JOURNAL OF ENDODONTICS Copyright 2003 by The American Association of Endodontists

Printed in U.S.A. VOL. 29, NO. 8, AUGUST 2003

Identification of Resected Root-end Dentinal Cracks: A Comparative Study of Visual Magnification

C. Cornelious Slaton, DMD, Robert J. Loushine, DDS, R. Norman Weller, DMD, MS, M. Harry Parker, DDS, MS, W. Frank Kimbrough, DDS, MS, and David H. Pashley, DMD, PhD

The purpose of this in vitro study was to evaluate and compare the effectiveness of visual enhancements as aids in identifying artificially created dentinal cracks in resected root ends. Fifty human maxillary central incisors were decoronated, and the root canals were instrumented to ISO size 50 at the working length. The apical 3 mm of the roots were resected, and cracks were artificially created in the apical dentin with an average load of 5.6 kg using a cylindrical wedge in a miniature drill press. A video microscope at 65 magnification was used to observe the cracks as they developed. Four independent examiners evaluated the root specimens using unaided/corrected vision (group 1), loupes at 3.3 magnification (group 2), a surgical operating microscope at 10 magnification (group 3), and the Orascope at 35 magnification (group 4). The examiners proficiency at correctly identifying root ends with and without cracks was evaluated. The data were compared to the predetermined standard (27 cracked, 23 not cracked) with a one-tailed Fishers exact test ( 0.05). Statistically, the Orascope (p 0.02) was significantly superior, whereas using unaided/corrected vision (p 0.99), loupes (p 0.88), or the microscope (p 0.14) was not significantly better than guessing. The accuracy of correct identification for unaided/ corrected vision, loupes, the microscope, and the Orascope was 39%, 45%, 53%, and 58%, respectively. A two-way analysis of variance of the accuracy of crack identification showed a significant difference among the four visualization techniques (p 0.0007) and also among the four evaluators (p 0.006).

Endodontic surgery may be indicated when it is not feasible to disassemble complex fixed prosthetic restorations, when calcification of the canal system precludes treatment by nonsurgical root canal therapy, or when nonsurgical root canal therapy fails to
519

alleviate the patients symptoms. Root-end resection and ultrasonic preparation are commonly performed during apical surgery. A possible sequelae of these procedures may be the development of cracks in the apical dentin. Saunders et al. (1) noted the presence of dentin cracks in resected root-ends prepared with ultrasonics while evaluating microleakage of root-end fillings. Layton et al. (2) found cracks in the apical dentin after root-end resection and after root-end preparation. A clinicians ability to diagnose fractures in the dentin of resected roots depends on the ability to identify their presence. Direct and indirect visualization of the root end is the primary means by which the diagnosis can be made. Factors such as the level of operator experience, the time of day an observation is made, the level of operator fatigue, and distractions may influence the clinicians visual interpretation. Unaided vision is inadequate, in most instances, to properly evaluate a resected root end for cracks, completeness of resection, and anatomical variations. Bellizzi and Loushine (3) advocated enhanced illumination and magnification as adjuncts for posterior surgery, as did Rubinstein and Kim (4). Carr (5) chronicled the introduction of aids to enhance vision in endodontics and found that they paralleled those of many medical specialties, particularly neurosurgery, ophthalmology, and microvascular surgery. The combination of improved lighting and magnification has been provided by several means: fiberoptic headlamps and loupes, the surgical operating microscope (SOM), and most recently the Orascope, which was introduced by Bahcall and Barss (6). The importance of visual enhancement in the delivery of endodontic care has been recognized and exemplified by the action of the American Dental Association (ADA). As of January 1998, the ADA requires all accredited postgraduate endodontic programs to provide instruction and clinical training in performing nonsurgical and surgical endodontic procedures using microscopy (7). The purpose of this in vitro study was to evaluate and compare the effectiveness of visual enhancements as aids in correctly identifying artificially created dentinal cracks in resected root ends using unaided/corrected vision, loupes at 3.3 magnification, a SOM at 10 magnification, and the Orascope at 35 magnification. The null hypothesis was that crack identification is not influenced by magnification.

