Incidence of False Positive Rotator Cuff Pathology in MRIs of Patients With Adhesive Capsulitis

Abstract

The purpose of this study was to compare the incidence of presumed rotator cuff pathology based on radiologist and surgeon interpretation of preoperative magnetic resonance images (MRIs) with actual rotator cuff pathology found intraoperatively in patients undergoing arthroscopic capsular release for adhesive capsulitis.

The medical records of 38 consecutive patients who underwent arthroscopic glenohumeral capsular release for adhesive capsulitis between September 2001 and May 2007 were retrospectively reviewed. Radiologist and surgeon predicted status of the rotator cuff by prospective MRI interpretation was compared to the actual status of the rotator cuff intraoperatively. Radiologists’ preoperative MRI interpretations predicted a 57.9% incidence of rotator cuff pathology, while operative findings revealed a true incidence of only 13.2% (P<.0001). Radiologists accurately predicted the absence or presence of a rotator cuff tear in 19 of 38 cases (50%), compared to the surgeon, who correctly interpreted 29 of 38 MRIs (76.3%).

Interpretations of shoulder MRIs in patients with adhesive capsulitis may provide misleading information regarding rotator cuff pathology. The actual findings of shoulder MRI scans may lead to the appearance of false rotator cuff pathology in this population, and a high percentage of false positive MRI reports of rotator cuff tears was observed in these patients. If MRI is chosen as a diagnostic adjunct in this patient population, careful consideration should be given to its interpretation to avoid unnecessary or possibly aggravating surgical intervention.

Idiopathic adhesive capsulitis is a condition of unknown etiology characterized by painful restriction of both active and passive glenohumeral and periscapular range of motion (ROM) that occurs in the absence of a known intrinsic shoulder disorder or preceding traumatic or surgical event.¹ Classically, the clinical presentation of idiopathic adhesive capsulitis is characterized by the spontaneous onset of diffuse and often severe shoulder pain, followed by a period of progressively increasing limitations in active and passive shoulder motion. Ultimately, some patients experience a thawing phase of clinical improvement.² However, some patients experience prolonged or permanent symptoms of shoulder pain or stiffness.³ In these cases of recalcitrant or severe adhesive capsulitis, arthroscopic capsular release may be performed to restore shoulder motion and minimize the morbidity associated with this condition.⁴

It has been suggested that rotator cuff disease and adhesive capsulitis can typically be distinguished based on a careful history and physical examination.¹ Adhesive capsulitis is most commonly found in women aged 40 to 60 years and may occur in association with diabetes mellitus, thyroid disorders, or other endocrine conditions.^5-7 In contrast, rotator cuff pathology is seen in both men and women with increasing frequency after the age of 40, and often results in limited active shoulder motion, while passive ROM is characteristically maintained. Although these 2 clinical entities may be present concomitantly, it is unusual for patients with idiopathic adhesive capsulitis to have rotator cuff tears and vice versa.

Although magnetic resonance imaging (MRI) is not routinely recommended in the evaluation of patients with adhesive capsulitis,¹ Carrillon et al⁸ and Connell et al⁹ have suggested that MRI with intravenous gadolinium may be used to specifically evaluate the rotator interval and axillary pouch in the diagnosis of unclear cases of adhesive capsulitis. The study of Connell et al⁹ included patients with surgically confirmed changes of fibrovascular scar tissue formation in the rotator interval and around the biceps anchor. Three of their 24 patients with clinical evidence of adhesive capsulitis had preoperative MRI interpretations of partial-thickness rotator cuff tears; however, they did not report findings of rotator cuff pathology at the time of surgery. Carrillon et al⁸ found evidence of partial- or full-thickness supraspinatus tears in 8 of 25 patients with a clinical diagnosis of adhesive capsulitis, but there was no mention of preoperative MRI findings in these patients. To our knowledge, no study of patients with idiopathic adhesive capsulitis has compared the accuracy of preoperative MRI interpretations of rotator cuff pathology with the actual status of the rotator cuff as noted intraoperatively.

