1 Expert Review of Anticancer Therapy

Emerging Role of Stereotactic Body Radiotherapy in the Treatment of Pancreatic Cancer

Betul Berber, Juan R Sanabria, Kelly Braun, Min Yao, Rodney J Ellis, Charles A Kunos, Jason Sohn, Mitchell Machtay, Bin S Teh, Zhibin Huang, Nina A Mayr, Simon S Lo
Disclosures

Expert Rev Anticancer Ther. 2013;13(4):481-487.

Abstract
The management of pancreatic cancer continues to be challenging. Despite surgical, genetic and molecular advances, its overall prognosis remains poor. Surgical resection is the only modality that offers a chance for a cure, with an overall survival rate of 1020% at 5 years. However, only 20% of the patients are surgical candidates because of locally advanced disease or systemic stage at presentation. Conventional radiotherapy, with or without chemotherapy, has been used to treat patients with advanced-stage pancreatic cancer, an approach with high rates of local recurrence. Stereotactic body radiation therapy, also known as stereotactic ablative radiotherapy has emerged as a treatment modality that allows the precise delivery of a large ablative radiation dose to a tumor volume while sparing surrounding organs and tissues. Phase I and II studies have shown good rates of local control of the disease but rates of distant metastasis remain significant. Strategies to combine novel systemic therapy and stereotactic body radiation therapy are to be explored.

Introduction
The management of pancreatic cancer continues to be challenging. Although it is the tenth most frequent malignancy, it is the fourth leading cause of cancer mortality for both sexes in the USA.[1] It is estimated that 43,920 new cases of pancreatic cancer were diagnosed in the USA in 2012, with 37,390 related deaths. Overall, despite the surgical, genetic and molecular advances, there has been little change in the overall prognosis over the last decades, with an overall 5-year survival of 5%. The onset of pancreatic cancer is often insidious. The frequent lack of symptoms at its onset allows the cancer to progress to a more advanced stage prior to its diagnosis. Surgical resection is the only modality providing a chance of cure, with a 5-year survival of 1020% with an R0 resection.[2,3] In a recent report based on Radiation Therapy Oncology Group 9704, where patients were randomized to either pre- and post-chemoradiotherapy gemcitabine or pre- and post-chemoradiotherapy 5-fluorouracil, the 5-year survival rates were 22 and 18%,

respectively.[4] Unfortunately, only 20% of the patients are candidates for a curative resection due to local invasion of the tumor or metastatic spread.[5] The treatment of patients with unresectable disease involves chemoradiotherapy, with a median survival of 7.818.8 months and few long-term survivors.[69] In these patients, local control is a goal, not only to slow the progression of the disease, but to prevent the obstruction and pain that quickly become detrimental to a patient's quality of life. Despite attempts to improve local control, local failure occurs in almost half of patients after chemoradiotherapy, affecting quality of life. [10]Furthermore, approximately 42% of the patients undergoing pancreatic resection have positive margins (R1 resection).[11] The Royal College of Pathologists currently recommend that microscopic evidence of tumor within 1 mm of a resection margin should be classified as R1 for pancreatoduodenectomy specimen reporting. In a study from University of Liverpool (Liverpool, UK), patients were found to have poorer survival if they had R1 resection, with incorporation of The Royal College of Pathologists recommendation. No difference in overall survival was found between patients with positive resection margins and those with microscopic evidence of tumor within 1 mm of a resection margin.[12] R2 resections have also been performed in patients with pancreatic cancer. With an increased risk of morbidity and mortality, a median survival of 8.2 months has been reported.[13] Conventional radiotherapy has drawbacks, including the limitation of the amount of radiation that can be delivered to the GI tract, the relatively high number of treatments involved (typically 2530 fractions) and the side effects. There had been some uncertainty regarding the efficacy of chemoradiotherapy compared with chemotherapy alone.[1418] Recently, a Phase III randomized study from the Eastern Cooperative Oncology Group (ECOG) showed that patients with locally advanced disease receiving concurrent radiotherapy with gemcitabine had improved survival compared with those receiving gemcitabine alone.[19] Local therapy evidently plays an important part in the management of locally advanced pancreatic cancer, although the local failure rate remains high. Trials of chemoradiotherapy with standard fractionation have shown a high local failure rate of up to 3060%.[2022] The role of local therapy may thus be to delay the local progression of the disease. Stereotactic body radiotherapy (SBRT) is a recent advancement that allows for the precise delivery of a large ablative radiation dose to the tumor in one to five fractions. SBRT has been investigated in unresectable pancreatic cancer for local tumor control.[3,6,15,23,24] The aim of this article is to review the emerging role of SBRT in the management of pancreatic cancer.

