Veterinary Surgery 37:631–638, 2008

Magnetic Resonance Imaging Findings in Dogs with Confirmed Shoulder Pathology SEAN E. MURPHY,

DVM, Diplomate ACVS,

ELIZABETH A. BALLEGEER, DVM, LISA J. FORREST, and SUSAN L. SCHAEFER, MS, DVM, Diplomate ACVS

VMD, Diplomate ACVR,

Objective—To evaluate the diagnostic potential of magnetic resonance imaging (MRI) compared with a reference standard, arthroscopic and/or open surgery, in dogs with soft tissue shoulder pathology. Study Design—Retrospective study. Animals—Dogs (n ¼ 21). Methods—Magnetic resonance (MR) images were retrospectively evaluated in 21 dogs that had surgically identified soft tissue shoulder pathology. The musculotendinous units of the biceps, infraspinatus, teres minor, supraspinatus, subscapularis, and the medial and lateral glenohumeral ligaments (MGHL and LGHL) were graded as either normal or abnormal. Abnormal structures were further classified as being either inflamed, partially torn, or fully torn. Impingement of the biceps tendon was also evaluated. Results were reported in terms of agreement and concordance between MRI findings and surgical findings. Agreement was defined as the percentage of times MRI findings concurred with surgical findings with respect to a structure being either normal or abnormal. Concordance was defined as the percentage of times MRI concurred with the exact surgically assessed pathology when abnormality was identified. Results—The findings were biceps tendon: 90% agreement with 100% concordance; subscapularis: 95% agreement with 62% concordance; MGHL: 84% agreement with 83% concordance; LGHL: 88% agreement with 100% concordance; infraspinatus: 100% both agreement and concordance; biceps tendon impingement: 90% agreement with 100% concordance. Conclusions—Soft tissue abnormalities of the canine shoulder were readily identified on preoperative MR images. Clinical Relevance—MRI shows great potential as a diagnostic tool in the evaluation of canine shoulder disease. r Copyright 2008 by The American College of Veterinary Surgeons

active stabilizers: the supraspinatus, infraspinatus, biceps, teres minor muscles, and the subscapularis muscle with associated MGHL pathology.3,5–8 Although the incidence of shoulder injury has not been fully reported, a recent abstract described 385 dogs that had 422 shoulder arthroscopies where the major cause of lameness was intra-articular in 421 (99.8%) with the top 3 diagnoses: instability (48%), osteochondritis dessicans (25.4%), and biceps tendon/sheath disease (9.4%).9

INTRODUCTION

S

HOULDER PATHOLOGY is a common cause of forelimb lameness in dogs. Normal shoulder joint function is dependent on both passive and active stabilizing structures.1,2 Specific reports have associated lameness with pathology of both the passive stabilizers: the lateral glenohumeral ligament (LGHL), medial glenohumeral ligament (MGHL), the joint capsule,1,3,4 and the

From the Department of Surgical Sciences, School of Veterinary Medicine, University of Wisconsin, Madison, WI. This study was presented in part at the Advanced Canine Arthroscopy Symposium, August 2007, Naples, FL. Supported by a grant from the Companion Animal Fund, School of Veterinary Medicine, University of Wisconsin, Madison, WI. Address reprint requests to Susan L. Schaefer, MS, DVM, Diplomate ACVS, Department of Surgical Sciences, School of Veterinary Medicine, University of Wisconsin, 2015 Linden Drive, Madison, WI 53706-1102. E-mail: [email protected]. Submitted December 2007; Accepted March 2008 r Copyright 2008 by The American College of Veterinary Surgeons 0161-3499/08 doi:10.1111/j.1532-950X.2008.00429.x

