- Email: [email protected]
Sensitivity of Preoperative Magnetic Resonance Imaging and Magnetic Resonance Arthrography in Detection of Panlabral Tears of the Glenohumeral Joint
Sensitivity of Preoperative Magnetic Resonance Imaging and Magnetic Resonance Arthrography in Detection of Panlabral Tears of the Glenohumeral Joint Eric T. Ricchetti, M.D., Michael C. Ciccotti, B.A., Michael G. Ciccotti, M.D., Gerald R. Williams Jr., M.D., and Mark D. Lazarus, M.D.
Purpose: The purpose of this study was to evaluate the sensitivity of preoperative magnetic resonance imaging (MRI) in detecting combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum. Methods: Preoperative MRI findings were available in 46 cases of arthroscopic repair of combined tears of the anterior, posterior, and superior labrum (mean age, 31 years), including 23 noncontrast MRI studies and 24 magnetic resonance (MR) arthrography studies (1 patient had both). MRI findings were compared with the status of the labrum at the time of arthroscopy, and the sensitivity of MRI was determined. Results: MRI showed evidence of combined lesions of the anterior, posterior, and superior labrum in only 10 of 47 studies (21.3%). Only 2 of 23 (8.7%) tears were detected by noncontrast MRI, compared with 8 of 24 (33.3%) by MR arthrography (P ¼ .07). Non-contrast MRI showed evidence of labral pathology in more than one direction in 10 of 23 studies (43.5%), compared with 20 of 24 MR arthrography studies (83.3%) (P ¼ .006). Noncontrast MRI showed no evidence of a labral tear in 3 of 23 patients (13.0%), whereas no MR arthrogram was completely negative for a labral tear (0%) (P ¼ .11). Conclusions: Combined tears of the anterior, posterior, and superior glenoid labrum are infrequent injuries that are typically not completely defined by either noncontrast MRI or MR arthrography. Level of Evidence: Level III, diagnostic study.
C
ombined lesions of the glenoid labrum involving an extensive continuous tearing inclusive of the anterior, posterior, and superior labrum have been infrequently reported in the literature.1-4 A circumferential labral lesion of the glenoid was first described by Powell et al., and classified as a type IX SLAP (superior labrum, anterior and posterior) tear.3 Subsequently, Lo and Burkart published a series of 7 cases of arthroscopic repair of combined lesions of the anterior, posterior, and superior labrum, an injury they termed a triple labral lesion.2 These injuries were reported to represent 2.4% of all labral tears in their series.2 Tokish et al. have the only reported series of circumferential panlabral
From the Department of Orthopaedic Surgery, The Rothman Institute at Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A. The authors report that they have no conflicts of interest in the authorship and publication of this article. Received January 18, 2012; accepted October 8, 2012. Address correspondence to Eric T. Ricchetti, M.D., Department of Orthopaedic Surgery, Cleveland Clinic Main Campus, Mail Code A40, 9500 Euclid Avenue, Cleveland, OH 44195, U.S.A.. E-mail: [email protected] Ó 2013 by the Arthroscopy Association of North America 0749-8063/1231/$36.00 http://dx.doi.org/10.1016/j.arthro.2012.10.005
274
lesions. They evaluated the outcomes of arthroscopic repair in 41 shoulders in a military population.4 Although significant improvements in functional scores were reported after repair, the authors noted even in the presence of advanced imaging techniques and equipment that preoperatively detecting these lesions and distinguishing them from unidirectional labral tears can be difficult. In this series, the interpretation of the preoperative magnetic resonance imaging (MRI) study was correct in only 59% of the shoulders, with a much lower accuracy rate for noncontrast MRI compared with magnetic resonance (MR) arthrography.4 Numerous studies in the literature have shown MRI, particularly MR arthrography, to be the imaging study of choice in diagnosing labral pathology, including anterior and posterior Bankart lesions, and SLAP tears.5-20 Sensitivities and specificities have been reported at greater than 90% in certain studies in detecting unidirectional tears of the glenoid labrum on noncontrast MRI, and even higher values have been reported with MR arthrography.5-20 Little has been written, however, regarding the use and accuracy of MRI in identifying labral tears involving more than 1 location of the glenoid.1,4 The purpose of this study was to evaluate the sensitivity of preoperative MRI, including both noncontrast MRI and
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 29, No 2 (February), 2013: pp 274-279
275
MRI SENSITIVITY IN PANLABRAL TEARS
MR arthrography, in detecting combined lesions of the anterior, posterior, and superior glenoid labrum. Our hypothesis was that preoperative MRI would have poor sensitivity in detecting these tears, as confirmed by arthroscopy, regardless of the presence or absence of arthrography.