We hypothesized that the intraoperative status of the rotator cuff would not consistently correlate with radiologists’ preoperative MRI interpretations of patients with idiopathic adhesive capsulitis. The purpose of this study was to compare the incidence of presumed rotator cuff pathology based upon radiologist and surgeon interpretations of preoperative MRIs with actual rotator cuff pathology found intraoperatively in patients undergoing arthroscopic capsular release for adhesive capsulitis.

Materials and Methods

Using an Institutional Review Board-approved protocol, the medical records of 78 consecutive patients who underwent arthroscopic glenohumeral capsular release between September 2001 and May 2007 were retrospectively reviewed. Patients were included in this study if (1) there was a preoperative diagnosis of primary adhesive capsulitis, (2) an MRI was obtained prior to referral for surgical intervention and an official radiologist’s interpretation of the MRI was available, and (3) arthroscopic glenohumeral capsular release was performed after failure of conservative treatment. Patients were excluded from the study if there was (1) a history of significant traumatic injury to the shoulder, (2) a pre- or postoperative diagnosis other than the idiopathic adhesive capsulitis causing shoulder stiffness (eg, osteoarthritis or avascular necrosis), or (3) a history of prior open or arthroscopic surgery on the affected shoulder. Thirty-eight of the 78 patients met criteria for inclusion. The records of these patients were then reviewed to determine demographics, history, radiologists’ MRI interpretations, the surgeon’s preoperative MRI interpretations, operative findings, and surgical procedures performed. All surgeries were performed by a single surgeon who prospectively documented specific intraoperative findings, including routine inspection of the bursal and articular surfaces of the rotator cuff, with both the operative report and intraoperative photographs.

The study cohort consisted of 19 men and 19 women with a mean age of 53.4 years (range, 36-70 years). Average duration of symptoms was 12.6 months (range, 3-30 months). The dominant arm was involved in 23 patients. Fourteen patients recalled a specific episode of minor trauma that was felt to subjectively initiate the painful process, while 24 patients reported an insidious onset of symptoms. Eight patients were diabetic (4 type I and 4 type II), and 2 patients were being treated for hypothyroidism. Combined glenohumeral and scapulothoracic preoperative elevation in the scapular plane averaged 114°, external rotation 35°, and internal rotation to the ipsilateral greater trochanter.

The MRIs were often obtained prior to our evaluation and at outside facilities, and therefore the MRIs were not standardized. There was 1 gadolinium-enhanced magnetic resonance arthrogram and no MRIs with intravenous gadolinium. There were no open MRI studies in this series. Seventeen of the MRIs were performed using a 1.5-T magnet, and none of the studies used a >3.0-T magnet. We were unable to determine the field strength of the remaining MRIs, which were obtained at outside institutions.

Twenty-three of the MRIs were interpreted by fellowship-trained musculo-skeletal radiologists, while the remainder of the studies were interpreted by radiologists with unknown levels of additional musculoskeletal training. Radiology reports described rotator cuff integrity as being intact or having a partial- or full-thickness tear. Two MRI reports provided equivocal descriptions of the rotator cuff integrity. In the first case, the report read, “high-grade tendinosis versus partial-thickness tear.” In the second case, the radiologist’s interpretation was a “partial versus full-thickness tear.” In these 2 cases, the more severe of the diagnoses was recorded for study purposes.

The surgeon’s preoperative MRI interpretations of all patients were divided into 2 groups. The first group was felt to have a partial- or full-thickness rotator cuff tear likely needing repair, while the second group was interpreted as having either insignificant partial-thickness involvement, tendinosis, or no tear.

The diagnostic accuracy of the radiologists’ MRI interpretations in detecting rotator cuff tears was evaluated by calculating the sensitivity, specificity, and positive and negative predictive values compared to the gold standard of direct examination at the time of arthroscopy. To determine the reliability of these diagnostic accuracy measures, 95% confidence intervals (CIs) were also calculated. Disagreements between the radiologists’ interpretations of the rotator cuff and the actual status of the rotator cuff intraoperatively were evaluated using McNemar’s test.

