devices are not now commercially available for this indication and no systematic clinical trials have been performed. Aneurysms When there is a disruption of the arterial wall because of trauma, resulting in a dilatation or out

Fig. 4. A 65-year-old female had recently been experiencing episodes of syncope. She had no history of recent trauma; however, she had a significant history of prior chiropractic manipulation. (A) Angiographic evaluation of the right vertebral artery demonstrates extensive rapid filling of draining veins (arrowheads). This represents an arteriovenous fistula, the site of which is demonstrated (arrow). (B) The arteriovenous fistula has created a steal phenomenon best demonstrated during arterial injection of the left vertebral artery. Contrast is seen proceeding in an antegrade direction in the left vertebral artery, which at the vertebrobasilar junction reverses flow (arrows). This fills the site of the fistula (open arrow). The draining veins are demonstrated (arrowhead ). A 6 20 mm Wallgraft is placed in the right vertebral artery, which demonstrates subsequent complete closure of the arteriovenous fistula with preservation of the parent artery and reversal of flow to the brain in the vertebral artery distal to the stent (C). Stent margins are outlined (arrowheads).

258

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

pouching of the arterial lumen, a traumatic intracranial aneurysm can develop [98]. These lesions can be divided into true and false traumatic aneurysms. True traumatic aneurysms occur when the intima, internal elastic lamina, and media are disrupted, leaving only adventitia [97]. False traumatic aneurysms are more common than true traumatic aneurysms, and result when there is disruption of all vessel layers. These lesions are actually hematomas that are contained by surrounding tissue, and are in contact with turbulent arterial blood at the site of the arterial defect [99,100]. As the false traumatic aneurysm matures; it forms a wall from collagenous organization of the periphery of the original hematoma [97]. The term pseudoaneurysm is more commonly used, as the distinction between true and false traumatic aneurysms cannot be made solely by angiography, but depends on a histologic diagnosis [76].

Clinical presentation In large series of intracranial aneurysms, traumatic aneurysms have been found to compose less than 1% of the total [99,101]. Kieck and de Villiers [102] found a 10% incidence of traumatic aneurysms in 109 patients with penetrating head injuries, most of which were stab wounds. Conversely, in 2187 patients with penetrating head traumas (mostly gunshot wounds) sustained in Vietnam, Ferry and Kempe found only two cases of traumatic intracranial aneurysm [103]; however, the most likely cause of traumatic intracranial aneurysm in the civilian population is severe closed head trauma [88,104, 105]. Unlike penetrating injuries, which can result in a pseudoaneurysm in whichever vessel is involved (Fig. 5), traumatic aneurysms from serve closed head trauma have a predilection for the distal cerebral arteries [93].

Fig. 5. (A) A 35-year old male who sustained a gunshot wound to the face. CT scan of the skull base demonstrates metallic shrapnel in the region of the carotid canal (arrow). (B) On lateral right internal carotid angiography, a pseudoaneurysm is demonstrated (arrow). (C) Balloon test occlusion during injection of the left internal carotid artery demonstrates excellent collateral supply across the anterior communicating artery. (D) In addition, during left vertebral artery injection, there is excellent collateral circulation from the posterior circulation to the basilar artery (open arrow) through the posterior communicating artery (arrow), reconstituting the right supraclinoid internal carotid artery (arrowhead ). Because of concerns of embolic sequelae of this pseudoaneurysm and possible enlargement, (E) carotid occlusion with detachable silicone balloons (arrows) was performed.

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

259

Fig. 5 (continued ).

260

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

261

Sudden intracranial hemorrhage is the most common presentation when pseudoaneurysms develop. Rarely, symptoms of mass effect due to an enlarging pseudoaneurysm may precede a rupture. It is common for traumatic intracranial aneurysms to present themselves in a delayed fashion two to eight weeks after a traumatic event [97] (Fig. 6). Pseudoaneurysms of the supraclinoid internal carotid artery are of particular concern, because they can result in lifethreatening epistaxis if they rupture through associated basal skull fractures or into the sphenoid sinus [106 108]. Treatment If traumatic pseudoaneurysms are identified before they rupture, they should be treated, as they are not likely to spontaneously resolve, and may progress to hemorrhage [109]. Treatment may consist of surgical or endovascular therapy, depending on the morphology of the lesion and the presence of possible comorbidities. Lempert et al [79] have described endovascular embolization of the pseudoaneurysms with detachable coils, with preservation of the parent artery. When possible, it is preferable that the pseudoaneurysm be somewhat non-acute, to assure a more mature wall of clot in pseudoaneurysm in which to support the coils. Where endovascular coil occlusion of the pseudoaneurysm is not feasible, parent vessel occlusions with detachable balloons and or coils may be necessary. It is always helpful to perform a balloon test occlusion of the intended artery before definitive permanent occlusion, to demonstrate tolerance to this deconstructive procedure [106,107]. In the future, covered stents (stent-grafts) may be available that can be deployed within the parent artery, across the neck of the pseudoaneurysm, effectively excluding it from the circulation while preserving the patency of the artery. Surgical treatment may consist of carotid ligation in the neck when sufficient collateral circulation is adequate, but may require a vascular bypass when parent artery ligation will not be tolerated. Some lesions may be amenable to direct surgical clipping or resection.

Fistulas Carotid-cavernous fistulas (CCF) are spontaneous or acquired connections between the carotid artery and the cavernous sinus. A direct CCF is created by direct connections between the internal carotid artery and the cavernous sinus, and may occur as result of trauma, a ruptured intracavernous carotid artery aneurysm, collagen deficiency syndromes, arterial dissection, fibromuscular dysplasia, or direct surgical trauma [110 118]. An indirect CCF is usually supplied by dural branches of the external carotid artery, but can be supplied by dural branches of the internal carotid artery. The exact cause of these lesions is unknown; however, factors associated with their development include pregnancy, sinusitis, trauma, surgical procedures, and cavernous sinus thrombosis [119]. A traumatic etiology of indirect CCF is relatively rare and is not discussed at length in this chapter. Clinical presentation Direct CCFs are usually solitary connections between the cavernous carotid artery and the cavernous sinus. Post-traumatic CCFs most often occur as a result of closed head trauma associated with basal skull fracture, but can also occur with penetrating trauma to the head and orbit. The internal carotid artery (ICA) is fixed by a dural attachment between the foramen lacerum and the anterior clinoid process. The shearing forces of severe head trauma, often accompanied by penetrating injury from bony spicules, can cause the ICA to be torn between its points of dural attachment. The laceration of the ICA is usually single, large (2 5 mm), and unilateral. The unusual bilateral traumatic CCF is generally associated with more severe head trauma, and is more commonly fatal [120]. Rarely, a unilateral CCF may present with bilateral orbital symptoms, by supplying the contralateral cavernous sinus with arterialized blood by way of the circular sinus, thus creating an intercavernous communication. Motor vehicle accidents represent the most common cause of traumatic CCFs, followed by trauma from falls and penetrating injuries from other

Fig. 6. A 2-year old female suffered from a head injury due to a motor vehicle accident. On subsequent follow-up CT scans 6 months later, an enlarging pseudoaneurysm was identified. (A) Angiographic assessment, lateral view demonstrates a 12-mm pseudoaneurysm (arrowhead ) arising from the supraclinoid internal carotid artery. There is no identifiable neck. (B) The AP view demonstrates the pseudoaneurysm (arrowhead ). (C) Excellent collateral circulation demonstrated by angiography of the right internal carotid artery demonstrates cross filling across the anterior communicating artery. The previously seen pseudoaneurysm was treated with GDC embolization coils. This angiograpic projection demonstrates a pseudoaneurysm projected over the middle cerebral artery. The middle cerebral artery, however, remained patent.

262

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

causes. The incidence of posttraumatic CCF is higher among males and probably reflects a male predominance for most traumatic types of injury. In children, CCFs are more commonly caused by falls and penetrating injuries [121]. Iatrogenic causes of direct CCFs are less common. These include injury to the ICA during transphenoidal hypophysectomy [118], Fogarty catheter manipulation for carotid angioplasty

[122], thrombendarterectomy [123], trigeminal rhizotomy, and nasopharyngeal biopsy [89]. The clinical presentation of direct CCFs is related to their size, duration, location, adequacy and route of venous drainage, and presence of arterial and venous collateral vessels. The most common presenting symptoms are similar to those of indirect CCFs, and reflect a similar pathophysiology related to the venous

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

263

Fig. 7. A 28-year-old male suffered a motor vehicle accident. He was noted to have a severe left eye proptosis. (A) An MRI scan of the brain demonstrates flow voids on the region of the left cavernous sinus (arrowheads). This represents a large varix of the left cavernous sinus due to a direct carotid-cavernous sinus fistula. (B) The fistula is best demonstrated during angiographic evaluation of the left internal carotid artery. The internal carotid artery (open curved arrow) immediately fills the cavernous sinus and the large varix (arrowheads). The varix then drains into the superior ophthalmic vein (open arrow), as well as filling of the middle cerebral arteries (arrows), consistent with cortical venous drainage. There is no evidence of filling beyond the level of the site of injury in the left internal carotid artery. (C) The injection of the left vertebral artery with imaging over the head, demonstrated on this lateral view with left carotid compression, demonstrates filling across the posterior communicating artery (arrow) with retrograde filling of the supraclinoid carotid artery filling the rent of the left internal carotid artery, demonstrating the jet of contrast into the fistula (arrowhead ). (D) Injection of the right internal carotid artery, AP view, demonstrates excellent filling of the circle of Willis and intracerebral vessels. There is also late filling of contrast into the left cavernous sinus varix (arrowhead ). Due to the excellent collateral circulation, the carotid artery was occluded and trapped above and below the fistula site with detachable silicone balloons.

drainage of the orbit. The superior and inferior ophthalmic veins provide normal venous drainage into the cavernous sinus. The cavernous sinus normally drains through the superior and inferior petrosal sinuses and

through emissary veins to the pterygoid plexus. Reversal of flow through the ophthalmic veins or sphenoparietal sinus is possible if an arteriovenous connection develops in the cavernous sinus, especially

264

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

if normal routes of venous drainage of the cavernous sinus are occluded. Elevated venous pressure in veins draining the orbit may produce orbital venous congestion, transudation of interstitial fluid into the orbit with resultant proptosis, and increased intraocular pressure because of impaired drainage of aqueous humor and secondary glaucoma [124,125]. Elevated venous pressure and increased intraocular pressure can compromise retinal perfusion and result in severely diminished visual acuity [125]. Diminished arterial pressure to the orbit may also result from a steal away from the ophthalmic artery into the fistula through reversal of flow in the supraclinoid carotid artery into the fistula. If severe compromise of visual acuity has developed with loss of light perception, the chances of visual recovery are unlikely. Edema of orbital contents because of impaired orbital venous drainage can result in mechanical limitation of extraocular muscle movement, and can contribute to the development of diplopia. Diplopia, however, may be a direct result of ophthalmoplegia from a cavernous sinus syndrome, caused by mass effect on the cavernous sinus from the fistula, or because of cranial nerve injury by direct trauma. Less frequently, there is reversal of venous drainage into the sphenoparietal sinus, with resultant cerebral cortical venous hypertension. These patients are at risk of intracerebral hemorrhage and should receive emergent treatment. In a review of 155 cases of CCFs, Halbach et al [126] reported thirteen patients (8.4%) with associated intracranial hemorrhage. Four (2.6%) of these patients had fatal subarachnoid hemorrhage (SAH), and three of the four had a varix of the cavernous sinus. Eleven patients, all of whom had cortical venous drainage, were symptomatic with signs and symptoms suggestive of increased intracranial pressure. Four of these eleven also suffered intracerebral hemorrhage. All of these patients had hypoplasia or occlusion of normal cavernous sinus outflow pathways, with diversion of outflow into cortical veins, and all had symptoms that abated following closure of the fistula [126]. Radiographic evaluation Contrast cerebral angiography is the imaging modality of choice to confirm the presence of and plan treatment for CCFs. Although the diagnosis is essentially a clinical one, MRI and CT may be useful in establishing the degree of brain injury. CT is especially useful to identify skull fractures that may compromise the carotid artery lumen and create bony encroachment on the optic canal. MRA can usually identify the presence of direct CCFs since the flow-rate is usually

rapid enough to allow detection; however, the slowerflow indirect CCF may be occult on MRA. Optimum selective cerebral angiography requires the use of high-resolution digital subtraction angiography equipment with rapid filming and roadmapping capability. The initial angiographic evaluation should be tailored to obtain the following information: Size and location of the fistula in the ICA Identification of any associated carotid-cavernous aneurysm that may have ruptured Differentiation of a direct CCF from an indirect CCF (dural arteriovenous fistula [AVF]) Identification and confirmation of patency of outflow pathways of the cavernous sinus Identification of high-risk features such as cortical venous drainage, pseudoaneurysm, and cavernous sinus varix Identification of associated vascular injuries in contralateral ICA and vertebral arteries (collateral circulation) In high-flow direct CCFs, it may not be possible to identify the morphology of the fistula on selective ICA angiography without specific maneuvers to slow flow through the fistula (Fig. 7). The MehringerHieshima maneuver consists of a gentle ipsilateral ICA injection and manual compression of the ipsilateral carotid artery, while filming over the fistula. By eliminating the high-flow arterial input from the ipsilateral carotid artery, the fistula fills at a slower rate, which allows for better delineation of the fistula site. The Heuber maneuver is a similar technique that opacifies the fistula through a patent posterior communicating artery (if one exists) during injection of the dominant vertebral artery, while manually compressing the ipsilateral carotid artery. With these maneuvers, the unusual case of a CCF with more than one tear or complete transection of the ICA may be revealed. Rarely, external carotid angiography demonstrates associated vascular injuries, including facial or scalp fistulas, pseudoaneurysms, or dissections.

Treatment Following a thorough evaluation of the patients condition from a clinical, anatomical, and hemodynamic standpoint, and with a full understanding of the therapeutic options available, including their risks and benefits, a treatment plan may be formulated. In situations where there is rapid visual decline, increasing intracranial pressure, or cortical venous drainage, treatment should be initiated urgently; however, if there are associated multiple injuries, more

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269

265

critical injuries may require medical attention before treating the fistula. The treatment of CCF has evolved over the past 40 years. The early treatment by proximal occlusion or trapping has fallen out of use because of the high incidence of stroke and blindness associated with these therapies, often without complete closure of the fistula [127,128]. Prolo and Hanberry [129] described the use of a fixed balloon to close both the fistula and the carotid artery. Serbinenko [86,130] and Debrun et al [131] described techniques to close fistulas with detachable balloons, often with preservation of the ICA. Large series have proven the effectiveness of transarterial balloon embolization [76,131 133], currently the treatment of choice for this disease. When transarterial routes are unsuccessful, transvenous embolization [128,131,134] is often successful at obliterating the fistula. Rarely, when these transvenous routes are unsuccessful, direct surgical exposure and embolization with copper wire [135], liquid adhesives, or balloons [136] have been performed. Future advances in covered stent technology (stent-graft) will allow development of devices that can bridge the rent in the carotid artery and preserve patency of the parent carotid artery.

Summary Traumatic vascular injury to the intracranial and extracranial circulation can be sequelae of blunt, penetrating, or iatrogenic insults to the head, face, or neck. Treatment options include conservative medical management, or more invasive surgical or endovascular therapy. The appropriate treatment depends on the risk-benefit ratio of each option considering the natural history of each. Injuries include mild intimal irregularities, intimal flaps, pseudoaneurysms, fistulas, and occlusions. Need for treatment is partly determined by the collateral circulation to the brain, and the degree to which the lesion is thrombogenic. Advances in endovascular devices and techniques provide us with less invasive alternatives to surgery intervention or allow the interventionalist to treat lesions not treatable by any other modality.

References
[1] Desfontaines P, Despland PA. Dissection of the internal carotid artery: aetiology, symptomatology, clinical and neurosonological follow-up, and treatment in 60 consecutive cases. Acta Neurol Belg 1995;95:226 34.

[2] Fabian TC, Patton JH, Jr., Croce MA, et al. Blunt carotid injury. Importance of early diagnosis and anticoagulant therapy. Ann Surg 1996;223:513 522 [discussion: 522 5]. [3] Steiger HJ. Treatment of traumatic carotid dissection. Neurochirurgia (Stuttg) 1988;31:128 33. [4] Watridge CB, Muhlbauer MS, Lowery RD. Traumatic carotid artery dissection: diagnosis and treatment. J Neurosurg 1989;71:854 7. [5] Cogbill TH, Moore EE, Meissner M, et al. The spectrum of blunt injury to the carotid artery: a multicenter perspective. J Trauma 1994;37:473 9. [6] Davis JM, Holbrook TL, Hoyt DB. Blunt carotid artery dissection: Incidence, associated injuries, screening, and treatment. J Trauma 1990;30:1514 7. [7] Parikh AA, Luchette FA, Valente JF, et al. Blunt carotid artery injuries. J Am Coll Surg 1997;185:80 6. [8] Berlot G, Nicolazzi G, Viviani M, et al. Traumatic blunt carotid injury: clinical experience and review of the literature. Eur J Emerg Med 1996;3:36 42. [9] Berne JD, Norwood SH, McAuley CE, et al. The high morbidity of blunt cerebrovascular injury in an unscreened population: more evidence of the need for mandatory screening protocols. J Am Coll Surg 2001;192:314 21. [10] Biffl WL, Moore EE, Elliott JP, et al. Blunt cerebrovascular injuries. Curr Probl Surg 1999;36:505 99. [11] Biffl WL, Moore EE, Elliott JP, et al. The devastating potential of blunt vertebral arterial injuries. Ann Surg 2000;231:672 81. [12] Biffl WL, Moore EE, Offner PJ, et al. Blunt carotid arterial injuries: implications of a new grading scale. Journal of Trauma-Injury Infection & Critical Care 1999;47:845 53. [13] Biffl WL, Moore EE, Offner PJ, et al. Optimizing screening for blunt cerebrovascular injuries. Am J Surg 1999;178:517 22. [14] Biffl WL, Moore EE, Ray C, et al. Emergent stenting of acute blunt carotid artery injuries: a cautionary note. Journal of Trauma-Injury Infection & Critical Care 2001;50:969 71. [15] Busch T, Aleksic I, Sirbu H, et al. Complex traumatic dissection of right vertebral and bilateral carotid arteries: a case report and literature review. Cardiovasc Surg 2000;8:72 4. [16] Colella JJ, Diamond DL. Blunt carotid injury: reassessing the role of anticoagulation. Am Surg 1996;62: 212 7. [17] Davis JM, Zimmerman RA. Injury of the carotid and vertebral arteries. Neuroradiology 1983;25:55 69. [18] Dragon R, Saranchak H, Lakin P, et al. Blunt injuries to the carotid and vertebral arteries. Am J Surg 1981;141:497 500. [19] Duke BJ, Partington MD. Blunt carotid injury in children. Pediatr Neurosurg 1996;25:188 93. [20] Krajewski LP, Hertzer NR. Blunt carotid artery trauma: report of two cases and review of the literature. Ann Surg 1980;191:341 6. [21] Kraus RR, Bergstein JM, DeBord JR. Diagnosis,

266

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269 treatment, and outcome of blunt carotid arterial injuries. Am J Surg 1999;178:190 3. Lew SM, Frumiento C, Wald SL. Pediatric blunt carotid injury: a review of the National Pediatric Trauma Registry. Pediatr Neurosurg 1999;30:239 44. Li MS, Smith BM, Espinosa J, et al. Nonpenetrating trauma to the carotid artery: seven cases and a literature review. Journal of Trauma-Injury Infection & Critical Care 1994;36:265 72. Marciani RD, Israel S. Diagnosis of blunt carotid injury in patients with facial trauma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1997;83:5 9. Martin RF, Eldrup-Jorgensen J, Clark DE, et al. Blunt trauma to the carotid arteries. J Vasc Surg 1991;14: 789 793 [discussion: 793 5]. McIntyre WB, Ballard JL. Cervicothoracic vascular injuries. Semin Vasc Surg 1998;11:232 42. Miller PR, Fabian TC, Bee TK, et al. Blunt cerebrovascular injuries: diagnosis and treatment. Journal of Trauma-Injury Infection and Critical Care 2001;51: 279 285 [discussion: 285 6]. Mokri B. Traumatic and spontaneous extracranial internal carotid artery dissections. J Neurol 1990;237: 356 61. Mokri B, Piepgras DG, Houser OW. Traumatic dissections of the extracranial internal carotid artery. J Neurosurg 1988;68:189 97. Mokri B, Schievink WI, Olsen KD, et al. Spontaneous dissection of the cervical internal carotid artery. Presentation with lower cranial nerve palsies. Arch Otolaryngol Head Neck Surg 1992;118:431 5. Mokri B, Sundt Jr. TM, Houser OW, et al. Spontaneous dissection of the cervical internal carotid artery. Ann Neurol 1986;19:126 38. Mulloy JP, Flick PA, Gold RE. Blunt carotid injury: a review. Radiology 1998;207:571 85. Perry MO, Snyder WH, Thal ER. Carotid artery injuries caused by blunt trauma. Ann Surg 1980;192: 74 7. Punjabi AP, Simon J, Plaisier B. Blunt carotid injury. Journal of Trauma-Injury Infection & Critical Care 1996;41:1077 8. Shintani S, Shiigai T, Tsuruoka S, et al. TIAs in a spontaneously dissecting aneurysm of the internal carotid artery a case report. Angiology 1992;43:621 4. Willis BK, Greiner F, Orrison WW, et al. The incidence of vertebral artery injury after midcervical spine fracture or subluxation. Neurosurgery 1994; 34:435 441 [discussion: 441 2]. Yamada S, Kindt GW, Youmans JR. Carotid artery occlusion due to nonpenetrating injury. Journal of Trauma-Injury Infection and Critical Care 1967;7: 333 42. Duke BJ, Ryu RK, Coldwell DM, et al. Treatment of blunt injury to the carotid artery by using endovascular stents: an early experience. J Neurosurg 1997;87: 825 9. Ramadan F, Rutledge R, Oller D, et al. Carotid artery trauma: a review of contemporary trauma center experiences. J Vasc Surg 1995;21:46 55 [discussion: 55 6]. Caplan LR, Zarins CK, Hemmati M. Spontaneous dissection of the extracranial vertebral arteries. Stroke 1985;16:1030 8. Mokri B, Houser OW, Sandok BA, et al. Spontaneous dissections of the vertebral arteries. Neurology 1988; 38:880 5. Showalter W, Esekogwu V, Newton KI, et al. Vertebral artery dissection. Acad Emerg Med 1997;4: 991 5. Thibodeaux LC, Hearn AT, Peschiera JL, et al. Extracranial vertebral artery dissection after trauma: a 5-year review. Br J Surg 1997;84:94. Malek AM, Halbach VV, Phatouros CC, et al. Endovascular treatment of a ruptured intracranial dissecting vertebral aneurysm in a kickboxer. Journal of Trauma-Injury Infection & Critical Care 2000;48: 143 5. Leys D, Lucas C, Gobert M, et al. Cervical artery dissections. Eur Neurol 1997;37:3 12. Munera F, Soto JA, Palacio D, et al. Diagnosis of arterial injuries caused by penetrating trauma to the neck: comparison of helical CT angiography and conventional angiography. Radiology 2000;216: 356 62. Rogers FB, Baker EF, Osler TM, et al. Computed tomographic angiography as a screening modality for blunt cervical arterial injuries: preliminary results. Journal of Trauma-Injury Infection & Critical Care 1999;46:380 5. Levy C, Laissy JP, Raveau V, et al. Carotid and vertebral artery dissections: three-dimensional timeof-flight MR angiography and MR imaging versus conventional angiography. Radiology 1994;190: 97 103. Brugieres P, Castrec-Carpo A, Heran F, et al. Magnetic resonance imaging in the exploration of dissection of the internal carotid artery. Journal of Neuroradiology. Journal de Neuroradiologie 1989; 16:1 10. Gelbert F, Assouline E, Hodes JE, et al. MRI in spontaneous dissection of vertebral and carotid arteries. 15 cases studied at 0.5 tesla. Neuroradiology 1991;33: 111 3. Kuroda S, Abumiya T, Takahashi A, et al. Magnetic resonance findings in spontaneous dissection of the cervical internal carotid artery case report. Neurol Med Chir (Tokyo) 1992;32:773 7. Nelson JR, Boundy KL. Magnetic resonance imaging in carotid dissection. Australas Radiol 1992;36: 40 3. Nguyen Bui L, Brant-Zawadzki M, Verghese P, et al. Magnetic resonance angiography of cervicocranial dissection. Stroke 1993;24:126 31. Prabhu VC, Patil AA, Hellbusch LC, et al. Magnetic resonance angiography in the diagnosis of traumatic vertebrobasilar complications: a report of two cases. Surg Neurol 1994;42:245 8.