Methods Institutional review board approval was obtained for this study (Thomas Jefferson University IRB Control No. 08D.360). No external source of funding was used for any aspect of this study. We retrospectively identified all cases of arthroscopic labral repair at our institution between January 2005 and December 2009. A total of 1,624 cases were identified by using the Current Procedural Terminology (CPT) codes 29806 (arthroscopy, shoulder, surgical capsulorrhaphy) and/or 29807 (arthroscopy, shoulder, repair of superior labrum). The operative reports on these patients were reviewed to identify all cases that involved tears of the anterior, posterior, and superior labrum (continuous or not) and underwent arthroscopic repair of lesions in all 3 regions, defined as a panlabral repair. All cases had a minimum of one suture anchor placed in each of the anterior, posterior, and superior labrum; and all tears were found to involve at least 270 of the glenoid labrum. Cases involving placement of suture anchors in a region of the labrum without an identified labral tear (e.g., capsulorrhaphy for instability without evidence of labral tearing, or labral fraying without labral detachment) were excluded. A total of 58 (3.6%) patients met these inclusion criteria. The medical records of these cases were then retrospectively reviewed to identify all patients who had a copy of the preoperative MRI report of the surgical shoulder on file. A total of 46 (3.0%) patients met this criterion. There were 39 (85%) men and 7 (15%) women in this group, with a mean age at the time of surgery of 31 years (range, 15 to 55). Arthroscopic repair was performed by 1 of 4 surgeons fellowship trained in shoulder reconstruction, in either the beach chair (n ¼ 31) or lateral decubitus (n ¼ 15) position. A mean of 8.3 anchors (range, 5 to 13) were used. On the basis of the reports, 22 (48%) patients underwent noncontrast MRI, 23 (50%) patients underwent contrast-enhanced MR arthrography, and one (2%) patient underwent both. Thirty different imaging locations were used for the 47 MRI studies. Magnet strength was listed in 14 of 47 radiology reports, with 13 studies employing a 1.5-T magnet and one using a 0.7-T magnet. Use of the abductioneexternal rotation (ABER) view was documented in 4 of the MR arthrograms. No other specific extremity positions were noted in the remainder of the reports. On the basis of the MRI reports, the studies were read by 39
different radiologists, with the radiologist not identifiable on one report. Twelve studies were performed at an academic facility, where it was possible to confirm that the study was read by a radiologist with musculoskeletal specialization. In addition to the MRI reports, the medical record was reviewed to determine the treating surgeons’ preoperative interpretation of the MRI studies. The findings on the preoperative MRI reports and the treating surgeons’ preoperative MRI interpretation were compared to determine if any discrepancies existed. The presence of a labral tear was based on the interpretation and comments of the reviewing radiologist and/or treating surgeon, rather than on specific criteria on the imaging sequences, because of the variability in imaging sequences that were obtained. Tears were classified into 3 regions (anterior, posterior, superior) to parallel the classification of repairs described at the time of arthroscopy and also because of the lack of consistency in reporting a separate inferior region. In 12 instances, the treating surgeon did not additionally comment on the MRI findings beyond what was noted in the radiology report. In 6 cases, a discrepancy existed between the radiology report and the treating surgeons’ interpretation. In one instance, the treating surgeon noted combined tears of the anterior, posterior, and superior labrum on MRI, whereas the radiology report described only anterior and superior tears. In another instance, the radiology report described combined tears of the anterior, posterior, and superior labrum on MRI, whereas the treating surgeon described only an anterior tear. In 4 other cases, the radiology report and treating surgeons’ interpretation did not completely agree, but neither described combined tears of the anterior, posterior, and superior labrum. For each case in which a discrepancy existed, the interpretation (radiologist or treating surgeon) with the largest tear involvement was used for reporting purposes for the results of this study. Labral findings at the time of arthroscopy were used as the gold standard in comparison to the results of preoperative MRI (Fig 1). The sensitivity of MRI in detecting combined lesions of the anterior, posterior, and superior labrum was determined, including comparison of noncontrast MRI with MR arthrography. Statistical analysis was performed using the Fisher exact test to compare sensitivity differences between noncontrast MRI and MR arthrography. The level of significance was set at P .05 for all analyses.
Results Overall, the MRI report stated there was evidence of a combined lesion of the anterior, posterior, and superior labrum in 10 of 47 studies (21.3%). Only 2 of 23 (8.7%) tears were detected by noncontrast MRI, whereas 8 of 24 (33.3%) were detected by MR
276
E. T. RICCHETTI ET AL.
Fig 1. (A-C) Coronal (A, B) and axial (C) images of a magnetic resonance arthrogram showing evidence of anterior and posterior labral tearing without definitive extension into the superior labrum. (D-I) At the time of arthroscopy, the patient was found to have tears of the anterior (D), posterior (E), and superior (F) labrum, and underwent arthroscopic repair in all 3 regions (G ¼ anterior repair, H ¼ posterior repair, I ¼ superior repair). Letters in images D to F identify the anterior (A), posterior (P), and superior (S) labrum. Arrows in images G to I identify representative sites of suture anchor fixation.
arthrography (P ¼ .07) (Table 1). Noncontrast MRI showed evidence of labral pathology in more than one direction in 10 of 23 studies (43.5%), compared with 20 of 24 MR arthrography studies (83.3%) (P ¼ .006). Noncontrast MRI showed no evidence of a labral tear in 3 of 23 studies (13.0%), whereas no MR arthrogram was completely negative for a labral tear (0%) (P ¼ .11) (Table 1).
Discussion Combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum have been infrequently reported in the literature.1-4 Detecting these lesions preoperatively and distinguishing them from unidirectional labral tears can be difficult, even in the presence of advanced imaging. The purpose of the current study, therefore, was to evaluate the sensitivity of preoperative MRI done at a wide range
of imaging facilities, including both noncontrast MRI and MR arthrography, in detecting combined lesions of the anterior, posterior, and superior glenoid labrum. We found that these tears were frequently missed on MRI, regardless of the presence or absence of arthrography, supporting our hypothesis. The sensitivity of MRI overall in detecting combined tears of the anterior, posterior, and superior glenoid labrum was approximately 20% in the current report, with MR arthrography showing a trend toward a higher, but still low, sensitivity in comparison to noncontrast MRI (33.3% v 8.7%). Although detecting the full degree of labral pathology was poor regardless of the use of intra-articular contrast, MR arthrography did reveal some degree of labral pathology in all instances and showed labral pathology in more than one direction in most of the cases (83.3%). Noncontrast MRI did not show any labral tears in 3 cases, and was
277
MRI SENSITIVITY IN PANLABRAL TEARS
Table 1. Detection of Combined Lesions of the Anterior, Posterior, and Superior Glenoid Labrum by Preoperative Magnetic Resonance Imaging Condition
MRI Overall (n ¼ 47)
Noncontrast MRI (n ¼ 23)
MR Arthrography (n ¼ 24)
P value
Detection of tears of anterior, posterior, and superior labrum Detection of tears in more than one location Detection of no labral pathology
10 (21.3%) 30 (63.8%) 3 (6.4%)
2 (8.7%) 10 (43.5%) 3 (13.0%)
8 (33.3%) 20 (83.3%) 0 (0%)
.07 .006 .11
NOTE. Frequency distributions (%) are reported and compared using the Fisher exact test. P values represent comparisons between noncontrast MRI and MR arthrography for the conditions listed.