Results

Of the 38 patients included in the study, radiologists interpreted the preoperative MRIs as having either a partial- or full-thickness rotator cuff tear in 22 patients (57.9%), while the surgeon predicted this in only 8 patients (21.1%). Radiologists interpreted 15 MRIs as having partial-thickness tears (39.5%), 7 as having full-thickness tears (18.4%), and 16 as having an intact rotator cuff (42.1%). The surgeon interpreted 1 MRI as having a partial-thickness tear (2.6%), 7 as having full-thickness tears (18.4%), and 30 as having an intact rotator cuff (78.9%). The accuracy of these interpretations is summarized in Table 1.

Intraoperatively, only 5 of the 38 patients (13.2%) were found to have objective rotator cuff pathology. Three patients (7.9%) with minimal articular-sided tears of the supraspinatus involving <3 mm of the medial footprint were treated with limited debridement, and 2 patients (5.3%) with significant partial-thickness tears (>6 mm) were treated with arthroscopic repair. These results, along with the accompanying interpretations of the radiologists and surgeon, are included in Table 2. No patient was found to have a full-thickness rotator cuff tear intraoperatively.

Of the 15 cases preoperatively interpreted by the radiologists as having partial-thickness rotator cuff tears, only 3 were confirmed intraoperatively. The surgeon predicted a partial-thickness tear in 1 case, and in this patient no tear was found intraoperatively. Radiologists predicted full-thickness rotator cuff tears in 7 patients; however, 6 of these 7 patients had no rotator cuff pathology and only 1 had a partial-thickness tear. The surgeon also interpreted 7 cases of full-thickness rotator cuff tears, and only 2 of these patients had partial-thickness tears. Both of these tears were significant, however, and both were repaired at the time of arthroscopic capsular release. The Figure demonstrates a representative case in which the radiologist interpreted a full-thickness rotator cuff tear, while intraoperative findings demonstrated the bursal and articular surfaces of the rotator cuff to be intact. One of the 16 patients with a radiologist reading of an intact rotator cuff was found to have a partial-thickness tear intraoperatively. This MRI had been interpreted as high-grade tendinopathy with an intact rotator cuff, and this patient had a minimal articular-sided tear of the supraspinatus involving <3 mm of the footprint.

	Figure: Preoperative coronal T2-weighted MRI interpreted by the radiologist as a small full-thickness rotator cuff tear (A). The same patient was found to have a normal rotator cuff on the articular (B) and bursal (C) surfaces intraoperatively.

The fellowship-trained shoulder surgeon preoperatively interpreted 8 cases of significant rotator cuff tears (7 full-thickness and 1 partial-thickness) in which repair was anticipated, while only 2 of these 8 cases (25%) were actually found to have significant rotator cuff tears, which were then repaired at the time of surgery. The surgeon correctly interpreted the presence or absence of a rotator cuff tear in 29 of 38 cases overall (76.3%), compared to 19 of 38 cases (50%) by the radiologists. These results are summarized in Table 3.

Overall, the radiologists’ and surgeon’s preoperative MRI interpretations predicted a 57.9% and 21.1% incidence of partial- and full-thickness rotator cuff tears, respectively, while operative findings revealed an incidence of only 13.2% (all of which were partial-thickness tears). There were 18 cases where the radiologists interpreted a rotator cuff tear and the surgical findings revealed no tear, while there was only 1 case where the radiologist interpreted no rotator cuff pathology and a rotator cuff tear was visualized intraoperatively (P<.0001). The surgeon predicted a rotator cuff tear in 6 cases where surgical findings revealed no tear, while in 3 of the 30 instances in which the surgeon predicted no tear, there was a partial-thickness tear (none of which required repair). There were 18 cases where the radiologists interpreted a rotator cuff tear and the surgical findings revealed no tear, while there was only 1 case where the radiologist interpreted no rotator cuff pathology and a rotator cuff tear was visualized intraoperatively. Compared to the gold standard of direct surgical visualization, the radiologists’ and surgeon’s preoperative MRI evaluations had sensitivities of 80% (95% CI, 28%-99.5%) and 40% (95% CI, 7%-83%), respectively; specificities of 45% (95% CI, 28%-64%) and 82% (95% CI, 64%-92%), respectively; positive predictive values of 18% (95% CI, 5%-40%) and 25% (95% CI, 44%-64%), respectively; and negative predictive values of 94% (95% CI, 70%-99.8%) and 90% (95% CI, 72%-97%), respectively, for detecting rotator cuff tears in patients undergoing arthroscopic capsular release for recalcitrant adhesive capsulitis.