SBRT Principles
SBRT enables the precise application of individualized high doses of radiation to extracranial tumors in one to five fractions, while preserving the normal surrounding tissues due to the rapid dose falloff beyond the treated volume. It is a development from stereotactic radiosurgery that was originally used for treatment of intracranial tumors or benign lesions, with very effective local control rates of greater than 8090%.[25]Although its use in extracranial tumors was hindered by the movement of abdominal organs with respiration, various methods can be used to account for respiratory motions. The target volume motion can be dampened, gated or tracked and has been discussed in other publications.[26,27] The treatment devices available for the delivery of SBRT to the target volume include various linear accelerators with conebeam computed tomography (CT) capability or stereoscopic x-ray such as Elekta Synergy S, Varian Trilogy and TrueBeam, and Novalis TX, TomoTherapy, and CyberKnife. The University Hospitals Seidman Cancer Center (OH, USA) utilizes a CyberKnife system, which enables real-time tracking of implanted fiducial markers, and when combined with respiratory motion modeling, submillimeter accuracy can be maintained by continuously detecting and correcting for tumor motion during treatment. Previous reports have described the use of SBRT for various abdominal tumors.[28] Regardless of the device used, the keys to safe delivery of high-quality SBRT are proper training of the treatment team and rigorous quality assurance. The linear-quadratic (LQ) model has been used to estimate the tumoricidal effects and normal tissue constraints based on conventional radiotherapy data. However, the LQ model most likely overestimates radiation cell kill in the ablative dose range. Alternative models such as the universal survival model[29]and generalized LQ model[30] have been proposed for better estimation of tumor and normal cell kill for the dose range used in SBRT.

SBRT in Pancreatic Cancer

Patient Selection
The studies in the literature reporting SBRT for pancreatic cancer have enrolled patients with nonmetastatic locally advanced pancreatic cancer.[6,23,24,3134] In pancreatic cancer, nodal and hepatic metastases develop early, requiring the administration of chemotherapy early to target microscopic disease. A potential advantage of SBRT over conventional fractionated radiotherapy is to allow for earlier delivery of chemotherapy after local therapy.[35] Inclusion criteria used in these studies are biopsy proven pancreatic adenocarcinoma, unresectable disease and a life expectancy of at least 12 weeks.[34] Other inclusion criteria of some of these

studies are ECOG status of 2,[6] tumor size less than 7.5 cm[23] and absence of extensive vascular involvement, portal vein occlusion and aorta or inferior vena cava invasion.[24] Leukocyte count of >3000/l, absolute neutrophil count of >1500/l, total bilirubin <1.5 upper limit of normal, transaminases <2.5 upper limit of normal and creatinine level within normal limits are other biochemical parameters used for inclusion of patients for SBRT.[24] Preprocedural work up includes standard staging with CT, MRI or PET, as well as routine blood work and CA 19-9. Although research continues to be ongoing to identify novel biomarkers, CA 19-9 remains the most commonly used marker for monitoring the disease process in patients with pancreatic cancer. Despite the low positive predictive value in the initial diagnosis of pancreatic cancer, CA 19-9 levels can be monitored throughout the course of the disease to indicate response to treatment or in conjunction with imaging studies to identify disease progression.[36]

Treatment Planning
Treatment planning technique is dependent upon the treatment system used. For linear accelerator (LINAC)-based systems, intensity-modulated radiation therapy treatment planning is typically used because the pancreas is in close proximity to various organs at-risk such as the duodenum, stomach, liver and kidneys. Efforts have to be made to spare these structures to avoid serious complications. Several centers, including the University Hospitals Seidman Cancer Center, utilize the CyberKnife system for treatment of pancreatic cancer with SBRT (Figure 1). Three to six cylindrical solid gold markers 35 mm in length are placed either endoscopically, laparoscopically, percutaneously under CT-guidance or via laparotomy within and around the tumor at a minimum distance of 2 cm between adjacent markers.[6,23,24,31
34]