significantly worse than MR arthography in detecting tears in more than one direction. Tokish et al. recently reported on outcomes of arthroscopic repair of circumferential lesions of the glenoid labrum in 41 shoulders from a military population.4 Findings on preoperative MRI were included. All 41 shoulders underwent preoperative MRI, with 30 undergoing MR arthrography. The radiologist correctly read a panlabral lesion on the MR image in 24 of 41 (59%) shoulders, with better sensitivity on the MR arthrogram (23 of 30, 77%) than on the noncontrast MR image (1 of 11, 9%).4 Although the findings on noncontrast MRI are comparable to the results in our study, the sensitivity of MR arthrography was substantially higher than the value we reported. This discrepancy may be caused by several factors. A wide range of imaging locations were used in the current study to obtain preoperative MR images, resulting in variability in equipment, imaging protocol, and number and training of radiologists used. In contrast, Tokish et al. examined a captured military population, with all surgical procedures performed at one of 2 military sites and all preoperative imaging likely limited to a small number of sites. This potentially led to a more consistent imaging protocol for the MRI performed in that study, with use of more uniform equipment, and interpretation of the scans by a fewer number of more uniformly trained radiologists. These factors could have potentially led to improved imaging and detection of labral pathology, resulting in improved sensitivity. In addition, a heightened awareness for these multidirectional labral injuries may have been present because of the young, highly active military population that was being examined. An older, less active civilian population was examined in the current study. This may represent a more commonly treated group of patients, however. In a radiological study in 2005, Lindauer et al. reported on the MRI appearance of labral tears involving greater than 180 of the glenoid labrum, noting that these lesions had not specifically been described in the orthopaedic or radiological literature.1 Preoperative MRI studies were available for 19 of 31 patients (10 of 19 MR arthrograms) with arthroscopically confirmed labral tears involving 180 to 360 of the glenoid labrum (mean arc, 220 ). All MRI studies were performed at the authors’ institution and were reviewed in
a blinded manner by 2 board-certified musculoskeletal radiologists: first at the time of preoperative evaluation and then retrospectively after the extent of labral pathology was confirmed arthroscopically. Correct preoperative MRI diagnosis was again noted to be poor in this study, with the complete labral injury detected in only 1 of 19 patients and no labral pathology noted in 4 of 19 patients.1 At the second-look retrospective review, all 19 patients who had MRI were diagnosed with some degree of labral pathology, but the complete extent of labral injury was still detected in only 1 of 19 patients. Labral pathology in more than one quadrant of the glenoid was noted on only 6 MR images preoperatively, compared with 17 after second-look retrospective review.1 Differences in evaluation of noncontrast MR images and MR arthrograms were not reported in this study. A strength of the current study is the generalizability of the findings in comparison to prior reports. The previous studies that have reported on the use of preoperative MRI in identifying labral tears involving more than one location of the glenoid have used more narrow patient populations, with the imaging studies performed at a limited number of centers or performed on a more select group of patients.1,4 The current report evaluated a more commonly treated group of patients from a civilian population, who used a variety of imaging centers for their MRI studies, with the studies interpreted by a number of different radiologists. We, therefore, believe it is more representative of a true-tolife scenario for practicing orthopaedic surgeons. Limitations There were several limitations of this study. As this was a retrospective chart review, there was no uniform imaging protocol for each of the MRI studies. Patients were free to undergo MRI at any location, and 30 different imaging locations were used for the 47 preoperative MRI studies. Therefore, variability existed in the type of equipment used, the imaging technique employed, and the radiologists’ interpretation of the study. For example, magnet strength has been shown to impact study resolution and the degree of anatomic detail on MRI, and can be variable across different imaging machines.21,22 In the current study, the magnet strength of the MRI scanner was listed in only
278
E. T. RICCHETTI ET AL.
14 of 47 radiology reports, with 13 noting utilization of a 1.5-T magnet and one using a 0.7-T magnet. In addition to the presence or absence of intra-articular contrast, the imaging technique can vary considerably in other ways, such as the image sequencing used and the position of the extremity while imaged. Certain MR sequences have been shown to be more reliable than others in detecting labral tears, and the sequencing employed for each MRI study in this study was not consistently included in the MRI reports.20,21,23 Certain extremity positions have been shown to more reliably exhibit pathology on MRI, such as the ABER view in detecting anteroinferior labral tears of the shoulder.24-28 In the current study, use of the ABER view was documented in only 4 MR arthrograms, and no other specific extremity positions were noted in the remainder of the MR images. Finally, the training level and skill of the radiologists who documented the MRI reports were not controlled for in this study. The MR images were read by 39 different radiologists, and it was possible to confirm that a study was read by a radiologist with musculoskeletal specialization in only 12 cases. Variability likely existed, therefore, in the training and experience of the radiologists interpreting these studies. Although hospitals in major academic centers typically employ radiologists with fellowship training in musculoskeletal radiology to read shoulder MRI studies, community hospitals or community-based imaging centers may or may not employ radiologists with such advanced training. Another limitation of the current study was the inability to perform a more detailed statistical analysis of the diagnostic utility of MRI in detecting panlabral tears. Although the number of true positive and false negative MRI findings could be determined to calculate sensitivity, the lack of knowledge of true negatives and false positives left us unable to calculate additional diagnostic statistics, such as specificity, accuracy, and positive and negative predictive values. Determination of all diagnostic test characteristics in a retrospective manner would require analysis of thousands of additional MRI reports because of the infrequency of panlabral lesions, which was beyond the scope of this study.