[22]

[40]

[23]

[41]

[42]

[24]

[43]

[25]

[44]

[26] [27]

[45] [46]

[28]

[29]

[47]

[30]

[31]

[48]

[32] [33]

[49]

[34]

[50]

[35]

[36]

[51]

[37]

[52]

[53]

[38]

[54]

[39]

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269 [55] Auer D, Karnath HO, Nagele T, et al. Non invasive investigation of pericarotid syndrome: role of MR angiography in the diagnosis of internal carotid dissection. Headache 1995;35:163 8. [56] Giyanani VL, Gerlock Jr. AJ, Mirfakhraee M, et al. Duplex sonography in embolotherapy of external carotid artery traumatic aneurysms. AJNR Am J Neuroradiol 1987;8:1131 3. [57] Montalvo BM, LeBlang SD, Nunez Jr. DB, et al. Color Doppler sonography in penetrating injuries of the neck. AJNR Am J Neuroradiol 1996;17:943 51. [58] Mullges W, Ringelstein EB, Leibold M. Non-invasive diagnosis of internal carotid artery dissections. Journal of Neurology. Neurosurgery & Psychiatry 1992; 55:98 104. [59] Rawat B, Gerroti M, Ashby P, et al. Application of duplex Doppler sonography in intracranial carotid dissection. J Ultrasound Med 1992;11:121 3. [60] Romner B, Sjoholm H, Brandt L. Transcranial Doppler sonography, angiography and SPECT measurements in traumatic carotid artery dissection. Acta Neurochir (Wien) 1994;126:185 91. [61] Sturzenegger M, Mattle HP, Rivoir A, et al. Ultrasound findings in carotid artery dissection: analysis of 43 patients. Neurology 1995;45:691 8. [62] Trattnig S, Rand T, Thurnher M, et al. Colour-coded Doppler sonography of common carotid artery dissection. Neuroradiology 1995;37:124 6. [63] Gomez CR, May AK, Terry JB, et al. Endovascular therapy of traumatic injuries of the extracranial cerebral arteries. Crit Care Clin 1999;15:789 809. [64] Srinivasan J, Newell DW, Sturzenegger M, et al. Transcranial Doppler in the evaluation of internal carotid artery dissection. Stroke 1996;27:1226 30. [65] Axisa BM, Loftus IM, Fishwick G, et al. Endovascular repair of an innominate artery false aneurysm following blunt trauma. Journal of Endovascular Therapy: Official Journal of the International Society of Endovascular Specialists 2000;7:245 50. [66] Babatasi G, Massetti M, Bhoyroo S, et al. Nonpenetrating subclavian artery trauma: management by selective transluminally placed stent device. Thorac Cardiovasc Surg 1999;47:190 3. [67] Bejjani GK, Monsein LH, Laird JR, et al. Treatment of symptomatic cervical carotid dissections with endovascular stents. Neurosurgery 1999;44:755 61. [68] Coldwell DM, Novak Z, Ryu RK, et al. Treatment of posttraumatic internal carotid arterial pseudoaneurysms with endovascular stents. Journal of Trauma-Injury Infection & Critical Care 2000;48: 470 2. [69] Coric D, Wilson JA, Regan JD, et al. Primary stenting of the extracranial internal carotid artery in a patient with multiple cervical dissections: technical case report. Neurosurgery 1998;43:956 9. [70] DeOcampo J, Brillman J, Levy DI. Stenting: a new approach to carotid dissection. J Neuroimaging 1997; 7:187 90. [71] Dorros G, Cohn JM, Palmer LE. Stent deployment

267

[72]

[73]

[74]

[75]

[76]

[77]

[78]

[79]

[80]

[81]

[82]

[83]

[84]

[85]

[86]

[87]

resolves a petrous carotid artery angioplasty dissection. AJNR Am J Neuroradiol 1998;19:392 4. du Toit DF, Strauss DC, Blaszczyk M, et al. Endovascular treatment of penetrating thoracic outlet arterial injuries. Eur J Vasc Endovasc Surg 2000; 19:489 95. Halbach VV, Higashida RT, Dowd CF, et al. Endovascular treatment of vertebral artery dissections and pseudoaneurysms. J Neurosurg 1993;79:183 91. Halbach VV, Higashida RT, Hieshima GB. Treatment of vertebral arteriovenous fistulas. AJR. Am J Roentgenol 1988;150:405 12. Hieshima GB, Mehringer CM, Grinnell VS, et al. Emergency occlusive techniques. Surg Neurol 1978; 9:293 302. Higashida RT, Halbach VV, Tsai FY, et al. Interventional neurovascular treatment of traumatic carotid and vertebral artery lesions: results in 234 cases. AJR. Am J Roentgenol 1989;153:577 82. Kerby JD, May AK, Gomez CR, et al. Treatment of bilateral blunt carotid injury using percutaneous angioplasty and stenting: case report and review of the literature. Journal of Trauma-Injury Infection & Critical Care 2000;49:784 7. Kwok PC, Cheung JY, Tang KW, et al. Endovascular treatment of acute carotid blow-out syndrome. J Vasc Interv Radiol 2001;12:895 6. Lempert TE, Halbach VV, Higashida RT, et al. Endovascular treatment of pseudoaneurysms with electrolytically detachable coils. AJNR Am J Neuroradiol 1998;19:907 11. Macdonald S, Gan J, McKay AJ, et al. Endovascular treatment of acute carotid blow-out syndrome. J Vasc Interv Radiol 2000;11:1184 8. Malek AM, Higashida RT, Halbach VV, et al. Patient presentation, angiographic features, and treatment of strangulation-induced bilateral dissection of the cervical internal carotid artery. Report of three cases. JNeurosurg 2000;92:481 7. Mehringer CM, Hieshima GB, Grinnell VS, et al. Therapeutic embolization for vascular trauma of the head and neck. AJNR Am J Neuroradiol 1983;4: 137 42. Miller RE, Hieshima GB, Giannotta SL, et al. Acute traumatic vertebral arteriovenous fistula: balloon occlusion with the use of a contralateral approach. Neurosurgery 1984;14:225 9. Miyachi S, Ishiguchi T, Taniguchi K, et al. Endovascular stenting of a traumatic dissecting aneurysm of the extracranial internal carotid artery case report. Neurol Med Chir (Tokyo) 1997;37:270 4. Redekop G, Marotta T, Weill A. Treatment of traumatic aneurysms and arteriovenous fistulas of the skull base by using endovascular stents. J Neurosurg 2001;95:412 9. Serbinenko FA. Balloon catheterization and occlusion of major cerebral vessels. J Neurosurg 1974;41: 125 45. Shames ML, Davis JW, Evans AJ. Endoluminal stent

268

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269 placement for the treatment of traumatic carotid artery pseudoaneurysm: case report and review of the literature. Journal of Trauma-Injury Infection & Critical Care 1999;46:724 6. Teitelbaum GP, Bernstein K, Choi S, et al. Endovascular coil occlusion of a traumatic basilar-cavernous fistula: technical report. Neurosurgery 1998;42: 1394 1397 [discussion:1397 98]. Teitelbaum GP, Dowd CF, Larsen DW, et al. Endovascular management of biopsy-related posterior inferior cerebellar artery pseudoaneurysm. Surg Neurol 1995;43:357 9. Teitelbaum GP, Halbach VV, Larsen DW, et al. Treatment of massive posterior epistaxis by detachable coil embolization of a cavernous internal carotid artery aneurysm. Neuroradiology 1995;37:334 6. Wolpert SM. In re: Serbinenko FA. Balloon catheterization and occlusion of major cerebral vessels. J Neurosurg 1974;41:1974. AJNR Am J Neuroradiol 2000;21:1359 60. Zimmerman MC, Mickel RA, Kessler DJ, et al. Treatment of impending carotid rupture with detachable balloon embolization. Arch Otolaryngol Head Neck Surg 1987;113:1169 75. Hemphill III JC, Gress DR, Halbach VV. Endovascular therapy of traumatic injuries of the intracranial cerebral arteries. Crit Care Clin 1999;15:811 29. Berger MS, Wilson CB. Intracranial dissecting aneurysms of the posterior circulation. Report of six cases and review of the literature. J Neurosurg 1984;61: 882 94. Yonas H, Agamanolis D, Takaoka Y, et al. Dissecting intracranial aneurysms. Surg Neurol 1977;8:407 15. Yonas H, Dujovny M. True traumatic aneurysm of the intracranial internal carotid artery: case report. Neurosurgery 1980;7:499 502. Mendel R, Carter L. Intracranial arterial injury. In: Carter L, Spetzler R, Hamilton M, editors. Neurovascular Surgery. New York: McGraw Hill; 1995. p. 1301. Holmes B, Harbaugh RE. Traumatic intracranial aneurysms: a contemporary review. Journal of Trauma-Injury Infection & Critical Care 1993;35:855 60. Acosta C, Williams Jr. PE, Clark K. Traumatic aneurysms of the cerebral vessels. J Neurosurg 1972;36: 531 6. Ventureyra EC, Higgins MJ. Traumatic intracranial aneurysms in childhood and adolescence. Case reports and review of the literature. Childs Nerv Syst 1994;10:361 79. Benoit BG, Wortzman G. Traumatic cerebral aneurysms. Clinical features and natural history. Journal of Neurology. Neurosurgery & Psychiatry 1973;36: 127 38. Kieck CF, de Villiers JC. Vascular lesions due to transcranial stab wounds. J Neurosurg 1984;60:42 6. Ferry DJ, Kempe LG. False aneurysm secondary to penetration of the brain through orbitofacial wounds. Report of two cases. J Neurosurg 1972;36:503 6. [104] Feldges A, Nau HE, Reinhardt V. Two cases of sudden death by rupture of traumatic and bacterial aneurysms. Intensive Care Med 1989;15:400 2. [105] Quintana F, Diez C, Gutierrez A, et al. Traumatic aneurysm of the basilar artery. AJNR Am J Neuroradiol 1996;17:283 5. [106] Chen D, Concus AP, Halbach VV, et al. Epistaxis originating from traumatic pseudoaneurysm of the internal carotid artery: diagnosis and endovascular therapy. Laryngoscope 1998;108:326 31. [107] Han MH, Sung MW, Chang KH, et al. Traumatic pseudoaneurysm of the intracavernous ICA presenting with massive epistaxis: imaging diagnosis and endovascular treatment. Laryngoscope 1994;104:370 7. [108] King MA, Barkovich AJ, Halbach VA, et al. Traumatic monocular blindness and associated carotid injuries. Pediatrics 1989;84:128 32. [109] DAlise M, Coimbra C, Batjer HH. Vascular complications of head injury. In: Batjer HH, Caplan LR, Friberg L, editors. Cerebrovascular disease. Philadelphia: Lippincott-Raven, 1997. p. 1169. [110] Dandy W. Carotid-cavernous aneurysms (Pulsating exophthalmos). Zentralbl Neurochir 1937;2:77 113. [111] Dany F, Fraysse A, Priollet P. Syndrome dysmorphique et dysplasie vasculaire: une forme atypique dEhlers-Danlos type IV. J Mal Vasc 1986;11:263 9. [112] Davie J, Richardson R. Distal internal carotid thrombo-embolectomy using a Fogarty catheter in total occlusion. Technical note. J Neurosurg 1967;27:171 7. [113] Eggers F, Lukin R, Chambers A. Iatrogenic carotidcavernous fistula following Fogarty catheter thromboendarterectomy. Case report. J Neurosurg 1979; 51:543 5. [114] Lister J, Sybert G. Traumatic false aneurysm and carotid-cavernous fistula: a complication of sphenoidectomy. Neurosurgery 1979;5:473 5. [115] Motarjeme A, Keifer J. Carotid-cavernous sinus fistula as a complication of carotid endarterectomy. A case report. Radiology 1973;108:83 4. [116] Pedersen R, Troost B, Schramm V. Carotid cavenous sinus fistula after external ethmoid-sphenoid surgery. Clinical course and management. Arch Otolaryngol 1981;107:307 9. [117] Song I, Bromberg B. Carotid cavernous sinus fistula occurring after a rhinoplasty. Case report. Plast Reconstr Surg 1975;55:92 6. [118] Takahashi M, Killeffer F, Wilson G. Iatrogenic carotid cavernous fistulas. Case report. J Neurosurg 1969;30: 498 500. [119] Halbach VV, Higashida RT, Larsen DW, et al. Treatment of dural arteriovenous fistulas. In: Maciunas RJ, editor. Endovascular neurologic intervention. Park Ridge: American Association of Neurologic Surgeons, 1995. p. 217 46. [120] Wilms G. Unilateral double carotid cavernous fistula treated with detachable balloons. AJNR Am J Neuroradiol 1990;11:517. [121] Kupersmith MJ, Berenstein A. Neurovascular neuroophthalmology. Berlin: Springer-Verlag, 1993.

[88]

[89]

[90]

[91]

[92]

[93]

[94]

[95] [96]

[97]

[98]

[99]

[100]

[101]

[102] [103]

D.W. Larsen / Neuroimag Clin N Am 12 (2002) 249269 [122] Flandroy P, Lenelle J, Collignon J. Carotid cavernous fistula associated with Fogarty catheter angioplasty. AJNR Am J Neuroradiol 1988;9:1242. [123] Barker WF, Stern WE, Krayenbuhl H, et al. Carotid endarterectomy complicated by carotid cavernous sinus fistula. Ann Surg 1968;167:568 72. [124] Jorgensen JS, Guthoff R. Ophthalmoscopic findings in spontaneous carotid cavernous fistula: an analysis of 20 patients. Graefes Arch Clin Exp Ophthalmol 1988;226:34 6. [125] Sanders MD, Hoyt WF. Hypoxic ocular sequelae of carotid-cavernous fistulae. Study of the caues of visual failure before and after neurosurgical treatment in a series of 25 cases. Br J Ophthalmol 1969;53: 82 97. [126] Halbach VV, Hieshima GB, Higashida RT, et al. Carotid cavernous fistulae: indications for urgent treatment. AJR. Am J Roentgenol 1987;149:587 93. [127] Halbach VV, Higashida RT, Hieshima GB, et al. Direct puncture of the proximally occluded internal carotid artery for treatment of carotid cavernous fistulas. AJNR Am J Neuroradiol 1989;10:151 4. [128] Halbach VV, Higashida RT, Hieshima GB, et al. Transvenous embolization of direct carotid cavernous fistulas. AJNR Am J Neuroradiol 1988;9:741 7. [129] Prolo DJ, Hanbery JW. Intraluminal occlusion of a

269

[130]

[131]

[132]

[133]

[134]

[135] [136]

carotid-cavernous sinus fistula with a balloon catheter. Technical note. J Neurosurg 1971;35:237 42. Teitelbaum GP, Larsen DW, Zelman V, et al. A tribute to Dr. Fedor A. Serbinenko, founder of endovascular neurosurgery. Neurosurgery 2000;46:462 9 [discussion: 469 470]. Debrun G, Lacour P, Vinuela F, et al. Treatment of 54 traumatic carotid-cavernous fistulas. J Neurosurg 1981;55:678 92. Debrun GM, Vinuela F, Fox AJ, et al. Indications for treatment and classification of 132 carotid-cavernous fistulas. Neurosurgery 1988;22:285 9. Norman D, Newton TH, Edwards MS, et al. Carotidcavernous fistula: closure with detachable silicone balloons. Radiology 1983;149:149 57. Halbach VV, Higashida RT, Barnwell SL, et al. Transarterial platinum coil embolization of carotidcavernous fistulas. AJNR Am J Neuroradiol 1991;12: 429 33. Hosobuchi Y. Electrothrombosis of carotid-cavernous fistula. J Neurosurg 1975;42:76 85. Batjer HH, Purdy PD, Neiman M, et al. Subtemporal transdural use of detachable balloons for traumatic carotid-cavernous fistulas. Neurosurgery 1988;22: 290 6.

Neuroimag Clin N Am 12 (2002) 271 294

Imaging of pediatric head trauma

Tina Young Poussaint, MD*, Karen K. Moeller, MD
Department of Radiology, Harvard Medical School, Boston, MA 02115, USA Division of Neuroradiology, Department of Radiology, Childrens Hospital, Boston, MA 02115, USA

Pediatric head trauma is one of the leading causes of injury, mortality, and morbidity in childhood, accounting for 95,000 hospital admissions and 600,000 emergency department visits in the United States per year, with an incidence from 0.2% to 0.3% [1 4]. Neuroimaging as an aid in the diagnosis and treatment of intracranial injury has revolutionized the management and care of these patients, improving outcome and prognosis. The mechanisms of injury in children vary depending on age. In children below age two, nonaccidental trauma accounts for more than 80% of deaths [1,5]. The younger the child, the higher the risk of injury [6], and the risk for asymptomatic intracranial injury is highest for infants younger than six months [7], because of their large heads, weak neck musculature, and relatively thin calvaria [8]. As the age of the child increases, falls become a less frequent cause of accidental trauma, whereas bicycle injuries and motor vehicle accidents become more common [8].

monitoring devices, and is easy to obtain. It is useful for the detection of acute or subacute hemorrhage, scalp injury, and pneumocephalus, as well as other neurosurgically significant conditionshydrocephalus, midline shift, masses and mass effect, ischemia, and herniationthat require rapid detection for potential treatment. CT helps to demonstrate linear calvarial skull fractures that are not in the plane of scanning, depressed, and basilar skull fractures as well as facial fractures. Limitations of CT are related to beam hardening streak artifact in the posterior fossa, patient motion, partial volume averaging, and detection of small extraaxial hematomas. Magnetic resonance imaging Magnetic resonance imaging (MRI) provides multiplanar capability, and superior sensitivity and specificity to abnormal brain anatomy and function in the injured pediatric brain. It is helpful for characterization and timing of hemorrhage and for anatomic localization in the extraaxial space or brain parenchyma [9,10]. MRI has increased sensitivity for detection of intraparenchymal injury such as edema, hematoma, contusion, diffuse axonal injury, and brain edema. It provides superior visualization of the brainstem and posterior fossa components, aiding in the detection of injuries there. The sequelae of head trauma are better evaluated with MR than CT imaging [9 12]. In addition to standard T1- and T2-weighted multiplanar imaging, gradient echo images with long echo times are sensitive for detecting blood products and should be performed in any patient with a history of head trauma [13,14]. Fluid attenuated inversion recovery (FLAIR) sequences have been reported to be sensitive to the detection of subarachnoid hemorrhage (SAH) [15 17], although CT remains the initial

Imaging evaluation Computed tomography Computed tomography (CT) is the modality of choice for the evaluation of acute neurologic presentations or hemorrhage in the child with trauma. It is rapid, widely available, and inexpensive. CT can accommodate life support equipment, traction and

* Corresponding author. Division of Neuroradiology, Department of Radiology, Childrens Hospital, 300 Longwood Avenue, Boston, MA 02115. E-mail address: tina.poussaint@tch.harvard.edu (T. Poussaint).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 5 - 9

272

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

screen for SAH. Magnetic resonance angiography (MRA) can provide information about the involvement of major vascular structures, such as the circle of Willis or the dural venous sinuses, after head trauma. MRA is used for the diagnosis of vascular injury such as dissection, occlusion, or pseudoaneurysm, which may be corroborated with conventional angiography, depending on the clinical setting. MR perfusion and diffusion MR perfusion can assess cerebral perfusion dynamics by analyzing hemodynamic parameters, including relative cerebral blood volume, relative cerebral blood flow, and transit time. Using echoplanar or line-scan spin echo techniques, diffusion imaging (DI) provides information based upon differences in the rate of diffusion of water molecules, and is especially sensitive to intracellular changes in patients with head trauma. The rate of diffusion, or apparent diffusion coefficient (ADC), is higher for free or pure water than for macromolecular bound water. The ADC varies according to the microstructural or physiologic state of a tissue. A particularly important application of DI is in the early detection of diffuse and focal ischemic injury [18,19]. Diffusion imaging within one to two hours of injury [20] in the patient with head trauma has been found to detect more lesions than conventional MR images [21 23]. In a study of children with nonaccidental head injury, early DI was found to reveal diffusion abnormalities in approximately 90% of the children, and was helpful in early detection of infarction, revealing more extensive brain injury or injury that had not been seen on conventional MRI [24]. The severity of the abnormalities on DI correlated with poor outcome (P < 0.005). Diffusion tensor imaging (DTI) has been found to provide information about brain ultrastructure by quantifying isotropic and anisotropic water diffusion. Measuring water diffusion anisotropy can yield information about structures such as white matter tracts [25]. Changes in diffusion anisotropy in traumatic brain injury and diffuse axonal injury have been reported [23,26]. In a study of 20 adult patients, changes in diffusion anisotropy within the white matter correlated with acute Glasgow coma scale and Rankin score at discharge (Thierry A.G.M. Huisman, MD, personal communication, 2001). There was a statistically significant correlation between fractional anisotropy values and severity of head injury in both the acute setting (Glasgow coma score) and long-term prognosis (Rankin score) in different predilection sites of diffuse axonal injury. Future studies using this

application in the pediatric population will likely yield more important information about head trauma severity and outcome in children who have had accidental or nonaccidental trauma. Magnetic resonance spectroscopy Magnetic resonance spectroscopy (MRS) offers a noninvasive approach to biochemical analysis and provides quantitative information regarding cellular metabolites, since signal intensity is linearly related to steady-state metabolic concentration. Elevated lactate has been found in cerebral contusions and regions of brain infarction in pediatric head trauma [27]. In babies who were shaken, a decrease in N-acetylaspartate (NAA) and increase in lactate between five and seven days after injury indicated poor prognosis and significant brain damage [28]. In adults with traumatic brain injury, proton MR spectroscopic findings have been found to correlate with neuropsychological function with reduced NAA in white matter, which suggests neuronal injury and inflammation [29,30]. Thus, MRS can predict patient outcome and has the capability to assess severity of brain injury in older pediatric patients. Functional MRI Functional MRI (fMRI) is the term applied to brain activation imaging in which local or regional changes in cerebral blood flow that accompany stimulation or activation of sensory (eg, visual, auditory), motor, or cognitive centers are displayed. One fMRI study involved the effects of trauma on cognition, neural circuitry, learning and memory [31], with the aim of using fMRI as an assessment and prognostic tool to evaluate trauma patients. In mild traumatic brain injury (TBI) in adults [32], fMRI has demonstrated effects on working memory. In other adult patient studies, impairment of working memory was associated with alterations in functional cerebral activity in moderate to severe TBI [33] and was used to predict brain function in a comatose head-injury [34]. Thus, fMRI should be useful to evaluate brain function and prognosis in nonresponsive brain trauma patients of any age. Magnetic source imaging Magnetic source imaging (MSI) integrates anatomic data from conventional MRI with electrophysiological data from magnetoencephalography. Magnetoencephalography is a technique that measures magnetic fields associated with intracellular cur-

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

273

rent flows within neurons [35] and has demonstrated brain dysfunction in patients with postconcussive symptoms [36]. It apparently has greater sensitivity for this purpose than conventional MR or electroencephalography (EEG). Magnetization transfer imaging Magnetization transfer imaging relies upon the large difference between the relaxation properties of free water and water bound to macromolecules and macromolecular protons [37]. The exchange of magnetization between these two pools has been applied to patients with traumatic brain injury. Abnormal magnetization transfer ratios were found in normal appearing white matter on conventional MRI prone to axonal injury and are an apparent predictor of poor outcome [38]. Nuclear medicine Single photon emission computed tomography (SPECT) imaging may demonstrate focal, multifocal, or regional areas of hypoperfusion in patients with brain injury in the acute (< 24 hours) and chronic phases when compared to CT [39 41]. In a study of SPECT, CT, and MRI, SPECT was compared to CT in the acute phase and showed slightly more abnormalities. However, MR showed more abnormalities than SPECT [42]. F-18 fluorodeoxyglucose positron emission tomography (FDG-PET imaging) has demonstrated a decrease in the regional cerebral metabolic rate of glucose in traumatic brain injuries [43]. Oxygen-15 positron emission tomography (O-15 PET) imaging in severely brain-injured patients demonstrated alterations in substrates in verbal recall. Frontal regional cerebral blood flow was decreased during free recall and enhanced during recognition in these patients compared with controls [44]. Ultrasonography Ultrasonography (US), with its real-time capabilities and ready access, has been used in the evaluation of extracerebral fluid collections in infants and in critically ill patients who are not able to be transported for CT or MR. Distinction of the subarachnoid spaces over the convexities from subdural collections is accomplished using standard and color Doppler techniques [45,46]. Transcranial doppler techniques have been used to correlate resistive indices with elevated intracranial pressure in patients with head trauma [47].

Birth trauma Scalp injury Three patterns of extracranial hemorrhage associated with birth trauma are: caput succedaneum, subgaleal hemorrhage, and cephalohematoma [48]. Caput succedaneum is the most common scalp injury during the birth process, frequently occurring following vaginal delivery. It consists of hemorrhage and edema within the fibrofatty layer just beneath the skin, localized at the vertex with associated molding. These hemorrhages are soft to palpation, are selflimited and typically resolve over days. Although imaging is not necessary to make the diagnosis, caput succedaneum is often demonstrated incidentally and appears as focal soft tissue swelling on CT or MRI. Subgaleal hemorrhage occurs between the epicranial aponeurosis (galea) and the periosteum of the calvaria. It remains superficial to the temporalis muscle in this region of the skull [49]. This space contains small veins and is also crossed by emissary veins that connect the dural sinus with the superficial veins of the scalp. These hemorrhages are firm to palpation and typically asymptomatic. The anatomy of the subgaleal space may allow pooled blood to become very large and extensively extend around the cranium leading to significant blood loss. This type of injury can occur following instrumentation used during delivery, especially vacuum extraction [50]. Subgaleal hematomas appear on imaging studies as diffuse soft tissue swelling that crosses suture lines. The hematoma typically resolves over two to three weeks. Cephalohematoma is subperiosteal in location and occurs in up to 1% of live births [51]. Cephalohematomas may lie beneath the temporalis muscle. They may occur following vaginal delivery, but they have an increased incidence in deliveries that employ forceps or use vacuum extraction. Because of the subperiosteal location of the hemorrhage, they are bounded by the sutures where the periosteum is firmly attached. These hemorrhages present as a palpably firm, tense mass that resolves over weeks to months and can calcify as they age and become incorporated in periosteal new bone. They are most common in the parietal region and can be unilateral or bilateral [49]. On imaging, cephalohematomas have a crescentic shape and are bound by sutures. Curvilinear peripheral calcification can develop within 14 days (Fig. 1). The appearance on MRI varies depending on the age of the hematoma. Cephalohematomas are usually hyperintense on T1-weighted imaging, and a hematocrit effect may be visualized on T2weighted imaging [52]. Rarely, cephalohematomas

274

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

delivery accompanied by instrumentation [12]. The most frequent site for depressed skull fractures in the neonate is the parietal bone, followed by the frontal bone. The degree of depression of the involved bone is best evaluated by CT, which may also show an underlying injury. The depressed fracture is often referred to as the ping-pong fracture, because of the inward buckling of the bone, which may be present without loss of bony continuity and resembles the same deformity in a ping-pong ball. Intracranial injury Intracranial injury can occur during delivery secondary to cephalopelvic disproportion, breech presentation, or from instrumentation involving forceps or vacuum extraction. The rate of intracranial hemorrhage is higher in infants delivered by vacuum extraction, forceps, or cesarean section during labor than among those delivered spontaneously [54] (Fig. 2). The four most common intracranial injuries associated with birth trauma with subdural hemorrhage are: tentorial laceration, occipital osteodiastasis, falcine laceration, and rupture of bridging cortical and meningeal veins [55]. A large tear in the tentorium can result in rupture of the vein of Galen, straight sinus, or transverse

Fig. 1. Cephalohematoma in a 4-month old infant with a left parietal bump. (A) AP scout CT tomogram and (B) axial CT bone window images demonstrate calcified left parietal cephalohematoma.

may become infected, leading to complications such as subdural empyema, epidural abscess and osteomyelitis [53]. Skull injury A type of skull injury unique to newborns is the overlapping of calvarial bones at the sutures. Often, the degree of overlap is minor and of no clinical significance. The calvarial deformity resolves over time with growth. On CT, overlapping of adjacent bones at the sutures is best visualized on the bone window images. Depressed skull fractures are rare in neonates, but can be encountered following vaginal

Fig. 2. Subdural hemorrhage after forceps and vacuum delivery. Axial CT image demonstrates acute subdural blood along the right tentorium and cerebral convexity with mass effect on the right lateral ventricle and midline shift to the left.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

275

sinus, leading to large posterior fossa subdural hematomas (SDH) (Fig. 3). These hematomas can become massive, leading to brain stem compression and even death [56,57]. Smaller posterior fossa subdural hematomas are being recognized more frequently with the increased utilization of MRI. These hematomas result

from smaller tears in the tentorium or rupture of small infratentorial veins. On CT, these subdurals present as increased attenuation and thickening along the tentorium. The SDH can also extend laterally along the cerebellar hemispheres. Follow-up imaging is recommended to ensure that hydrocephalus does not occur secondary to fourth ventricle compression. Occipital osteodiastasis occurs when there is a traumatic tear and separation along the synchondrosis between the squamosal and the exo-occipital portions of the occipital bone [58]. It is probably secondary to pressure on the occipital bone from the pubic symphysis of the mother during breech delivery, or results from forced engagement of the fetal head. Feto-pelvic disproportion is probably also an important factor contributing to this injury [59]. There also may be displacement of the anterior margin of the squamosal portion of the occipital bone into the posterior fossa, which may result in a tear of the dural sinuses or inferior cerebellar veins. Lateral skull films show overriding of the squamosal portion of the occipital bone on the condylar portions of the occipital bone. On CT, large subdural hemorrhages may be visualized in the posterior fossa. Parenchymal hemorrhage and cerebellar edema may also occur. The SDH usually occurs along the inferior aspect of the tentorium, with extension laterally along the cerebellar hemispheres. These hemorrhages can be seen on US as extraxial collections that are mild to moderately echogenic. The MR appearance varies depending on the age of the hematoma, with coronal imaging helpful for demonstrating these SDHs. Laceration of the falx is less common than laceration of the tentorium. This injury often occurs at the junction of the falx and tentorium, with bleeding occurring from a tear in the inferior sagittal sinus, which leads to SDH along the inferior aspect of the interhemispheric fissure adjacent to the corpus callosum. SDH over the cerebral convexity may occur from rupture of superficial cortical veins that bridge the dura. These SDHs are often unilateral and can be accompanied by subarachnoid hemorrhage (SAH) or adjacent cerebral contusion.

Skull fractures The calvarium in a child is softer and thinner than an adults and is therefore susceptible to fracture. Skull fractures from minor trauma are more common in children than adults, especially in children less than two years old [60]. In addition, under the age of four, the calvarium is unilaminar and lacks diploe [49]. Therefore the skull offers less protection to the

Fig. 3. Tentorial laceration. (A,B) Axial CT images demonstrate severe hydrocephalus, blood in the fourth ventricle, along the tentorium, and in the region of the vein of Galen and straight sinus.

276

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

childs brain than it does in the adult, and children with skull fractures are at increased risk of having intracranial injury [6]. Skull fractures may be diagnosed by CT or plain radiography. Skull fractures in neonates may be of the following types: linear, diastatic, depressed, compound, or buckled. Older children may sustain comminuted fractures as the skull becomes more rigid with age. Linear fractures are by far the most common type of fracture in all age groups. Most linear fractures are associated with an overlying hematoma or scalp soft-tissue swelling. These fractures are more often diastatic in newborns and in infants that are less than six months old. On plain films, these are seen as lucent linear defects that tend to stop at the suture. They may be seen only on one view and are often better visualized on the lateral view of the skull. If a child is less than six months old, the central portion of the fracture line can be wider. Linear skull fractures may also be detected on CT but may be missed, especially if the fracture line is parallel to the plane of scanning. In this situation, three-dimensional (3D) CT maximum intensity projections (MIP) of the calvarium may be helpful to better visualize the fracture [61]. In infants and older children approximately 30% of patients with a depressed skull fracture have an associated brain injury [6,62] (Fig. 4). Complications from depressed skull fractures include dural tear, cerebral contusion, retained osseous fragments, and cosmetic deformity. The degree of depression correlates with underlying brain injury, so that if the outer portion of the depressed fragment is not depressed beyond the inner table of the adjacent calvarium, the fracture is considered less serious. Leptomeningeal cyst is a unique pediatric lesion that occurs as a delayed complication in approximately 0.05% to 1.6% of infants and children with skull fractures [63,64]. Ninety percent of these lesions (also called growing fractures) occur in children who are younger than three [65]. These are skull fractures that widen over time secondary to the torn meninges, becoming interposed between the fracture fragments. Cerebral spinal fluid (CSF) pulsation causes increased herniation of the meninges and progressive widening of the skull defect. There are three factors that are probably responsible for the increased incidence of this complication in children: an actively growing skull, dura that is more tightly adherent to the calvaria, and decreased thickness of the pediatric calvaria [63,66,67]. A leptomeningeal cyst may complicate any skull fracture, but is often associated with a diastatic parietal fracture. A dural tear occurs at the fracture site and arachnoid herniates into the defect [68]. This prevents

Fig. 4. Depressed skull fracture with underlying contusion. (A) Lateral scout CT tomogram (B) demonstrates depressed comminuted right parietal skull fracture with underlying contusion.

osteoblasts from migrating across the fracture line, with subsequent impaired healing. CSF pulsations lead to further herniation of the arachnoid into the defect. CSF becomes entrapped within the defect secondary to fibrosis and adhesions. Pathologically, the wall of the cyst consists of a fibrocollagenous membrane. Finally, slow erosion of the skull occurs at the site of the growing cyst. To recognize this complication, follow-

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

277

up imaging is recommended in all young children with diastatic skull fractures. Risk factors for the development of a leptomeningeal cyst include age less than three and fracture diastasis of 4 mm or greater [63]. On plain films, leptomeningeal cysts appear as oval areas of bone erosion. The margins of the lesion are smooth and the inner table is eroded more than the outer table. The edges of the defect are elevated and thickened. CT (Fig. 5) and MRI confirm the skull defect and demonstrate the cyst, which parallels CSF in attenuation and signal characteristics [69]. In some cases, there is

herniation of brain tissue at the defect. Underlying encephalomalacia and ipsilateral hemiatrophy may also occur. Early diagnosis of these lesions is important because brain damage is progressive. Surgery is the treatment of choice. Since the main factor responsible for leptomeningeal cyst formation is a dural tear, some authors recommend US at the initial time of injury in patients thought to be at risk for this complication. US in experienced hands can be used to diagnose a dural tear [70]. Normally, the dura appears as an echogenic line

Fig. 5. Leptomeningeal cyst. (A) Axial CT image (B) demonstrates comminuted left parietal fracture with contusion. (C) Lateral CT scout tomogram 2 months later demonstrates fracture with leptomeningeal cyst (white arrow). (D) Axial CT image (E) demonstrates left parietal encephalomalacia and bulging of intracranial contents through bony defect.