Discussion

The diagnosis of adhesive capsulitis is made clinically and is characterized by a loss of both active and passive glenohumeral motion. Adhesive capsulitis may be subdivided into 3 distinct entities for descriptive purposes: idiopathic adhesive capsulitis, secondary adhesive capsulitis, and postoperative stiffness. The present study examined only patients with idiopathic adhesive capsulitis to exclude potentially confounding data. As has been commonly accepted, treatment options for adhesive capsulitis include benign neglect, physical therapy (home-based or supervised), intra-articular steroid injections, manipulation under anesthesia, and arthroscopic capsular release.^{1,3,4,6,10-12} As adhesive capsulitis is a clinical diagnosis, preoperative MRIs are often not necessary to validate the diagnosis. There are times, however, when the diagnosis of idiopathic versus secondary adhesive capsulitis may be elusive. It is in these instances where MRI may be used by some clinicians as a diagnostic adjunct to provide additional objective insight into the shoulder condition. It has been consistently emphasized in the literature that treating clinicians must always carefully scrutinize radiologists’ MRI interpretations, correlate them to clinical findings, and be aware of the diagnostic accuracy of MRIs in different clinical settings.^13-17

Numerous studies have demonstrated the efficacy of MRI in characterizing rotator cuff tears, with sensitivities and specificities ranging from 85% to 100% for both partial- and full-thickness tears.^18-23 The accuracy of detecting rotator cuff pathology by MRI has, however, been shown to be less reliable in some more complex clinical situations.^17,24 Wnorowski et al¹⁷ studied 39 shoulders that were considered difficult cases (ie, referrals from other surgeons or diagnostic dilemmas) and compared MRI interpretations with arthroscopic findings. They evaluated both community radiologists and musculoskeletal-trained radiologists’ interpretations and compared each group’s interpretations with surgical findings. The sensitivity, specificity, positive predictive value, and negative predictive value were calculated for each group to be 85%, 52%, 50%, 87% (community radiologists) and 71%, 71%, 59%, and 81% (musculoskeletal-trained radiologists), respectively. The authors concluded that MRIs were not reliably as accurate as arthroscopy in these clinical situations. Magnetic resonance imaging was most helpful in cases where no rotator cuff pathology was interpreted, and an MRI interpretation of a partial tear of the rotator cuff was of little value in these difficult clinical settings.

Motamedi et al²⁴ studied 37 shoulders in patients with continued pain and dysfunction after rotator cuff repair by correlating musculoskeletal-trained radiologists’ MRI interpretations with surgical findings.²⁴ The sensitivity and specificity for MRI diagnosis of a recurrent tear was 91% and 25%, respectively, and radiologists incorrectly reported a recurrent tear in 75% of the intact rotator cuffs. These studies suggest that radiologists’ interpretations of MRIs in complex and/or revision situations have decreased diagnostic value.

Other studies have documented MRI findings in asymptomatic shoulders, thereby warning of potential hazards in relying on MRI alone as a determinant of operative intervention.^9,13-16 Sher et al¹⁶ studied 96 asymptomatic shoulders and found MRI evidence of rotator cuff tear in 54% of patients older than 60 years, 28% of patients aged 40 to 60 years, and 4% of patients aged 19 to 39 years. The overall rate of MRI findings suggesting rotator cuff tears was 34%. Miniaci et al¹⁵ evaluated 30 asymptomatic shoulders with an average age of 29 years with MRIs and found a 100% incidence of grade 1 signal changes (focal, linear, or diffuse intermediate signal within the tendon) and a 23% incidence of grade 2 signal changes (high-signal intensity less than full thickness); no shoulders had full-thickness tears. Connor et al¹³ performed an MRI study of asymptomatic elite overhead athletes younger than 40 years and found a 40% incidence of partial- or full-thickness rotator cuff tears in dominant shoulders compared to a 0% incidence in nondominant shoulders. In addition, none of these athletes demonstrated subjective shoulder symptoms or required any further evaluation or treatment of their dominant shoulders at the study’s 5-year follow-up.