631

632

MRI FINDINGS IN DOGS WITH SHOULDER PATHOLOGY

Orthopedic examination, radiographs, and arthroscopy were used as diagnostic tools. A multi-institutional abstract on shoulder pathology, which also incorporated diagnostic ultrasound, reported the same top 3 diagnoses with a slightly lower incidence of intra-articular pathology as the major cause for lameness, 90 of 96 reported cases (94%).10 Determining the exact causes of shoulder lameness is often challenging. Both an intra-articular and extraarticular assessment may be required. Preoperative diagnostic tests such as a refined shoulder examination, radiographs, standard arthrography, ultrasound (US), magnetic resonance imaging (MRI), and magnetic resonance (MR) arthrography may improve diagnostic competence and effect subsequent treatment options with the ultimate goal of improving patient outcomes. MRI, a powerful diagnostic modality, is used commonly for diagnosis of human musculoskeletal disease providing information on intra- and extra-articular structures simultaneously with high soft tissue contrast, high-resolution and multiplanar imaging capabilities.11 Normal musculotendinous and capsuloligamentous structures of the canine shoulder are readily identified by MRI in cadavers.12 The only reported clinical use of MRI for evaluation of canine shoulder pathology has been limited to osteochondrosis assessment and a single case report of supraspinatus tendonopathy.13–15 Our purpose was to identify whether MRI has the potential to play a greater role in the diagnosis of canine shoulder disease. Our first aim was to determine whether soft tissue abnormalities identified at surgery in clinically affected dogs could be retrospectively identified as abnormalities on preoperative MR images. Our second aim was to determine whether the specific types of abnormalities noted could be distinguished from one another on MRI in these same cases. MATERIALS AND METHODS Inclusion Criteria All dogs admitted (January 2004–June 2007) for arthroscopic examination of the shoulder because of joint pain and forelimb lameness had a preoperative MRI performed. Dogs that had a radiographic diagnosis of shoulder osteochondritis dissecans were excluded from the study. Medical records, including all diagnostic images, were reviewed retrospectively.

MRI Examination Examinations were performed with the dog in lateral recumbency with the affected shoulder up. Shoulders were positioned in 100–1101 flexion and the distal aspect of the limbs supported with a foam block to maintain the limb in a normal

stance position. A dedicated shoulder coil was placed either around or on top of the shoulder joint depending on the individual dog size. A 1.0 T superconducting magnet (GE Medical Systems, Milwaukee, WI) was used for all scans. Nineteen shoulders were scanned in 3 planes: dorsal, transverse, and sagittal.12 Two shoulders were scanned in the transverse and sagittal planes only. A minimum of 2 sequences were collected in each plane. Sequences used were T1-weighted (T1), T2-weighted gradient echo (GE) and proton density fat saturation (PD). Direct magnetic resonance arthrography was performed in 11 dogs using 1 mL/4.5 kg bodyweight of intra-articular gadodiamide (Gd-DTPA-BMA; Omniscan, GE Healthcare, Princeton, NJ). The gadodiamide was diluted before intraarticular injection with saline (0.9% NaCl) solution to 1:800 in the first 5 dogs and 1:1200 in the final 6 dogs. After gadodiamide injection, all shoulders were re-scanned in all 3 planes with either a T1-weighted fat saturation sequence or a PD sequence.

Arthroscopy Arthroscopy was performed using either a 2.7 or 2.5 mm 301 oblique arthroscope (Linvatec, Utica, NY). All procedures were initially performed in lateral recumbency using a caudal/ lateral telescope portal with a cranial/lateral instrument portal.16 In 2 dogs, an additional arthroscopic examination from a cranial/medial telescope portal was required. These dogs were repositioned into dorsal recumbency and a 2.7 mm switching stick was used to accurately transfer the arthroscope to the cranial/medial position.

Retrieved Data Reviewed surgical data included a detailed surgery report (21 dogs), multiple arthroscopic still images (21), and arthroscopic video (19). Each surgery report included a completed modified shoulder assessment form evaluating biceps tendon, intertubercular groove, subscapularis, MGHL, LGHL, and the caudal/medial aspect of the supraspinatus insertion.10 Extra-articular structures were evaluated in dogs that had open surgical procedures after arthroscopy. All surgical data were reviewed and graded by the same individual (S.M.). MRI scans were retrospectively evaluated by a radiologist (L.F.), a senior radiology resident (E.B.), and a surgeon (S.S.). Given that MR imaging of canine shoulder pathology is not well described, evaluators were not blinded to the case data. The following soft tissue structures were initially identified as either normal or abnormal on each MRI: musculotendinous units of the biceps, infraspinatus, teres minor, supraspinatus, subscapularis muscles, and capsuloligamentous structures including the MGHL and LGHL. If structures were identified as abnormal, then they were further classified as inflamed, partially torn, or fully torn. Structures were considered inflamed when enlargement was identified, often with a heterogeneous signal intensity or generalized hyperintensity. Partial tears were identified by minor disruptions of the tendon/ ligament silhouette. More complete tearing was identified by