278

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

shape, and does not cross suture lines. There may be an associated fracture (Fig. 6). The appearance on MR

Fig. 5 (continued ).

on ultrasonography and is not visualized if there is a defect. The fracture itself is used as an acoustic window. Color Doppler sonography has been helpful for demonstrating arterial flow through the defect [71].

Extraaxial hemorrhage Epidural hematoma Epidural hematomas (EDH) occur between the inner table of the skull and the outer layer of the dura. Epidural hematomas are less common in infants and young children than in adults and are seen in 1 to 3% of children with head trauma [72,73]. In children, the dura is more firmly adherent to the inner table of the skull and the groove for the middle meningeal artery is shallow, allowing for more mobility of the vessel. For these reasons EDH is less common, and when it does occur, is more often from venous bleeding than arterial. The source of the blood can be in the diploe or the dura [49]. Skull fractures are less commonly associated with EDH in children because of the increased plasticity of the childs skull. Since the bleeding is often venous, these hematomas evolve slowly, and the clinical presentation of acute EDH in the young child can be less dramatic than in an adult. Often, there is an asymptomatic interval followed by rapid deterioration. Aggressive early imaging is recommended when this injury is suspected. On imaging, EDH appears identical to the findings seen in adults. The EDH appears hyperdense on CT, has a biconvex

Fig. 6. Epidural hematoma in an 11-month-old infant who fell from a walker. (A) Axial CT image (B) demonstrates epidural hematoma, midline shift to right, mass effect on left lateral ventricle, and associated nondisplaced left parietal skull fracture.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

279

varies with the age of the hematoma, with acute hematomas appearing isointense on T1-weighted images and dark on T2-weighted images. There have been recently described cases of epidural hematomas decompressing into the subgaleal space through a skull fracture. These were shown to resolve as the hematoma redistributed into the subgaleal space [74]. Subdural hematoma

subarachnoid spaces associated with macrocephaly, thought to be secondary to immature CSF absorption at the level of the arachnoid villi (Fig. 7). Often vessels can be seen traversing the extra-axial space, confirming the subarachnoid location. This condition often resolves by age two without sequelae [76,77]. These patients may be predisposed to extra-axial hemorrhage after minor head trauma [78] (Fig. 8). Subarachnoid hemorrhage

Subdural hematomas are collections of blood located between the inner dural layer and the arachnoid. These may occur following accidental trauma, but are also associated with child abuse. Other predisposing factors include prematurity and blood dyscrasias. These collections are more common in infants and younger children than in adolescents. In infants, the cortical veins are more easily torn, because of the plasticity of the skull, the softness of the underlying unmyelinated brain, and the expanded extraaxial space, which places more tension on the bridging cortical veins when trauma occurs. Unlike in adults, where the SDH is often unilateral, SDH in children is bilateral in up to 80% of cases [51]. In addition, interhemispheric SDH is more common in children than in adults. Often SDH in the pediatric age group is extensive, with involvement of the temporal, frontal and parietal regions. This results from the lack of adhesions in the subdural space that are present in the adult. Imaging findings in children with subdural hematoma are similar to those in adults. In the acute phase (< 3 days) a crescentic shaped, high-attenuation extraaxial collection is visualized on CT. In 1 to 3 weeks, the hematoma gradually becomes isodense relative to the adjacent brain. Gradually, after two to three weeks the SDH decreases to an attenuation that is similar to CSF. On MR, the SDH appearance will vary depending on the age of the blood products within the hematoma. The evolution of blood products in an extraxial collection is similar to that of an intraparenchymal hematoma, but tends to occur over a longer time span because of differences in oxygen tension in the subdural space [12]. In addition, unlike parenchymal hematomas, ferritin and hemosiderin are not deposited in the wall of a chronic SDH [75]. This is because the blood-brain barrier inhibits resorption of the blood products in intraparenchymal hematomas. Chronic SDH can develop outer and inner membranes that enhance following contrast medium administration. Varying ages of blood products may appear as fluid-fluid levels. In infants, a mimic of chronic subdural hematoma is benign external hydrocephalus. These are prominent

Subarachnoid hemorrhages frequently accompany head injury and the imaging findings are identical to adults. The sylvian fissure and interpeduncular cistern are common sites for SAH to be visualized on imaging. CT is the imaging modality of choice for detecting acute SAH. If MRI is being performed, FLAIR sequences are more sensitive in detecting SAH than other routine MR sequences. SAH will appear as high signal within the sulci, whereas normal cerebrospinal fluid will be low in signal intensity [15 17]. Intraventricular hemorrhage Intraventricular hemorrhage may be associated with parenchymal hematomas, contusions, and diffuse axonal injury (DAI). DAI involving the corpus callosum is often associated with intraventricular hemorrhage from rotational forces that lead to tearing of the subependymal veins on the ventral surface of the corpus callosum [79].

Intraaxial hemorrhage Parenchymal injuries encountered during childhood include contusions, DAI, and intracerebral hematomas. Cortical contusions occur when there is direct impact of the brain against bone during a deceleration force and represent  45% of intra-axial injuries [80]. Contusions often occur in the inferior frontal lobes and anterior temporal lobes, where the adjacent inner table of the skull is irregular. Because the inner table of the skull is relatively smooth in children compared with adults, contusions are less common in them. Contusions can be adjacent to the point of impact (coup) or on the opposite side of the head (contracoup). They involve the superficial gray matter and usually spare the underlying subcortical white matter unless large. On CT, contusions appear as illdefined regions of high attenuation in characteristic locations, namely the anterior inferior frontal lobe

280

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

Fig. 7. External hydrocephalus. (A) Sagittal T1 MR image demonstrates prominent subarachnoid spaces with displacement of vessels away from cortex. (B) Axial proton density and (C) T2 MR images demonstrate prominent subarachnoid spaces that follow CSF on all sequences.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

281

Fig. 8. External hydrocephalus with subdural hematoma. (A) Sagittal T1 MR image demonstrates prominent subarachnoid spaces and hyperintense left subdural hematoma. (B) Axial proton density MR image demonstrates hyperintense left subdural hematoma and enlarged subarachnoid spaces. (C) Axial T2 MR image demonstrates left subdural hematoma hypointense to CSF.

and the anterior temporal lobe. MRI is more sensitive than CT in detecting nonhemorrhagic contusions. If a trauma patient is being evaluated by MRI, gradient echo images are an important sequence also, because of increased sensitivity for detecting hemorrhage [13,14].

Intracerebral hemorrhages may be difficult to distinguish from hemorrhagic contusions or diffuse axonal injury. They are commonly seen in the frontotemporal white matter or basal ganglia [73]. On MRI, the appearance of the intraparenchymal hemorrhage varies with the age of the lesion. Acute hema-

282

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

tomas will appear isointense to the brain on T1weighted images and hypointense on T2-weighted images, because of the presence of deoxyhemoglobin

in the red blood cells. With time, deoxyhemoglobin is converted to intracellular methemoglobin and the hematoma will appear bright on T1-weighted images

Fig. 9. Diffuse axonal injury. A 15-year-old boy with traumatic brain injury 2 months prior. (A,B) Axial T2 MR images demonstrate bilateral chronic subdural hematomas and abnormal hyperintense signal in the periventricular white matter and left midbrain. (C,D) Axial gradient echo images demonstrate numerous hypointense lesions in the left cerebral peduncle, bilateral frontal, parietal and temporal lobe white matter, and splenium of corpus callosum.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

283

and dark on T2-weighted images. High signal develops later on the T2-weighted images as the red blood cells lyse and the methemoglobin becomes extracellular. Chronic hemorrhage appears hypointense on T1-weighted images and T2-weighted images as hemosiderin and ferritin accumulate in macrophages. Subcortical gray matter injury is relatively uncommon. It consists of petechial hemorrhages in the thalami, basal ganglia, and regions around the third ventricle and likely occurs because of injury to the small perforating blood vessels [80]. Diffuse axonal injury follows severe head injury when there is sudden acceleration and deceleration combined with rotational forces. It is one of the most common types of severe head injury [73]. With this mechanism, portions of the brain move relatively slowly compared to adjacent areas, which leads to a shearing stress with subsequent axonal injury. This pattern of injury involves the subcortical white matter, the corpus callosum, the brainstem, and the internal capsule. Pathologically, there is extensive damage, with multiple torn and retracted white matter tracts (retraction balls) and perivascular hemorrhage [81]. Compared to older children and adults, infants are more susceptible to this injury, because of their relatively large heads, weaker neck musculature, increased CSF spaces, and increased compliance of the calvarium [49]. It is most often associated with high-velocity motor vehicle accidents, but may also be seen in blunt trauma and nonaccidental trauma. CT in the setting of DAI may be normal or may only show small foci of high attenuation. These patients may have profound impairment of consciousness that suggests the diagnosis even with a near normal CT. MRI is more sensitive for the detection of DAI lesions that can be hemorrhagic or nonhemorrhagic [10,80] (Fig. 9). On gradient echo images, these lesions appear hypointense. They may be round or elliptical in shape, and parallel to the axonal tracts involved. Milder forms of DAI consist of lesions near the gray white junction. More severe injury results in lesions in deeper structures within the brain, such as the centrum semiovale and corpus callosum, particularly the posterior body and splenium [79]. Severe DAI may affect the brain stem, including the dorsolateral midbrain, pons, and cerebellar peduncles. [82] As mentioned earlier, diffusion tensor imaging may be helpful in the evaluation the brain parenchyma DAI [23,26].

hypoxic-ischemic injury, developmental delay, posttraumatic seizures, cognitive deficits, encephalomalacia, infection, cranial nerve injury, and pituitary/ hypothalamic injury. Vascular injury, diffuse cerebral edema, and hydrocephalus will be discussed below. Vascular injury Vascular injuries that may occur in traumatic head injury are similar to those found in adults and include arterial dissections, venous occlusions or ruptures, and carotid-cavernous fistula. Dissections of the carotid artery or vertebral artery may occur as a result of penetrating or blunt trauma to the neck [83]. In children, carotid dissections may also result from falls when pencils or similar objects are in the mouth [84,85]. Carotid artery dissection may also be associated with fractures involving the skull base and extending into the carotid canal [86]. These vascular injuries have also been described in the setting of seemingly trivial traumatic forces, such as sudden flexion or extension of the neck [87]. In this situation, the dissection is often classified as being spontaneous, although usually a history of minor trauma or sudden neck movement can be elicited. In extracranial carotid dissection, the internal carotid artery (ICA) is often involved just distal to its bifurcation, with cephalad extension to the skull base. Intracranial carotid dissection may involve the supraclinoid ICA. In nonpenetrating trauma, vertebral artery dissection usually occurs at the C1 C2 level or at the site of cervical spine fractures [88]. Cerebral infarction demonstrated on conventional or diffusion MR images results from emboli or arterial occlusion. On MRI, narrowing or occlusion of the vessel may be present, characterized by narrowing of the vascular flow void or abnormal signal within the lumen of the vessel (Fig. 10). In addition, intramural hemorrhage may be seen in the wall of the vessel, causing a periarterial collar of abnormal signal that may be asymmetric on one side of the lumen [89,90]. The intramural hematoma may be better demonstrated using T1-weighted imaging with fat saturation through the neck and skull base of the affected vessel [91]. MRA may also demonstrate lumen narrowing, vessel occlusion, or pseudoaneurysm characterized by bright signal projecting beyond the lumen of the artery [92]. Conventional angiography, which may demonstrate stenosis, occlusion, or pseudoaneurysm formation, may serve as an adjunct to MRA. Carotid-cavernous fistulas may present with pulsatile exopthalmos or cranial nerve palsies. Conventional angiography is the imaging modality of choice for diagnosis, because it demonstrates the fistula

Sequelae of trauma Sequelae of traumatic injury may include vascular injury, diffuse cerebral edema, hydrocephalus,

284

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

Fig. 10. Carotid dissection. A 9-year-old girl who suffered laceration of her right oropharynx when a handlebar entered her mouth as she fell from a bicycle. (A) Axial CT image demonstrates right MCA sign with increased attenuation in the right middle cerebral artery (white arrow) and (B) infarct in the right posterior basal ganglia. (C) Diffusion MR image demonstrates restricted diffusion in right subinsular region, basal ganglia, and posterior limb of internal capsule. (D) 2D-time of flight (TOF) MRA image of the neck and (E) 3D-TOF MRA image of the circle of Willis demonstrate occlusion of RICA from carotid dissection.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

285

and blood flow, and restricted diffusion in brain regions with neuronal injury [51]. As in adults, transtentorial herniation and subfalcine herniation may occur in children from increased ICP. Transtentorial herniation may be complicated by posterior cerebral infarcts secondary to compression of the posterior cerebral artery by the free edge of the tentorium in the ambient cistern. The anterior cerebral artery may be similarly compressed during subfalcine herniation. Hydrocephalus Hydrocephalus frequently develops after intracranial injury, and may be related to adhesions forming after SAH that likely block CSF absorption at the arachnoid villi. It may not be possible to differentiate post-traumatic hydrocephalus from cerebral atrophy, which may also occur following trauma [96,97]. Correlation with head circumference may aid in distinguishing the two entities in young children.

Fig. 10 (continued ).

Nonaccidental head injury The incidence of child abuse in the United States is estimated to be 42 per one thousand children. The Third National Incidence Study of Child Abuse and Neglect estimated that 2,815,600 children were victims of child abuse [98]. The number of children seriously harmed by abuse or neglect was approximately 569,000 in 1993. These estimates are likely low, for it is generally accepted that many cases of child abuse go unreported [46]. Head trauma is the leading cause of morbidity and mortality in abused children less than two years old [5,46,99]. The incidence is even greater in infants that are less than one year old. Over 95% of life-threatening head injuries in infants are the result of abuse. On clinical followup, the majority of infants who are victims of abuse have significant neurologic impairment. Child abuse is a known cause of mental retardation, developmental delay, behavorial disturbances, learning disabilities, and cerebral palsy. Abused children are also at increased risk for the development of depression and anxiety disorders. The clinical presentation in infants who have sustained nonaccidental head injury (NAHI) includes decreased consciousness, irritability, apnea, respiratory difficulty, and seizures. Caretakers may describe a history of lethargy or decreased appetite. Nonaccidental injury should be suspected in the infant or young child who presents with no history of injury; with a discrepancy between the explanation

between the cavernous ICA and cavernous sinus. CT findings include prominence of the cavernous sinus and superior opthalmic vein on the affected side. Diffuse cerebral edema Post-traumatic cerebral swelling is more common in the pediatric population than in adults. Loss of autoregulation of blood flow to the brain following injury in children may lead to hyperemic cerebral swelling [93 95]. Because children also have a lower mean arterial pressure, this can lead to decreased blood flow to the brain as swelling becomes more severe. Cerebral swelling may also occur in the setting of DAI. Accurate diagnosis is important to begin therapy for treatment of increased intracranial pressure (ICP). On imaging, diffuse cerebral swelling is usually manifested 24 to 48 hours following the initial traumatic event. CT and MRI demonstrate compressed lateral ventricles and sulci, loss of grey/ white differentiation, and effacement of the basilar cisterns. Acutely, it may not be possible to distinguish cerebral edema related to increased blood flow from that secondary to DAI. Follow-up imaging should demonstrate atrophy in children with neuronal damage. Diffusion imaging may also be useful in distinguishing these two forms of cerebral edema. Diffusion imaging demonstrates increased diffusion in areas of the brain with increased intersititial edema

286

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

and nature of lesions; with injuries of different ages or multiple injuries; with associated retinal hemorrhages; or with a change or inconsistency in the history, delay in medical care, repeated injuries, or evidence of overall poor care [6,46].

Mechanisms associated with abuse that may lead to head injury include direct impact, shaking, asphyxia, or nutritional deprivation. Whiplash or shaking is a specific and well known mechanism of injury that was initially described by Guthkelch and later defined more

Fig. 11. Nonaccidental head injury in a 3-month-old infant shaken by the father. (A,B) Axial CT images demonstrate interhemispheric (black arrow) and right convexity subdural hematomas and diffuse cerebral edema. (C) Axial T2 MR image demonstrates mild loss of cortical ribbon in the parieto-occipital regions. (D) Axial MR diffusion image and apparent diffusion coefficient (ADC) image (E) demonstrate restricted diffusion in the parietooccipital regions bilaterally, consistent with infarction.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

287

Fig. 11 (continued ).

specifically by Caffey [100 102]. It occurs when an infant is held by the caretaker, usually with two hands, and violently shaken. A wide variety of head injuries associated with child abuse may be demonstrated on CT, MRI, and skull radiographs. Some of these injuries are specific and others overlap with injuries sustained during accidental trauma. The injuries specific for child abuse will be discussed in detail here. Skull fractures The incidence of skull fractures is approximately 45% in known child abuse cases [99]. Nonaccidental skull fractures can be linear, depressed, diastatic, or complex. The parietal bone and occipital bone are most commonly involved [103]. In a comparison of fractures from abuse with fractures from accidental trauma, child abuse may be indicated by multiple or complex fractures, bilateral fractures, and fractures that cross suture lines or involve more than one bone [104,105]. Conventional radiographs are still considered the gold standard for the diagnosis of skull fractures in children; therefore skull films should still be routinely obtained as part of the radiographic skeletal survey in suspected child abuse cases. Subdural hematoma Since Caffeys original description of SDH in the setting of child abuse in 1946, SDH has come to be

regarded as one of the most characteristic CNS lesions encountered in nonaccidental head trauma [106]. In patients under the age of two, subdural hematomas are more strongly associated with inflicted head injury than with accidental trauma [107,108]. SDH is much more common than EDH in the setting of child abuse. Angular rotation of the brain within the calvarium is the mechanism of injury that has been widely accepted as the cause of subdural hematoma in whiplash-shaking injury. These hematomas most often occur posteriorly along the interhemispheric fissure [109] or in the parietooccipital region (Fig. 11A,B). Anterior interhemispheric and SDH along the tentorium may also occur. An interhemispheric hematoma appears as increased attenuation along the falx on CT imaging. Features which aid in distinguishing these lesions from the normally hyperdense falx include asymmetric thickening of the falx, and collections that are flat medially and convex laterally. In patients with an interhemispheric hematoma, MRI often shows that the hematoma extends laterally along the convexity. Large acute SDH is easily demonstrated on CT, because the hematoma appears hyperdense. It is important to recognize that in infants and children the hyperacute/acute SDH appears as a mixed high

Fig. 12. Shaken baby with subdural hematomas of different ages. Axial FLAIR MR image demonstrates bilateral subdural hematomas of different signal intensities, which was confirmed on gradient echo imaging (not shown).

288

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

indicate injuries of different ages [46] (Fig. 12). On CT, chronic SDH in the setting of nonaccidental trauma, is demonstrated by bifrontal collections that are isodense to CSF. MRI is more sensitive in the detection of small hematomas and nonacute hematomas because of its multiplanar capabilities and lack of artifact from the adjacent calvarium. MRI and ultrasound with color Doppler (Fig. 13) are superior to CT in distinguishing chronic SDH from prominent subarachnoid spaces as seen in benign external hydrocephalus. Subarachnoid hemorrhage Subarachnoid hemorrhages can be seen in patients who were shaken [101]. In the context of nonaccidental trauma, the most common location for SAH is along the sulci focally within the cerebral hemispheres or along the falx [46]. The hemorrhages along the falx can usually be distinguished from SDH by noting extension into the adjacent sulci. As previously mentioned, CT is the imaging modality of choice for SAH, although FLAIR imaging may be helpful for their detection on MR images.

Fig. 13. Ultrasound in a shaken infant with prominent extraaxial spaces on CT. (A) Coronal ultrasound images without and (B) with color Doppler are helpful in distinguishing subdural collections from enlarged subarachnoid (a) spaces.

and low density collection on CT. The high-density portion of the hematoma represents clotted blood and the low attenuation portion, unclotted blood. Reasons for this appearance may include active hemorrhaging, or an arachnoid tear with CSF extrusion into the subdural space. Until recently these SDH were presumed to be chronic with acute rebleeding, a pattern frequently seen in adults [110]. Subdural collections may have areas of different density and intensity characteristics with levels, loculations, or septations on CT and MR that may

Fig. 14. A 9-month-old strangled and shaken infant. Axial CT image demonstrates large right cerebral hemisphere and left frontal infarct with right convexity subdural hematoma and mild mass effect on the right lateral ventricle.

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

289

Fig. 15. Nonaccidental head injury in a 4-month-old infant shaken by the mothers boyfriend. (A,B) Axial CT images demonstrate cerebral edema with white cerebellar sign and interhemispheric (white arrow) and tentorial subdural hematoma. (C) Axial T2 MR image demonstrates loss of gray-white differentiation in parieto-occipital regions. (D) Axial MR diffusion and (E) ADC images demonstrate restricted diffusion in frontal and parieto-occipital regions.

290

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

with NAHI [24] (Figs. 11 and 15). In the study of Suh et al, diffusion abnormalities were noted in 90% of children with NAHI, and were multifocal or involved multiple lobes in > 90% of those patients. These injuries involved the posterior aspects of the cerebral hemispheres and spared the anterior temporal lobes and frontobasal frontal lobes (see Figs. 11C E, 15C E). This pattern differed from the accidental trauma group, in which there were more white matter shearing injuries or injuries that were focal or unilateral, restricted to a vascular distribution, or associated with an overlying fracture. Cerebral edema Cerebral edema is an important injury that commonly occurs in the setting of NAHI. Cerebral edema may be diffuse or focal. Younger infants are more likely to develop diffuse cerebral edema in the setting of trauma than older children and adults. The exact cause and pathologic events leading to diffuse cerebral edema are poorly understood. In children, the brain often responds to injury with an increase in cerebral blood flow [93]. This posttraumatic cerebral hypertension often occurs within hours of the traumatic episode. Brain swelling may subsequently lead to decreased blood flow and hypoxic-ischemic injury can ensue. Tissue hypoxia leads to cytotoxic edema, because adenosine triphosphate (ATP) is needed in order to maintain the sodium potassium pump and cell membrane homeostasis. Apnea and hypotension occuring in the immediate postinjury period are also thought to be factors that contribute to hypoxia and the development of cerebral edema in the setting of severe inflicted head injury [112,113]. Early CT findings in cerebral edema include loss of the graywhite differentiation and decreased conspicuity of the CSF spaces. The ventricles also may appear small. Over time, diffuse low attenuation is visualized, with loss of the gray white junction in both cerebral hemispheres. Often the cerebellum appears hyperdense relative to the rest of the brain. This has been called the CT reversal sign or the white cerebellar sign [99,114] (see Fig. 14A). The pathogenesis of this sign is not completely understood, but probably correlates with less edema in these regions. Infants with diffuse cerebral edema may subsequently develop border zone infarctions and cortical necrosis. The border zone infarctions occur most commonly between the anterior and middle cerebral arteries [112]. In addition, children with more focal cerebral swelling often demonstrated ipsilateral to an acute SDH may develop cortical necrosis in the involved hemisphere.

Fig. 15 (continued ).

Intra-axial injuries Intra-axial injuries that can be seen in NAHI include contusions, hematomas, infarcts, DAI, and white matter shear injuries. The location and appearance of cortical contusion do not differ significantly from those seen in accidental trauma. Contusions tend to occur in the inferior frontal lobes and anterior temporal lobes. Although DAI and intraparenchymal hematoma may occur in NAHI, they are more often present in accidental head injury [111]. Cerebral infarction may occur secondary to numerous mechanisms, including shaking; however, when the mechanism of injury includes strangulation, it is likely secondary to occlusion of the carotid artery in the neck (Fig. 14). Smothering is also a known mechanism of injury in NAHI. It can lead to cerebral infarction and hypoxic-ischemic injury. Infarctions can also occur when cerebral swelling leads to transfalcine or transtentorial herniation that compresses the anterior or posterior cerebral arteries respectively. Cerebral infarction unrelated to strangulation is a complication in NAHI that is likely underestimated [112]. Diffusion imaging compared to conventional MRI, as previously mentioned, shows increased sensitivity in the early detection of infarct in children

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294

291

Mimickers of child abuse Certain disorders have been recognized as diseases that may mimic child abuse. These include skeletal dysplasias such as osteogenesis imperfecta, which may present clinically with multiple long bone fractures or skull fractures. Other skeletal dysplasias with an increased tendency for fractures include Schmidlike metaphyseal chondrodysplasia and Menkes disease (copper deficiency) [115]. The radiographic skeletal survey usually contributes to the diagnosis of these disorders. Accidental trauma in infants that may result in significant head trauma and may mimic child abuse includes injuries associated with walkers, stairway falls, and improper immobilization in car seats [46]. In these settings, obtaining a careful clinical history may help distinguish between accidental and nonaccidental trauma. Patients with coagulation disorders may develop repeated intracranial hemorrhages that may be confused with nonaccidental head injury. Other conditions that have raised the suspicion of child abuse include diffuse fulminant encephalitides and mitochondrial diseases that simulate diffuse hypoxia-ischemia [115], as well as metabolic diseases such as glutaric aciduria type 1 that have been associated with subdural hemorrhage [116].

Acknowledgments We would like to thank Virginia Grove for manuscript preparation and Donald Sucher for photography.

References
[1] Centers for Disease Control and Prevention. Childhood injuries in the United States. Division of Injury Control, Center for Environmental Health and Injury Control, Centers for Disease Control. [see comments] Am J Dis Child 1990;144:627 46. [2] Kraus JF, Fife D, Conroy C. Pediatric brain injuries: the nature, clinical course, and early outcomes in a defined United States population. Pediatrics 1987;79: 501 7. [3] Kraus JF, Fife D, Cox P, et al. Incidence, severity, and external causes of pediatric brain injury. Am J Dis Child 1986;140:687 93. [4] Schutzman SA, Barnes P, Duhaime AC, et al. Evaluation and management of children younger than two years old with apparently minor head trauma: proposed guidelines. Pediatrics 2001;107:983 93. [5] Bruce DA, Zimmerman RA. Shaken impact syndrome. Pediatr Ann 1989;18:482 4.