520

Slaton et al.

Journal of Endodontics

MATERIALS AND METHODS Fifty, extracted, human, permanent, maxillary central incisors were selected for this in vitro study. All specimens had mature, intact apices and no previous endodontic therapy. They were kept in 100% humidity until used. The teeth were decoronated at the CEJ and an ISO size 10 Flexofile (Dentsply/Maillefer, Tulsa, OK) was inserted to the apical foramen. The working length was defined as 0.5 mm shorter than this length. Gates Glidden drills (Dentsply/Maillefer) were used to prepare the root canal to a length 5 mm short of the established working length. The remaining 5 mm was instrumented with Flexofiles, in a serial fashion, to ISO size 50 at the working length. The apical 3 mm of each root were resected perpendicular to the long axis with a multipurpose bur (Dentsply/ Maillefer) in a high-speed handpiece using water spray. The resected root ends were evaluated at 65 magnification with a video microscope (Micro Enterprises, Inc., Norcross, GA) to ensure that no cracks were present at the conclusion of the root-end resection. Dentinal cracks were created in 27 experimental teeth by the following method: a 15-mm long cylindrical wedge with a 1.12-mm maximum diameter was placed in the chuck of a drill press, leaving 10 mm and the tip exposed. A top-loading balance (15-kg capacity) was placed on the platform of a miniature drill press to record the load necessary to create the cracks. The tooth was oriented vertically with the apex up, and then the wedge was placed into the canal space. The video microscope at 65 magnification was focused on the tooth apex. Load was slowly applied to the tooth until dentinal cracks were observed. When cracks were identified, the load was removed. The load required to produce the cracks was recorded for each tooth. The specimens were prepared for clinically relevant viewing by mounting each root in the center of a 4-cm diameter sheet of rubber dam material in which a 0.5-mm hole had been punched. The rubber dam with the tooth was placed over the opening of a plastic canister, 3 cm in diameter and 5 cm in height. With the rubber dam stretched tightly and the tooth suspended in the center of the canister, the rubber dam was secured by an elastic band. The root was adjusted with digital pressure until approximately 0.5 mm of the apex was extending above the rubber dam. Five plastic trays were used, each designed with 10 numbered wells, to accommodate 10 canisters. The trays served as a platform on which the observations were made. The trays containing the canisters and specimens were placed in a humidor until the viewing sessions. The teeth were randomized before each viewing method and placed into the numbered wells within the trays. All 50 specimens, 27 of which contained cracks, were evaluated with each viewing method. A minimum of 24 h elapsed between viewing sessions. Four independent examiners viewed the teeth using unaided/corrected vision (group 1), loupes (Orascoptic Research, Madison, WI) at 3.3 magnification (group 2), a SOM (Global, St. Louis, MO) at 10 magnification (group 3), and the Orascope (Sitca, Ann Arbor, MI) at 35 magnification (group 4).

The trays containing the specimens were placed on a bench top and viewed from a seated position. The examiner was not permitted to pick up the tray or change its orientation. A dental operatory light was used for illumination of the specimens in groups 1 and 2. Viewing with the Orascope was performed with the 1.8-mm diameter probe mounted on a test tube stand 5 mm above the root surface to standardize the level of magnification for each examiner. Each specimen was assessed and a diagnosis of cracked, not cracked, or unsure was made and recorded. In recording the data for statistical analysis, the answer was deemed correct or diagnostic if the tooth was correctly identified as having or not having a crack. If the examiner selected yes or no incorrectly or if they chose unsure, the response was deemed incorrect or not diagnostic. To determine the effectiveness of each visualization technique, the data were collected and compared to the predetermined standard (27 cracked, 23 not cracked) with a one-tailed Fishers 0.05) using the NCSS statistical software package exact test ( (Kaysville, UT). The correct and incorrect responses were then utilized to calculate the sensitivity, specificity, and accuracy as described by Brunette (8). Sensitivity is the proportion of the roots containing cracks that were correctly diagnosed as having cracks (9). Specificity is the proportion of the roots without cracks that were correctly diagnosed as not having cracks. Accuracy is the proportion of the diagnoses that agreed with the known root condition (9). To compare the visualization techniques, a two-way analysis of variance was performed.