Although MRIs are not routinely used in the evaluation of patients with adhesive capsulitis, studies evaluating MRI arthrography and MRI with intravenous gadolinium have described the shoulder MRI findings associated with this condition. Manton et al²⁵ evaluated 28 shoulders (9 with adhesive capsulitis and 19 controls) with MRI arthrography and found no consistent differences in the adhesive capsulitis group as compared to controls. Mengiardi et al,²⁶ however, found thickening of the coracohumeral ligament and rotator interval capsule, as well as loss of the subcoracoid fat triangle to be characteristic of adhesive capsulitis in 22 MRI arthrographies. Jung et al²⁷ studied 28 shoulders (14 with adhesive capsulitis and 14 controls), also using MRI arthrography, and suggested that thickness of the capsule and synovium in the axillary recess .3 mm is highly accurate in diagnosing adhesive capsulitis. Abnormal signal within the rotator interval tissue was found to be highly sensitive but nonspecific in the diagnosis of adhesive capsulitis.²⁷ No studies using MRI arthrography of the shoulder have compared the MRI evaluation of rotator cuff status with surgical findings.

Shoulder MRI with intravenous gadolinium has been advocated by some as a means to evaluate patients’ shoulders with adhesive capsulitis and specifically to help differentiate adhesive capsulitis from other causes of shoulder pain and/or stiffness.^8,9,28 Carrillon et al⁸ found post-gadolinium enhancement of the joint capsule and synovium in the rotator interval in 25 of 25 patients and enhancement in the axillary recess in 22 of 25 of patients with adhesive capsulitis. The same study reported 8 of 25 patients with MRI evidence of rotator cuff tears, including 6 with partial-thickness and 2 with full-thickness tears.⁸ Connell et al⁹ found rotator interval enhancement after intravenous gadolinium in 22 of 24 shoulders with a clinical diagnosis of adhesive capsulitis. They also reported 3 of 24 shoulders had MRI evidence of partial-thickness rotator cuff tears.⁹ Neither of these studies, however, reported surgical findings of rotator cuff pathology.

The findings of the present study do not definitively answer the question as to whether the incidence of false positive rotator cuff tears are a reflection of the limitations of MRI in and of itself in the setting of adhesive capsulitis or the result of a tendency to over-report rotator cuff tears in this setting. Clark and Harryman’s²⁹ gross and histologic study of the rotator cuff demonstrated that the deepest surface of the 5 layers of the rotator cuff is the capsule. Thus, with adhesive capsulitis, the pathologic process affecting the capsule may potentially cause increased signal in the rotator cuff and thereby create an appearance on the MRI of rotator cuff pathology that does not exist. This close anatomic relationship of the rotator cuff-capsule complex may be a potential explanation for the findings of this study.

The results of the present study suggest that radiologists’ MRI interpretations of shoulders in patients with adhesive capsulitis may provide misleading information in regard to rotator cuff pathology. In this series, radiologists’ MRI interpretations substantially overestimated the presence of rotator cuff pathology (57.5% vs 13.2% found intraoperatively; positive predictive value, 18%). Although the surgeon more accurately interpreted the presence or absence of rotator cuff tears on preoperative MRIs, there were still 6 false positive interpretations of rotator cuff tears (15.8% of all patients) by the surgeon. Therefore, the question is whether the findings of this study reflect merely a tendency to over-report rotator cuff tears in this patient population, an issue with the actual findings of shoulder MRI scans in this setting, or a combination of these 2 factors. Although further research is needed to elucidate potential reasons for this phenomenon in this patient population, clinicians should be wary of making a diagnosis and providing treatment based solely on the radiologist’s interpretation of rotator cuff integrity in patients with adhesive capsulitis to avoid a potentially unnecessary or aggravating surgical intervention.