[6] Schutzman SA, Greenes DS. Pediatric minor head trauma. Ann Emerg Med 2001;37:65 74. [7] Greenes DS, Schutzman SA. Clinical indicators of intracranial injury in head-injured infants. Pediatrics 1999;104:861 7. [8] Zimmerman RA, Bilaniuk LT. Pediatric head trauma. Neuroimaging Clin N Am 1994;4:349 66. [9] Kelly AB, Zimmerman RD, Snow RB, et al. Head trauma: comparison of MR and CT experience in 100 patients. AJNR Am J Neuroradiol 1988;9:699 708. [10] Zimmerman RA, Bilaniuk LT, Hackney DB, et al. Head injury: early results of comparing CT and highfield MR. AJR Am J Roentgenol 1986;147:1215 22. [11] Bruce DA. Imaging after head trauma: why, when and which. Childs Nerv Syst 2000;16:755 9. [12] Woodcock RJ, Davis PC, Hopkins KL. Imaging of head trauma in infancy and childhood. Semin Ultrasound CT MR 2001;22:162 82. [13] Atlas SW, Mark AS, Grossman RI, et al. Intracranial hemorrhage: gradient-echo MR imaging at 1.5 T. Comparison with spin-echo imaging and clinical applications. Radiology 1988;168:803 7. [14] Unger EC, Cohen MS, Brown TR. Gradient-echo imaging of hemorrhage at 1.5 Tesla. Magn Reson Imaging 1989;7:163 72. [15] Noguchi K, Ogawa T, Inugami A, et al. MR of acute subarachnoid hemorrhage: a preliminary report of fluid-attenuated inversion-recovery pulse sequences. AJNR Am J Neuroradiol 1994;15:1940 3. [16] Noguchi K, Ogawa T, Inugami A, et al. Acute subarachnoid hemorrhage: MR imaging with fluid-attenuated inversion recovery pulse sequences. Radiology 1995;196:773 7. [17] Noguchi K, Ogawa T, Seto H, et al. Subacute and chronic subarachnoid hemorrhage: diagnosis with fluid-attenuated inversion-recovery MR imaging. Radiology 1997;203:257 62. [18] Chien D, Kwong KK, Gress DR, et al. MR diffusion imaging of cerebral infarction in humans. AJNR Am J Neuroradiol 1992;13:1097 1102 [discussion: 1103 1095]. [19] Cowan FM, Pennock JM, Hanrahan JD, et al. Early detection of cerebral infarction and hypoxic-ischemic encephalopathy in neonates using diffusion-weighted magnetic resonance imaging. Neuropeds 1994;25: 172 4. [20] Assaf Y, Beit-Yannai E, Shohami E, et al. Diffusionand T2-weighted MRI of closed-head injury in rats: a time course study and correlation with histology. Magn Reson Imaging 1997;15:77 85. [21] Liu AY, Maldjian JA, Bagley LJ, et al. Traumatic brain injury: diffusion-weighted MR imaging findings. [see comments] AJNR Am J Neuroradiol 1999;20: 1636 41. [22] Rugg-Gunn FJ, Symms MR, Barker GJ, et al. Diffusion imaging shows abnormalities after blunt head trauma when conventional magnetic resonance imaging is normal. J Neurol Neurosurg Psych 2001;70: 530 3.

292

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294 99m-HMPAO SPECT in the evaluation of patients with a remote history of traumatic brain injury: a comparison with x-ray computed tomography. J Nucl Med 1992;33:52 8. Roper SN, Mena I, King WA, et al. An analysis of cerebral blood flow in acute closed-head injury using technetium-99m-HMPAO SPECT and computed tomography. [erratum 2070] J Nucl Med 1991;32: 1684 7. Mitchener A, Wyper DJ, Patterson J, et al. SPECT, CT, and MRI in head injury: acute abnormalities followed up at six months. J Neurol Neurosurg Psych 1997;62: 633 6. Bergsneider M, Hovda DA, Lee SM, et al. Dissociation of cerebral glucose metabolism and level of consciousness during the period of metabolic depression following human traumatic brain injury. J Neurotrauma 2000; 17:389 401. Ricker JH, Muller RA, Zafonte RD, et al. Verbal recall and recognition following traumatic brain injury: a [0 15]-water positron emission tomography study. J Clin Exp Neuropsychol 2001;23:196 206. Chen CY, Chou TY, Zimmerman RA, et al. Pericerebral fluid collection: differentiation of enlarged subarachnoid spaces from subdural collections with color Doppler US. Radiology 1996;201:389 92. Kleinman PK, Barnes PD. Head Trauma. In: Kleinman PK, editor. Diagnostic imaging of child abuse. St. Louis: Mosby; 1998. p. 285 342. Goraj B, Rifkinson-Mann S, Leslie DR, et al. Correlation of intracranial pressure and transcranial Doppler resistive index after head trauma. AJNR Am J Neuroradiol 1994;15:1333 9. Volpe JJ. Injuries of extracranial, cranial, intracranial, spinal cord and peripheral nervous system structures. In: Volpe JJ, editor. Neurology of the newborn. 3rd edition. Philadelphia: W. B. Saunders; 1995. p. 769 92. Gean A. Imaging of head trauma. New York: Raven Press; 1994. p. 368 426. Anton J, Pineda V, Martin C, et al. Posttraumatic subgaleal hematoma: a case report and review of the literature. Pediatr Emerg Care 1999;15:347 9. Barkovich A. Pediatric neuroimaging. 3rd edition. Philadelphia: Lippincott Williams and Wilkins; 2000. Rorke LB, Zimmerman RA. Prematurity, postmaturity, and destructive lesions in utero. AJNR Am J Neuroradiol 1992;13:517 36. Goodwin MD, Persing JA, Duncan CC, et al. Spontaneously infected cephalohematoma: case report and review of the literature. J Craniofac Surg 2000;11: 371 5. Towner D, Castro MA, Eby-Wilkens E, et al. Effect of mode of delivery in nulliparous women on neonatal intracranial injury. [see comments] N Engl J Med 1999; 341:1709 14. Volpe JJ. Intracranial hemorrhage: subdural, subarachnoid, intracerebellar, intraventricular (term infant) and miscellaneous. In: Volpe JJ, editor. Neurology of

[23] Wieshmann UC, Symms MR, Clark CA, et al. Blunthead trauma associated with widespread water-diffusion changes. Lancet 1999;353:1242 3. [24] Suh DY, Davis PC, Hopkins KL, et al. Nonaccidental pediatric head injury: diffusion-weighted imaging findings. Neurosurgery 2001;49:309 18. [25] Pierpaoli C, Jezzard P, Basser PJ, et al. Diffusion tensor MR imaging of the human brain. Radiology 1996;201:637 48. [26] Ulug AM, Moore DF, Bojko AS, et al. Clinical use of diffusion-tensor imaging for diseases causing neuronal and axonal damage. AJNR Am J Neuroradiol 1999;20:1044 8. [27] Sutton LN, Wang Z, Duhaime AC, et al. Tissue lactate in pediatric head trauma: a clinical study using 1H NMR spectroscopy. Pediatr Neurosurg 1995;22: 81 7. [28] Haseler LJ, Arcinue E, Danielsen ER, et al. Evidence from proton magnetic resonance spectroscopy for a metabolic cascade of neuronal damage in shaken baby syndrome. Pediatrics 1997;99:4 14. [29] Friedman SD, Brooks WM, Jung RE, et al. Quantitative proton MRS predicts outcome after traumatic brain injury. Neurology 1999;52:1384 91. [30] Friedman SD, Brooks WM, Jung RE, et al. Proton MR spectroscopic findings correspond to neuropsychological function in traumatic brain injury. AJNR Am J Neuroradiol 1998;19:1879 85. [31] Ricker JH, Hillary FG, DeLuca J. Functionally activated brain imaging (O-15 PET and fMRI) in the study of learning and memory after traumatic brain injury. J Head Trauma Rehabil 2001;16:191 205. [32] McAllister TW, Sparling MB, Flashman LA, et al. Differential working memory load effects after mild traumatic brain injury. Neuroimage 2001;14:1004 12. [33] Christodoulou C, DeLuca J, Ricker JH, et al. Functional magnetic resonance imaging of working memory impairment after traumatic brain injury. J Neurol Neurosurg Psych 2001;71:161 8. [34] Moritz CH, Rowley HA, Haughton VM, et al. Functional MR imaging assessment of a non-responsive brain injured patient. Magn Reson Imaging 2001; 19:1129 32. [35] Otsubo H, Snead III OC. Magnetoencephalography and magnetic source imaging in children. J Child Neurol 2001;16:227 35. [36] Lewine JD, Davis JT, Sloan JH, et al. Neuromagnetic assessment of pathophysiologic brain activity induced by minor head trauma. AJNR Am J Neuroradiol 1999;20:857 66. [37] Santyr GE, Mulkern RV. Magnetization transfer in MR imaging. Relaxometry and Biophysics Committee. J Magn Reson Imaging 1995;5:121 4. [38] Bagley LJ, McGowan JC, Grossman RI, et al. Magnetization transfer imaging of traumatic brain injury. J Magn Reson Imaging 2000;11:1 8. [39] Camargo EE. Brain SPECT in neurology and psychiatry. J Nucl Med 2001;42:611 23. [40] Gray BG, Ichise M, Chung DG, et al. Technetium-

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49] [50]

[51] [52]

[53]

[54]

[55]

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294 the newborn. 3rd edition. Philadelphia: W. B. Saunders; 1995. p. 373 402. Govaert P, Calliauw L, Vanhaesebrouck P, et al. On the management of neonatal tentorial damage. Eight case reports and a review of the literature. Acta Neurochir (Wien) 1990;106:52 64. Perrin RG, Rutka JT, Drake JM, et al. Management and outcomes of posterior fossa subdural hematomas in neonates. Neurosurgery 1997;40:1190 1199 [discussion: 1199 1200]. Hemsath FA. Birth injury of the occipital bone with a report of thirty-two cases. Am J Obstet Gynecol 1934; 27:194 203. Currarino G. Occipital osteodiastasis: presentation of four cases and review of the literature. Pediatr Radiol 2000;30:823 9. Gruskin KD, Schutzman SA. Head trauma in children younger than 2 years: are there predictors for complications? [erratum appears in Arch Pediatr Adolesc Med 1999 May;153(5):453] Arch Pediatr Adolesc Med 1999;153:15 20. Medina LS. Three-dimensional CT maximum intensity projections of the calvaria: a new approach for diagnosis of craniosynostosis and fractures. AJNR Am J Neuroradiol 2000;21:1951 4. Ersahin Y, Mutluer S, Mirzai H, et al. Pediatric depressed skull fractures: analysis of 530 cases. Childs Nerv Syst 1996;12:323 31. Ersahin Y, Gulmen V, Palali I, et al. Growing skull fractures (craniocerebral erosion). Neurosurg Rev 2000;23: 139 44. Naim Ur R, Jamjoom Z, Jamjoom A, et al. Growing skull fractures: classification and management. Br J Neurosurg 1994;8:667 79. Lende RA, Erickson TC. Growing skull fractures in childhood. J Neurosurg 1961;18:479 89. Gugliantini P, Caione P, Fariello G, et al. Post traumatic leptomeningeal cysts in infancy. Pediatr Radiol 1980;9:11 4. Rahimizadeh A, Haddadian K. Bilateral traumatic leptomeningeal cysts. Neurosurgery 1986;18:386 7. Taveras J, Ransohoff J. Leptomeningeal cysts of the brain following trauma with erosion of the skull: a study of seven cases treated by surgery. J Neurosurg 1953;10:233 41. Husson B, Pariente D, Tammam S, et al. The value of MRI in the early diagnosis of growing skull fracture. Pediatr Radiol 1996;26:744 7. Decarie JC, Mercier C. The role of ultrasonography in imaging of paediatric head trauma. Childs Nerv Syst 1999;15:740 2. Yoshioka H, Sakoda K, Kohno H, et al. Usefulness of color Doppler sonography in a growing skull fracture: case report. J Trauma 1997;42:144 6. Choux M, Grisoli F, Peragut JC. Extradural hematomas in children. 104 cases. Childs Brain 1975;1:337 47. Gentry LR. Head trauma. In: Atlas SW, editor. Magnetic resonance imaging of the brain and spine. Philadelphia: Lippincott-Raven; 1996. p. 611 47.

293

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65] [66]

[67] [68]

[69]

[70]

[71]

[72] [73]

[74] Malek AM, Barnett FH, Schwartz MS, et al. Spontaneous rapid resolution of an epidural hematoma associated with an overlying skull fracture and subgaleal hematoma in a 17-month-old child. [see comments] Pediatr Neurosurg 1997;26:160 5. [75] Barkovich AJ, Atlas SW. Magnetic resonance imaging of intracranial hemorrhage. Radiol Clin North Am 1988;26:801 20. [76] Barlow CF. CSF dynamics in hydrocephalus with special attention to external hydrocephalus. Brain Dev 1984;6:119 27. [77] Nickel RE, Gallenstein JS. Developmental prognosis for infants with benign enlargement of the subarachnoid spaces. Dev Med Child Neurol 1987;29:181 6. [78] Papasian NC, Frim DM. A theoretical model of benign external hydrocephalus that predicts a predisposition towards extra-axial hemorrhage after minor head trauma. Pediatr Neurosurg 2000;33:188 93. [79] Gentry LR, Thompson B, Godersky JC. Trauma to the corpus callosum: MR features. AJNR Am J Neuroradiol 1988;9:1129 38. [80] Gentry LR, Godersky JC, Thompson B. MR imaging of head trauma: review of the distribution and radiopathologic features of traumatic lesions. AJR Am J Roentgenol 1988;150:663 72. [81] Adams JH, Graham DI, Murray LS, et al. Diffuse axonal injury due to nonmissile head injury in humans: an analysis of 45 cases. Ann Neurol 1982;12: 557 63. [82] Gentry LR, Godersky JC, Thompson BH. Traumatic brain stem injury: MR imaging. Radiology 1989;171: 177 87. [83] Mokri B, Piepgras DG, Houser OW. Traumatic dissections of the extracranial internal carotid artery. J Neurosurg 1988;68:189 97. [84] Graham CJ, Schwartz JE, Stacy T. Stroke following oral trauma in children. Ann Emerg Med 1991;20: 1029 31. [85] Woodhurst WB, Robertson WD, Thompson GB. Carotid injury due to intraoral trauma: case report and review of the literature. Neurosurgery 1980;6:559 63. [86] Tartara F, Regolo P, Servadei F, et al. Fatal carotid dissection after blunt head trauma. J Neurosurg Sci 2000;44:103 6. [87] Camacho A, Villarejo A, Martinez de Aragon A, et al. Spontaneous carotid and vertebral artery dissection in children. Pediatr Neurol 2001;25:250 3. [88] Willis BK, Greiner F, Orrison WW, et al. The incidence of vertebral artery injury after midcervical spine fracture or subluxation. Neurosurgery 1994; 34:435 441 [discussion: 441 432]. [89] Levy C, Laissy JP, Raveau V, et al. Carotid and vertebral artery dissections: three-dimensional time-offlight MR angiography and MR imaging versus conventional angiography. Radiology 1994;190:97 103. [90] Provenzale JM. Dissection of the internal carotid and vertebral arteries: imaging features. AJR Am J Roentgenol 1995;165:1099 104. [91] Ozdoba C, Sturzenegger M, Schroth G. Internal carotid

294

T.Y. Poussaint, K.K. Moeller / Neuroimag Clin N Am 12 (2002) 271294 artery dissection: MR imaging features and clinicalradiologic correlation. Radiology 1996;199:191 8. Provenzale JM. Centennial dissertation. Honoring Arthur W. Goodspeed, MD and James B. Bullitt, MD. CT and MR imaging and nontraumatic neurologic emergencies. AJR Am J Roentgenol 2000;174:289 99. Bruce DA, Alavi A, Bilaniuk L, et al. Diffuse cerebral swelling following head injuries in children: the syndrome of malignant brain edema. J Neurosurg 1981;54:170 8. Sharples PM, Matthews DS, Eyre JA. Cerebral blood flow and metabolism in children with severe head injuries. Part 2: Cerebrovascular resistance and its determinants. J Neurol Neurosurg Psych 1995;58:153 9. Sharples PM, Stuart AG, Matthews DS, et al. Cerebral blood flow and metabolism in children with severe head injury. Part 1: Relation to age, Glasgow coma score, outcome, intracranial pressure, and time after injury. J Neurol Neurosurg Psych 1995;58:145 52. Guyot LL, Michael DB. Post-traumatic hydrocephalus. Neurol Res 2000;22:25 8. Licata C, Cristofori L, Gambin R, et al. Post-traumatic hydrocephalus. J Neurosurg Sci 2001;45:141 9. US Department of Health and Human Services, The Third National Incidence Study of Child Abuse and Neglect. Washington, DC: US Government Printing Office; 1996. Harwood-Nash DC. Abuse to the pediatric central nervous system. AJNR Am J Neuroradiol 1992;13: 569 75. Caffey J. On the theory and practice of shaking infants. Its potential residual effects of permanent brain damage and mental retardation. Am J Dis Child 1972; 124:161 9. Caffey J. The whiplash shaken infant syndrome: manual shaking by the extremities with whiplash-induced intracranial and intraocular bleedings, linked with residual permanent brain damage and mental retardation. Pediatrics 1974;54:396 403. Guthkelch AN. Infantile subdural haematoma and its relationship to whiplash injuries. BMJ 1971;2:430 1. Chen CY, Zimmerman RA, Rorke LB. Neuroimaging in child abuse: a mechanism-based approach. Neuroradiology 1999;41:711 22. [104] Hobbs CJ. Skull fracture and the diagnosis of abuse. Arch Dis Child 1984;59:246 52. [105] Meservy CJ, Towbin R, McLaurin RL, et al. Radiographic characteristics of skull fractures resulting from child abuse. AJR Am J Roentgenol 1987;149: 173 5. [106] Caffey J. Multiple fractures in the long bones of infants suffering from chronic subdural hematoma. Am J Roentgenol 1946;56:163 73. [107] Duhaime AC, Alario AJ, Lewander WJ, et al. Head injury in very young children: mechanisms, injury types, and ophthalmologic findings in 100 hospitalized patients younger than 2 years of age. Pediatrics 1992;90:179 85. [108] Hadley MN, Sonntag VK, Rekate HL, et al. The infant whiplash-shake injury syndrome: a clinical and pathological study. Neurosurgery 1989;24:536 40. [109] Zimmerman RA, Bilaniuk LT, Bruce D, et al. Interhemispheric acute subdural hematoma: a computed tomographic manifestation of child abuse by shaking. Neuroradiology 1978;16:39 40. [110] Barnes PD, Robson CD. CT findings in hyperacute nonaccidental brain injury. Pediatr Radiol 2000;30: 74 81. [111] Ewing-Cobbs L, Prasad M, Kramer L, et al. Acute neuroradiologic findings in young children with inflicted or noninflicted traumatic brain injury. Childs Nerv Syst 2000;16:25 33 [discussion: 34]. [112] Gilles EE, Nelson Jr. MD. Cerebral complications of nonaccidental head injury in childhood. Pediatr Neurol 1998;19:119 28. [113] Johnson DL, Boal D, Baule R. Role of apnea in nonaccidental head injury. Pediatr Neurosurg 1995;23: 305 10. [114] Han BK, Towbin RB, De Courten-Myers G, et al. Reversal sign on CT: effect of anoxic/ischemic cerebral injury in children. AJNR Am J Neuroradiol 1989;10:1191 8. [115] Beckett WW, Ball WS. Craniocerebral trauma. In: Ball WS, editor. Pediatric neuroradiology. Philadelphia: Lippincott-Raven; 1997. p. 443 87. [116] Woelfle J, Kreft B, Emons D, et al. Subdural hemorrhage as an initial sign of glutaric aciduria type 1: a diagnostic pitfall. Pediatr Radiol 1996;26:779 81.

[92]

[93]

[94]

[95]

[96] [97] [98]

[99]

[100]

[101]

[102] [103]

Neuroimag Clin N Am 12 (2002) 295 309

Imaging of facial trauma

Julie K. Sun, MDa,*, Daniel R. LeMay, MD, PhDb
Department of Radiology, Drew University and University of California-Los Angeles, 12021 South Wilmington Avenue, Los Angeles, CA 90059, USA b Department of Neuroscience, Drew University and University of California-Los Angeles, 12021 South Wilmington Avenue, Los Angeles, CA 90059, USA
a

Facial trauma Facial trauma seen in our emergency department, an urban level-one trauma center, is usually the result of a motor vehicle accident, fall, assault, or gunshot injury. Many of the patients have imaging studies for injuries not apparent by direct inspection. The most common reason for utilizing facial imaging in facial trauma is to identify fractures and to define their direction, extent, and displacement. Whenever possible, the fracture patterns are categorized to better convey the information to the treating physicians and to help them with making decisions. Accurate depiction of bony injuries is important in facilitating successful treatment of facial trauma, since many surgeons believe early definitive treatment improves overall outcome [1 3].

Normal anatomy and facial buttresses The facial skeleton is a system of pneumatized sinus cavities supporting and protecting the mouth, dentition, nasal fossae, and orbits, which function as attachments for the facial muscles. There are several different classifications to define the facial bones. Gentry et al [4] (Fig. 1) described the facial bones as being five sagittal, three horizontal, and two coronal supporting buttresses. The five sagittal buttresses are a midline strut formed by the cartilaginous nasal septum, the vomer, and the perpendicular ethmoid

* Corresponding author E-mail address: kwangping@hotmail.com (J.K. Sun).

plate; two parasagittal buttresses comprised of the right and left medial orbital wall, the medial wall of the maxillary sinus, and the pterygoid plates; and two lateral struts formed from the right and left lateral orbital walls, the lateral maxillary sinus walls, and the lateral maxillary alveolus. The superior horizontal strut consists of the cribriform plate, the fovea ethmoidalis, and the orbital roof. The middle horizontal buttress comprises the orbital floor and the zygomatic arch. The hard palate makes up the inferior horizontal strut. There also are two coronally oriented buttresses. The anterior coronal strut makes up the anterior facial contour and is composed of the anterior walls of the frontal and maxillary sinuses, the zygomaticofrontal buttress, the nasofrontal complex, and the anterior maxillary alveolar ridge. The posterior strut is composed of symmetrical components consisting of the posterior wall of the maxillary sinus and the pterygoid plate. The facial supporting buttresses also have been described by Sicher and Debrul [5] and are reemphasized by Manson et al [6,7]. The three sets of principal maxillary buttresses are the medial buttress (the nasomaxillary buttress), the lateral buttress (the zygomaticomaxillary buttress), and the posterior buttress (the pterygomaxillary buttress). The medial buttress extends from the anterior maxillary alveolus up the lateral wall of the pyriform aperture and the nasal process of the maxilla to the frontal cranial attachment. The lateral buttress is formed by the lateral maxillary alveolus to the zygomatic process of the frontal bone and laterally to the zygomatic arch. The posterior buttress attaches the maxilla posteriorly to the pterygoid plate of the sphenoid bone.

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 2 - 3

296

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Fig. 1. Major struts and buttresses of the face. Abbreviations: SHS, superior horizontal strut; MHS, middle horizontal strut; IHS, inferior horizontal strut; ACS, anterior coronal strut; PCS, posterior coronal strut; MSS, median sagittal strut; PS, parasagittal strut; LSS, lateral sagittal strut.

The nasomaxillary and the zygomaticomaxillary buttresses are most important in the treatment of facial bony injuries. The exact anatomic reconstruction of the buttresses, or the frontal and lateral bony contour

of the face, permits correct anteroposterior position of the face in relation to the cranial base and reconstruction of the exact vertical height and horizontal projection of the maxilla [2] (Fig. 2). An intact mandible

Fig. 2. Diagrammatic representation of the maxillary buttresses showing the two anterior buttresses (medial or nasomaxillary and lateral or zygomaticomaxillary) and the posterior buttress (peterygomaxillary).

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

297

provides an additional stabilizing force in relation to the cranial base and the reestablishment of normal maxillomandibular occlusion.

Imaging techniques High-resolution computed tomography (CT) has been the most important imaging modality in facial trauma due to its high tissue contrast and clear demonstration of bony fractures. The information gained from a CT scan is often of greater value than that from a combination of clinical examination and routine radiography for all facial fractures except mandibular fractures, which can be well demonstrated by Panorex [21]. In a stable patient, typically 2 3 mm axial images are obtained, which are reviewed in both the soft tissue window (approximately 300/25) and using a high detail/bone algorithm (1500 2000/300 500) for enhanced fracture detection. Coronal images can be helpful in better visualization of fractures through the horizontally oriented bony structures such as the orbital roof, cribriform plate, and the orbital floor; and can give a different perspective in maxillary, nasal, and pterygoid fractures. If the patient cannot be positioned for coronal imaging, a two-dimensional multiplanar reconstruction can be helpful. A threedimensional surface rendering may also provide additional information to a treating physician. In a patient with possible intracranial injury or other life threatening injury, imaging of these injuries takes precedence. With the development of the multidetector CT scanner, the images can now be obtained at thinner slices with higher resolution and less artifact. Fracture detection by conventional radiography alone is limited because of its relatively low tissue contrast and the presence of overlying structures. The use of plain film in facial trauma is usually limited to nasal bone fracture, orbital floor fracture, and an isolated zygomatic arch injury. Mandibular fractures are often imaged by conventional radiography and Panorex, which can include the entire bone on one film. High-resolution facial CT is clearly superior to plain film in discerning facial fractures, and has mostly replaced tomography. Maganetic resonance imaging (MRI) is not usually used for facial fracture.

separated into low-energy, middle-energy, and highenergy injuries [14]. By anatomic location of bony injuries, facial fractures can be classified into limited fractures, transfacial fractures, and smash fractures [8] (Table 1). Limited fractures can involve only one buttress strut (simple fractures) or two adjacent buttress struts (complex fractures). A transfacial fracture involves the pterygoid plates by definition. A smash fracture is a severely comminuted fracture that does not follow the pattern of a traditional transfacial fracture. Low-energy injuries usually cause fractures with minimum comminution or displacement, and are accompanied by subtle symptoms, requiring simple and usually more conservative treatment. Middle-energy injuries comprise the majority of injuries seen in the emergency department. They often demonstrate mild to marked displacement, and many of them require standard methods of open reduction and internal fixation. A small fraction of injuries are categorized as high-energy fractures with dramatic instability and marked alterations in facial architecture that justify an extended exposure for reduction and fixation.

Table 1 Classification of facial fractures Limited fractures Simple fracture Nasal fracture Maxillary alveolar fracture Zygomatic arch Localized sinus wall Frontal Maxillary Mandibular fracture Complex fracture Nasofrontal Nasomaxillary Nasoethmoidal Zygomaticomaxillary Malar (tripod) fracture Transfacial fractures (require pterygoid plate fractures) Le Fort I Le Fort II Le Fort III Le Fort II and Zygomaticomaxillary complex fracture (most common) Combination of Le Fort fractures Smash fractures Naso-ethmoidal-orbital Midface Craniofacial Modified from Kassell EE, Gruss JS. Imaging of midface fractures. Neuroimaging Clin N Am 1991;1:259 83.

Facial fractures There are several classifications for facial fractures [8 14]. Facial fractures are commonly defined by the anatomic location involved. The injuries can also be

298

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Fig. 3. An assaulted victim with bilateral zygomatic arch fractures.

Fig. 4. Status post motor vehicle accident (MVA) with fractures of both tables of the frontal sinus. The left orbital roof and the squamosa portion of the right frontal bone are also fractured. There is small pneumocephalus.

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

299

Fig. 5. (a). Bilateral parasymphyseal fractures of the mandible in an MVA patient. (b) Bilateral mandibular condylar fractures in another patient after MVA.

300

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Limited fractures Nasal fractures Nasal injuries are the most common fractures of the face; approximately 50% of facial fractures are isolated fractures of the nasal pyramid [8 12,14]. Two-thirds of nasal fractures are the result of a lateral force, and only 13% result from a frontal impact [15]. The majority of nasal fractures involve the thinner, distal third of the nasal bones [16]. Nasal fractures are usually visualized by a coneddown lateral view of the nasal bones. Waters view may also help. Any transverse lucent line across the nasal bones is suggestive of a fracture. Both the nasal bones and the nasal cartilage can fracture. The nasal septum and the frontal process of the maxilla may also be involved. Clinically, the patient can have a saddle nose deformity, splaying of the nose, hypertelorism, or telecanthus. Hemorrhage from ruptured anterior or posterior ethmoidal arteries may also happen. There is poor correlation (6.6 10%) between a fracture demonstrated by the plain film and the clinical indication for surgical reduction [15].

If the fracture is predominantly unilateral, with or without septal fracture, a simple closed reduction usually is sufficient. For bilateral injuries with both lateral and posterior dislocation, incomplete fractures may be surgically extended until complete, and then closed reduction and external nasal splinting are often the treatment of choice. In a highly fragmented isolated nasal fracture, external bolsters or an open reduction with or without dorsal nasal bone grafting may be required [13]. Maxillary alveolar fractures Maxillary alveolar fractures are the most common isolated maxillary fractures. The force is usually an upward blow pushing the mandible into the maxilla and fracturing the alveolus. The injury can also occur with a direct blow to the anterior maxilla, where the fracture line extends down to the alveolus. With highenergy injury, the palate can be fractured along the sagittal plane. These fractures can also be part of the Le Fort I fracture. Teeth may fracture or become devitalized in maxillary alveolar fractures. In children, the tooth germs can be damaged [16].

Fig. 6. Illustration of a malar (tripod) fracture.

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

301

Fig. 7. (a) Three of the four elements fractured in a patient with a malar (tripod) fracture: the right zygomatic arch, anterior inferior right orbital rim extending to the anterior maxillary sinus wall, and lateral right maxillary sinus wall. (b) The right lateral orbital wall is also fractured in the same patient. (c) The right tripod fracture is seen in the coronal plane, demonstrating fractures of the right lateral orbital wall, right inferior orbital rim involving the anterior aspect of the orbital floor, and right lateral sinus wall.