633

MURPHY ET AL major disruptions of the tendon/ligament silhouette. Additionally, the relationship between the biceps tendon and supraspinatus tendon was evaluated for impingement syndrome.15 Cases of biceps impingement were further classified subjectively as mild, moderate, or severe based on degree of tendon compression.

Data Analysis Results are reported in terms of agreement and concordance between MRI findings and surgical findings. Agreement was defined as the percentage of times MRI findings concurred with surgical findings with respect to a structure being normal or abnormal. In structures where there was agreement that pathology was present, concordance was defined as the percentage of times MRI findings concurred with the exact surgically assessed pathology.

RESULTS Twenty dogs met the inclusion criteria. Mean (
SD) age at admission was 4.9
2.8 years. Breeds were Labrador Retriever (n ¼ 7), mixed breed (5), Vizsla (2), German Shorthair Pointer (2), and 1 each of Rottweiler, Flat Coated Retriever, American Bulldog, King Charles Cavalier Spaniel, and Miniature Poodle. Shoulder radiographs were available for review in 20 dogs. Radiographic findings were normal in 7 dogs. Abnormal findings were mild degenerative disease (n ¼ 7), mineralization within the supraspinatus tendon (3), osteophytosis of the bicipital intertubercular groove (2), mineralization of the infraspinatus tendon insertion (1), ununited infraglenoid tubercle (1), lateral luxation (1), and ectopic calcification in distal biceps tendon sheath (1). MRI examination (21 dogs) and positive contrast MR arthrography (11 dogs) were performed without complications. MR arthrography was considered essential in defining pathology in 3 dogs and substantially enhanced the pathologic lesions in 3 other dogs. After arthroscopic surgery, 10 dogs also had open surgery for therapeutic reasons, allowing for evaluation of additional extraarticular structures. Findings from the physical examination, shoulder arthroscopy, and/or an open surgical procedures and biopsy yielded the following overall diagnoses: primary biceps brachii tendon pathology (n ¼ 3), biceps brachii tendon pathology with concurrent supraspinatus tendinopathy (5), medial instability (6), medial compartment pathology without instability (1), osteochondroma with biceps tendon impingement (1), lateral instability (3), infraspinatus muscle contracture (1), and osteonecrosis of the humeral head (1). Values for agreement and concordance for each structure are listed in Table 1. More specific descriptions of the findings are given in the following sections.

Table 1. Agreement and Concordance Between MRI and Surgical Findings in 21 Dogs

Structure

No. in No. in Agreement/No. % Agree- Concordance/ % ConcorExamined ment No. Examined dance

Biceps Subscapularis MGHL LGHL Infraspinatus Impingement

19/21 20/21 16/19 7/8 4/4 19/21

90 95 84 88 100 90

16/16 8/13 10/12 3/3 4/4 5/5

100 62 83 100 100 100

Total

85/94

90

46/53

87

All structures listed had both an MRI evaluation and a surgical assessment conducted. The number examined varied between structures as not all tendons and ligaments were seen at every surgery. MRI, magnetic resonance imaging; MGHL, medial glenohumeral ligaments; LGHL, lateral glenohumeral ligaments.