The present study has several limitations. Magnetic resonance imaging studies were not standardized, thus resulting in variability in the quality of the studies performed. In addition, the MRI interpretations were performed by many different radiologists with variability in their level of experience in interpreting these studies. However, this variability is representative of this center’s tertiary referral practice, as well as that of many other orthopedic practices, and therefore is of practical value to orthopedic surgeons whose patients’ MRIs are performed at different locations and interpreted by various radiologists.

In the present study, it may be difficult to directly compare the radiologists’ and surgeon’s preoperative MRI interpretations, as the surgeon’s interpretations may have been influenced by the patients’ history and physical examination findings. A future study evaluating a standardized group of blinded radiologist and surgeon MRI interpretations may further elucidate whether it is the MRI itself or the interpretations thereof that contribute to the high false positive rate of rotator cuff tears in patients with adhesive capsulitis observed in this study.                                                           

References

1. Brue S, Valentin A, Forssblad M, Werner S, Mikkelsen C, Cerulli G. Idiopathic adhesive capsulitis of the shoulder: a review [published online ahead of print February 28, 2007]. Knee Surg Sports Traumatol Arthrosc. 2007; 15(8):1048-1054.
2. Lundberg BJ. The frozen shoulder. Clinical and radiographical observations. The effect of manipulation under general anesthesia. Structure and glycosaminoglycan content of the joint capsule. Local bone metabolism. Acta Orthop Scand Suppl. 1969; (119):1-59.
3. Shaffer B, Tibone JE, Kerlan RK. Frozen shoulder. A long-term follow-up. J Bone Joint Surg Am. 1992; 74(5):738-746.
4. Holloway GB, Schenk T, Williams GR, Ramsey ML, Iannotti JP. Arthroscopic capsular release for the treatment of refractory postoperative or post-fracture shoulder stiffness. J Bone Joint Surg Am. 2001; 83(11):1682-1687.
5. Bridgman JF. Periarthritis of the shoulder and diabetes mellitus. Ann Rheum Dis. 1972; 31(1):69-71.
6. Grey RG. The natural history of “idiopathic” frozen shoulder. J Bone Joint Surg Am. 1978; 60(4):564.
7. Wohlgethan JR. Frozen shoulder in hyperthyroidism. Arthritis Rheum. 1987; 30(8):936-939.
8. Carrillon Y, Noel E, Fantino O, Perrin-Fayolle O, Tran-Minh VA. Magnetic resonance imaging findings in idiopathic adhesive capsulitis of the shoulder. Rev Rhum Engl Ed. 1999; 66(4):201-206.
9. Connell D, Padmanabhan R, Buchbinder R. Adhesive capsulitis: role of MR imaging in differential diagnosis [published online ahead of print March 7, 2002]. Eur Radiol. 2002; 12(8):2100-2106.
10. Hannafin JA, Chiaia TA. Adhesive capsulitis. A treatment approach. Clin Orthop Relat Res. 2000; (372):95-109.
11. Griggs SM, Ahn A, Green A. Idiopathic adhesive capsulitis. A prospective functional outcome study of nonoperative treatment. J Bone Joint Surg Am. 2000; 82(10):1398-1407.
12. Hill JJ Jr, Bogumill H. Manipulation in the treatment of frozen shoulder. Orthopedics. 1988; 11(9):1255-1260.
13. Connor PM, Banks DM, Tyson AB, Coumas JS, D’Alessandro DF. Magnetic resonance imaging of the asymptomatic shoulder of overhead athletes: a 5-year follow-up study. Am J Sports Med. 2003; 31(5):724-727.