302

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Fig. 7 (continued ).

Because of the strong surrounding soft tissue support, isolated maxillary alveolar fractures rarely have significant displacement, and surgery is usually not required. Frontal sinus fractures Frontal sinus fractures are the result of either direct trauma over the forehead, or are a part of complex fractures involving the adjacent bony structures (Fig. 3). The majority of frontal sinus fractures are limited to the anterior table. The anterior table fracture often extends down to the superior orbital rim. Isolated posterior table involvement is rare [16]. Both table fractures involving the frontal sinus are often associated with other midface osseous injuries. When the posterior table of the frontal sinus is fractured, there may be cerebrospinal fluid (CSF) leak or communication with the dural spaces, through which intracranial infection can spread. Frontal sinus fractures are best visualized by CT using the bone window. Sinus mucosal thickening and an air-blood level in the frontal sinus in a posttraumatic patient can direct ones attention to search for possible fractures. If only plain films are available, lateral projection may better demonstrate the fractured fragment and the air-blood level.

If only the anterior table of the frontal sinus is fractured, elevation of depressed fragments may be done for aesthetic purpose. When both tables of the frontal sinus are fractured, the sinus may be obliterated or cranialized, depending on the status of the posterior table. Mandibular fractures Mandibular fractures are very common, exceeded only by nasal and zygomatic fractures in adults, and nasal fractures in children (Fig. 4). The mandible is the strongest bone in the face, but has several weak areas that are predisposed to fracture. These areas include the level of the mental foramen, the tooth sockets, crypts of impacted teeth, and the condylar neck. The body and condylar process of the mandible account for two-thirds of mandibular fractures. Since it is a rigid structure firmly attaching to the skull base at the temporomandibular joint, up to 50% of the fractures occur bilaterally. The mandibular body and angle fractures can be due to a blow to the mandibular symphysis or to the side of contralateral mandible. Communication with a tooth root converts a mandibular fracture into an open injury. Clinically, mandibular fractures can be diagnosed by malocclusion or displacement of the teeth. Radio-

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

303

Fig. 8. (a) Left malar (tripod) fracture, fracturing the left zygomatic arch, anterior, and posterolateral left maxillary sinus walls. There is overlying facial soft tissue swelling. (b) The left lateral orbital wall is fractured in the same patient. The left orbital floor and skull base are also fractured in the same patient, with minimal pneumocephalus.

304

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

graphically, a Panorex view can clearly demonstrate the injuries from condyle to condyle. The fractures can also be readily shown by CT and conventional radiography. Significant complications include mandibular growth arrest, condylar avascular necrosis, and TMJ ankylosis. Treatment includes closed reduction using arch bars, splints, and intermaxillary fixation devices; or open reduction and internal fixation using microplates, versus external fixation. Lateral midface fracture Lateral midface fractures include the zygomatic arch fractures (Fig. 5), malar (tripod) fractures, zygomaticomaxillary fractures, zygomaticomandibular fractures, and fractures of the orbital floor (blowout fractures). Malar (tripod) fractures are second only to nasal fractures in frequency [12] (Figs. 6 8). The name malar fractures may be more appropriate than tripod fractures, since the fractures involve four sutures: the zygomaticofrontal, the zygomaticosphenoid, the zygo-

maticomaxillary, and the zygomaticotemporal sutures. On CT imaging, four structures are usually seen fractured: the lateral orbital wall, the zygomatic arch, the anterior maxillary sinus wall involving the anterior orbital floor, and the posterolateral maxillary sinus wall. The fracture line usually involves the infraorbital rim and the orbital floor lateral to the infraorbital canal, in contrast with Le Fort II fractures, which usually occur medially. Since the fractures can be very close to the infraorbital foramen, the infraorbital nerve is involved in 94.2% of the cases [15]. In one study, up to one-third of all patients suffering comminuted malar fractures had an ocular problem [17]. Malar (tripod) fractures can also occur in conjunction with other fractures, such as Le Fort or smash injuries, when the fracture is caused by a high-energy injury. When the fracture includes a maxillary segment, the injury becomes a zygomaticomaxillary fracture. The fracture line usually involves the orbital floor, extending down the anterior inferior orbital rim near the infraorbital foramen, through the premolar region, and across the palate to the pterytgoid plates.

Fig. 9. Illustration of Le Fort fractures. Le Fort I fracture involves the maxillary alveolar process and all walls of the maxillary sinus creating a floating palate. Le Fort II fracture is also known as a pyramidal fracture separating the midface from the skull and the lateral midface creating a dish face deformity. Le Fort III fracture causes craniofacial disassociation. All Le Fort fractures involve the pterygoid plates.

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

305

Transfacial fractures Transfacial fractures include Le Fort I, Le Fort II, Le Fort III, or combinations of Le Fort fractures and Le Fort fractures plus zygomaticomaxillary complex fractures (Figs. 9 11). Le Fort I fractures The Le Fort I fracture is also known as the Guerins fracture, and results from a blow over the upper lip region, causing a horizontally oriented fracture separating the hard palate and maxillary alveolus from the remainder of the midface. The fracture courses from the lateral border of the pyriform sinus across the lateral maxillary sinus wall, behind the tuberosity, and across the pterygoid junction. The fracture extends posteriorly to the pterygoid plate, like all Le Fort fractures. All walls of the maxillary sinus and the nasal septum are involved, and patients are described as having floating palates. The fracture is better visualized by coronal CT if the patient can tolerate the position.

Le Fort II fractures A strong, broad blow to the midface can cause the Le Fort II fracture, which is also known as the pyramidal midface fracture because of its triangular configuration. It is the most common type of Le Fort fracture, separating the midface from the skull and the lateral midface. The fracture line typically starts at the root of the nose, extending across the lacrimal bones and the medial orbital walls, turning anteriorly along the orbital floor, across the anterior and posterolateral maxillary sinus walls, and terminating at the pterygoid plate. The fracture causes a dish face deformity. The Le Fort II fracture is better demonstrated by axial CT, except for the pterygoid fracture, which is better seen by coronal CT. Le Fort III fractures The Le Fort III fracture is usually the result of a high-energy impact causing complete separation of the face from the skull base. This fracture involves both the central and lateral mid face bony structures,

Fig. 10. A patient sustained right Le Fort I, II, zygomaticomaxillary complex (ZMC) fracture, and left Le Fort I fracture in an MVA. The lateral and medial maxillary sinus walls and the nasal septum are fractured. There is also sagittal fracture of the palate on the left. Fractures of the right orbital floor and lateral orbital wall are also seen as part of right Le Fort II and right zygomatical complex fractures.

306

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Fig. 11. Axial view of bilateral Le Fort I fracture. All maxillary sinus walls, nasal septum, and bilateral pterygoid plates are fractured (right pterygoid fracture is not seen in the picture).

resulting in complete craniofacial separation. The fracture begins similarly from the root of the nose, extending across the lacrimal bones and the medial orbital walls. The fracture line then travels posterolaterally to involve the lateral orbital wall superiorly. The zygomatic arch is also fractured. The second component of the fracture line passes posteriorly through the posterior wall of the maxillary sinus to involve the pterygoid plate. This fracture has high association with intracranial injury. The lacrimal apparatus and the infraorbital nerve may also be involved. Midface elongation and retrusion are the classical clinical characteristics of untreated Le Fort fractures. The goal of Le Fort fracture treatment is to reestablish the height and projection of the midface using an arch bar, or open reduction and internal fixation using the microplates, miniplates, and arch bars as necessary. Smash fractures This category of injuries includes fractures of the face that do not fall into other facial fracture classi-

fications. They are usually highly comminuted fractures resulting from a high-energy impact. They have been further subdivided into nasoethmoid smash, frontal smash, central midface smash [18], and total craniofacial smash [12]. These fractures have a high association with intracranial injury. The patients are often medically unstable. Life threatening injuries need to be evaluated and treated first, and conventional or CT angiography may be necessary to exclude a carotid artery injury before full evaluation of the facial fractures. Nasoethmoidal-orbital fractures The nasoethmoidal-orbital fracture is often the result of a direct medium- or high-energy blow over the nasal bridge (Figs. 12,13). Five main types of nasoethmoidal-orbital fracture patterns have been described by Gruss, based on the injury extent, displacement, orbital complication, and associated facial fractures [10]. The nasal pyramid is displaced posteriorly, causing multiple fractures in an accordion-like fashion. The nasal bones, nasal septum, ethmoid bones (including the cribriform plate), lac-

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

307

Fig. 12. A patient sustained nasoeithmoidal-orbital fracture in an MVA. The nose, nasal septum, bilateral ethmoid bones, and bilateral medial orbital walls are fractured. He also has multiple other injuries, including left lateral orbital wall and right superior orbital rim.

rimal bones, frontal process of the maxilla, and walls of the frontal sinus are often involved. The fractures often dissipate sufficient energy and provide a protective mechanism for the orbits and the central skull base. There can be adjacent soft tissue injuries, such as to the lacrimal apparatus (in 20%), the medial canthal tendon, and the dura adherant to the frontal bone as well as to the cribriform plate. Hypertelorism, telecanthus, visual loss (in 30%), CSF rhinorrhea (in 40%) [19], nasofrontal duct injury (in 67%) [20], and intracranial spread of infection may also occur. The fractures are best visualized by CT, which can also detect associated soft tissue injury to the orbits, cranium, or the intracranial contents. Pediatric facial fractures Pediatric facial structures are more elastic, with a thick periosteum. They behave differently in injury, in the pattern of fracture, healing, and treatment of choice.

The fractures often are incomplete fractures or green stick fractures. Healing is faster, which means if there is delay in treatment or the bones are not properly set, the bones tend to heal in a post-traumatic deformity. The tooth germs can also be injuried, and the problem may not manifest until permenent teeth erupt. The facial ossification centers can also be affected, resulting in osseous hypoplasia, functional abnormalities, and cosmetic deformities.

Summary Facial trauma is a commonly encountered injury in the emergency department. Facial fractures can be categorized into limited, transfacial, and smash fractures. Limited fractures may involve one or two adjacent facial supporting struts. Transfacial fractures are classically the Le Fort fractures. Smash fractures are comminuted fractures that do not follow classical facial fracture patterns. Depending on the severity of injury, treatment may be conservative or surgi-

308

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309

Fig. 13. Another patient with nasoethmoidal-orbital fracture. The nose, nasal septum, bilateral ethmoid bones, and medial orbital walls are fractured. The right lateral orbital wall is also fractured, which is not part of the nasoethmoidal-orbital fracture.

cal. Computed tomography is superior to conventional radiography, tomography, and MRI in detecting facial fractures, defining their direction, extent, and displacement.

[5] [6]

Acknowledgments
[7]

Thank you to Lorraine M. Smith, MD, MPH, FACS, and Tom Sanders, MD, for their editorial assistance.

[8]

References
[9] [1] Gruss JS, Kassel EE, Bubak P. Clinical, surgical and treatment perspectives in the management of craniomaxilaofacial trauma. Neuroimaging Clin N Am 1991; 1:341 55. [2] Gruss JS, Mackinnon SE. Complex maxillary fractures: role of buttress reconstruction and immediate bone grafts. Plast Reconstru Surg 1986;78:9 22. [3] Marciani RD, Gonty AA. Principles of management of complex craniofacial trauma. J Maxillofac Surg 1993; 51:535 42. [4] Gentry LR, Manor WF, Turskip A, Strother CM. High-

[10]

[11]

[12] [13]

resolution CT analysis of facial struts in trauma: 1. Normal anatomy. AJR Am J Roentgenol 1983;140: 523 32. Sicher H, DeBrul EL. Oral Anatomy. 5th edition. St. Louis: Mosby; 1970. p. 78. Manson PN, Hoopes JE, Su CT. Structural pillars of the facial skeleton: an approach to the management of Le Fort fractures. Plast Reconstr Surg 1980;66:54 61. Manson PN, Crawley WA, Varemchuk MJ, et al. Midface fractures: advantages of immediate extended open reduction and bone grafting. Plast Reconstr Surg 1985; 76(1):1 12. Arden RL, Mathog RH. Nasal fractures. In: Cummings CW, Fredrickson JM, Harker LA, et al., editors. Otolaryngologyhead and neck surgery. 2nd Edition. St. Louis: Mosby Year Book; 1993. p. 737 53. Gentry IR, Smoker WRK. Computed tomography of facial trauma. Semin Ultrasound CT MR 1985;6: 129 45. Gruss JS. Naso-ethmoid-orbital fractures: classification and the role of primary bone grafting. Plast Reconstruct Surg 1985;75:303 15. Jackson IT. Classification and treatment of orbitozygomatic and orbitoethmoid fractures. Clin Plast Surg 1989;16:77 91. Kassel EE, Gruss JS. Imaging of midface fractures. Neuroimaging Clin N Am 1991;1:259 83. Manson PN, Markowitz B, Mirvis S, et al. Toward CT-

J.K. Sun, D.R. LeMay / Neuroimag Clin N Am 12 (2002) 295309 based facial fracture treatment. Plasti Reconstr Surg 1990;85:2. Zigg M, Laedrach K, Chen J, et al. Classification and treatment of zygomatic fractures: a review of 1025 cases. J Oral Maxillofac Surg 1992;50:778 90. Rowe NL, Williams JL, editors. Maxillofacial injuries. vol. 1. Edinburgh, Scotland: Churchill Livingstone Inc; 1985. p. 363 558. Schwenzer N, Kruger E. Midface fracture. In: Kruger E, Kruger E, Schilli W, Worthington P, editors. Oral and maxillofacial traumatology, vol. 2. Chicago: Quintessence Publishing Co, Inc.; 1986. p. 107 36. al-Qurainy IA, Stassen LF, Dutton GN, et al. The char-

309

[14]

[18] [19] [20] [21]

[15]

[16]

[17]

acteristics of midface factures and the association with ocular injury: a prospective study. Br J Oral Maxillofac Surg 1991;29:291 301. Dolan K, Jacoby C, Smoker W. The radiology of facial fractures. Radiographics 1984;4:577 663. Cruse CW, Blevins PK, Luce EA. Naso-ethmoidalorbit fractures. J Trauma 1980;20:551 6. Harris L, Marano GD, McCorkle D. Nasofrontal duct: CT in frontal sinus trauma. Radiology 1987;165:195 8. Stanley Jr RB. Maxillofacial trauma. In: Cummings CW, Fredrickson JM, Harker LA, et al., editors. Otolaryngology-head and neck surgery. 2nd Edition. St. Louis: Mosby Year Book; 1993. p. 374 402.

Neuroimag Clin N Am 12 (2002) 311 324

Orbital trauma
John L. Go, MD*, Vy Nguyen Vu, MD, Kichun Jason Lee, MD, Terry S. Becker, MD
Division of Neuroradiology, Department of Radiology, Keck School of Medicine, University of Southern California, 1200 North State Street, Room 3740F, Los Angeles, CA 90033, USA

Postacceleration injury, motor vehicular collisions, and violent crimes may result in a myriad of injuries involving the bony orbit or its contents. In a study by Rootman et al, 4.5% of all orbital pathologies were related to trauma [1]. In the United States, traumatic accounts for over 30,000 hospital admissions per year, as well as 2.5 million eye injuries every year alone [2]. In assessing the patient with suspected orbital trauma, a basic understanding of anatomy in this region is necessary to determine the presence and extent of disease. A comprehensive review of the anatomy of the orbit and its contents are beyond the scope of this article, and the reader is advised to refer to the many excellent texts on the subject. A brief review of the anatomy in this location is provided below.

Anatomy The bony orbit is a conical structure, with its base facing anterolaterally and its apex originating posteromedially. It is comprised of seven bones: (1) ethmoid, (2) frontal, (3) lacrimal, (4) maxillary, (5) palatine, (6) sphenoid, and (7) zygomatic. The anterior rim of the bony orbit, the orbital rim, is formed by orbital processes of the maxilla, zygoma, and the frontal bone. The orbital roof is comprised of the orbital process of the frontal bone and the lesser wing of the sphenoid bone. The medial wall of the bony orbit is comprised of the ethmoid bone, body of the sphenoid bone, and frontal process of the maxillary

* Corresponding author. E-mail address: jlgo@hsc.usc.edu (J.L. Go).

bone. The ethmoid sinus parallels the medial orbital wall (lamina papyracea). The anterior aspect of the medial orbital wall is formed by the lacrimal bone, encasing the nasolacrimal duct. The orbital floor is comprised of the orbital processes of the zygomatic bone, maxillary bone, and small component of the palatine bone. The lateral orbital wall is unique in that it is not bordered by a sinus and is the thickest of the orbital walls. It is comprised of the orbital processes of the greater wing of the sphenoid bone, zygomatic bone, and the orbital process of the frontal bone. The zygomaticofrontal suture is seen superiorly, and the zygomaxillary suture inferiorly. Not directly associated with the orbit, the temporal process of the zygoma extends posterolaterally to form the anterior component of the zygomatic arch. The posterior third of the orbit, the orbital apex, is comprised by the optic canal and superior orbital fissure. The optic canal is bounded by the superior and inferior roots of the lesser wing of the sphenoid bone. The superior orbital fissure, found inferolateral to the optic canal, is created by the greater and lesser wings of the sphenoid bone. The inferior orbital fissure, found in the inferolateral portion of the orbit, is bounded by the lesser wing of the sphenoid above, inferiorly by the palatine and maxillary bone, and laterally by the zygomatic bone. Two foramina are also found in the medially wall of the orbit through which the superior and inferior ethmoidal nerves enter the orbit. The periosteum covering the orbit, the periorbita, loosely attaches to the bony orbit but is more adherent to the bony orbit at the orbital rim. Anteriorly, the preorbital soft tissues are separated from true orbital contents by a physical barrier produced by a sheet of collagenous tissue, the orbital septum. The periosteum of the orbit merges with the

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 1 2 - 6

312

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

orbital septum. The orbital septum separates the contents of the orbit posteriorly, the postseptal space, from the soft tissue anterior to the septum, the preseptal space. The tissue found in the preseptal space includes the eyelids, orbicularis oculi muscles, tarsal plates, and tarsal glands. The globe is nearly spherical in shape and can be divided into an aqueous and vitreous chamber by the lens and ciliary body. Anteriorly, the aqueous chamber can be divided into a small anterior and posterior chamber by the iris. The space between the cornea and the iris is the anterior chamber. The posterior chamber lies between the iris and the lens and ciliary body. The lens is held in place by the zonular fibers arising from the ciliary body. The cornea constitutes the transparent outer layer of the aqueous chamber and, with the sclera, comprises the outer layer of the globe. The uveal tract, which includes the choroid, iris, and ciliary body, comprises the middle layer. The deepest layer is the retinal layer. These individual layers are usually not differentiated on CT or MR imaging in the normal globe. The globe is enclosed within a fibrocartilageneous sheet called Tenons capsule. The optic nerve extends from the posterior pole of the globe at the fovea centralis and courses posteromedially back toward the optic canal, where it exits the orbit toward the optic chiasm. The optic nerve may pursue a sinusoidal course, reflecting its normal 0.7 cm length beyond the direct globe-toorbital apex distance, which allows for ocular motion. Though the optic nerve and its sheath may not be visually separable on CT, MR imaging nicely displays cerebral spinal fluid circumferentially contained within the optic nerve sheath [3]. The superior, inferior, medial, and lateral recti muscles originate from the annulus of Zinn at the orbital apex and insert on the globe just beyond its widest transverse dimension. These extraocular muscles form a cone, dividing the postseptal space into an intraconal and extraconal compartment. From the orbital apex just medial to the annulus of Zinn along the lesser wing of the sphenoid bone, the superior oblique muscle originates and courses superomedially toward the supramedial orbital rim. At the trochlea anteriorly, the superior oblique tendon turns laterally and inserts along the superolateral aspect of the globe past the equator. The levator palpebrae superioris muscle travels superior to the superior rectus muscle, and the inferior oblique muscle is seen inferior to the inferior rectus muscle. The levator palpebrae superioris muscle and superior rectus muscle cannot be readily separated on CT and may be referred to as the levator palpebrae superiorissuperior rectus muscle complex. On MR imaging,

these two muscles can easily be distinguished from each other. The inferior oblique muscle arises from the inferomedial portion of the bony orbit and travels below the inferior rectus muscle to insert along the posterolateral portion of the globe just past the equator. The lacrimal gland is seen as a homogeneous density on CT at the superolateral aspect of the orbit, just anterior to the globe within the lacrimal fossa. The superior opthalmic vein is readily identified at the superior aspect of the orbit, coursing from lateral to medial toward the apex and is easily identified on both CT and Mr imaging [3].

Imaging Plain films of the orbits are no longer advocated for the diagnosis of fractures in todays era of available high-speed helical CT machines. In a recent series of 59 orbital blow-out fractures collected by Brady et al from 1994 1998, the number of false negatives and nondiagnostic results from plain radiographs were 50% and 30%, respectively. [4] In daily practice, however, plain films of the orbits continue to be ordered by referring clinicians. In that case, orbits plain film study should optimally include: direct frontal (PA or AP), Caldwell, Waters, Towne, and lateral views. Classically, plain film findings for orbital floor blow-out fractures have been extensively described, particularly the trapdoor sign. The floor fragment typically remains attached medially, similar to a hinge so that the fragment is characteristically lateralsloping. Often though, the entire floor fragment can be depressed into the subjacent maxillary sinus. If mildly displaced, it may lie parallel (inferior) to its original position (Figs. 1 and 2). Depending on orientation, the depressed floor fragment can be seen as a nonanatomic density in the maxillary antrum, or the displaced floor segment could be seen to disappear altogether if it were oriented in a way that inadequately attenuated the x-ray beam. An orbital floor blow-out fracture with frank enophthalmos would appear as a bulbous softtissue mass extending from the expected level of the orbital floor into the maxillary antrum beneath [3]. CT is the study of choice for orbital fractures. Of all radiologic imaging modalities, CT (slices 3 mm or less) best illustrates fine bony structures of the midface and orbits. Axial images are performed with slices parallel to the infraorbital meatal line [5]. Coronal slices are obtained perpendicular to the axial plane and extend posteriorly to include the optic chiasm. Slice thickness of 3 mm or less is desired. Coronal images provide very good assessment of all

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

313

Fig. 1. Orbital anatomy. (A) Coronal CT scan of the orbit obtained in the posterior orbit at the orbital apex. (B) Coronal CT scan through the mid orbit posterior to the globe, demonstrating the major intraorbital contents. (C) Fat suppressed post contrast T1-weighted coronal image of the orbit obtained at the equator of the globe anterior to the insertion of the inferior rectus muscle. Note that there is separation of the superior rectus and levator palpebrae superioris muscles in the superior orbit. These two muscles are not clearly separated on coronal CT. (D) Axial CT scan through the orbit at the level of the optic nerve. Abbreviations: LW, lesser wing of the sphenoid bone; GW, greater wing of the sphenoid bone; SOF, superior orbital fissure; IOF, inferior orbital fissure; SRM, superior rectus muscle; SOM, superior oblique muscle; IOM, inferior oblique muscle; MRM, medial rectus muscle; IRM, inferior rectus muscle; LRM, lateral rectus muscle; COR, cornea; AC, aqueous chamber; CB, ciliary body; VC, vitreous chamber; LG, lacrimal gland; LPS, levator palpebrae superioris muscle; ON, optic nerve and nerve sheath.

orbital walls, especially the orbital floor and roof. False negative interpretation of coronal images could be from an unusual fracture directed superior-inferiorly in the same plane of section. Axial plane imaging, though more comfortable for the patient as performed easily in the supine position, is very good for fractures of the medial and lateral walls. But it is suboptimal for evaluation of the orbital roof and floor because fractures along those structures and any displaced fragments (superior-inferiorly) are commonly in the plane of section and thus not ideally defined. Sagittal reconstructions obtained from the axial data set should be performed when the slice thickness is less than 3 mm to prevent stair-step artifacts. Current multislice detector CT scanners can acquire axial images with an effective slice thickness of less than 1 mm. As a result, sagittal reconstructions are possible without stepladder artifacts.

Indirect findings on CT and even plain films include asymmetrical opacification by hemorrhage

Fig. 2. Waters view of the skull demonstrating a depressed orbital floor fracture on the left side.

314

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

of a paranasal sinus adjacent to a particular orbital surface. For example, an air-fluid level in the maxillary antrum raises suspicion of orbital floor injury. Unilateral opacification of the ethmoid air cells would raise suspicion of possible medial wall fracture. Another indirect plain film and CT finding is orbital emphysema. This pathologic collection of air is seen as a lucency at the superior-posterior aspect of the orbit. Though passage of air into the orbit can theoretically occur via communication with any violated adjacent paranasal sinus cavity, orbital emphysema when detected on plain films is most frequently from a medial wall blow-out fracture. Medial orbital wall and orbital roof fractures are poorly visualized by plain films. They are best defined by computed tomography (Figs. 3 4). MR imaging is not applicable for the initial assessment of orbital injuries in the setting of trauma. The most important reason for this is that the possibility of a metal foreign body within the orbit is an absolute contraindication. The potential for move-

Fig. 4. Axial CT scan at the level of the optic nerve on the left, demonstrating orbital emphysema around the left optic nerve. A small amount of air is also seen adjacent to the globe medially.

ment of a ferromagnetic foreign body within the fluctuating magnetic fields of a MR machine could result in blindness. This catastrophic potential is particularly worrisome in the trauma setting, where the patients history is frequently not known. Once metallic fragments have been screened by plain films of the orbit (or CT), however, MR imaging can image injury to secondary soft-tissue structures such as the optic nerve or globe.

Orbital fractures Orbital fractures may result from a direct blow to the orbit or be the result from injuries sustained to the midface. Orbital fractures sustained from midface fractures such as Le Fort and Tripod fractures are discussed elsewhere in this issue. This next section will primarily deal with blows directed toward the bony orbit. Blow-out fractures are produced by a nonpenetrating trauma to the anterior periorbital region, resulting in outward transmission of force to the orbital walls. The fracture fragment in a blow-out fracture is directed away from the bony orbit. Orbital blow-in fractures are produced by blunt trauma, most commonly against the frontal bone or the maxilla, with resultant energy transmission toward the orbital roof or floor, respectively. Fracture fragments in this case are thus characteristically displaced toward the orbital space. Pure forms of blow-out and blow-in fractures refer to the internal orbital skeleton, [6] with the orbital rim remaining intact. Of course, with impure complex orbit fractures, the rim is also fractured. The two accepted mechanisms of orbital blow-out fracture are: (1) the hydraulic theory and 92) the

Fig. 3. (A) AP view of the skull, demonstrating orbital emphysema within the right orbit secondary to an inferior orbital floor fracture. The gas (arrows) outlines the right globe. (B) AP view of the orbit, demonstrating orbital emphysema outlining the optic nerve sheath (arrow) on the right side.