Conclusions Combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum are infrequent injuries that are typically not completely defined by either noncontrast MRI or MR arthrography.
References 1. Lindauer KR, Major NM, Rougier-Chapman DP, Helms CA. MR imaging appearance of 180-360 degree labral tears of the shoulder. Skeletal Radiol 2005;34:74-79. 2. Lo IK, Burkhart SS. Triple labral lesions: Pathology and surgical repair techniquedReport of seven cases. Arthroscopy 2005;21:186-193.
3. Powell SE, Nord KD, Ryu RKN. The diagnosis, classification, and treatment of SLAP lesions. Oper Tech Sports Med 2004;12:99-110. 4. Tokish JM, McBratney CM, Solomon DJ, Leclere L, Dewing CB, Provencher MT. Arthroscopic repair of circumferential lesions of the glenoid labrum. J Bone Joint Surg Am 2009;91:2795-2802. 5. Chandnani VP, Yeager TD, DeBerardino T, et al. Glenoid labral tears: Prospective evaluation with MRI imaging, MR arthrography, and CT arthrography. AJR Am J Roentgenol 1993;161:1229-1235. 6. Flannigan B, Kursunoglu-Brahme S, Snyder S, Karzel R, Del Pizzo W, Resnick D. MR arthrography of the shoulder: Comparison with conventional MR imaging. AJR Am J Roentgenol 1990;155:829-832. 7. Gross ML, Seeger LL, Smith JB, Mandelbaum BR, Finerman GA. Magnetic resonance imaging of the glenoid labrum. Am J Sports Med 1990;18:229-234. 8. Iannotti JP, Zlatkin MB, Esterhai JL, Kressel HY, Dalinka MK, Spindler KP. Magnetic resonance imaging of the shoulder: Sensitivity, specificity, and predictive value. J Bone Joint Surg Am 1991;73:17-29. 9. Jahnke AH Jr, Petersen SA, Neumann C, Steinbach L, Morgan F. A prospective comparison of computerized arthrotomography and magnetic resonance imaging of the glenohumeral joint. Am J Sports Med 1992;20:695-700. 10. Palmer WE, Caslowitz PL. Anterior shoulder instability: Diagnostic criteria determined from prospective analysis of 121 MR arthrograms. Radiology 1995;197:819-825. 11. Sanders TG, Tirman PF, Linares R, Feller JF, Richardson R. The glenolabral articular disruption lesion: MR arthrography with arthroscopic correlation. AJR Am J Roentgenol 1999;172:171-175. 12. Seeger LL, Ruszkowski JT, Bassett LW, Kay SP, Kahmann RD, Ellman H. MR imaging of the normal shoulder: Anatomic correlation. AJR Am J Roentgenol 1987;148:83-91. 13. Sher JS, Iannotti JP, Williams GR, et al. The effect of shoulder magnetic resonance imaging on clinical decision making. J Shoulder Elbow Surg 1998;7:205-209. 14. Legan JM, Burkhard TK, Goff WB II, et al. Tears of the glenoid labrum: MR imaging of 88 arthroscopically confirmed cases. Radiology 1991;179:241-246. 15. Palmer WE, Brown JH, Rosenthal DI. Labraleligamentous complex of the shoulder: Evaluation with MR arthrography. Radiology 1994;190:645-651. 16. Bencardino JT, Beltran J, Rosenberg ZS, et al. Superior labrum anterioreposterior lesions: Diagnosis with MR arthrography of the shoulder. Radiology 2000;214:267-271. 17. Karzel RP, Snyder SJ. Magnetic resonance arthrography of the shoulder: A new technique of shoulder imaging. Clin Sports Med 1993;12:123-136. 18. Tirman PF, Applegate GR, Flannigan BD, Stauffer AE, Crues JV III. Magnetic resonance arthrography of the shoulder. Magn Reson Imaging Clin North Am 1993;1:125-142. 19. Magee T. 3-T MRI of the shoulder: Is MR arthrography necessary? AJR Am J Roentgenol 2009;192:86-92. 20. Gusmer PB, Potter HG, Schatz JA, et al. Labral injuries: Accuracy of detection with unenhanced MR imaging of the shoulder. Radiology 1996;200:519-524. 21. Magee T. Can isotropic fast gradient echo imaging be substituted for conventional T1 weighted sequences in
MRI SENSITIVITY IN PANLABRAL TEARS
22.
23. 24.
25.
shoulder MR arthrography at 3 Tesla? J Magn Reson Imaging 2007;26:118-122. Magee TH, Williams D. Sensitivity and specificity in detection of labral tears with 3.0-T MRI of the shoulder. AJR Am J Roentgenol 2006;187:1448-1452. Chaipat L, Palmer WE. Shoulder magnetic resonance imaging. Clin Sports Med 2006;25:371-386. Chiavaras MM, Harish S, Burr J. MR arthrographic assessment of suspected posteroinferior labral lesions using flexion, adduction, and internal rotation positioning of the arm: Preliminary experience. Skeletal Radiol 2010;39:481-488. Cvitanic O, Tirman PF, Feller JF, Bost FW, Minter J, Carroll KW. Using abduction and external rotation of the
279
shoulder to increase the sensitivity of MR arthrography in revealing tears of the anterior glenoid labrum. AJR Am J Roentgenol 1997;169:837-844. 26. Harish S, Nagar A, Moro J, Pugh D, Rebello R, O’Neill J. Imaging findings in posterior instability of the shoulder. Skeletal Radiol 2008;37:693-707. 27. Song HT, Huh YM, Kim S, et al. Anterioreinferior labral lesions of recurrent shoulder dislocation evaluated by MR arthrography in an adduction internal rotation (ADIR) position. J Magn Reson Imaging 2006;23: 29-35. 28. Kwak SM, Brown RR, Trudell D, Resnick D. Glenohumeral joint: Comparison of shoulder positions at MR arthrography. Radiology 1998;208:375-380.