RESULTS From the 200 responses for each visualization technique, the number of unsure responses was 20 for unaided/corrected vision, 25 for loupes, 20 for the microscope, and 6 for the Orascope. These were treated as incorrect responses. A one-tailed Fishers exact test revealed a significant difference in crack identification with the use of the Orascope (group 4, p 0.02) but not for unaided/corrected vision (group 1, p 0.99), loupes (group 2, p 0.88), or the operating microscope (group 3, p 0.14). The reliability of these methods of crack diagnosis is represented by their sensitivity (correct identification of cracks), specificity (correct identification of the absence of cracks), and accuracy (proportion of diagnoses that agreed with the known root condition). The values were calculated for the four groups and are summarized in Table 1. The Orascope demonstrated the greatest sensitivity at 53%, followed by the microscope, loupes, and unaided/corrected vision at 35%, 33%, and 19%, respectively. The microscope demonstrated the greatest specificity at 73%. The specificity of loupes, unaided/corrected vision, and the Orascope was 60%, 61%, and 63%, respectively (Table 1). The greatest accuracy was demonstrated by using the Orascope at 58% followed by the microscope at 53%, loupes at 45%, and unaided/corrected vision at 39% (Table 1).

TABLE 1. Results of root crack identification Unaided/Corrected (%) Sensitivity Specificity Accuracy 19 61 39 Loupes (%) 33 60 45 Microscope (%) 35 73 53 Orascope (%) 53 63 58
proportion of

Sensitivity proportion of roots with cracks that were correctly diagnosed; Specificity diagnoses that agreed with known tooth condition.

proportion of roots without cracks that were correctly diagnosed; Accuracy

Vol. 29, No. 8, August 2003

Comparative Study of Visual Magnification

521

FIG 1. The trend toward greater accuracy is seen as the level of magnification increases.

Scores were recoded as 1 for a diagnostic or correct answers and 0 for nondiagnostic, incorrect or unsure answers. The average of these recoded scores represents the percentage of correct answers or accuracy. A two-way analysis of variance of accuracy showed no higher order interaction between the two categories: examiner and visualization technique (p 0.85). Analysis of variance revealed that a significant difference was found among accuracy of the four visualization techniques (p 0.0007, power 0.86) and also among the four evaluators (p 0.006, power 0.95). The trend in accuracy for each examiner to do better with increasing magnification is seen in Fig. 1. For all examiners, the accuracy was best with the orascopic magnification. Multiple comparison tests were performed using Fishers LSD test to identify which visualization techniques gave significantly superior accuracy in crack identification. There was no significant difference between unaided/corrected vision and the use of loupes (p 0.05). There was a significant difference between unaided/ corrected vision or loupes and the Orascope (p 0.05). Although there was a significant difference in accuracy between unaided/ corrected vision and the use of the microscope (p 0.05), there was no significant difference between the use of loupes and the use of the microscope (p 0.05). DISCUSSION Observations were made with the root apex positioned in a specimen holder constructed from a plastic canister and rubber dam material. This method was designed to isolate the root apex while leaving only 0.5 mm of the lateral root surface exposed. Concealment of the lateral root surface was necessary because propagation of cracks along this surface made identification of complete dentinal cracks very obvious. Clinically, it is uncommon to have much apical root exposed for visualization of cracks on the lateral surface of the root. Although the dentinal cracks in our study were created via a different mechanism than those observed by Layton et al. (2), the types of cracks that we observed were consistent with their findings. Three types of cracks were seen: canal cracks (both complete and incomplete), intradental cracks, and cemental cracks. The average load required to produce the cracks in the experimental specimens was 5.6 kg. An interesting finding was that as the load was being applied, the dentin around the area of strain became opaque or frosted in appearance before a crack developed. We