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

315

buckling theory. In their classic experiment, Smith and Reagan [7] placed a Gaelic hurling ball over intact cadaver orbits and struck it with a hammer. This produced classic blow-out fractures of the orbital floor, with the anterior bony rim remaining intact. Their work has traditionally been cited in support of the hydraulic theory, wherein compression and intraorbital soft-tissues produces generalized outward pressure on all orbital walls [8]. The weakest portion of the orbit, the floor, would fracture most often. In conjunction with this theory, Erling et al suggest that after a blow, retropulsion of the globe itself toward the orbital apex can initiate fractures by direct globe-to-wall contact to the floor and the medial orbital wall (then rebounding) [9]. They cited examples (75%) from 20 randomly selected CT scans where orbital wall/floor fracture contours appear to accommodate the curvature and dimensions of their patients globe [9]. In contrast, the buckling theory stipulates direct trauma to the inferior orbital rim, which would then transmit energy posteriorly to create a compressiontype fracture of the orbital floor. Fujino and et al in the early and mid 1970s set up experiments where they struck the globe alone, the infraorbital rim alone, or in combination [10]. All three methods produced orbital wall fractures. Using high-speed cineradiography, Fujino et al showed elastic motion of the inferior orbital rim during blow-out fracture and subsequent return of the rim to its original position [10]. They concluded that three times as much force was required to fracture their orbital floor model when the globe alone was struck as compared with the other two methods, and felt that traumatic contact to the globe alone was actually rare. They finally concluded that direct trauma to the inferior orbital rim resulted in most of the clinical blow-out fractures [9]. Though these experiments support the bone conduction (buckling) theory, other investigators [11] have pointed out that associated extraocular muscle entrapment sometimes complicating blow-out fractures would occur much less frequently by this theory alone. Erling and et al question this theorys degree of contribution when there are associated medial wall blow-out fractures because initial disruption of the orbital floor would be expected to dissipate the energy from external trauma [9]. Surely, in clinical settings both the hydraulic and the buckling theories do have validity, though proportionately based on the individual circumstance of injury. The energy required to rupture a human globe has been reported to be 2.2 joules [12]. To determine the energy required to fracture the orbital floor directly, Warwar et al conducted an experiment in which a

metal rod was dropped from successive height increments onto exenterated orbital floor bones (after the thinnest spot to visual inspection was marked) until fracture. They calculated the mean energy required to fracture the human cadaver orbital floor directly was 78 mJ (millijoules) [8]. This supports the classical teaching that the orbital floor is weaker than the globe, and that by being more easily disrupted, floor fracture would provide an outlet for energy from the initial impact. It is not surprising that blow-out fractures are most commonly directed inferiorly, given that the orbital floor is the weakest portion of the orbit. The infraorbital nerve (V2 branch of the trigeminal nerve) traverses this region via the infraorbital groove and canal to provide sensation to the ipsilateral cheek as well as to the gingiva and teeth of the superior alveolar ridge. With inferior displacement of the floor fracture fragment, intraorbital soft-tissues such as fat and even extraocular muscle not infrequently herniate into the maxillary antrum. The inferior rectus muscle is considered hooked if only one portion of it is caught on fracture edges [1]. If the inferior rectus muscle is observed to be acutely buckled and frankly herniated between fracture edges, entrapment can be suggested by radiographic criteria. Strictly speaking though, the diagnosis of muscle entrapment in orbital fractures should be based on clinical examination. Small fibers of the inferior rectus muscle can be caught along the fracture margins and result in diplopia, though this may not be appreciated even by CT examination [13]. Common signs encountered on physical examination are periorbital ecchymosis, diplopia with upward gaze, hypoesthesia in cranial nerve V2 distribution, enophthalmos, and intraorbital emphysema. Surgical interventions are based on criteria such as diplopia, enophthalmos, or entrapment of extraocular muscles. Rarely, incarceration of the inferior rectus muscle and orbital soft tissues in an orbital floor trapdoor fracture can trigger an oculocardiac reflex (triad): bradycardia, nausea, and syncope. This occurs via afferent impulses passing from the ophthalmic division of the trigeminal nerve to the visceral motor nucleus of the vagus nerve. Efferent impulses are then transferred down the vagus nerve to the stomach and heart. In this specific clinical scenario, Sires et al recommend same-day fracture repair and release of entrapped muscle/soft tissues to avoid possible development of life-threatening arrhythmias as well as ischemic necrosis of the extraocular muscle [14]. Ocular injury in association with blow-out fractures has been reported as 14 40% [4]. The spectrum of associated ocular injury includes ruptured globe,

316

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

retro-orbital hemorrhage, hyphema, anterior chamber angle recession, lens dislocation, (secondary) glaucoma, vitreous hemorrhage, and choroidal/retinal tear or detachment, and commotio retinae. Interestingly, two investigators since 1981 have reported complete subluxation of the globe into the maxillary sinus following trauma, though the globe remained intact [15,16]. It has been suggested that the soft-tissue glove of orbital fat, ligamentous and fascial tissues, and extraocular muscles protects the globe by absorbing considerable energy during impact [17]. Blow-out fractures involving the paper-thin medial orbital wall (lamina papyracea) do occur as isolated events, but more frequently they occur in combination with orbital floor fractures. While the lamina papyracea is truly thinner than the orbital floor, it is actually buttressed by perpendicular elements of the ethmoid sinus bony lattice [2]. Orbital emphysema is most commonly seen with medial wall

orbit fractures. This can be exaggerated by nose blowing or Valsalva maneuver. Orbital contents can uncommonly prolapse into the ethmoid sinus if the fracture is large enough, though actual medial rectus muscle entrapment is rare [2]. On axial CT images, the entire length of the medial rectus can be assessed adjacent to a lamina papyrecea fracture. Should the trochlea of the orbit be involved as part of the fracture, the superior oblique muscle would be compromised [2]. Though medial wall fractures rarely develop significant complications, enophthalmos can also occur with this type of fracture [18,19]. In a series by Burm et al, the most common facial fracture associated with medial wall fracture was fracture of the nasal bone(s) (51%). They suggest a buckling force from the medial orbital rim as an important mechanism in medial wall injuries [19]. Rare superior orbital blow-out fractures involve the orbital roof, with an intact superior orbital rim. In some cases, the

Fig. 5. Orbital blow-out fractures. (A) Coronal CT scan, demonstrating an inferior floor fracture with herniation of fat and the inferior rectus muscle through the defect. Note the extensive orbital emphysema outlining the optic nerve sheath (arrow). (B) Coronal CT scan, demonstrating an inferior orbital floor fracture on the right. Note the fragment of bone within the superior aspect of the maxillary antrum and herniation of fat through the defect (arrows). (C) Axial CT scan of the orbits in another patient, demonstrating a fracture of the medial wall of the right orbit with herniation of the medial rectus muscle through the defect. (D) Coronal CT scan in the same patient as C again demonstrates the medial wall fracture and herniation of the medial rectus muscle through the defect on the right.

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

317

bone fragments may be displaced into the frontal sinus (Fig. 5). As described previously, orbital blow-in fractures result in fracture fragments being displaced inwards toward the orbit. The orbital roof is especially weak posteriorly, near the optic canal and superior orbital fissure. The oculomotor nerve is vulnerable at this fracture point. Inwardly displaced orbital roof fragments can impress into superior rectus and levator palpebrae muscles. Interestingly, 14 29% of reported orbital roof blow-in fractures are associated with intraocular injury (globe) [2]. Within the skull cavity, associated injuries include frontal lobe contusion, dural tear, and even formation of acquired encephalocele within the fracture defect (Fig. 6). Blunt trauma to the anterior maxilla is a typical cause of inferior blow-in fractures of the orbit, purportedly from a rapid increase in pressure within the maxillary antrum. Upwardly displaced bone fragments may impinge the inferior rectus and inferior oblique muscles. Medial blow-in fractures have been described, typically from a direct blow fragmenting the nasal bones and the medial orbital wall.

Fig. 7. Hyphema. Axial CT scan through the orbits demonstrates hyperdense material filling the anterior chamber of the left globe (arrow). Also note the preseptal soft tissue swelling.

Intraocular injury The initial assessment of intraocular injury to the globe can be performed by CT. Hemorrhage into the anterior chamber, a hyphema, may be demonstrated by increased attenuation within the anterior chamber with widening of the AP distance from the lens to the anterior margin of the globe (Fig. 7). Increased

hemorrhage in the vitreous humor, seen as increased attenuation in this location, is indicative of intravitreous hemorrhage (Fig. 8). Choroidal detachment and retinal detachment can be seen as bilentiform collections along the margins of the globe within the vitreous humor. The retinal detachment with subretinal hemorrhage will posteriorly extend to the optic disk. The posterior margin of a choroidal detachment will extend some distance away from the optic disk, typically a third of the distance from the posterior margin of the globe, where the vorticeal veins along the surface of the globe penetrate into the globe. These types of intraocular hemorrhage are typically hypderdense to muscle on a noncontrast CT scan and, depending on the blood products present, may have variable signal intensity on MR imaging. Traumatic lens subluxation occurs secondary to rapid equatorial expansion of the globe which may

Fig. 6. Blow in fracture of the orbit. The patient is status post MVA with frontal bone fracture. A blow in fracture of the superior orbital roof is present with associated extraconal hemorrhage and inferior displacement of the orbit. Note that the direction of the fragment is toward the bony orbit. There is also a foreign body within the orbit proven to be glass at time of surgery.

Fig. 8. Intravitreous hemorrhage secondary to foreign body. Axial CT scan of the orbits demonstrates a piece of glass within the right globe. Note the hyperdense material within the vitreous humor representing hemorrhage.

318

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

occur with blow-in fractures. The zonular fibers that normally hold the lens in place may be incompletely torn resulting in lens subluxation, or completely torn resulting in lens dislocation. Subluxation is more common then dislocation, but correctable vision is still possible [20]. The imaging finding of a lens subluxation is suggested by abnormal biconcavity or maligned position of the lens. A lens dislocation is seen as a bilentiform or elliptical hyperdensity within the vitreous humor, away from its normal position (Figs. 9 10). Globe rupture may occur if there is a laceration or perforation of the globe with leakage of the vitreous humor. As a result, there is a decrease in the intraocular pressure. Signs of globe rupture include an asymmetrical size or contour of the globe. In addition, stranding of the retrobulbar fat is also suggestive

Fig. 10. Lens dislocation. (A) Axial CT scan of the orbits demonstrates a round hyperdensity within the vitreous humor of the right globe that is isoattenuating with the lens on the left. Also note that no lens is present in its expected location within the anterior right globe. (B) In a different patient, the left lens is dislocated and found within the vitreous humor. Also note the apparent thickening of the posterior margin of the globes bilaterally representing subTenon hemorrhages (arrows).

of globe rupture. The resulting chronic sequala of globe rupture is involution and calcification of the globe, termed phthisis bulbi (Fig. 11).

Orbital hemorrhage
Fig. 9. Lens subluxation. (A) Axial CT scan of the orbits at the level of the optic disk demonstrates subtle posterior displacement and flattening of the lens on the left side. There is also slight widening of the AP distance of the anterior chamber (arrow). (B) In another patient who sustained blunt trauma to the left orbit, again note the widening of the anterior chamber (arrow) and posterior displacement of the left lens. There is also a retinal detachment and subretinal hemorrhage within the posterior left globe (arrowheads).

Hemorrhage within the orbit may occur in the retrobulbar space, in the extraconal space, around the globe (sub Tenon space between the sclera and Tenons capsule), or in the preseptal space. Hemorrhage within the extraconal space typically is linear in appearance and conforms with the bony structure that it abuts. The periosteum that covers the bony orbit is adherent to the underlying bone, with a looser attachment along the orbital roof. As a result, subperiosteal

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

319

Retrobulbar hemorrhage may be related to the dura surrounding the optic nerve. Subdural hemorrhage is seen as a tram-track density paralleling the course of the optic nerve, whereas epidural hemorrhage is a focal hematoma associated with the nerve sheath. Retrobulbar hemorrhage not associated with the optic nerve sheath may be seen as surrounding the optic nerve, resulting in a four-leaf clover appearance of the hemorrhage with coronal imaging [22]. Sub-Tenon hemorrhage results in a hematoma between Tenons capsule and the sclera. On imaging, these hematomas are curvilinear hyperdensities conforming to the contour of the globe extending from

Fig. 11. Globe rupture. (A) Axial CT scan of the orbits in a patient status post stab wound to the right orbit. There is a discrepancy in size between the globes, with increased attenuation within the vitreous humor representing intravitreous hemorrhage. Globe rupture is present with discontinuity of the medial border of the globe. A sub-Tenon hemorrhage is also present medially (arrow). (B) Coronal image in the same patient demonstrating the abnormal contour and size of the globe compared with the normal left side.

bleeds may occur in this location because of tearing of the traversing subperiosteal vessels. These subperiosteal bleeds may result in the formation of an orbital hematocyst and may persist for some time after the initial trauma. Patients may present with delayed dipoplia and proptosis with CT and MR findings of an extraconal mass [2,21]. On CT these lesions may be hyperdense relative to muscle and do not demonstrate significant enhancement. On MR imaging, the signal intensity of these orbital hematocysts may be variable depending on the age of the hematoma but because of their chronicity tend to be hyperintense in signal intensity with a surrounding rim of marked hypointense signal on both T1- and T2-weighted images relative to muscle (Figs. 12 13).

Fig. 12. Extraconal hemorrhage. (A) Axial CT scan of the orbit in a patient status post stab wound to the orbit. A hyperdense extraconal hematoma is present within the lateral extraconal space on the right and displaces the intraorbital contents medially resulting in an exolphthalmos. Note that there is tenting of the posterior aspect of the globe because of tension on the optic nerve. A preseptal hematoma is also present. (B) In another patient who sustained penetrating trauma to the orbit, coronal CT scan demonstrates an extraconal hematoma superiorly displacing the globe inferiorly. Note the small amount of orbital emphysema associated with the hematoma.

320

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

the optic nerve for possible bony fragments utilizing thin section CT with bone algorhythms in the axial and coronal planes. With the advent of multislice detector CT, isotropic sagittal reconstructions are possible. Because of the proximity of the optic nerve to the body of the sphenoid and cavernous sinuses, assessment of these regions as well as the bony orbit and optic canal are warranted. In the setting of orbital trauma, transient distortion of the bony orbit and orbital contents may result in stretch injury to the optic nerve or compromise of the vascular supply with resulting necrosis of the optic nerve [2,23,24]. This type of injury to the optic nerve is not well evaluated by CT if there is no resulting impingement of the optic nerve. If there are no contraindications to MR imaging, visualization of the optic nerves with

Fig. 13. Hematic cyst. (A) Patient with history of previous history of trauma to the right orbit with long-standing proptosis of the right globe. Coronal CT scan of the orbit demonstrates a bilobed cystic lesion within the superior orbit and extending into the anterior cranial fossa. The superior orbital roof was not identified. (B) Sagittal T1-weighted axial image through the right orbit demonstrates the lesion to be primarily hyperintense in signal intensity with a peripheral rim of signal drop off. The T2-weighted images (not shown) also demonstrated similar signal characteristics. Final pathologic diagnosis after removal and repair of the orbital roof was hematic cyst.

the ciliary body anteriorly and ending at the optic nerve (Figs. 10, 11, and 14).

Optic nerve injury Optic nerve injury may result from direct or indirect injury. Fractures involving the orbital apex or optic canal may result in laceration or transection of the optic nerve. Penetrating trauma to the orbit may also result in optic nerve injury. In patients with immediate or rapidly declining visual acuity after the time of injury should be rapidly assessed for possible surgical intervention in the event of impingement of
Fig. 14. Subdural hematoma. (A) Coronal CT scan through the orbital apex in a patient who sustained blunt trauma to the right orbit with rapid onset of monocular blindness. A fracture of the lateral orbital wall (arrow) is seen with a small bone fragment projecting medially into the orbital apex. (B) Postcontrast T1-weighted coronal image with fat suppression demonstrating a round hyperintense lesion, representing a subdural hematoma (arrows) compressing the optic nerve (arrowhead).

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

321

high resolution T2-weighted images in the coronal plane may be used to demonstrate abnormal increased signal intensity within the optic nerve (Fig. 15).

Penetrating injury and foreign bodies It has been reported that penetrating injury is a component of 50% of all trauma to the eye [3]. The gamut of objects involved in penetrating injuries is broad, whether from occupational hazard (eg, nail gun, glass workers), crime (eg, knives, bullets, BB pellets), or motor vehicular collisions (eg, glass, metal fragments). Orbital fractures and possible globe injury are directly anticipated in the path of a projectile or other introduced object. The greater concern, however, is that the orbits provide access to the cranial vault, with higher associated morbidity and mortality. A BB pellet is typically 4.5 mm in diameter, weighs 0.35 gm, and is fired at velocities of 250 750 ft/sec. In contrast, bullets are heavier and travel at velocities of 755 3250 ft/sec. With their higher energies, bullets and other high-energy projectiles are more likely to injure structures adjacent to the orbits, such as the paranasal sinuses and the brain [25] (Fig. 16). The detection and localization of foreign bodies is a critical component of preoperative planning. The examining clinician relies on imaging techniques to evaluate the retrobulbar space, and also for detection of intraocular foreign bodies in situations that limit fundoscopic examination such as cataracts, hyphema, or other intraocular hemorrhage. Metal is seen as a radiodensity on plain films. On CT, bright streaks from metallic beam-hardening artifact

Fig. 16. BB pellet within the globe. Axial CT scan of the orbits in a patient shot in the orbit with a BB gun. A round metallic foreign body is seen within the posterior aspect of the globe with associated intravitreous hemorrhage. Note the metallic artifact related to the globe. There is also associated gas within the anterior chamber and posterior subluxation of the lens (arrows).

Fig. 15. Stretch injury of the optic nerve. The patient was struck in the periorbital region with sudden deterioration of vision. FSE T2-weighted axial image with fat suppression demonstrates increased signal within the right optic nerve within the optic canal (arrows).

allow identification of metal fragments less than 1 mm in size [3]. At the risk of stating the obvious, it must be emphasized that the possibility of a metal foreign body within orbit is an absolute contraindication to Mr imaging as this may result in blindness. This is of particular concern in trauma settings, where full clinical (and occupational) history is frequently not available (Fig. 17). In their series on metallic foreign bodies within the orbit, Finkelstein et al found associated globe injuries of varying degrees in virtually all of their cases [26]. These ocular injuries were more common, however, and of greater severity where there were (metal) foreign bodies located posteriorly in the orbit. Finkelstein et al recommended conservative observation of patients with posterior foreign bodies caused by inherent risks of surgical exploration deep in the orbit. But even when deep in the orbit, they did advocate consideration for removal of the foreign body in the event of marked orbital inflammation, infection, or evidence of nerve impingement with associated optic neuropathy [26]. A foreign body at the orbital apex and adjacent to the optic nerve might require optic canal decompression. At the same time, they did recommend removal of metal foreign bodies located anteriorly in the orbit or epibulbar area to avoid risks of fistula formation, infection, motility impairment, or ptosis. In a recent comparison of CT, sonography, and MR imaging for detection of intraocular glass, Gor et al found helical CT to be the most sensitive modality [27]. The density of glass subtypes is directly correlated with detection rates. The highest

322

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

Fig. 17. Nail within the orbit. (A) AP view of the right orbit demonstrates a metallic foreign body within the orbit. Orbital emphysema is present as the optic nerve is visualized (arrow). (B) Lateral view of the skull demonstrates the foreign body to be a nail. The patient received penetrating trauma by being shot in the orbit with a nail gun. (C) Coronal CT scan of the orbit at the level of the mid globe demonstrates the metallic artifact associated with the nail within the left globe.

density of glass in their series (green beer bottle glass) was easiest to identify, whereas the least dense (spectacle glass) was most difficult to detect.

Size of glass fragments was also critical. They found that helical CT could detect 1.5 mm fragments at a 96% rate, but 0.5 mm fragments were detected at a

Fig. 18. Orbital glass. (A) Axial CT scan of the upper orbit demonstrating multiple hyperdense fragments consistent with glass within the superior left orbit. Note the geometric appearance of the fragments. (B) In another patient who fell through a pane of glass, a polygonal homogeneous hyperdensity is seen within the lateral orbit.

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324

323

rate < 50%. Finally, location of the glass fragment was influential. Fragments in the anterior chamber of the globe were easiest to identify, followed by those in the posterior chamber, with glass fragments on the corneal surface most difficult to detect [26] (Fig. 18). Although helical CT is accepted as accurate for detecting and localizing intraocular and orbital metal, glass, and stone foreign bodies [28], the detection of wood in the orbit is less certain. Orbital plain films are rarely sensitive for wooden fragments where wood is seen as air [29]. On both CT and Mr imaging, wood is also seen as air [30]. Suspicion for wood is increased if a particular collection of air displays a geometric margin. Bone fragments also appear as signal void on MR imaging, however [30]. Indeed, McGuckin et al found in their in vitro model that MR did not differentiate dry wood foreign bodies from air and bone fragments [31]. They found that, with optimized window settings, CT could differentiate dry ( 656 Hounsfield units) and fresh ( 24 HU) wood from air and bone fragments, and concluded that CT was the preferred imaging modality for wooden foreign bodies [31]. Detection by CT is improved if a wooden fragment is associated with metallic paint or a granuloma [29], though this relies on fortuitous circumstances. McGulkin et al did find that T1-weighted images provided better contrast between wood and fat than other spin-echo MR pulse sequences. Some investigators advocate a role for MR imaging in cases where orbital penetration by a wooden foreign body is suspected but CT scan does not show one [29]. The difficulty of detecting wood is critical because organic foreign bodies can incite an inflammatory reaction and lead to cellulitis, abscess, and death, whereas metallic (exception of copper) and glass particles of themselves are relatively well tolerated [24]. In their series on retained organic foreign bodies, Nasr et al encountered a spectrum of complications: purulent inflammation, abscess, gangrene, tetanus, granulomatous tissue reaction, orbitocutaneous fistula, and osteomyelitis. With intracranial extension of trauma, Miller et al reported high rates of infection (65%) and mortality (25%) when associated with wooden foreign bodies.

has proven to be an adjunct modality in the assessment of orbital injury.

References
[1] Rogers LF. Radiology of skeletal trauma, 2nd edition. New York: Churchill Livingstone Inc.; 1992. pp. 412 419. [2] Gean A. Imaging of head trauma. New York: Raven Press; 1994. [3] Harris JH Jr, Harris WH, Novelline RA, Castillo M. The Radiology of emergency medicine. 3rd ed. Baltimore: Williams & Wilkins Co; 1993. [4] Brady SM, McMann MA, Mazzoli RA, et al. The diagnosis and management of orbital blowout fractures: update 2001. Am J Emerg Med 2001;19:147 54. [5] Mafee MF, Pruzansky S, Corrales MM, et al. CT in the evaluation of the orbit and the bony interorbital distance. AJNR 1986;7:265 9. [6] Mauriello JA, Lee HJ, Nguyen L. CT of soft tissue injury and orbital fractures. Radiol Clin N Am 1999; 37(1):241 52, xii. [7] Smith B, Regan WF. Blowout fractures of the orbit: mechanism and correction of internal orbital fractures. Am J Ophthalmol 1957;44:733 9. [8] Warwar RE, Bullock JD, et al. Mechanisms of orbital fractures: a clinical, experimental, and theoretical study. Ophthalmic, Plastic, and Reconstructive Surgery 2000;16(3):188 200. [9] Erling BF, Iliff N, Robertson B, Manson PN. Footprints of the globe: a practical look at the mechanism of orbital blowout fractures, with a revisit to the work of Raymond Pfeiffer. Plastic and Reconstructive Surgery 1998;103(4):1313 9. [10] Fujino T, Sugimoto C, Tajima S, et al. Mechanism of orbital blowout fracture II: analysis by high speed camera in two dimensional eye model. Keio J Med 1974; 23:115 24. [11] Converse JM. The blow-out fracture of the orbit: some common sense. In: R.J. Brockhurst, editor. Controversy in ophthalmology. Philadelphia: Saunders;1977. [12] Bullock JD, Johnson DA, Ballal DR, Bulock RJ. Ocular and orbital trauma from water balloon slingshots: a clinical, epidemiological, experimental, and theoretical study. Transactions of the American Opthalmology Society 1996;XCIV:105 34. [13] Ochs MW, Beatty R. Complex facial fractures in the periorbital region. Semin Opthalmology 1994;9: 218 28. [14] Sires BS, Stanley RB, Levine LM. Oculocardiac reflex caused by orbital floor trapdoor fracture: an indication for urgent repair. Arch Opthalmol. 1998;116 (7):955 6. [15] Berkowitz RA, Putterman AM, Patel DB. Prolapse of the globe into the maxillary sinus after orbital floor fracture. Am J Ophthalmol 1981;91:253 7. [16] Ziccardi VB, Patterson GT, Ramasastry S, Sottereanos GC. Management of a severe zygomatico-orbital frac-

Summary Assessment of the orbit for orbital trauma is best achieved expeditiously with CT in the determination of extent of injury and the presence of foreign body. MR imaging has a limited role but is valuable in examining the optic nerve and globe for injury and

324

J.L. Go et al / Neuroimag Clin N Am 12 (2002) 311324 ture with dislocation of the globe into the antrum. J Craniofac Surg 1993;4:95 101. Pelton RW, Rainey AM, Lee AG. Traumatic subluxation of the globe into the maxillary sinus. AJNR Am J Neuroradiol 1998;19(8):1450 1. Segrest DR, Dortzbach RK. Medial orbital wall fractures: complications and management. Opthalmic Plastic and Reconstructive Surgery 1989;52(2):75 80. Burm JS, Chung CH, Oh SJ. Pure orbital blowout fracture: New concepts and importance of medial orbital blowout fracture. Plast Reconstr Surg 1999;103: 1839 49. McDonald PR, Purnell JE. The dislocated lens. JAMA 1951;145:220 6. Milne III HL, Leone CR, Kincaid MC, Brennan MW. Chronic hematatic cyst of the orbit. Ophthalmology 1987;94:271 7. Lustrin ES, Brown JH, Novelline R, Weber AL. Radiologic assessment of trauma and foreign bodies of the eye and orbit. Neuroimaging Clin N Am 1996;6 (1):219 37. Atlas SW, Galetta S. The orbit and visual system. In: Atlas SW, editor. Magnetic resonance imaging of the brain and spine. Philadelphia: Lippincott-Raven; 1996. Nasr AM, Barrett G, Fleming JC, et al. Penetrating orbital injury with organic foreign bodies. Opthalmology 1999;106(3):523 32. Chu A, Levine MR. Gunshot wounds to the eye and orbit. Opthalmic Surg Lasers 1989;20:729 36. Finkelstein M, Legmann A, Rubin PA. Projectile metallic foreign bodies in the orbit. opthalmology 1997; 104(1):96 103. Gor DV, Kirsch CF, Leen J. et al. Radiologic differentiation of intraocular glass: evaluation of imaging techniques, glass types, size, and effect of intraocular hemorrhage. AJR 2001;177:1199 203. Latkis A, Prokesch R, Scholda C, Bankier A. Orbital helical computed tomography in the diagnosis and management of eye trauma. Ophthalmology 1999;106: 2330 5. Green BF, Kraft SP, Carter KD, Buncic JR, et al. Intraorbital wood: detection by magnetic resonance imaging. Opthalmology 1990;97:608 11. Atlas SW. Magnetic resonance imaging of brain and spine, ed 2. Philadelphia: Lippincott-Raven Publishers; 1996. pp, 1061 3. McGuckin JF, Akhtar N, Ho VT, Smergel EM, Kubacki EJ, Villafana T. CT and MR evaluation of a wooden foreign body in an in vitro model of the orbit. AJNR 1996;17:129 33.

[17]

[25] [26]

[18]

[27]

[19]

[28]

[20] [21]

[29]

[22]

[30]

[23]

[31]

[24]

Neuroimag Clin N Am 12 (2002) 325 338

Imaging of sequelae of head trauma

Chi-Shing Zee, MD*, Armen Hovanessian, MD, John L. Go, MD, Paul E. Kim, MD
Department of Radiology, Keck School of Medicine, University of Southern California, 1500 San Pablo Street, Los Angeles, CA 90033, USA

Head trauma is a major public health problem in the United States, responsible for 200 to 300 hospital admissions per 100,000 population each year. Most of the admissions last only a few days, and the patients are admitted for clinical observation. Care and treatment of patients who suffer head trauma can be a serious financial burden to society, with lost labor and reduced productivity compounding the negative impact. Severe traumatic brain injury is undoubtedly associated with poor clinical outcome; however, even minor head injury can be associated with significant morbidity. Sequelae of head trauma can mean entirely different things for a neuroscience clinician and a neuroradiologist. From a clinical perspective, sequelae of head trauma include post-traumatic epilepsy, postconcussion syndrome, or intellectual impairment [1]. Post-traumatic epilepsy can be of early onset or late onset. Early onset of seizure is characterized by onset within one week following head trauma, whereas late onset of seizure occurs more than one week after injury. The primary significance of an early onset of post-traumatic epilepsy is that it can be an indicator of an increased risk for late onset seizures, which occur in approximately 25% of the patients with early onset of seizure. Most cases of late onset of seizures occur approximately two months after head trauma, as it takes eight weeks for an epileptic focus to form [1 3]. Brain injury seems to initiate a series of biochemical, electrical, and structural changes that lead to the formation of an epileptogenic focus [3]. The seizure focus is usually seen adjacent to the

* Corresponding author. E-mail address: chishing@hsc.usc.edu (C.-S. Zee).

injury site. New imaging techniques, including diffusion weighted imaging (DWI), fluid attenuated inversion recovery (FLAIR) technique, and magnetic resonance spectroscopy, can provide valuable information in the evaluation of patients with head trauma. These imaging modalities make it possible to detect biochemical and molecular changes before gross anatomical changes in the brain. Nevertheless, despite the power of magnetic resonance imaging (MRI), current imaging techniques still cannot definitively determine which patient will eventually develop seizure following head trauma, or which traumatic lesion may evolve into a seizure focus. Intellectual impairment occurs after severe head trauma. Levin et al [4] suggested that characteristic neuropsychological and psychiatric deficits are responsible for the chronic disability seen following severe head trauma. The characteristic pattern includes impairment in performance IQ and memory as well as behavioral changes. These findings can vary from minimal to marked, depending on a particular patients recovery. Post-concussion syndrome occurs following minor head trauma and exhibits a variety of clinical symptoms, including headache, dizziness, forgetfulness, anxiety, irritability, and impaired concentration. The cause of postconcussion syndrome is widely debated. One cannot be certain whether the symptoms are the result of psychological trauma or structural damage, or a combination of both. To date, there is no definitive correlation between imaging findings and post-concussion syndrome. This is because most of the studies have focused on correlation of clinical outcome with MR imaging findings of lesional anatomy [5]. With the availability of new MR imaging techniques such as volumetric measurement, MR spec-

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 0 4 - 7

326

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

troscopy, magnetization transfer, and diffusion/perfusion imaging, several authors have found some correlation between neuroimaging findings and neurocognitive outcome [6 11]. Traditionally, neuroradiologists have focused their interest in defining the anatomical changes related to head trauma on imaging studies. Computed tomography (CT) scans were thought to be adequate to evaluate patients with head trauma. However, there is

now increasing evidence that magnetic resonance imaging, with its newer techniques, is an indispensable modality in the evaluation of head trauma patients, particularly in the subacute and chronic stages [1,7 10]. Based on the descriptive anatomy of the lesions, neuroradiological imaging studies reveal that patients may develop a variety of sequelae of head trauma, which could be early onset or late onset. Early onset

Fig. 1. Delayed intracranial hematoma evolving into encephalomalacia in a child. (A) Axial CT scan obtained shortly after head trauma demonstrates intraventricular hemorrhage with high density blood seen in the atrium of the lateral ventricle. (B) Axial CT scan obtained at the same time as Fig. 1A demonstrates subtle low density contusion in the left frontal and left posterior temporal region. (C) A follow-up CT scan obtained two days after head trauma shows the presence of delayed intracerebral hematoma in the left frontal, posterior temporal region. Another small hematoma is seen in the right frontal lobe. (D) A follow-up CT scan obtained 2 months later demonstrates low-density encephalomalacia in the left frontal and temporal region. Dilatation of the lateral ventricles and cortical sulci is seen, consistent with diffuse volume loss.