Purpose: The purpose of this study was to evaluate the sensitivity of preoperative magnetic resonance imaging (MRI) in detecting combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum. Methods: Preoperative MRI findings were available in 46 cases of arthroscopic repair of combined tears of the anterior, posterior, and superior labrum (mean age, 31 years), including 23 noncontrast MRI studies and 24 magnetic resonance (MR) arthrography studies (1 patient had both). MRI findings were compared with the status of the labrum at the time of arthroscopy, and the sensitivity of MRI was determined. Results: MRI showed evidence of combined lesions of the anterior, posterior, and superior labrum in only 10 of 47 studies (21.3%). Only 2 of 23 (8.7%) tears were detected by noncontrast MRI, compared with 8 of 24 (33.3%) by MR arthrography (P ¼ .07). Non-contrast MRI showed evidence of labral pathology in more than one direction in 10 of 23 studies (43.5%), compared with 20 of 24 MR arthrography studies (83.3%) (P ¼ .006). Noncontrast MRI showed no evidence of a labral tear in 3 of 23 patients (13.0%), whereas no MR arthrogram was completely negative for a labral tear (0%) (P ¼ .11). Conclusions: Combined tears of the anterior, posterior, and superior glenoid labrum are infrequent injuries that are typically not completely defined by either noncontrast MRI or MR arthrography. Level of Evidence: Level III, diagnostic study.
C
ombined lesions of the glenoid labrum involving an extensive continuous tearing inclusive of the anterior, posterior, and superior labrum have been infrequently reported in the literature.1-4 A circumferential labral lesion of the glenoid was first described by Powell et al., and classified as a type IX SLAP (superior labrum, anterior and posterior) tear.3 Subsequently, Lo and Burkart published a series of 7 cases of arthroscopic repair of combined lesions of the anterior, posterior, and superior labrum, an injury they termed a triple labral lesion.2 These injuries were reported to represent 2.4% of all labral tears in their series.2 Tokish et al. have the only reported series of circumferential panlabral
From the Department of Orthopaedic Surgery, The Rothman Institute at Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A. The authors report that they have no conflicts of interest in the authorship and publication of this article. Received January 18, 2012; accepted October 8, 2012. Address correspondence to Eric T. Ricchetti, M.D., Department of Orthopaedic Surgery, Cleveland Clinic Main Campus, Mail Code A40, 9500 Euclid Avenue, Cleveland, OH 44195, U.S.A.. E-mail: [email protected] Ó 2013 by the Arthroscopy Association of North America 0749-8063/1231/$36.00 http://dx.doi.org/10.1016/j.arthro.2012.10.005
274
lesions. They evaluated the outcomes of arthroscopic repair in 41 shoulders in a military population.4 Although significant improvements in functional scores were reported after repair, the authors noted even in the presence of advanced imaging techniques and equipment that preoperatively detecting these lesions and distinguishing them from unidirectional labral tears can be difficult. In this series, the interpretation of the preoperative magnetic resonance imaging (MRI) study was correct in only 59% of the shoulders, with a much lower accuracy rate for noncontrast MRI compared with magnetic resonance (MR) arthrography.4 Numerous studies in the literature have shown MRI, particularly MR arthrography, to be the imaging study of choice in diagnosing labral pathology, including anterior and posterior Bankart lesions, and SLAP tears.5-20 Sensitivities and specificities have been reported at greater than 90% in certain studies in detecting unidirectional tears of the glenoid labrum on noncontrast MRI, and even higher values have been reported with MR arthrography.5-20 Little has been written, however, regarding the use and accuracy of MRI in identifying labral tears involving more than 1 location of the glenoid.1,4 The purpose of this study was to evaluate the sensitivity of preoperative MRI, including both noncontrast MRI and
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 29, No 2 (February), 2013: pp 274-279
275
MRI SENSITIVITY IN PANLABRAL TEARS
MR arthrography, in detecting combined lesions of the anterior, posterior, and superior glenoid labrum. Our hypothesis was that preoperative MRI would have poor sensitivity in detecting these tears, as confirmed by arthroscopy, regardless of the presence or absence of arthrography.