Approximately 12 weeks later, a pancreas protocol CT scan or PET/CT is obtained for CT

simulation. The patient is immobilized using a custom-made cradle. The scans are then imported into the Multiplan treatment planning systems and are digitally fused. Target volumes and critical structures, such as stomach, duodenum, kidneys, liver and spinal cord are contoured.[34] No expansion margin is used when the tumor is in contact with the stomach or duodenum; otherwise a 5 mm expansion margin is included.[15]

5
Figure 1

Locally advanced pancreatic cancer treated with CyberKnife-based stereotactic body radiotherapy to a dose of 30 Gy in three fractions. The orange line (inner) represents the prescribed isodose line (30 Gy at 70%) and the yellow line (outer) represents the 10 Gy line.

SBRT Dose Calculation

According to Mahedevan et al., the relationship between tumor location and duodenum is used to determine the prescribed dose.[15] The suggested maximal point tolerance dose of the duodenum is 30 Gy in three fractions. If the tumor approximated a third or more of the circumference of the duodenum or stomach, a dose of 24 Gy in three fractions was used. If the tumor abutted the bowel only in one area, and/or the space between the tumor and bowel wall was <3 mm, a dose of 30 Gy in three fractions was prescribed. If the gap between the tumor and duodenum was 3 mm, a dose of 36 Gy in three fractions was used. The dose was prescribed to the isodose line covering 95% of the planning target volume. These authors also used normal tissue constraints during treatment planning, where the volume of liver receiving 21 and 15 Gy was kept to <30 and 50%, respectively. The volume of each kidney receiving 12 Gy was kept to <25%. The total maximal spinal cord limit was limited to 12 Gy and the maximal point dose to the bowel <30 Gy in three fractions. The other studies have used 25 Gy prescribed to the isodose line that completely covered 95% of the planning target volume.[23,24]

Treatment Delivery
Patients are premedicated with dexamethasone and ondansetron and monitored during the treatment. Pretreatment verification of the patient's position is performed with image guidance, which can be in the form of a stereoscopic x-ray, as available in the Novalis TX and the CyberKnife or through volumetric imaging using cone-beam CT, CT-on-rail or megavoltage CT from TomoTherapy devices or other adapted LINACs. Treatment techniques with LINAC-based systems include 3D conformal radiotherapy, intensity modulated radiotherapy or volumetricmodulated arc therapy. The CyberKnife system tracks tumors during respiration and automatically adjusts during patient positioning. It has the following three key components: an advanced, lightweight LINAC, a robotic arm which can point the LINAC from a wide variety of angles and a tumor tracking system. The tracking of the tumor during the patient's normal respiratory cycle is performed by internal fluoroscopic monitoring of fiducial markers placed in or around the tumors, a computer software that creates an algorithm linking the tumor movement with chest wall movement and infrared light emitting diodes placed on the patient's chest and wall-mounted infrared detector which allows for the construction of a patient's breathing model. The accuracy for spinal lesions was reported to be 0.3 0.1 mm.[37] Figure 1 shows a CyberKnife-based SBRT plan for pancreatic cancer.

Follow-up
Patients are typically followed-up with a CT of the pancreas with thin cuts at 46 weeks, 1012 weeks and then every 3 months.[15]Some studies have incorporated PET-CT and/or MRI scans into the follow-up protocol.[23,34] CA 19-9 levels are also obtained at these time points. Acute toxicity is defined as occurring within 3 months after SBRT and long-term toxicity after 3 months.