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

327

sequelae of head trauma may develop from a few hours to a few weeks following head trauma. Late onset sequelae of head trauma may develop from a few months to years following head trauma. It is imperative to be aware of and to be able to detect the early sequelae of head trauma, which can have serious consequences if untreated. Late onset sequelae may range from mild treatable lesions to those associated with severe neurological deficits. Early onset sequelae include delayed intracerebral hematoma, delayed subdural hematoma, delayed epidural hematoma, pneumocephalus, cerebrospinal fluid leak, or fistula. Late onset sequelae include cerebral atrophy, cerebellar atrophy, atrophy of the corpus callosum, cerebrospinal fluid leak or fistula, acquired cephalocele, or leptomeningeal cyst, cerebral infarction, and post-traumatic hypopituitarism. CT is the imaging modality of choice in patients with acute head trauma, because of its ability to detect skull fractures and acute intra- or extra-axial hemorrhage [12 17]. Furthermore, severe head trauma patients frequently require life support devices that are routinely placed in the CT suite, whereas magnetic resonance imaging requires specialized MR-compatible life support devices. CT is generally adequate and probably preferred in the evaluation of patients for early onset of sequelae of head trauma. Generally speaking, MRI is the imaging modality of choice for evaluation of subacute and late onset of delayed sequalae of head trauma.

coagulation changes may result in cerebral hemorrhage or infarction [21 23]. The imaging modality of choice for evaluation of delayed or recurrent intracerebral hematoma is computed tomography. Delayed intracerebral hematomas are almost always lobar, are frequently multiple, and often occur in areas that show cerebral contusion initially (Fig. 1). Delayed intracerebral hematomas,

Delayed intracerebral hematoma Delayed intracerebral hematomas occur in up to 7% of patients with severe head trauma and are usually seen within the first week of the injury, most frequently within the first 24 to 48 hours. Recurrent hematomas may also occur within the first 10 days in those patients who had emergency surgery for evacuation of an acute intracranial hematoma. The incidence of delayed intracerebral hematoma is significant and may require emergency craniotomy. A high index of clinical suspicion is required to detect these delayed intracerebral hematomas [18,19]. The neurological abnormalities associated with delayed intracerebral hematomas that occur within the first week following head trauma may be difficult to distinquish from those that resulted from cerebral edema [20]. Clinical findings may include deteriorating level of consciousness, a recent onset of third nerve palsy, or an increase in intracranial pressure. Several reports have noted that brain injury as a result of head trauma can cause abnormalities of blood coagulation. These

Fig. 2. Delayed intracerebral hematoma in a patient with head trauma. (A) Axial CT scan obtained shortly following trauma demonstrates a small amount of subarachnoid hemorrhage along the tentorium on the left side. (B) Axial CT scan obtained 24 hours later because of the change in patients neurological status demonstrates the presence of a large delayed intracerebral hematoma in the left temporal lobe.

328

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

however, can occasionally occur in areas that appear normal on the initial CT (Fig. 2). We routinely rescan patients who have sustained severe head trauma and do not recover clinically in the first 2 to 5 days.

Delayed subdural and epidural hematomas Delayed epidural and subdural hematomas occur in up to 7% of patients following head trauma, similar to intracerebral hematoma [24]. The clinical signs are often insidious, and a high index of suspicion is required to detect these delayed hematomas. They usually occur within the first 10 days following a severe head injury, especially during the first 24 to 48 hours (Fig. 3). Therefore, a follow-up CT scan should be routinely performed in a severe head trauma patient whose clinical signs do not return to normal within the first 48 hours, even if the initial CT

scan did not show any significant space-occupying lesion such as intra-axial or extra-axial hematoma. If a patient is under observation for a small extra-axial hematoma, a follow-up CT scan should be performed in three to five days regardless of the patients clinical recovery. A smaller but significant number of patients who have initial emergency craniotomy for evacuation of an extra-axial hematoma can develop a delayed contralateral hematoma in up to two weeks [25] (Fig. 4). The authors reviewed eight cases of delayed epidural hematoma at their institution and found hematoma occurred from 8 to 72 hours after initial injury, or subsequent to surgery for decompression of another epidural, subdural, or intracerebral hematoma because of the reduction of tamponade effect [16,17]. While CT or MRI can be used to detect delayed extra-axial hematomas, MRI is more sensitive in detecting small subdural hematomas.

Relationship between disseminated intravascular coagulopathy and fibrinosis and head trauma Several previous reports have discussed the association between head injury and clinically obvious bleeding abnormalities [21,26 28]. The majority of the patients discussed died shortly after clinical observation. However, there are reports that some patients with massive head trauma and laboratory evidence of disseminated intravascular coagulopathy and fibrinosis (DICF) survived following clinical observation [29]. Patients who suffer minor head trauma can also develop DICF. DICF has been implicated as a cause of delayed and recurrent intracranial hematomas [30,31]. Several studies have demonstrated a high percentage of abnormal coagulation studies drawn 24 hours following head trauma, even with minor injuries [23,26 28,32 34]. Not only are coagulation abnormalities common shortly after head injury, but they also are more common after severe head injuries. The elevation of fibrin split products and the decrease in fibrin levels in head trauma patients indicate that fibrinolysis is a serious problem. Although screening clotting tests show that abnormalities of the thromboplastin-activated extrinsic clotting system occur more frequently than those of the intrinsic clotting system, it is possible to see both types of clotting abnormalities in the same patient [27,28]. DICF has significant impact on the morbidity and mortality of the patients with head trauma. Van der Sande et al [35] compared CT scans, neurological findings, and coagulation test results on hospital admission of head injury patients and found

Fig. 3. Delayed subdural hematoma. (A) Axial CT scan obtained shortly after head trauma shows a minimal amount of subarachnoid hemorrhage. Mild atrophy of the cerebrum is seen. (B) Axial CT scan obtained 2 weeks later because of increasing headache demonstrates bilateral isodense subdural hematoma.

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

329

Fig. 4. Delayed subdural hematoma. (A) Axial T1-weighted image obtained shortly after surgical evacuation of a large traumatic right subdural hematoma demonstrates a small residual subdural hematoma on the right side. (B) Axial fluid attenuated inversion recovery (FLAIR) image shows also shows the residual subdural hematoma. (C) A routine follow-up study was performed 3 months later. Axial T1-weighted image demonstrates the resolution of the right subdural hematoma, but there is now a new left hyperintense subdural hematoma seen. (D) Axial FLAIR images also show the new left subdural hematoma and the resolution of the right residual subdural hematoma.

that coagulation test results were a more sensitive indicator of brain contusion than CT scans.

Pneumocephalus Pneumocephalus usually results from fractures involving the paranasal sinuses and mastoid air cells, but fractures elsewhere that involve the entire thickness of the calvarium (scalp, skull, and meninges)

may permit air to enter the intracranial space. Intracranial air collections can be seen in any compartment, including epidural, subdural, subarachnoid, paranchymal, and intraventricular locations. Intracranial air collections can be diffuse or localized and are called pneumatoceles when localized. Air is absorbed over time; hence, if pneumocephalus persists, a continuing cerebrospinal fluid (CSF) leak should be suspected. Rarely, the collection of the air may expand because of a check valve mechanism pro-

330

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

ducing tension pneumocephalus. Pneumocephalus usually occurs immediately following trauma; however, it can occur delayed from a few days to a few weeks following initial head trauma (Fig. 5). A follow-up CT scan is necessary to exclude the possibility of delayed pneumocephalus in patients with basilar skull fractures (Fig. 6) [16,17,36]. Air is easily identified on CT scan of the head as present in areas of very low attenuation, lower than fat in the orbits and similar to air in the paranasal sinuses.

Fig. 6. Delayed pneumatocele formation following head trauma. (A) Axial CT scan demonstrates right frontal low density with a small high-density area posteriorly, consistent with hemorrhagic contusion of the right frontal lobe. The patient also had basilar skull fracture on the right side (not shown). (B) Axial CT scan obtained 2 days later shows the presence of a large air collection in the right frontal region, consistent with a pneumatocele.

On magnetic resonance imaging, air shows as an area of signal void. Epidural air tends to be localized and does not move with change in head position. Subdural air is confluent in the subdural space and often forms an air-fluid level. Subarachnoid air is typically seen in the cerebral sulci and is often multifocal in a dropletlike appearance. Intraventricular air is usually seen in the frontal horns of the lateral ventricles; as opposed to intraventricular hemorrhage, which is commonly seen in the occipital horns (Fig. 7).
Fig. 5. Pneumoncephalus because of depressed skull fracture. (A) Axial CT shows a depressed skull fracture in the left frontoparietal region. No pneumocephalus was seen on this study. (B) Axial CT obtained 24 hours later demonstrates a large amount of intracranial air in the left frontal region, consistent with pneumocephalus.

Traumatic CSF rhinorrhea The vast majority of cases of cerebrospinal fluid rhinorrhea are the result of trauma and are frequently

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

331

Fig. 7. Delayed pneumocephalus with intraventricular air. (A) Sagittal T1-weighted image shows a left frontal porencephalic cyst that resulted from a previous head trauma. An air fluid level is seen in the porencephalic cyst, which is communicating with the frontal horn of the left lateral ventricle. (B) Axial T1-weighted image demonstrates the air in the left frontal porencephalic cyst with an air fluid level. (C) Axial T2-weighted image shows the air fluid level again.

a complication of basilar skull fractures. Before localization of suspected CSF leakage, it is necessary to identify the fluid as CSF, with biochemical tests showing glucose values greater than 30 mg/dl to be diagnostic. While most cases will cease spontaneously without treatment, patients with persistent CSF leaks may be at increased risk for meningitis. The onset of meningitis may be fulminant and lifethreatening. The reported incidence of meningitis

following post-traumatic CSF leak varies from 16% to 21% [37,38]. Lewin reported that in two-thirds of the patients in his extensive series, CSF leaks occurred within 48 hours of the head trauma [39]. Most of the remaining leaks occurred later, usually within the first three months; however, in a few cases, the onset of CSF leak was delayed by many months to years. The longest published delayed onset of CSF leak as a cause of recurrent meningitis is 25 years

332

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

[40]. Spontaneous closure of such fistulas occurs by adhesions or herniations of brain tissue into the bony defects, or by the formation of granulation tissue as a result of meningitis. The quality of spontaneous closure is frequently inadequate, and recurrent meningitis occurs as a result [36,41]. It is imperative to be vigilant regarding CSF leakage in all patients with head trauma, especially those with basilar skull fracture. Special attention should be given to identifying fractures of the frontal, ethmoid sinuses, and the petrous bone. In acute traumatic rhinorrhea, the CSF is invariably bloody initially and becomes clear later on. Delayed CSF leaks may present with a fairly sudden copious gush or as an onset of meningitis without obvious leak [42]. Continuous leakage of CSF is less likely to be associated with meningitis than is a scanty, stop and go leak. The overall incidence of CSF leaks in head-injured patients varies between 0.25% to 2%. In addition to the clinical evidence of meningitis or CSF leaks, the skull radiograph should be studied for air-fluid levels in the paranasal sinuses, especially the sphenoid sinus. Any detection of intracranial air should lead to the suspicion of a CSF leak. When CSF leaks are suspected, a radionuclide scintigraphy using 111-indium-DTPA can be performed to identify the origin of leak. Computed tomographic cisternography is required to demonstrate the exact site of CSF leak [43]. Myelographic contrast can be injected through lumbar puncture or C1 C2 puncture, then the

contrast material is maneuvered into the patients frontal fossa by tilting the patients head down and feet up. The patient is asked to keep the head straight on the flurouscopic table. With the patient in prone position and chin up, a series of coronal CT scans using 1.5 mm or 2 mm thin slices are obtained for evaluation. Frequently, a small fracture of the cribriform plate is seen with leakage of the myelographic contrast material through the defect into the paranasal sinuses (Fig. 8). More recently, 3D-constructive interference steady-state sequence (CISS) magnetic resonance imaging technique has been reported to have high sensitivity and specificity in the detection of CSF rhinorrhea after head injury [44].

Leptomeningeal cyst Most linear skull fractures in childhood ordinarily heal without sequelae. A growing skull fracture is a rare complication of linear skull fracture and occurs predominantly in infancy and early childhood. It is characterized by progressive widening of the fracture line. It is frequently associated with a cystic mass filled with cerebrospinal fluid at the enlarging fracture site and, therefore, is also called a leptomeningeal cyst. The incidence of growing skull fracture is reported to be from 0.5% to 1% of all pediatric skull fractures [45]. Continuous pulsatile wedge pressure from the leptomeninges herniating through the skull fracture line causes resorption of the adjacent bone, and thus a growing skull fracture. Sometimes brain tissue may herniate through the defect along with the cerebrospinal fluid and leptomeninges [46]. Growing skull fractures usually occur in infancy, but are frequently seen in the first three years of life. The rapid growth of the brain tissue during the early years of life is an essential factor in the development of the growing skull fracture. A growing skull fracture may remain stable for many years into adulthood [47]. Plain radiographs demonstrate diastasis of the fracture line. Growth of the fracture can usually be confirmed on the follow-up radiograph films [48]. CT or MRI provides more information than skull radiographs. In addition to the skull fracture, the contents of the growing skull fracture, such as CSF or brain tissue, are better seen. Initial CT obtained at the time of trauma may demonstrate the presence pf cerebral contusion, hemorrhage, or extra-axial hematoma in addition to the skull fracture. When a growing skull fracture is diagnosed on CT, an area of hypodensity representing encephalomalacia may be seen adjacent to the fracture site and the ipsilateral ventricle may show focal dilatation.

Fig. 8. CT cisternogram performed to detect CSF leak. Coronal CT scan performed following the intrathecal injection of myelographic contrast in the lumbar region demonstrates the leakage of contrast into the right ethmoid sinus through a bony defect at the cribriform plate.

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

333

Post-traumatic atrophy of the cerebrum, cerebellum, and corpus callosum Atrophy of the cerebrum may occur focally or diffusely in patients with previous head trauma [49,50]. In such cases, both CT and MRI show widening of the cortical sulci and concordant ventriculomegaly in the affected areas. Cerebellar atrophy is demonstrated by the prominence of subarachnoid and cisternal space in the posterior fossa. Time-dependent atrophic changes occurring after traumatic brain injury can be quantified using MR volumetric studies and, in chronic stages, MR volumetric studies may predict eventual cognitive outcome [49]. Focal atrophy may present as focal areas of encephalomalacia or porencephalic cyst. Sometimes it may be difficult to differentiate encephalomalacia from porencephalic cyst on CT, as both entities show low density. Since porencephalic cyst contains CSF, it is generally of lower density than areas of encephalomalacia. On MR imaging, they can be easily differentiated from each other on the FLAIR sequence, as porencephalic cyst shows low signal intensity, whereas encephalomalacia demonstrates high signal intensity (Fig. 9). In patients with long-standing severe closed head injury and diffuse white matter injury, atrophy of the corpus callosum can occur. The degree of corpus callosal atrophy correlates significantly with the chronicity of the injury (Fig. 10). MR imaging provides an in vivo determination of corpus callosal atrophy that may reflect the severity of diffuse axonal injury [4]. The MR findings of corpus callosal atrophy following closed head trauma appear to correlate clinically with post-traumatic hemispheric disconnection effects [51]. Reduction in fornix size and hippocampal volume has also been reported in patients with traumatic brain injury [52]. We have recently observed a case of post-traumatic seizure in a patient with previous temporal lobe injury. MR imaging of the temporal lobe demonstrated temporal lobe encephalomalacia and bilateral mesial temporal sclerosis (Fig. 11).

vascular injury, such as in carotid artery dissection. Neuroimaging studies can demonstrate cerebral infarction secondary to cerebral herniation that is caused by intracerebral hematoma, cerebral contusion, cerebral edema, subarachnoid hemohrrage, subdural hematoma, or epidural hematoma. Diffusion-weighted MR techniques are useful to distinquish cerebral infarction

Post-traumatic cerebral infarction Post-traumatic cerebral infarction can occur as a result of cerebral herniation from mass effect of a traumatic intracerebral, subdural, or epidural hematoma, or from cerebral contusion. Cerebral infarction occurs because of gross mechanical shift of the brain across the falx or tentorium, with kinking and compression of the anterior or posterior cerebral arteries [53]. Cerebral infarction may also occur secondary to

Fig. 9. Left frontal encephalomalacia and porencephaly. (A) Axial T2-weighted image shows an area of high signal intensity in the left frontal region, which could be attributed to encephalomalacia or porencephaly. (B) Axial FLAIR image clearly demonstrates the small low signal intensity area of porencephaly adjacent to the high signal intensity encephalomalacia.

334

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

Fig. 10. Bilateral frontal volume loss because of porencephaly and encephalomalacia. (A) Axial T1-weighted image shows bilateral frontal low signal intensity areas, consistent with focal volume loss. (B) Axial FLAIR image demonstrates the presence of porencephaly and encephalomalacia. (C) Sagittal T1-weighted image reveals significant volume loss in the genu of the corpus callosum.

secondary to cerebral herniation from post-traumatic cerebral edema.

thalamic failure more often than pituitary gland failure. Magnetic resonance imaging shows an intrasellar arachnoidoceae with extinction of the focal high signal in the posterior pituitary gland.

Post-traumatic hypopituitarism Post-traumatic hypopituitarism is a rare disease, accounting for 3% of all cases of hypopituitarism reported in the literature [54]. Most traumatic cases occur after severe head injury. Post-traumatic hypopituitarism exceptionally involves only one hormone: growth hormone. The pathogenesis involves hypoCorrelation of neuroimaging and neurotraumatic outcomes The use of advanced neuroimaging techniques in traumatic brain injury management is undergoing a fundamental change. Historically, the radiographic

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

335

Fig. 11. Post-traumatic right temporal volume loss and seizure. (A) Sagittal T1-weighted image shows right temporal volume loss. (B) Coronal FLAIR image confirms the right temporal volume loss. (C) Coronal FLAIR image demonstrate abnormal high signal intensity involving both hippocampal formation.

findings in neurotrauma have focused on the descriptive anatomy of lesions with little regard to correlation with clinical outcomes of patients. Given the wide spectrum of traumatic mechanisms, coupled with the complexity of neurophysiological autoregulation, prognostication based solely on location, number, and size of lesions has been poor at best; however, recent advances in neuroimaging have opened new opportunities for understanding the biology of neuro-

trauma and stratifying the clinical outcomes based on physiological, functional, and anatomic imaging correlates. With the growing use of MRI, single-photon emission CT, and positron emission tomography scanning, new techniques that aim to improve clinical diagnosis and prognosis are being applied to trauma situations. With the widespread adoption of advanced neuroimaging studies, neuroradiologists have gained the

336

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338

ability to correlate subtle changes in neurophysiology and map them anatomically. MR imaging can detect punctate areas of hemorrhage, differentiate between vasogenic and cytotoxic edema, and demonstrate areas of ischemia/infarct with much greater precision and speed than earlier generation neuroimaging tools. While several studies have examined the link between traumatic brain injury and routine MRI findings, these investigations have focused primarily on lesional anatomy [8,55 57]. However, with the development of MR spectroscopy, magnetization transfer, and diffusion/perfusion imaging, our understanding of the neurophysiology of traumatic brain injury has been greatly enhanced. For example, both MR spectroscopy and magnetization transfer imaging have been shown by Sinson et al to quantify damage after traumatic brain injury [11]. Posttraumatic differences in N-acetyl aspartate (NAA)/ Creatine (Cr) ratios between patients with good and bad outcomes were observed but were not statistically significant. Furthermore, McGowan et al have reported that quantitative magnetization transfer imaging can be used to detect abnormalities associated with mild traumatic brain injury that are not apparent on routine CT or MR scanning [10]. Although there was only a weak correlation between the MR and neurophysiological data, refinements in the technique may allow development of a grading system to predict extent of injury. In fact, mild traumatic brain injury is becoming an intense area of focus, given its high prevalence in the population and the greater sensitivity of advanced MR techniques. Hofman et al recently reported the largest prospective study to date correlating neuroimaging findings and neurocognitive tests in patients with mild traumatic brain injury [9]. Their results suggest that even mild trauma to the brain results in abnormalities identified on single photon emission computed tomography (SPECT) and MRI that were previously inapparent. Unfortunately, the correlation between the two imaging techniques and neurocognitive tests was poor. Nevertheless, given the sensitivity of MRI and SPECT imaging to post-traumatic brain lesions, the data support the further application of these techniques to patients with neurotrauma. The future of traumatic brain injury research and neuroimaging is bright. We are no longer limited to simple anatomic descriptive analysis, but can extend ourselves into the sphere of outcomes research. The role of the neuroradiologist as prognosticator is becoming more important as the tools at our disposal allow better understanding of the link between what we see and what clinicians observe.

References
[1] Young B. Sequelae of head injury. In: Wilkins RH, Rengachary SS, editors. Neurosurgery. New York: McGraw-Hill; 1996. p. 2841 6. [2] Jennet B. Epilepsy after non-missile head injuries. 2nd edition. Chicago Year Book, 1975. [3] Rapport RL, Ojemann GA. Prophylactically administered phenytoin: effects on the development of chronic cobalt-induced epilepsy in the cat. Arch Neurol 1975; 32:539 48. [4] Levin HS, Grossman RG, Rose JE, Teasdale G. Longterm neuropsychological outcome of closed head injury. J Neurosurg 1979;50:412 22. [5] Clifton GL, Grossman RG, Makela ME, Miner ME, Handel S, Sadhu V. Neurological course and correlated computerized tomography findings after severe closed head injury. J Neurosurg 1980;52:611 24. [6] Bigler ED. Quantitative magnetic resonance imaging in traumatic brain injury. Head Trauma Rehabil 2001; 16(2):117 34. [7] Bigler ED, Johnson SC, Blatter DD. Head trauma and intellectual status: relation to quantitative magnetic resonance imaging findings. Appl Neuropsychol 1999;6 (4):217 25. [8] Hoelper BM, Soldner F, Chone L, Wallenfang T. Effect of intracerebral lesions detected in early MRI on outcome after acute brain injury. Acta Neurochirurgica Suppl 2000;76:265 7. [9] Hofman P, Stapert SZ, van Kroonenburgh M, Jolles J, de Kruijk J, Wilmink JT. MR imaging, single-photon emission CT, and neurocognitive performance after mild traumatic brain injury. AJNR Am J Neuroradio 2001;22:441 9. [10] McGowan J, Yang JH, Plotkin RC, Grossman RI, Umile EM, Cecil KM, et al. Magnetization transfer imaging in the detection of injury associated with mild head trauma. AJNR Am J Neuroradio 2000;21:875 80. [11] Sinson G, Bagley LJ, Cecil KM, Torchia M, McGowan J, Lenkinski RE, et al. Magnetization transfer imaging and proton MR spectroscopy in the evaluation of axonal injury: correlation with clinical outcome after traumatic brain injury. AJNR Am J Neuroradiol 2001; 22:143 51. [12] Hans JS, Kaufman B, Alfidi RJ, et al. Head trauma evaluated by magnetic resonance and computed tomography: a comparison. Radiology 1984;150:71 7. [13] Kim KS, Hemmati M, Weinberg PE. Computed tomography in isodense subdural hematoma. Radiology 1978;128:71 4. [14] Tsai FY, Huprich JE, Gardner FC, et al. Diagnostic and prognostic implications of computed tomography of head trauma. J Comput Assist Tomogr 1978;2:323 31. [15] Udstuen GJ, Claar JM. Imaging of acute head injury in the adult. Neuroimaging Clinics of North Am. 2001; vol 11(3):433 45. [16] Zee CS, Go JL. CT of head trauma. Neuroimaging Clinics of North Am. 1998;vol 8(3):525 39. [17] Zee CS, Segall HD, Destian S, et al. Radiologic eval-

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338 uation of head trauma. In: Wilkins R, Rengachary S, editors. Neurosurgery. New York: McGraw-Hill, 1996. p. 2675 87. Diaz FG, Yock Jr. DH, Larson D, Rockswald GL. Early diagnosis of delayed posttraumatic intracerebral hematomas. J Neurosurg 1979;50:217 23. Gentleman D, Nath F, MacPherson P. Diagnosis and management of delayed traumatic intracerebral haematomas. Br J Neurosurg 1989;3:367 72. Gudemann SK, Kishore PRS, Miller JD, Girevendulis AK, Lipper MH, Becker DP. The genesis and significance of delayed traumatic intracerebral hematoma. Neurosurgery 1979;5:309 13. Clark JA, Finelli RE, Netsky MG. Disseminated intravascular coagulation following cranial trauma: case report. J Neurosurg 1980;52:266 9. Drayer BP, Poser CM. Disseminated intravascular coagulation and head traumatwo cases studies. JAMA 1975;231:174 5. McGauley JL, Miller CA, Penner JA. Diagnosis and treatment of diffuse intravascular coagulation following cerebral trauma: case report. J Neurosurg 1975;43: 374 6. Poon WS, Rehman SU, Poon CY, Li AKC. Traumatic extradural hematoma of delayed onset is not a rarity. Neurosurgery 1992;30:681 6. Sweet RC, Miller JD, Lipper M, et al. Significance of bilateral abnormalities on the CT scan in patients with severe head injury. Neurosurgery 1978;3:16 21. Keimowitz RM, Annis BL. Disseminated intravascular coagulation associated with massive brain injury. J Neurosurg 1973;39:178 80. Miner ME. Delayed and recurrent intracranial hematomas and post-traumatic coagulopathies. In: Wilkins R, Rengachary S, editors. Neurosurgery. New York: McGraw-Hill, 1996. p. 2809 12. Miner ME, Kaufman HK, Graham SH, et al. Disseminated intravascular coagulation, a fibrinolytic syndrome following head injury in children: frequency and prognostic implications. J Pediatr 1982;100:687 91. Pondaag W. Disseminated intravascular coagulation related to outcome in head injury. Acta Neurochir (Wien) 1979;28:98 102. Kaufman HH, Moake JL, Olson JD, Miner ME, duCret RP, Pruessner JL, et al. Delayed and recurrent intracranial hematomas related to disseminated intravascular clotting and fibrinolysis in head injury. Neurosurgery 1980;7:445 50. Pretorius ME, Kaufman HH. Rapid onset of delayed traumatic intracerebral haematoma with diffuse intravascular coagulation and fibrinolysis. Acta Neurochair 1982;65:103 9. Goodnight SH, Kenoyer G, Rapaport SI, Patch MJ, Lee JA, Kurze T. Defibrination after brain tissue destruction: a serious complication of head injury. N Engl J Med 1974;290:1043 7. Vecht CJ, Sibinga CTS, Minderhoud JM. Disseminated intravascular coagulotion and head injury. J Neurol Neurosurg Psychiatry 1975;38:567 71.