Methods Institutional review board approval was obtained for this study (Thomas Jefferson University IRB Control No. 08D.360). No external source of funding was used for any aspect of this study. We retrospectively identified all cases of arthroscopic labral repair at our institution between January 2005 and December 2009. A total of 1,624 cases were identified by using the Current Procedural Terminology (CPT) codes 29806 (arthroscopy, shoulder, surgical capsulorrhaphy) and/or 29807 (arthroscopy, shoulder, repair of superior labrum). The operative reports on these patients were reviewed to identify all cases that involved tears of the anterior, posterior, and superior labrum (continuous or not) and underwent arthroscopic repair of lesions in all 3 regions, defined as a panlabral repair. All cases had a minimum of one suture anchor placed in each of the anterior, posterior, and superior labrum; and all tears were found to involve at least 270 of the glenoid labrum. Cases involving placement of suture anchors in a region of the labrum without an identified labral tear (e.g., capsulorrhaphy for instability without evidence of labral tearing, or labral fraying without labral detachment) were excluded. A total of 58 (3.6%) patients met these inclusion criteria. The medical records of these cases were then retrospectively reviewed to identify all patients who had a copy of the preoperative MRI report of the surgical shoulder on file. A total of 46 (3.0%) patients met this criterion. There were 39 (85%) men and 7 (15%) women in this group, with a mean age at the time of surgery of 31 years (range, 15 to 55). Arthroscopic repair was performed by 1 of 4 surgeons fellowship trained in shoulder reconstruction, in either the beach chair (n ¼ 31) or lateral decubitus (n ¼ 15) position. A mean of 8.3 anchors (range, 5 to 13) were used. On the basis of the reports, 22 (48%) patients underwent noncontrast MRI, 23 (50%) patients underwent contrast-enhanced MR arthrography, and one (2%) patient underwent both. Thirty different imaging locations were used for the 47 MRI studies. Magnet strength was listed in 14 of 47 radiology reports, with 13 studies employing a 1.5-T magnet and one using a 0.7-T magnet. Use of the abductioneexternal rotation (ABER) view was documented in 4 of the MR arthrograms. No other specific extremity positions were noted in the remainder of the reports. On the basis of the MRI reports, the studies were read by 39
different radiologists, with the radiologist not identifiable on one report. Twelve studies were performed at an academic facility, where it was possible to confirm that the study was read by a radiologist with musculoskeletal specialization. In addition to the MRI reports, the medical record was reviewed to determine the treating surgeons’ preoperative interpretation of the MRI studies. The findings on the preoperative MRI reports and the treating surgeons’ preoperative MRI interpretation were compared to determine if any discrepancies existed. The presence of a labral tear was based on the interpretation and comments of the reviewing radiologist and/or treating surgeon, rather than on specific criteria on the imaging sequences, because of the variability in imaging sequences that were obtained. Tears were classified into 3 regions (anterior, posterior, superior) to parallel the classification of repairs described at the time of arthroscopy and also because of the lack of consistency in reporting a separate inferior region. In 12 instances, the treating surgeon did not additionally comment on the MRI findings beyond what was noted in the radiology report. In 6 cases, a discrepancy existed between the radiology report and the treating surgeons’ interpretation. In one instance, the treating surgeon noted combined tears of the anterior, posterior, and superior labrum on MRI, whereas the radiology report described only anterior and superior tears. In another instance, the radiology report described combined tears of the anterior, posterior, and superior labrum on MRI, whereas the treating surgeon described only an anterior tear. In 4 other cases, the radiology report and treating surgeons’ interpretation did not completely agree, but neither described combined tears of the anterior, posterior, and superior labrum. For each case in which a discrepancy existed, the interpretation (radiologist or treating surgeon) with the largest tear involvement was used for reporting purposes for the results of this study. Labral findings at the time of arthroscopy were used as the gold standard in comparison to the results of preoperative MRI (Fig 1). The sensitivity of MRI in detecting combined lesions of the anterior, posterior, and superior labrum was determined, including comparison of noncontrast MRI with MR arthrography. Statistical analysis was performed using the Fisher exact test to compare sensitivity differences between noncontrast MRI and MR arthrography. The level of significance was set at P .05 for all analyses.
Results Overall, the MRI report stated there was evidence of a combined lesion of the anterior, posterior, and superior labrum in 10 of 47 studies (21.3%). Only 2 of 23 (8.7%) tears were detected by noncontrast MRI, whereas 8 of 24 (33.3%) were detected by MR
276
E. T. RICCHETTI ET AL.
Fig 1. (A-C) Coronal (A, B) and axial (C) images of a magnetic resonance arthrogram showing evidence of anterior and posterior labral tearing without definitive extension into the superior labrum. (D-I) At the time of arthroscopy, the patient was found to have tears of the anterior (D), posterior (E), and superior (F) labrum, and underwent arthroscopic repair in all 3 regions (G ¼ anterior repair, H ¼ posterior repair, I ¼ superior repair). Letters in images D to F identify the anterior (A), posterior (P), and superior (S) labrum. Arrows in images G to I identify representative sites of suture anchor fixation.
arthrography (P ¼ .07) (Table 1). Noncontrast MRI showed evidence of labral pathology in more than one direction in 10 of 23 studies (43.5%), compared with 20 of 24 MR arthrography studies (83.3%) (P ¼ .006). Noncontrast MRI showed no evidence of a labral tear in 3 of 23 studies (13.0%), whereas no MR arthrogram was completely negative for a labral tear (0%) (P ¼ .11) (Table 1).
Discussion Combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum have been infrequently reported in the literature.1-4 Detecting these lesions preoperatively and distinguishing them from unidirectional labral tears can be difficult, even in the presence of advanced imaging. The purpose of the current study, therefore, was to evaluate the sensitivity of preoperative MRI done at a wide range
of imaging facilities, including both noncontrast MRI and MR arthrography, in detecting combined lesions of the anterior, posterior, and superior glenoid labrum. We found that these tears were frequently missed on MRI, regardless of the presence or absence of arthrography, supporting our hypothesis. The sensitivity of MRI overall in detecting combined tears of the anterior, posterior, and superior glenoid labrum was approximately 20% in the current report, with MR arthrography showing a trend toward a higher, but still low, sensitivity in comparison to noncontrast MRI (33.3% v 8.7%). Although detecting the full degree of labral pathology was poor regardless of the use of intra-articular contrast, MR arthrography did reveal some degree of labral pathology in all instances and showed labral pathology in more than one direction in most of the cases (83.3%). Noncontrast MRI did not show any labral tears in 3 cases, and was
277
MRI SENSITIVITY IN PANLABRAL TEARS
Table 1. Detection of Combined Lesions of the Anterior, Posterior, and Superior Glenoid Labrum by Preoperative Magnetic Resonance Imaging Condition
MRI Overall (n ¼ 47)
Noncontrast MRI (n ¼ 23)
MR Arthrography (n ¼ 24)
P value
Detection of tears of anterior, posterior, and superior labrum Detection of tears in more than one location Detection of no labral pathology
10 (21.3%) 30 (63.8%) 3 (6.4%)
2 (8.7%) 10 (43.5%) 3 (13.0%)
8 (33.3%) 20 (83.3%) 0 (0%)
.07 .006 .11
NOTE. Frequency distributions (%) are reported and compared using the Fisher exact test. P values represent comparisons between noncontrast MRI and MR arthrography for the conditions listed.