14. Miniaci A, Mascia AT, Salonen DC, Becker EJ. Magnetic resonance imaging of the shoulder in asymptomatic professional baseball pitchers. Am J Sports Med. 2002; 30(1):66-73.
15. Miniaci A, Dowdy PA, Willits KR, Vellet AD. Magnetic resonance imaging evaluation of the rotator cuff tendons in the asymptomatic shoulder. Am J Sports Med. 1995; 23(2):142-145.
16. Sher JS, Uribe JW, Posada A, Murphy BJ, Zlatkin MB. Abnormal findings on magnetic resonance images of asymptomatic shoulders. J Bone Joint Surg Am. 1995; 77(1):10-15.
17. Wnorowski DC, Levinsohn EM, Chamberlain BC, McAndrew DL. Magnetic resonance imaging assessment of the rotator cuff: is it really accurate? Arthroscopy. 1997; 13(6):710-719.
18. Balich SM, Sheley RC, Brown TR, Sauser DD, Quinn SF. MR imaging of the rotator cuff tendon: interobserver agreement and analysis of interpretive errors. Radiology. 1997; 204(1):191-194.
19. 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(1):17-29.
20. Palmer WE, Brown JH, Rosenthal DI. Rotator cuff: evaluation with fat-suppressed MR arthrography. Radiology. 1993; 188(3):683-687.
21. Quinn SF, Sheley RC, Demlow TA, Szumowski J. Rotator cuff tendon tears: evaluation with fat-suppressed MR imaging with arthroscopic correlation in 100 patients. Radiology. 1995; 195(2):497-500.
22. Teefey SA, Rubin DA, Middleton WD, Hildebolt CF, Leibold RA, Yamaguchi K. Detection and quantification of rotator cuff tears. Comparison of ultrasonographic, magnetic resonance imaging, and arthroscopic findings in seventy-one consecutive cases. J Bone Joint Surg Am. 2004; 86(4):708-716.
23. Yamakawa S, Hashizume H, Ichikawa N, Itadera E, Inoue H. Comparative studies of MRI and operative findings in rotator cuff tear. Acta Med Okayama. 2001; 55(5):261-268.
24. Motamedi AR, Urrea LH, Hancock RE, Hawkins RJ, Ho C. Accuracy of magnetic resonance imaging in determining the presence and size of recurrent rotator cuff tears. J Shoulder Elbow Surg. 2002; 11(1):6-10.
25. Manton GL, Schweitzer ME, Weishaupt D, Karasick D. Utility of MR arthrography in the diagnosis of adhesive capsulitis. Skeletal Radiol. 2001; 30(6):326-330.
26. Mengiardi B, Pfirrmann CW, Gerber C, Hodler J, Zanetti M. Frozen shoulder: MR arthrographic findings [published online ahead of print September 9, 2004]. Radiology. 2004; 233(2):486-492.
27. Jung JY, Jee WH, Chun HJ, Kim YS, Chung YG, Kim JM. Adhesive capsulitis of the shoulder: evaluation with MR arthrography [published online ahead of print October 14, 2005]. Eur Radiol. 2006; 16(4):791-796.
28. Lefevre-Colau MM, Drapé JL, Fayad F, et al. Magnetic resonance imaging of shoulders with idiopathic adhesive capsulitis: reliability of measures [published online ahead of print July 8, 2005]. Eur Radiol. 2005; 15(12):2415-2422.
29. Clark JM, Harryman DT II. Tendons, ligaments, and capsule of the rotator cuff. Gross and microscopic anatomy. J Bone Joint Surg Am. 1992; 74(5):713-725.

Authors

Dr Loeffler is from the Department of Orthopedic Surgery, Carolinas Medical Center, Charlotte, North Carolina; Dr Brown is from Rome Orthopaedic Clinic, Rome, Georgia; and Drs D’Alessandro, Fleischli, and Connor are from OrthoCarolina Sports Medicine Center, Charlotte, North Carolina.

Drs Loeffler, Brown, D’Alessandro, Fleischli, and Connor have no relevant financial relationships to disclose.

Correspondence should be addressed to: Patrick M. Connor, MD, OrthoCarolina Sports Medicine Center, 1915 Randolph Rd, Charlotte, NC 28207 (patrick.connor@orthocarolina.com).

doi: 10.3928/01477447-20110317-14