337

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34] Preston FE, Malia RG, Sworn MJ, Timperley WR, Blackburn EK. Disseminated intravascular coagulation as a consequence of cerebral damage. J Neurol Neurosurg Psychiatry 1974;37:241 8. [35] Van der Sande JJ, Emeis JJ, Lindeman J. Intravascular coagulation: a common phenomenon in minor experimental head injury. J Neurosurg 1981;54:21 5. [36] Ommaya AK. Cerebrospinal fluid fistula and pneumocephalus. In: Wilkins R, Rengachary S, editors. Neurosurgery. New York: McGraw-Hill, 1996. p. 2773 82. [37] Friedman JA, Ebersold MJ, Quast LM. Post-traumatic cerebrospinal fluid leakage. World J Surg 2001;25(8): 1062 6. [38] Wehrle PF, Mathies Jr. AW, Leedom JM. Management of bacterial meningitis. Clin Neurosurg 1967;14:72 85. [39] Lewin W. Cerebrospinal fluid rhinorrhea in nonmissile head injuries. Clin Neurosurg 1966;12:237 52. [40] Kastner S, Schroth I, Deinsberger W, Boker DK. Delayed appearance of posttraumatic cerebrospinal fluid fistula as a cause of recurrent meningitis. Nervenrzt 2001;72(4):307 11. [41] MacGee EE, Cauthen JC, Brackett CE. Meningitis following acute traumatic cerebrospinal fluid fistula. J Neurosurg 1970;33:312 6. [42] Schneider RC, Thompson JM. Chronic and delayed traumatic cerebrospinal rhinorrhea as a source of recurrent attacks of meningitis. Ann Surg 1957;145:517 29. [43] Lund VJ, Savy L, Lloyd G, Howard D. Optimum imaging and diagnosis of cerebrospinal fluid rhinorrhea. J Laryngol Otol 2000;114(12):988 92. [44] Klein S, Woischneck D, Firsching R, Heinrichs T. Magnetic resonance imaging diagnosis of craniobasal cerebrospinal fluid fistulas by 3D-CISS sequence. Zentral Neurochir 2000;61(3):150 4. [45] Husson B, Pariente D, Tammam S, et al. The value of MRI in the early diagnosis of growing skull fracture. Pediatr Radiol 1996;26:744 7. [46] Taveras J, Ransohoff J. Leptomeningeal cysts of the brain following trauma with erosion of the skull: a study of seven cases treated by surgery. J Neurosugery 1953;10:233 41. [47] Rahimizadeh A. Growing fracture of the skull in the elderly. Neurosurgery 1986;19(4):675 6. [48] Lende RA, Erickson TC. Growing skull fractures of childhood. J Neurosurg 1961;18:479 89. [49] Blatter DD, Bigler ED, Gale SD, et al. MR- based brain and cerebrospinal fluid measurement after traumatic brain injury: correlation with neuropsychological outcome. Am J Neuroradiol 1997;18(1):1 10. [50] Soto-Ares G, Vinchon M, Delmaire C, et al. Cerebellar atrophy after severe traumatic head injury in children. Childs Nerv Syst 2001;17(4 5):263 9. [51] Benavidez DA, Fletcher JM, Hannay HJ, Bland ST, Caudle SE, Mendelsohn DB, et al. Corpus callosum damage and interhemispheric transfer of information following closed head injury in children. Cortex 1999; 35(3):315 36. [52] Tate DF, Bigler ED. Fornix and hippocampal atrophy in traumatic brain injury. Learn Mem 2000;7(6):442 6.

338

C.-S. Zee et al / Neuroimag Clin N Am 12 (2002) 325338 [56] Pierallini A, Pantano P, Fantozzi LM, Bonamini M, Vichi R, Zylberman R, et al. Correlation between MRI findings and long-term outcome in patients with severe brain trauma. Neuroradiology 2000;42(12): 860 7. [57] Wedekind C, Fischbach R, Pakos P, Terhaag D, Klug N. Comparative use of magnetic resonance imaging and electrophysiologic investigation for the prognosis of head injury. J Trauma 1999;47(1):44 9.

[53] Server A, Dullerud R, Haakonsen M, Nakstad PH, et al. Post-traumatic cerebral infarction. Neuroimaging findings, etiology and outcome. Acta Radiol 2001;42(3): 254 60. [54] Charfi N, Abid M, Mnif M, et al. Post-traumatic hypopituitarism. Presse Med 2001;30(2):59 61. [55] Firsching R, Woischneck D, Klein S, Reissberg S, Dohring W, Peters B. Classification of severe head injury based on magnetic resonance imaging. Acta Neurochirurgica (Wien) 2001;143(3):263 71.

Neuroimag Clin N Am 12 (2002) 339 343

Surgical management of head trauma

Peter Gruen, MD
Department of Neurosurgery, Keck School of Medicine, University of Southern California, 1200 North State Street, Room 5046, Los Angeles, CA 90033, USA

Imaging provides essential information, both direct and inferential, that neurosurgeons use to guide resuscitation and determine who is or is not a candidate for intracranial pressure (ICP) monitoring and/or evacuation of a traumatic hematoma mass. An important limitation of currently available neuroimaging in the management of head trauma patients is its inadequate depiction of the dynamic evolution of traumatic intracranial pathology. Another is that there is still no imaging modality that approaches the accuracy of direct measurement (usually through ventriculostomy) in diagnosing increased intracranial pressure. Another limitation of imaging is its inability to accurately predict pressure elevation. Imaging is good for pressure elevation prediction in the presence of certain signs: obliteration of cisterns, midline shift, mass effect. Pressure can be elevated, however, without these findings. The CT appearance of a lesion is an important factor in surgical candidate determination [1] Most neurosurgeons are uncomfortable just observing (and not surgically evacuating) a mass larger than 2 cm in or near elegant (functionally distinct and important) brain regardless of the patients neurologic status. Neurosurgeons rely on anatomic information for determining whether surgery is indicated, and, importantly, on the location and extent of cranial opening necessary to evacuate a lesion [2 4].

Imaging findings Imaging findings are important in management, monitoring, surgery, and prognostication of head injury [5]. Although essential in patient selection and surgical planning, imaging findings (by whatever modality) must be assessed in the context of the patients history and exam and an ever-expanding range of monitoring modalities that include tissue oxygen saturation, pH, lactate, pCO2, and temperature, in addition to intracranial (and cerebral perfusion) pressure.

Increased intracranial pressure Though CT does not always accurately predict intracranial pressure elevation, any or all of the radiographic findings that follow should prompt the neurosurgeon to serious consideration of intracranial pressure monitoring and intracranial pressure control management. Loss of gray-white junction indicates the presence of cerebral edema. This finding correlates highly with increased intracranial pressure.

Midlline shift Shift across the midline is an ominous finding not just because it is frequently associated with enough distortion of normal brainstem anatomy to cause dysfunction clinically manifest as extensor posturing with pupillary nonreactivity, but also because it is frequently associated with increased intracranial pressure. A 1982 review of the CT scans of 40 closed head injury patients showed no difference in the frequency or magnitude pressure elevation between

E-mail address: jpgruen@hsc.edu (P. Gruen).

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 1 3 - 8

340

P. Gruen / Neuroimag Clin N Am 12 (2002) 339343

those patients whose CT scans had midline shift and those that did not.

Hematoma mass An extra-axial hematoma thicker than 2 or 3 mm is in contact with the cortical surface. All additional thickness comes with increasing pressure on the brain. Each of three intracranial compartments contributes to the total intracranial pressure in proportion to its volume: (1) soft tissue (brain, hematoma, tumor, etc.), (2) intravascular (arteries, veins, capillaries), and (3) cerebrospinal fluid (ventricles, cisterns, sulci) Additional soft tissue volume is tolerated with minimal incremental elevation in ICP until a critical threshold is reached after which subsequent increment in volume leads to a steep and accelerating climb. Because the point of accelerated ICP increase is somewhat variable, the ICP of some patients decompensates with even a relatively small volume hematoma; the surgeons evacuation-indication-size threshold needs to be relatively low. The same 1982 study that found midline shift did not correlate with increased ICP, found no statistically significant correlation between hematoma mass size and elevated intracranial pressure.

Most herniation syndromes, however, occur in the context of diffuse cerebral swelling (edema) with midline shift, mass effect, and obliteration of the basal cisterns. The correlation between the finding of cistern obliteration and fixed and dilated pupils is high. Patients with obliterated cisterns are frequently posturing (Glascow Coma Score [GCS] 4 or 5) or not moving at all (GCS = 3), frequently also with one or both pupils fixed and dilated. The prognosis for meaningful neurologic recovery with CT evidence of cistern obliteration is poor.

Contusion Contusions can increase in size (blossom) suddenly and unpredictably. A cluster of small hematomas separated by normal brain can coalesce into large single hematomas of volume sufficient to raise intracranial pressure. Frequently, coalescence and blossoming of contusions is associated with an increase in surrounding parenchymal edema which further elevates intracranial pressure.

Fractures Elevation of open depressed skull fractures is indicated when there is bone displaced more than the full thickness of the skull. Elevation of less depressed fractures should be considered in the presence of air in the subarachnoid space, especially in proximity to an aerated sinus (frontal, ethmoid, mastoid) in which case the elevation is done to allow for water-tight dural closure to prevent leakage of cerebrospinal fluid (CSF).

Mass effect Reduction in size with change in shape of the ventricles is an indicator of mass effect frequently associated with elevated ICP.

Subdural hematoma A thin layer of blood in the subdural space is often associated with disproportionate swelling of the underlying brain, mass effect, and (almost always) increased ICP. Larger (thicker, more expansive) hematomas are associated with increased ICP because increased intracranial volume are caused by both clot and brain parenchymal swelling.

Surgery CT findings that are important for surgery include those that are surgical indications for operation and those used for planning and execution (Table 1).

Indications Indications for surgery to evacuate a hematoma can include its size and location. Size of clot Volume and extent of clot are important indications for or against surgery. Small- to moderate-size

Herniation (cistern obliteration) A clot of volume too small to elevate ICP can nevertheless cause herniation if located at or near the border between two intracranial compartments (such as medial temporal [uncal], cingulate, and cerebellar tonsillar).

Table 1 Head trauma CT imaging findings and management implications CT finding Loss of gray-white differentiation Obliteration of cisterns Pathology Cerebral edema, probable increased ICP Probable herniation, increased ICP Resuscitation Control ICP Control ICP, ICP monitoring Surgery Possible decompressive craniectomy Possible decompressive craniectomy Prognosis Difficult to control ICP, guarded Poor. Signs of brainstem compression (posturing, blown pupils) uncommonly reversible. Variable Better with timely diagnosis and treatment.

Midline shift Hydrocephalus

Increased ICP, possible herniation Increased ICP

Intraventricular hemorrhage Intracerebral hematoma Epidural hematoma

Hydrocephalus, increased ICP Increased ICP, focal areas dead brain Increased ICP, pressure on cortical surface. Clinical lucid interval. Cerebral edema, increased ICP Cerebral edema, increased ICP, pressure (and irritation) cortical on cortical surface acute Cerebral edema, increased ICP, pressure (and irritation) cortical on cortical surfaceinsidious Increased ICP from volume of fluid, membranes predisposed to re-bleed Risks of vasospasm Possible sagittal laceration risk of exsanguination on elevation of bone fragments. Possible sagittal sinus thrombosis with venous hypertension/hemorrhage Bleeding from transverse sinus laceration

Control ICP, ICP monitoring Consider ventriculostomy if acute by history. ICP control measures (diuresis, hyperventilation, etc.) do not decrease hydrocephalus Consider ventriculostomy for CSF drainage Ventriculostomy Control ICP, ICP monitoring only if suspicion of cerebral edema associated with hematoma ICP monitoring, control ICP ICP monitoring, control ICP

Possible decompressive craniectomy Consider ventriculostomy if acute by history

P. Gruen / Neuroimag Clin N Am 12 (2002) 339343

Consider ventriculostomy for CSF drainage Large hematoma may need surgery Possible craniotomy for evacuation

Poor Variable. Timely (soon after onset deterioration) Variable. Early evacuation essential. Guarded Variable

Rim subdural hematoma Acute subdural hematoma

ICP monitoring only Craniotomy

Subacute subdural hematoma

Evacuation, ICP monitoring

Craniotomy, burr holes

Variable

Chronic subdural hematoma Subarachnoid hemorrhage Depressed skull fracture midline convexity

Evacuation, possible ICP monitoring Consider triple H therapy and/or Nimodipine None

Craniostomy No surgery Avoid disturbing midline bone fragments

Variable Variable Good

Posterior fossa epidural

None

Suboccipital craniectomy

Variable 341

Abbreviations: CSF, cerebrospinal fluid; ICP, intracranial pressure.

342

P. Gruen / Neuroimag Clin N Am 12 (2002) 339343

(2 5 cm diameter, 5 10 cc) hematomas usually are nonoperative unless in the medial temporal or cerebellar region (see below). Larger clots, especially when associated with mass effect and anatomic shift, require evacuation to control (decrease) ICP.

Localization Hematoma size and location (regional and layer) imaging gives information about the location and size of a surgical hematoma (Table 2). The surgeon plans the craniotomy according to the CT scan. The importance of localization of the hematoma to the epidural, subdural, or intraparenchymal space is that the size of the cranial opening (amount of exposure) is not simply equal to the extent of the clot. Evacuation of a subdural requires a bone opening overlapping the entire hematoma so it can be completely evacuated. It can be difficult and dangerous to reach clot under a bone flap that does not uncover the entire extent of the clot. Identification of cortical bleeders that can continue to bleed and cause reaccumulation of clot after evacuation requires lifting clot off of the brain surface and inspecting which can be problematic with inadequate exposure.

Location of clot The surgeons size threshold for operating to evacuate a hematoma depends on it location. As noted above, some locations (such as medial temporal) are more predisposed to herniation, and moderate-size clots that in other locations (frontal, parietal, occipital convexity) can be observed expectantly are viewed as more ominous, and therefore more strongly mandating evacuation if they occur temporally or in the caudal posterior fossa. In the posterior fossa, mass effect from a moderate-size clot in proximity to the fourth ventricle may threaten to compress the ventricle to the point that CSF circulation is blocked with resultant acute hydrocephalus and, with further expansion, brainstem compression. Swelling of the brain under a just-evacuated epidural or subdural hematoma may indicate that a small contralateral hematoma has enlarged in response to the intracranial decompression and decreased intracranial (and temponade) pressure. Such paradoxical brain swelling following extra-axial clot evacuation may prompt the surgeon to reposition or reprep and open the contralateral side of the head to evacuate what is suspected to be an enlarged contralateral clot.

Objectives Epidural hematomas result from bleeding from branches of the middle meningeal artery running through the dura. Pulling clot off the dura and inspecting it under an overlying edge of bone is much easier and safer than from the brain surface.

Hazards Imaging helps the surgeon avoid intracranial hazards. Traumatic hematomas may overlie vascular and/ or neural structures that encountered potentially disastrous consequences.

Cerebral edema Decompressive craniectomy is currently an unproven method for reduction of ICP that cannot be brought down by standard medical interventions such as head elevation, CSF drainage, diuresis, hyperventilation, and/or pharmacologically induced coma.

Intracranial pressure monitoring A limitation of imaging is its inability to accurately predict pressure elevation. Imaging is good for pressure elevation prediction in the presence of certain signs: obliteration of cisterns, midline shift, mass
Table 2 Hematoma size and location indications for evacuation Location Volume (cc) > 25 >10 >25 >25 >10

Planning In planning a craniotomy to evacuate a traumatic hematoma, the surgeon uses imaging information to optimize craniotomy location, safety, and extent of evacuation, and to minimize the risk of reaccumulation.

Frontal Temporal Parietal Frontotemporoparietal Posterior fossa

P. Gruen / Neuroimag Clin N Am 12 (2002) 339343

343

effect [6]. Pressure can be elevated without these findings, however. Indications for intracranial pressure monitoring include findings on imaging and neurologic exam (Glascow Coma Score is the most important). Intracranial pressure monitoring is also important for determining the timing of follow up CT scanning in patients with no or minimal findings on their admission study [7]. Two types of intracranial pressure monitoring devices are available: bolts and ventricular catheters. Bolt Bolts are indicated in patients with cerebral edema and/or hematoma mass sufficient to compress the ventricles to the point that placement of a ventriculostomy tube into one of the anterior horns is improbable if not impossible using external cranial landmarks (standard ventriculostomy insertion procedure). Bolts can be inserted into the epidural, subdural, or intraparenchymal (brain tissue) space at the bedside in the ICU or ER, or in the OR pre-, intra-, or postop. Ventricular catheter A tube in the ventricle can perform two very important functions in the management of head injury: (1) measurement of intracranial pressure, and (2) drainage of cerebrospinal fluid for control of ICP.

12-hour intervals, cannot adequately portray the dynamic, sometimes rapid evolution (usually growth) of a traumatic hematoma. These limitations aside, CT scanning provides adequate imaging for rational surgical treatment of head injury. Whether for monitoring, diagnostic, or therapeutic purposes, cranial procedures for traumatic pathology are guided by CT.

References
[1] Wong CW. CT and clinical criteria for conservative treatment of supratentorial traumatic intracerebral haematomas. Acta Neurochir 1995;135(3 4):131 5. [2] Stein SC, Ross SE. Moderate head injury: a guide to initial management. J Neurosurg 1992;77(4):562 4. [3] Brain Trauma Foundation and American Association of Neurological Surgeons. Guidelines for the management of severe traumatic head injury. New York: Brain Trauma Foundation, Inc.; 2000. [4] Soloniuk D, Pitts LH, Lovely M, Bartkowski H. Traumatic intracerebral hematomas: timing of appearance and indications for operative removal. J Trauma 1986; 26(9):787 94. [5] Zee CS, Go JL. CT of head trauma. Neuroimaging Clin N Am 1998;8(3):525 39. [6] Tabaddor K, Danziger A, Wisoff HS. Estimation of intracranial pressure by CT scan in closed head trauma. Surg Neurol 1982;18(3):212 5. [7] Zahari M, Mohd Ali AK, Chandrasekharan S. Delayed intracranial haemorrhage in head injury. Singapore Med J 1996;37(3):285 7.

Summary In addition to its inability to predict pressure elevation accurately, head CTs done serially even at

Neuroimag Clin N Am 12 (2002) 345 349

Index
Note: Page numbers of article titles are in boldface type.

A Aneurysms, following cranial gunshot wounds, 245 246 intracranial, 259 263 clinical presentation of, 260 263 treatment of, 263 Angiography, in cranial gunshot wounds, 236 237 Axonal injury(ies), diffuse, 194 197 clinical presentation of, 207 208 imaging of, 208 215 prognostic factors in, 208 histopathology of, 207 in children, 285 mechanism of injury in, 205 207 pathophysiology of, 205 208 and imaging of, 205 216 B Birth trauma, head trauma in, 275 277 Blunt neck injuries. See Neck, injuries to, blunt. Brain, contusions of, 177 178 injury to, classification of, 176 179 diffuse, 179 180 dementia following, 179 development of, 179 diagnosis of, 179 180 gross and microscopic abnormalities in, 179, 180 histologic appearance of, 180 hemorrhage in, 180 183 in head trauma, 165, 175 in trauma, mechanisms of, 175 176 swelling of, in brain injury associated with head trauma, 182 183 Brain death, in head trauma, 170 Brainstem, hemorrhage within, 222 223 injury to, in head trauma, 197

C Caput succedaneum, during birth process, 275 Carotid-cavernous fistulas, clinical presentation of, 263 266 direct, 263 indirect, 263 radiographic evaluation of, 266 treatment of, 266 267 Cephalohematoma, during birth process, 275 276 Cerebellar examination, in head trauma, 167 168 Cerebellum, post-traumatic atrophy of, 335, 337 Cerebral edema, diffuse, in pediatric head trauma, 287, 292 Cerebral herniation(s), acquired, imaging of, 217 229 ascending transtentorial, 223 224 definition of, 217 descending transtentorial, 222 223 external, 225, 227 intracranial hypotension as cause of, 226, 227 patterns of, 217, 218 posterior(parahippocampal) tentorial, 221 222 subfalcine, 219 220 tonsillar(foramen magnum), 224 226 transphenoidal(transalar), 225, 226 227 transtentorial, 220 uncal, 220 221 Cerebral infarction, post-traumatic, 335 336 Cerebral lacerations, 178 179 Cerebrospinal fluid fistulae, following cranial gunshot wounds, 245 Cerebrospinal fluid rhinorrhea, traumatic, 333 334 Cerebrum, post-traumatic atrophy of, 335, 337 Child abuse, mimickers of, 293

1052-5149/02/$ see front matter D 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 2 - 5 1 4 9 ( 0 2 ) 0 0 0 2 4 - 2

346

Index / Neuroimag Clin N Am 12 (2002) 345349

Children, diffuse axonal injury in, 285 diffuse brain swelling in, 182 183 extraaxial hemorrhage in, 280 281 head trauma in, imaging of, 273 296 hemorrhages in, in head trauma, 280 285, 290 Cistern obliteration, in head trauma, 342 Computed tomography, in cranial gunshot wounds, 233 235 in diffuse axonal injury, 209, 210 in extracranial vascular trauma, 252 in pediatric head trauma, 273 in traumatic intercranial hemorrhage, 189 190 Contusions, classification of, 177 contusion index and, 178 cortical, definition of, 177 formation of, 177 178, 179 in head trauma, 192 194 of brain, 177 178 Corpus callosum, post-traumatic atrophy of, 335, 336, 337 Cranial cavity, anatomy of, 217 219 Cranial gunshot wounds. See Gunshot wounds, cranial. Cranial nerve examination, in head trauma, 166 Cranial vault, gunshot wounds of, radiographic assessment of, 234 235, 236 237, 240, 241 244 Cyst, hematic, 322 leptomeningeal, following skull fracture, 278 280, 334 335

F Face, buttresses of, 297 299 fractures of, 299 309 classification of, 299 limited, 302 306 pediatric, 309 normal anatomy of, 297 trauma to, vascular, clinical presentation of, 256 257 treatment of, 257 259 Facial trauma, causes of, 299 imaging of, 297 311 techniques of, 299 Fistulas, carotid-cavernous. See Carotid-cavernous fistulas. Fluid attenuated inversion recovery, in diffuse axonal injury, 210 211 Frontal sinus, fractures of, 300, 304 Functional neuroimaging, in diffuse axonal injury, 213 215 G Gait examination, in head trauma, 168 Glasgow coma scale, 166, 169 170 Globe, rupture of, 321 Gray matter, deep cerebral, in head trauma, 197 Gunshot wounds, cranial, delayed complications of, 244 246 outcomes of, radiographic analysis of, 246 penetrating, 231, 234 235 perforating, 231, 236 237, 238 239 plain skull films in, 232 233 radiographic assessment of, 231 250 superficial, 231, 232 tangential, 231, 233 types of, 231 H Head trauma, bolts in, 345 brain injury in, 165 causes of, 165 cerebral edema in, 344 chronic symptoms of, 171 172 cistern obliteration in, 342 classification of, 169 170, 176 179 clinical evaluation of patients with, 165 174 clot in, location of, 344 size of, 342 computed tomography in, 175, 343

D Diffusion weighted imaging, in diffuse axonal injury, 211 212 Disseminated intravascular coagulopathy, and fibrinosis, and head trauma, 330 331 Durets hemorrhage, 222 223

E Electroencephalogram, in head trauma, 168 Encephalomalacia, and porencephaly, 335, 336 intracranial hematoma evolving into, 328, 329 Extraaxial hemorrhage, traumatic, 197 202 Eye, foreign bodies in, 323 325 penetrating injury of, 323

Index / Neuroimag Clin N Am 12 (2002) 345349

347

contusions in, 192 194, 342 delayed intracerebral hematomas in, 329 330 fractures in, 342 future treatment of, 173 hematoma in, hazards of, 344 mass of, 341 342 size of, indications for evacuation of, 344 herniation in, 342 imaging findings in, 341 imaging of sequelae of, 327 340 in seizure, 165 166 incidence of, 189 increased intracranial pressure in, 341 intracranial pressure monitoring in, 344 345 laboratory and diagnostic testing in, 168 loss of consciousness in, 165 magnetic resonance imaging in, 175 management of, 189 medical evaluation in, 165 166 medical-legal issues in, 172 midline shift in, 341 movement disorders following, 171 172 neurologic examination in, 165, 166 168 neuropsychologic changes of, 172 pathology of, 175 187 pathophysiology of, 168 169 pediatric, imaging of, 273 296 in birth trauma, 275 277 incidence of, 273 mechanisms of, 273 nonaccidental, 287 293 sequelae of, 285 287 prevention of, 172 173 seizures following, 171 signs of, 166 subdural hematoma in, 342 surgical management of, 341 345 therapy for, 170 171 types of injuries in, 175 ventricular catheter in, 345 Hematic cyst, 322 Hematoma(s), epidural, in children, 280 281 in head trauma, 181, 200 201 intracerebral, delayed, in head trauma, 329 330 in head trauma, 181 182 intraparenchymal, in head trauma, 194 subdural, 322 in children, 281, 282, 283, 289 290 in head trauma, 181, 189 190, 191, 197 199, 342 subdural and epidural, delayed, following head trauma, 330, 331

Hemorrhage, extraaxial, in children, 280 281 traumatic, 197 202 extraconal, 321 in injury to brain, 180 183 intercranial, traumatic, herniation following, 191 192 imaging of, 189 204 imaging techniques in, 189 190 mass effect of, 190 192 intraaxial, in children, 281 285 intraventricular, in children, 281 in head trauma, 202 intravitreous, 319 orbital, 320 322 subarachnoid, in children, 281, 290 in traumatic brain injuries, 201, 202 subdural, during birth process, 276 277 subgaleal, during birth process, 275 within brainstem, 222 223 Herniations, cerebral. See Cerebral herniation(s). Hydrocephalus, following cranial gunshot wounds, 245 in pediatric head trauma, 283, 287 Hygroma, subdural, computed tomography in, 183 following head trauma, 183 Hyphema, 319 Hypopituitarism, post-traumatic, 336 337 Hypotension, intracranial, as cause of cerebral herniations, 226, 227

I Infection(s), following cranial gunshot wounds, 244 Intraaxial hemorrhage, in children, 281 285 Intraaxial injury(ies), in pediatric head trauma, 292 traumatic, 192 194 Intracranial pressure, increased, in head trauma, 183 monitoring of, in head trauma, 170 171 Intraocular injury, 319 320 Intraventricular hemorrhage, in children, 281 in head trauma, 202 Ischemic/hypoxic damage, in head trauma, 183

L Laceration(s), cerebral, 178 179 LeFort fractures, 306, 307 308

348

Index / Neuroimag Clin N Am 12 (2002) 345349

Lens, dislocation of, 320 subluxation of, 320 Leptomeningeal cyst, following skull fracture, 278 280, 334 335

Neurologic examination, in head trauma, 165, 166 168 Neurotraumatic outcomes, and neuroimaging, correlation of, 337 338 Nose, fractures of, 302

M Magnetic resonance angiography, in extracranial vascular trauma, 252 Magnetic resonance imaging, in cranial gunshot wounds, 235 in diffuse axonal injury, 210 in pediatric head trauma, 273 274 in traumatic intercranial hemorrhage, 190 quantitative, in diffuse axonal injury, 213, 214 Magnetic resonance perfusion and diffusion, in pediatric head trauma, 274 Magnetic resonance spectroscopy, in diffuse axonal injury, 213 in pediatric head trauma, 274 Magnetic source imaging, in pediatric head trauma, 274 275 Magnetization transfer imaging, in diffuse axonal injury, 213 in pediatric head trauma, 275 Malar fracture, 302, 303 304, 305, 306 Mandible, fractures of, 301, 304 306 Maxilla, alveolar, fractures of, 302 304 Meningitis, following cranial gunshot wounds, 245 Mental status examination, in head trauma, 166 Midface, lateral, fracture of, 306 Motor examination, in head trauma, 167 Movement disorders, following head trauma, 171 172

Nuclear medicine, in pediatric head trauma, 275 O Optic nerve, injury to, 322 323 Orbit, bony, anatomy of, 313 314, 315 fractures of, 316 319 imaging of, 314 316 Orbital trauma, 313 326 P Plain skull films, in cranial gunshot wounds, 232 233 Pneumocephalus, following head trauma, 331 333 Porencephaly, encephalomalacia and, 335, 336 R Reflex examination, in head trauma, 167 Rhinorrhea, cerebrospinal fluid, traumatic, 333 334 S Scalp, gunshot wounds of, radiographic assessment of, 237 240 injury to, during birth process, 275 277 Seizures, following head trauma, 171 Sensory examination, in head trauma, 167 Skull, fracture(s) of, 176 177 classification of, 177 in children, 277, 289 leptomeningeal cyst following, 278 280, 334 335 gunshot wounds of, radiographic assessment of, 232, 234 235, 236 237, 240 241 injury to, during birth process, 276 Smash fractures, 308 Spectroscopy, magnetic resonance, in diffuse axonal injury, 213 in pediatric head trauma, 274 Subarachnoid hemorrhage, in children, 281, 290 in traumatic brain injuries, 201, 202

N Nasoethmoidal-orbital fractures, 308 309, 310 Neck, injuries to, blunt, vascular, 251 253 causes of, 251 252 diagnosis of, 252 incidence of, 251 symptoms of, 252 penetrating, 253 256 clinical presentation of, 253 self-recanalization of, 256 treatment of, 253 256

Index / Neuroimag Clin N Am 12 (2002) 345349

349

T T2*-weighted gradient recall echo, in diffuse axonal injury, 211 Transfacial fractures, 306, 307 309 U Ultrasound, duplex, in extracranial vascular trauma, 252 253 in pediatric head trauma, 275 in traumatic intercranial hemorrhage, 190

V Vascular injury(ies), extracranial, 251 259 in pediatric head trauma, 285 287 intracranial, 259 267 dissections, clinical presentation of, 259 treatment of, 259 traumatic, and management of, 251 271 Visual field tests, in head trauma, 166 167