Intra-Articular Pathology Biceps Brachii Tendon. Biceps brachii tendon pathology was observed arthroscopically in 16 dogs. Primary inflammation was noted in 3 dogs. Five dogs had biceps inflammation and/or tendon disruption because of impingement by an enlarged and/or calcified supraspinatus tendon (Fig 1). The remaining dogs had biceps tendon inflammation (6), or partial tearing (2) secondary to other pathologies including instability, osteochondroma, and humeral head osteonecrosis. All 21 biceps tendons could be readily identified with MRI. Agreement occurred in 19 of 21 dogs. In the 2 dogs without agreement, the MRI was interpreted as mild inflammation in the biceps tendon that could not be confirmed surgically. Concordance occurred in 16 of 16 dogs with documented surgical pathology. On MRI examination, inflammation of the biceps tendon appeared as an enlarged tendon with heterogeneous signal intensity on all 3 sequences, and as an increase in signal intensity on GE and PD images. Figure 2 illustrates an MR arthrogram and arthroscopic images of a partially avulsed biceps tendon. The diagnostic image of partial tendon avulsion was enhanced with MR arthrography. Subscapularis Tendon. Subscapularis tendon pathology was identified surgically in 13 dogs, with a full tendon avulsion in 1, partial tearing in 10, and tendon inflammation in 2 dogs. Arthroscopic examination was limited to the tendon insertion and the distal aspect of the tendon whereas the complete musculotendinous unit could be evaluated on all 21 MRI examinations. Agreement between surgical and MRI findings occurred in 20 of 21 dogs. In the 1 dog without agreement, changes thought to be within the cranial aspect of the MGHL on MRI were, in fact, within the overlaying subscapularis tendon at surgery. Concordance occurred in 8 of 13 abnormal

634

MRI FINDINGS IN DOGS WITH SHOULDER PATHOLOGY

Fig 1. Impingement of a biceps tendon by an enlarged supraspinatus tendon (ST). (A) Arthroscopic image of a cranial/lateral compartment of a left shoulder joint showing an enlarged ST impinging on the biceps tendon (BT). Note the associated biceps tendonitis at the site of contact. (B) Sagittal gradient echo (GE) magnetic resonance imaging (MRI) of the same case with a grossly enlarged ST (blue arrowheads) pressing down on the biceps tendon (white arrow). Magic angle artifact is identified by the red arrow. (C) Sagittal GE MRI of a normal ST (white arrowheads) and biceps tendon (white arrow). Sc, scapula; H, humeral head.

joints. In 4 abnormal joints without concordance, the MRI diagnosis was inflammation whereas partial tearing was identified surgically. In the fifth joint without concordance, a lesion in the subscapularis tendon was interpreted as being in the overlaying MGHL as noted earlier. On MRI examination, subscapularis tendon pathology appeared as heterogenous signal intensity on PD and GE sequences with enlargement of the tendon in cases of inflammation and fiber disruption, particularly along the tendon border in cases of tearing (Fig 3). MR arthrography further enhanced the finding of irregular tendon borders in cases with partial tearing and helped to confirm the single case of full tendon avulsion by highlighting the void between tendon and bone. Medial Glenohumeral Ligament. MGHL pathology was identified surgically in 12 dogs. Three dogs had inflammation of the capsuloligamentous region, 6 had partial tearing of the ligament, and 3 had complete or near complete tearing of the ligament. The MGHL could be identified on 19 MRI studies. Two early MRI studies did not include the dorsal plane, which prevented evaluation of MGHL. Agreement occurred in 16 of 19 dogs. In 2 dogs without agreement, MRI findings were considered normal when mild inflammation and partial tearing were

surgically identified, respectively. The dog with the undiagnosed partial tear was a 4.5 kg Miniature Poodle. In the third dog without agreement, MRI findings were suggestive of inflammation but this was not observed surgically. Concordance occurred in 10 of 12 abnormal dogs. MRI findings consisted of capsuloligamentous thickening in inflammatory cases with heterogeneous signal intensity on both PD and GE sequences. Partial tears were identified by minor disruptions of the tendon silhouette. More complete tearing was seen as the major disruption of the tendon silhouette often with an irregular, undulating pattern on positive contrast MR arthrography (Fig 4). Lateral Glenohumeral Ligament. The LGHL was observed arthroscopically in 8 dogs. Pathology was present in 3 dogs with partial tearing of the ligament in 2 dogs, and full tearing with lateral humeral head luxation in 1 dog. On MRI examination, the LGHL was identified in all the 8 dogs. Agreement occurred in 7 of 8 dogs. In the 1 dog without agreement, the MRI diagnosis of inflammation could not be confirmed surgically. Concordance occurred in 3 of 3 abnormal dogs. MRI findings consisted of capsuloligamentous thickening in inflammatory cases with concurrent ligament/joint