Review of the Literature

One of the first studies to demonstrate the feasibility of SBRT for locally advanced pancreatic cancer was reported by Koonget al. from Stanford University (CA, USA).[6] In this Phase I study, 15 patients with an ECOG status of 2 were enrolled.[6]The patients received single fraction SBRT dose of either 15, 20 or 25 Gy to the primary pancreatic tumor. The gross tumor volume (GTV) ranged from 19.2 to 71.9 cc (mean: 32.9 cc). The dose was prescribed to the isodose lines ranging from 64 to 85%, and the 50% isodose line was only allowed to cover the proximal duodenal wall. There was no significant gastrointestinal toxicity within 12 weeks of the treatment. Grade 1 and 2 toxicities were observed in two and three patients, respectively. There were no significant alterations of blood counts and liver function tests. Among the 13 patients

with follow-up imaging, six developed metastatic disease at 46-week CT scans. The median survival was 11 months for all patients in the study. In the six evaluable patients who received the highest dose of 25 Gy, the median overall survival was 8 months. Local tumor control was accomplished in all of these patients until death or last follow-up. The same group from Stanford University subsequently reported the results of a Phase II trial evaluating the efficacy of 45 Gy of conventional fractionated radiotherapy with concurrent fluorouracil, followed by a boost with SBRT of 25 Gy in one fraction to the primary pancreatic tumor in 16 patients.[38] In this trial, two patients developed grade 3 toxicity and 15 out of the 16 patients were free from local tumor progression until time of death. The median overall survival was 8.3 months. A subsequent study from Stanford University evaluated the use of single-fraction SBRT and sequential gemcitabine for the same patient population.[24] A single SBRT dose of 25 Gy was delivered to an internal target volume with a 23 mm margin. All patients completed SBRT and a median of five cycles of gemcitibine-based chemotherapy. No acute grade 3 toxicity was observed, but one patient developed late grade 4 toxicity with duodenal perforation. Three patients developed grade 2 duodenal ulcers. The overall median survival was 11.8 months, with a 1-year survival of 50% and a 2-year survival of 20%. The freedom from local tumor progression on CT was 94% at 1 year. Colleagues from Europe reported the results of a Phase II trial on 22 patients with unresectable, locally advanced pancreatic carcinoma.[3] Three of these patients had a recurrence after a Whipple operation, two with local recurrence and one with lymph node recurrence. SBRT was administered as a central dose of 15 Gy 3 within 510 days. Two patients had a partial response and in six patients, the tumor recurred locally. The actuarial local tumor control rate was 57%. The median time to local progression was 4.8 months. The progression-free survival was 9% and the overall survival was 5% at 1 year. Acute toxicity was significant, with four patients developing severe mucositis or ulceration of the stomach or duodenum and one patient developing a nonfatal perforation of the stomach. The authors concluded that SBRT was associated with a poor outcome, unacceptable toxicity and questionable palliative effect. Investigators from Harvard University (MA, USA) examined the effect of SBRT and gemcitabine in the treatment of locally advanced pancreatic cancer in 36 patients.[15] All patients had a follow-up of at least 12 months and received a total dose of 2436 Gy in three fractions. With a follow-up of 24 months, the local control rate was 78%, and median overall survival time was 14.3 months. Seventy eight percent eventually developed distant metastases and 17% were free of disease at final follow-up. Grade 2 and 3 toxicities related to SBRT were observed in nine and five patients, respectively. The authors concluded that SBRT could be delivered