speculate that these were caused by the formation of many microscopic cracks that have not yet formed a macro-crack. This finding should heighten the clinicians suspicion that a crack may be present. A low sensitivity of 19% for group 1 was the result of the high number of false negatives, with 87 false negatives out of 108 diagnoses. The specificity was 61% for that group, with 36 false positives out of 92 diagnoses. The accuracy of group 1 was the lowest of the four groups, at 39% (Table 1). This result suggests that the examiners could not see the cracks and that their responses were no more accurate than random guessing. Even when using the 35 magnification of the Orascope, correct diagnoses were made only 58% of the time. Two factors, specifically light reflection and irregularities in the resected root end, made identification of dentinal cracks difficult and may have resulted in greater observer variability. Irregularities in the resected root surface caused light, which was directed perpendicular to the resected root end, to be reflected in a manner that may have obscured the presence of some cracks. In other specimens, this reflected light gave the appearance of cracks where none were present. An analysis of each specimen across the spectrum of visual aid enhancements and observers revealed that no tooth was correctly diagnosed by all four observers by any viewing method. Neither dyes nor transillumination were used in this study to enhance visualization of the cracks. A number of investigators have recommended transillumination and dyes, finding them to be important aids in identifying cracks in teeth (10 14). DISCUSSION Our study found a trend of improved accuracy for each examiner with increasing magnification. However, even with 35 magnification, the sensitivity, specificity, and accuracy of identifying dentinal cracks were lower than expected. Future studies are needed with transillumination and dyes, using this same model, to determine if their use can increase the sensitivity, specificity, and accuracy of crack detection.
Dr. Slaton is a postgraduate student, Dr. Loushine is associate professor and program director, Postgraduate Endodontics, Dr. Weller is professor and chairman, Dr. Kimbrough is associate professor, Department of Endodontics; Dr. Parker is associate professor, Department of Oral Rehabilitation; and Dr. Pashley is regents professor, Department of Oral Biology and Maxillofacial Pathology, School of Dentistry, Medical College of Georgia, Augusta, GA. Address requests for reprints to Robert J. Loushine, DDS, Associate Professor and Program Director, Postgraduate Endodontics, Department of Endodontics, School of Dentistry, Medical College of Georgia, Augusta, Georgia 30912-1244.

References 1. Saunders WP, Saunders EM, Gutmann JL. Ultrasonic root-end preparation. Part 2. Microleakage of EBA root-end fillings Int Endod J 1994;27: 3259. 2. Layton CA, Marshall JG, Morgan LA, Baumgartner JC. Evaluation of cracks associated with ultrasonic root-end preparation. J Endodon 1996;22: 157 60. 3. Bellizzi R, Loushine R. Adjuncts to posterior endodontic surgery. J Endodon 1990;16:604 6. 4. Rubinstein RA, Kim S. Short-term observation of the results of endodontic surgery with the use of a surgical operation microscope and SuperEBA as root-end filling material. J Endodon 1999;25:43 8. 5. Carr GB. Microscopes in endodontics. J Calif Dent Assoc 1992;20:55 61. 6. Bahcall JK, Barss JT. Fiberoptic endoscope usage for intracanal visualization. J Endodon 2001;27:128 9. 7. Commission on Dental Accreditation. Accreditation standards for ad-

522

Slaton et al.

Journal of Endodontics
11. Pitts DL, Natkin E. Diagnosis and treatment of vertical root fractures. J Endodon 1983;9:338 46. 12. Cameron CE. Cracked-tooth syndrome. J Am Dent Assoc 1964;68: 40511. 13. Liewehr FR. An inexpensive device for transillumination. J Endodon 2001;27:130 1. 14. Viener AE. Fractured teeth: a cause of odontalgia. Oral Surg Oral Med Oral Pathol 1965;20:594 5.

vanced specialty education programs in endodontics. Chicago: ADA, 1998: 15. 8. Brunette DM. Critical thinking understanding and evaluating dental research. Chicago: Quintessence Publishing Co. Inc., 1996:99 101. 9. Pagano M, Gauvreau K. Principles of biostatistics. Belmont: Duxbury Press, 1993:124. 10. Cameron CE. The cracked tooth syndrome: additional findings. J Am Dent Assoc 1976;93:9715.