Fig 2. Partial avulsion of a biceps tendon. (A) Arthroscopic image of the cranial/medial compartment of a left shoulder joint showing a partial avulsed biceps tendon (arrowhead). The remaining tendon, identified by the yellow arrow, is still attached intact. (B) Sagittal proton density fat saturation (PD) magnetic resonance (MR) arthrogram with enlargement of the biceps tendon and a filling defect evident along the scapular attachment site (blue arrowheads). (C) Normal sagittal PD MR arthrogram for comparison with the white arrow at the biceps tendon insertion. Sc, scapula; H, humeral head.

MURPHY ET AL

635

Fig 3. Enlargement and partial tearing of the subscapularis tendon. (A) Arthroscopic view of the medial compartment of a left shoulder joint showing disruption of the normal subscapularis fiber pattern (blue arrowheads). Mild fraying of the medial glenohumeral ligament (MGHL) is also evident. (B) Dorsal proton density fat saturation (PD) magnetic resonance imaging (MRI) showing gross enlargement, heterogeneity, and loss of the finite silhouette of the subscapularis tendon (blue arrowheads). Sc, scapula; H, humeral head. (C) Dorsal PD MRI of a normal shoulder. Note the size and homogeneity of the subscapularis tendon (white arrows). Both (B) and (C) MR images were prearthrogram.

capsule disruption in partial or full tears usually at or near the scapular attachment (Fig 5). With MR arthrography, joint distension further enhanced observation of the ligament/capsular disruption. Supraspinatus Tendon. The caudal/medial aspect of the supraspinatus insertion was assessed arthroscopically; although this tendon is not intra-articular, enlargement of this tendon can be identified on arthroscopy. Primary tendon enlargement, causing compression or impingement of the biceps tendon was identified surgically in

5 dogs; 2 dogs were considered severe, 1 moderate, and 2 mild. When evaluating solely for arthroscopically evident tendon enlargement/impingement, agreement occurred in 19 of 21 dogs. In the 2 dogs without agreement, the MRI diagnosis of mild impingement could not be confirmed at surgery. Concordance occurred in 5 of 5 dogs. Supraspinatus pathology other than impingement was noted on MRI in 8 dogs. These changes were confirmed with open surgery in only 1 dog. The diagnosis in these 8 dogs included dystrophic mineralization and

Fig 4. Severe tearing of a medial glenohumeral ligament (MGHL). Arthroscopic images of the (A) cranial/medial, (B) medial, and (C) caudal/medial compartments of a left shoulder of a dog with moderate disruption of the cranial (Cr) aspect of the ligament and complete tearing of the caudal (Cd) aspect of the ligament. (D) Dorsal gradient echo (GE) magnetic resonance imaging (MRI) showing enlargement and heterogeneity of the MGHL (blue arrowhead). Also note enlargement and heterogeneity of the infraspinatus tendon (white ) and hyperintensity of a portion of the infraspinatus muscle (black ). Sc, scapula; H, humeral head. (E) Dorsal T1 magnetic resonance (MR) arthrogram of the same shoulder showing enlargement, heterogeneity, and an undulating pattern to the MGHL (red arrowhead). (F) Dorsal T1 MR arthrogram of a normal shoulder. The white arrow identifies a normal MGHL.