quickly and effectively in patients with nonmetastatic locally advanced pancreatic cancer with acceptable side effects and minimal interference with gemcitabine chemotherapy. The utility of SBRT in locally advanced pancreatic cancer, applied in a boost fashion, was also reported by Korean investigators.[33] In that retrospective review, a total dose of 40 Gy was delivered in 20 fractions using a conventional 3D technique, followed by a single fraction of 14, 15, 16 or 17 Gy SBRT as a boost without a break. Chemotherapy was administered to 21 patients. One-year survival was 60% and local progression-free survival was 70.2%. One patient (3%) developed grade 4 toxicity. CA 19-9 response was found to be an independent prognostic factor for survival. The conclusion was that a boost with SBRT provided a safe means of escalating radiation dose, thus indicating that a well-controlled Phase II study should be conducted. The largest study to date has been reported by Chang and colleagues on 77 patients with unresectable adenocarcinoma of the pancreas treated with SBRT involving 25 Gy in a single fraction.[23] The patients with a tumor size >7.5 cm were excluded as these tumors were deemed to be too large to be treated with SBRT in one session. Seventy two percent of the patients had either locally advanced or medically inoperable disease, while 19% had metastatic and 8% had locally recurrent disease. Twenty one percent also received fractionated radiotherapy and 96% received gemcitabine-based chemotherapy. With a median follow-up of 6 months for all patients and 12 months for surviving patients, the rate of freedom from local disease progression at 1 year was 84%. The local tumor control rate at 1 year was 95%. The overall progression-free survival at 1 year was 9% and the overall 1-year survival was 21%. Five percent of the patients developed grade 2 acute toxicity, 4% developed late grade 2 toxicity and 9% developed grade 3 late toxicity. At 1 year, the rate of grade 2 toxicity was 25%. The authors concluded that although their study achieved effective local tumor control with SBRT, there was a certain associated risk of toxicity; therefore, there was a need to reduce morbidity associated with SBRT. Investigators from University Hospitals Seidman Cancer Center have also reported their experience with SBRT for unresectable pancreatic cancer.[34] In their report of the first 20 patients receiving SBRT for unresectable pancreatic adenocarcinomas and one patient receiving SBRT for a neuroendocrine tumor, gold fiducial markers were placed within and around the tumor endoscopically in 13 patients, surgically in four patients, and percutaneously under CT-guidance in three patients. The mean radiation dose was 25 Gy (range: 2230) delivered in three fractions. Sixty eight percent of the patients received chemotherapy. The mean GTV was 57.2 cm3. The mean total GTV reduction at 3 and 6 months after SBRT was 21

and 38%, respectively. With a median follow-up of 15 months, the overall freedom from local progression at 6 and 12 months was 88 and 65%, respectively. The overall 1-year survival was 56%. There were no complications related to the placement of fiducial markers, and the rate of SBRT-related complications was 11% for grade 12 toxicity and 16% for grade 3 toxicity. These results have demonstrated that SBRT is a safe and effective local treatment modality for primary pancreatic malignancy with acceptable rate of adverse events. Table 1 summarizes the literature on SBRT for pancreatic cancer ( Table 2 ).

Expert Commentary
There is a large patient population affected by pancreatic adenocarcinoma destined to have a poor prognosis. Despite all the advancements in the management of these patients, the outcome of patients with locally advanced, unresectable disease is still dismal. SBRT is a valid option for achieving local tumor control in these patients. The potential advantages of SBRT include the potential for the delivery of a higher biological effective dose, the reduction of the number of patient visits, particularly for patients who need to travel long distances to specialized cancer centers for radiotherapy, and the reduction of cost. There have been a handful of studies reported to date which have reported local tumor control rates ranging between 57 and 100%.[3,5,6,15,23,24,34,38] Nevertheless, reflecting their aggressive tumor biology, overall survival has remained low, ranging between a median of 5.414 months in these studies, underscoring the need for more effective systemic therapy in these patients. Acute toxicities have been reported, consisting mostly of grade 2 and 3 gastrointestinal side effects, with a rare incidence of grade 3 and 4 toxicity, mainly due to gastrointestinal perforations. The incidence of complications has decreased in the most recent series performed by large-volume centers. Of note, none of the studies used relief of pain or biliary obstruction as end points. As a result, based on accumulating experience, it is fair to state that SBRT is a safe and effective local treatment option for unresectable pancreatic adenocarcinoma. In the multidisciplinary management of pancreatic cancer, SBRT adds to our armamentarium. The biggest challenge is to tackle the high risk of distant metastasis. Investigation of the combination of systemic therapy and SBRT is warranted. Currently, all of the clinical trials and studies have only one arm. A randomized Phase III trial comparing standard therapies such as chemotherapy alone or concurrent chemoradiotherapy to SBRT is needed to define the role of SBRT in the management of locally advanced pancreatic cancer.

10

Five-year View
Further prospective studies are necessary to determine long-term response and survival after SBRT for pancreatic. As we accrue more experience with SBRT for pancreatic malignancies, the clinical scenarios that incorporate SBRT into the management of these patients will probably also evolve. Despite improved local control, the biggest challenge is to tackle the high risk of distant metastasis. The combination of novel and more effective systemic therapy with SBRT should continue to be explored. Investigators from the University of Texas Southwestern Medical Center (TX, USA) are currently conducting a Phase I trial of margin intensive SBRT with or without gemcitabine for potentially resectable pancreatic cancer and the results of this trial are eagerly awaited. [101]