significantly worse than MR arthography in detecting tears in more than one direction. Tokish et al. recently reported on outcomes of arthroscopic repair of circumferential lesions of the glenoid labrum in 41 shoulders from a military population.4 Findings on preoperative MRI were included. All 41 shoulders underwent preoperative MRI, with 30 undergoing MR arthrography. The radiologist correctly read a panlabral lesion on the MR image in 24 of 41 (59%) shoulders, with better sensitivity on the MR arthrogram (23 of 30, 77%) than on the noncontrast MR image (1 of 11, 9%).4 Although the findings on noncontrast MRI are comparable to the results in our study, the sensitivity of MR arthrography was substantially higher than the value we reported. This discrepancy may be caused by several factors. A wide range of imaging locations were used in the current study to obtain preoperative MR images, resulting in variability in equipment, imaging protocol, and number and training of radiologists used. In contrast, Tokish et al. examined a captured military population, with all surgical procedures performed at one of 2 military sites and all preoperative imaging likely limited to a small number of sites. This potentially led to a more consistent imaging protocol for the MRI performed in that study, with use of more uniform equipment, and interpretation of the scans by a fewer number of more uniformly trained radiologists. These factors could have potentially led to improved imaging and detection of labral pathology, resulting in improved sensitivity. In addition, a heightened awareness for these multidirectional labral injuries may have been present because of the young, highly active military population that was being examined. An older, less active civilian population was examined in the current study. This may represent a more commonly treated group of patients, however. In a radiological study in 2005, Lindauer et al. reported on the MRI appearance of labral tears involving greater than 180 of the glenoid labrum, noting that these lesions had not specifically been described in the orthopaedic or radiological literature.1 Preoperative MRI studies were available for 19 of 31 patients (10 of 19 MR arthrograms) with arthroscopically confirmed labral tears involving 180 to 360 of the glenoid labrum (mean arc, 220 ). All MRI studies were performed at the authors’ institution and were reviewed in
a blinded manner by 2 board-certified musculoskeletal radiologists: first at the time of preoperative evaluation and then retrospectively after the extent of labral pathology was confirmed arthroscopically. Correct preoperative MRI diagnosis was again noted to be poor in this study, with the complete labral injury detected in only 1 of 19 patients and no labral pathology noted in 4 of 19 patients.1 At the second-look retrospective review, all 19 patients who had MRI were diagnosed with some degree of labral pathology, but the complete extent of labral injury was still detected in only 1 of 19 patients. Labral pathology in more than one quadrant of the glenoid was noted on only 6 MR images preoperatively, compared with 17 after second-look retrospective review.1 Differences in evaluation of noncontrast MR images and MR arthrograms were not reported in this study. A strength of the current study is the generalizability of the findings in comparison to prior reports. The previous studies that have reported on the use of preoperative MRI in identifying labral tears involving more than one location of the glenoid have used more narrow patient populations, with the imaging studies performed at a limited number of centers or performed on a more select group of patients.1,4 The current report evaluated a more commonly treated group of patients from a civilian population, who used a variety of imaging centers for their MRI studies, with the studies interpreted by a number of different radiologists. We, therefore, believe it is more representative of a true-tolife scenario for practicing orthopaedic surgeons. Limitations There were several limitations of this study. As this was a retrospective chart review, there was no uniform imaging protocol for each of the MRI studies. Patients were free to undergo MRI at any location, and 30 different imaging locations were used for the 47 preoperative MRI studies. Therefore, variability existed in the type of equipment used, the imaging technique employed, and the radiologists’ interpretation of the study. For example, magnet strength has been shown to impact study resolution and the degree of anatomic detail on MRI, and can be variable across different imaging machines.21,22 In the current study, the magnet strength of the MRI scanner was listed in only
278
E. T. RICCHETTI ET AL.
14 of 47 radiology reports, with 13 noting utilization of a 1.5-T magnet and one using a 0.7-T magnet. In addition to the presence or absence of intra-articular contrast, the imaging technique can vary considerably in other ways, such as the image sequencing used and the position of the extremity while imaged. Certain MR sequences have been shown to be more reliable than others in detecting labral tears, and the sequencing employed for each MRI study in this study was not consistently included in the MRI reports.20,21,23 Certain extremity positions have been shown to more reliably exhibit pathology on MRI, such as the ABER view in detecting anteroinferior labral tears of the shoulder.24-28 In the current study, use of the ABER view was documented in only 4 MR arthrograms, and no other specific extremity positions were noted in the remainder of the MR images. Finally, the training level and skill of the radiologists who documented the MRI reports were not controlled for in this study. The MR images were read by 39 different radiologists, and it was possible to confirm that a study was read by a radiologist with musculoskeletal specialization in only 12 cases. Variability likely existed, therefore, in the training and experience of the radiologists interpreting these studies. Although hospitals in major academic centers typically employ radiologists with fellowship training in musculoskeletal radiology to read shoulder MRI studies, community hospitals or community-based imaging centers may or may not employ radiologists with such advanced training. Another limitation of the current study was the inability to perform a more detailed statistical analysis of the diagnostic utility of MRI in detecting panlabral tears. Although the number of true positive and false negative MRI findings could be determined to calculate sensitivity, the lack of knowledge of true negatives and false positives left us unable to calculate additional diagnostic statistics, such as specificity, accuracy, and positive and negative predictive values. Determination of all diagnostic test characteristics in a retrospective manner would require analysis of thousands of additional MRI reports because of the infrequency of panlabral lesions, which was beyond the scope of this study.
Conclusions Combined lesions of the glenoid labrum involving tears of the anterior, posterior, and superior labrum are infrequent injuries that are typically not completely defined by either noncontrast MRI or MR arthrography.