636

MRI FINDINGS IN DOGS WITH SHOULDER PATHOLOGY

Fig 5. Tearing of the lateral glenohumeral ligament (LGHL). (A) Arthroscopic image of the cranial/lateral compartment of a right shoulder as seen from a caudal lateral portal. The arrowhead identifies frayed portions of the LGHL. (B) Dorsal proton density fat saturation (PD) magnetic resonance imaging (MRI) showing distension of the lateral joint capsule with disruption of the LGHL (blue arrowheads). Also note the hyperintensity surrounding the infraspinatus tendon (white ) and within the bicipital sheath (black ). Sc, scapula; H, humeral head. (C) Dorsal PD MRI of a normal shoulder. The white arrow identifies a normal LGHL.

tendon enlargement that was not evident arthroscopically. MRI findings for supraspinatus pathology were heterogeneous signal intensity on all 3 sequences and tendon enlargement with intermixed hypointensity if mineralization was present (Fig 1). On positive contrast arthrography, a lack of contrast filling between the biceps and the supraspinatus tendon on the sagittal view was observed when there was impingement along with medial displacement of the tendon on transverse images. Extra-Articular Pathology Infraspinatus. Infraspinatus pathology was identified on MRI in 4 dogs. All dogs had an open surgical procedure. These procedures were tenotomy for contracture (1), debulking of tendon mineralization that impinged on the acromion with limb abduction (1), and medial and lateral joint stabilization (2). In 3 of these dogs, the tendon was directly observed and MRI findings had 100% agreement and concordance with the surgical findings. Specific MRI findings for contracture were heterogeneous signal intensity within the muscle and at the muscle– tendinous junction with minimal extension into the tendon, best appreciated on the sagittal imaging planes. Acute injury appeared as an enlarged tendon with heterogeneous signal intensity and a hyperintensity within the muscle on GE or PD sequences because of suspected edema (Fig 4). Teres Minor. The teres minor was identified on all 21 MRI examinations. MRI-assessed pathology included muscle inflammation in 2 dogs, both of which had medial instability. Inflammation appeared as an enlargement with a hyperintensity on GE or PD sequences; however, no open or arthroscopic procedure was used to validate these findings. The teres minor was noted to be normal in 1 open procedure and MRI results were in agreement with this finding.

DISCUSSION In 21 dogs that had surgical exploration for shoulder pain, we found a high level of agreement and concordance between MRI and surgical findings. Of the 94 structures examined at surgery, there was 90% agreement between MRI and surgical findings, and overall concordance for 53 abnormal structures was 87%. Although we did not examine the direct relationship between diagnostic MRI and therapeutic outcome, our results show that preoperative MRI can define normal and abnormal soft tissue structures in the shoulder joint of dogs with lameness attributable to the shoulder region. The effect of preoperative MRI on treatment modification in canine shoulder lameness is unknown. By comparison, in a meta-analysis (265 papers) of the efficacy of MRI of the human shoulder, MRI altered the primary diagnosis in 23–68% of cases and management plans were subsequently modified in 15–61% of cases.17 When considering the clinical application of a new diagnostic tool, concern should be raised for the potential to misinterpret the information obtained. For biceps tendon inflammation, MRI findings were over interpreted in 2 dogs; in both cases, dogs were considered to have mild inflammatory changes on MRI. These changes were either incorrectly identified or so mild as to remain within the body of the tendon and thus were not evident during arthroscopy. Over interpretation was also noted in 1 dog with MGHL inflammation, 1 dog with LGHL inflammation, and 2 dogs with biceps tendon impingement. Discrepancies for the diagnosis of impingement between MRI and arthroscopy may be a factor of the reference standard. In our study, the ‘‘gold standard’’ for impingement classification was arthroscopic examination; however, joint distension during arthroscopy may change the spatial relationship between these structures creating a false negative on examination.

MURPHY ET AL

Although abnormalities in the MGHL and the subscapularis tendon were easily identified on MRI, there was a tendency to under diagnose the severity of the lesion in these structures. In 1 dog where a partial tear was not identified in the subscapularis tendon, the changes seen on MRI were attributed to pathology in the overlying MGHL. These structures are in close proximity and a lack of appropriate joint distension during arthrography may lead to misinterpretation of lesion location. Evaluation of MR images can also be challenging in smaller dogs. A partial tear of the MGHL was not identified by MRI in a toy poodle; however, a subscapular tendon avulsion was correctly identified in this same dog. Caution should also be used with interpretation of MRI of the supraspinatus tendon because there is a tendency for it to appear hyperintense in both normal and abnormal cases on PD and GE images. The cause of this hyperintensity is unknown but may be because of significantly higher collagen content at the insertion relative to the infraspinatus and subscapularis tendons.18 The role that supraspinatus tendon disease may play in thoracic limb lameness is also debatable because supraspinatus tendon calcification and tendonosis have been identified in dogs without clinical signs.5,19 Caution should also be used when interpreting changes in the biceps tendon as creation of a magic angle artifact is common. The magic angle artifact occurs as the tendon crosses the intertubercular groove and the tendon fibers approach a 551 angle to the static magnetic field. An increase in signal is appreciated on T2-weighted images with a low echo time and should not be misinterpreted as pathology (Fig 1).11