References 1. Lindauer KR, Major NM, Rougier-Chapman DP, Helms CA. MR imaging appearance of 180-360 degree labral tears of the shoulder. Skeletal Radiol 2005;34:74-79. 2. Lo IK, Burkhart SS. Triple labral lesions: Pathology and surgical repair techniquedReport of seven cases. Arthroscopy 2005;21:186-193.
3. Powell SE, Nord KD, Ryu RKN. The diagnosis, classification, and treatment of SLAP lesions. Oper Tech Sports Med 2004;12:99-110. 4. Tokish JM, McBratney CM, Solomon DJ, Leclere L, Dewing CB, Provencher MT. Arthroscopic repair of circumferential lesions of the glenoid labrum. J Bone Joint Surg Am 2009;91:2795-2802. 5. Chandnani VP, Yeager TD, DeBerardino T, et al. Glenoid labral tears: Prospective evaluation with MRI imaging, MR arthrography, and CT arthrography. AJR Am J Roentgenol 1993;161:1229-1235. 6. Flannigan B, Kursunoglu-Brahme S, Snyder S, Karzel R, Del Pizzo W, Resnick D. MR arthrography of the shoulder: Comparison with conventional MR imaging. AJR Am J Roentgenol 1990;155:829-832. 7. Gross ML, Seeger LL, Smith JB, Mandelbaum BR, Finerman GA. Magnetic resonance imaging of the glenoid labrum. Am J Sports Med 1990;18:229-234. 8. Iannotti JP, Zlatkin MB, Esterhai JL, Kressel HY, Dalinka MK, Spindler KP. Magnetic resonance imaging of the shoulder: Sensitivity, specificity, and predictive value. J Bone Joint Surg Am 1991;73:17-29. 9. Jahnke AH Jr, Petersen SA, Neumann C, Steinbach L, Morgan F. A prospective comparison of computerized arthrotomography and magnetic resonance imaging of the glenohumeral joint. Am J Sports Med 1992;20:695-700. 10. Palmer WE, Caslowitz PL. Anterior shoulder instability: Diagnostic criteria determined from prospective analysis of 121 MR arthrograms. Radiology 1995;197:819-825. 11. Sanders TG, Tirman PF, Linares R, Feller JF, Richardson R. The glenolabral articular disruption lesion: MR arthrography with arthroscopic correlation. AJR Am J Roentgenol 1999;172:171-175. 12. Seeger LL, Ruszkowski JT, Bassett LW, Kay SP, Kahmann RD, Ellman H. MR imaging of the normal shoulder: Anatomic correlation. AJR Am J Roentgenol 1987;148:83-91. 13. Sher JS, Iannotti JP, Williams GR, et al. The effect of shoulder magnetic resonance imaging on clinical decision making. J Shoulder Elbow Surg 1998;7:205-209. 14. Legan JM, Burkhard TK, Goff WB II, et al. Tears of the glenoid labrum: MR imaging of 88 arthroscopically confirmed cases. Radiology 1991;179:241-246. 15. Palmer WE, Brown JH, Rosenthal DI. Labraleligamentous complex of the shoulder: Evaluation with MR arthrography. Radiology 1994;190:645-651. 16. Bencardino JT, Beltran J, Rosenberg ZS, et al. Superior labrum anterioreposterior lesions: Diagnosis with MR arthrography of the shoulder. Radiology 2000;214:267-271. 17. Karzel RP, Snyder SJ. Magnetic resonance arthrography of the shoulder: A new technique of shoulder imaging. Clin Sports Med 1993;12:123-136. 18. Tirman PF, Applegate GR, Flannigan BD, Stauffer AE, Crues JV III. Magnetic resonance arthrography of the shoulder. Magn Reson Imaging Clin North Am 1993;1:125-142. 19. Magee T. 3-T MRI of the shoulder: Is MR arthrography necessary? AJR Am J Roentgenol 2009;192:86-92. 20. Gusmer PB, Potter HG, Schatz JA, et al. Labral injuries: Accuracy of detection with unenhanced MR imaging of the shoulder. Radiology 1996;200:519-524. 21. Magee T. Can isotropic fast gradient echo imaging be substituted for conventional T1 weighted sequences in
MRI SENSITIVITY IN PANLABRAL TEARS
22.
23. 24.
25.
shoulder MR arthrography at 3 Tesla? J Magn Reson Imaging 2007;26:118-122. Magee TH, Williams D. Sensitivity and specificity in detection of labral tears with 3.0-T MRI of the shoulder. AJR Am J Roentgenol 2006;187:1448-1452. Chaipat L, Palmer WE. Shoulder magnetic resonance imaging. Clin Sports Med 2006;25:371-386. Chiavaras MM, Harish S, Burr J. MR arthrographic assessment of suspected posteroinferior labral lesions using flexion, adduction, and internal rotation positioning of the arm: Preliminary experience. Skeletal Radiol 2010;39:481-488. Cvitanic O, Tirman PF, Feller JF, Bost FW, Minter J, Carroll KW. Using abduction and external rotation of the
279
shoulder to increase the sensitivity of MR arthrography in revealing tears of the anterior glenoid labrum. AJR Am J Roentgenol 1997;169:837-844. 26. Harish S, Nagar A, Moro J, Pugh D, Rebello R, O’Neill J. Imaging findings in posterior instability of the shoulder. Skeletal Radiol 2008;37:693-707. 27. Song HT, Huh YM, Kim S, et al. Anterioreinferior labral lesions of recurrent shoulder dislocation evaluated by MR arthrography in an adduction internal rotation (ADIR) position. J Magn Reson Imaging 2006;23: 29-35. 28. Kwak SM, Brown RR, Trudell D, Resnick D. Glenohumeral joint: Comparison of shoulder positions at MR arthrography. Radiology 1998;208:375-380.