MR Arthrography Direct MR arthrography is used commonly in people to determine the integrity of intra-articular ligamentous and fibrocartilaginous structures.20 In the only clinical report of direct MR arthrography of the canine shoulder, the technique was considered unrewarding for evaluation of osteochondrosis lesions. It should be noted that the gadopentetate dimeflumine used as a contrast agent in that study was used undiluted. Dilution of the contrast agent may be necessary to prevent the hyperintensity of the agent from obscuring subtle lesions on the images. In a cadaver study, a dilution of 1:1200 was optimal for direct MR arthrography of the canine shoulder.21 We found that MR arthrography was helpful for identifying tendon avulsion (partial or full) and severe tendon/ligament tears. Arthrography was also useful in defining the potential space between the biceps and supraspinatus tendons when impingement was suspected. Our greatest challenge with arthrography was achieving consistent filling of the medial compartment.

637

All dogs in our study had more than 1 abnormality in the shoulder joint as identified by both surgery and MRI; many had multiple pathologic changes. For instance, subscapularis tendon pathology occurred in 57% of cases. Whereas 1 dog had a primary tendon avulsion, tearing of the cranial aspect of the subscapularis tendon was noted in all 5 dogs diagnosed with MGHL-related medial instability. These findings were similar to a report by Cook et al where subscapularis tendon tears were noted in 86% of 43 dogs with medial instability.3 In our study, subscapularis tendon pathology was also present with biceps tendonitis (2), osteonecrosis (1), infraspinatus contracture (1), medial compartment pathology (1), and lateral instability (1). The relationship of subscapularis pathology with causation of lameness or clinical outcome has not been determined. Extra-Articular Pathology MRI permitted more complete evaluation of extraarticular pathology than arthroscopic surgery alone. The primary pathologic lesion was considered extra-articular in 15% of our dogs, which is much higher than previously reported. When cases of joint fracture and OCD were eliminated, Bardet identified only 0.4% cases with an extra-articular diagnosis whereas Cook reported 9% with extra-articular tendonopathy.9,10 It is important to note that musculoskeletal ultrasonography was used in the second study. Ultrasonography is useful in evaluating extra-articular structures on the lateral and cranial/ medial aspect of the shoulder joint22; however, ultrasonography is limited in its ability to completely evaluate the medial compartment. We identified secondary extraarticular pathology of the supraspinatus, infraspinatus, subscapularis, or teres minor muscles in conjunction with primary intra-articular injury in 12 of 21 dogs. Injury to more than one structure, as seen in Fig 4, illustrates the complexity of the shoulder joint and the challenge in determining both the exact cause of pain/lameness and appropriate therapeutic modality. Study Limitations Our study is limited by its retrospective, nonblinded nature, and small sample size. Even though human trials of MRI-assessed rotator cuff pathology had little change when radiologists were unblinded and provided with arthroscopy results,23 we chose not to calculate predictive values as we felt our protocol was inadequate for validation. Our study was not blinded because of a lack of publications describing MRI findings in canine shoulder disease and therefore lack of current knowledge in interpreting canine shoulder MRI. Histopathology of tendon and ligament pathology may have helped to more

638

MRI FINDINGS IN DOGS WITH SHOULDER PATHOLOGY

accurately understand and correlate MRI findings with the surgically documented changes. This study demonstrates the potential of MRI in the field of canine musculoskeletal disease. MR imaging allows for an accurate assessment of both intra- and extraarticular joint pathology. Future studies are needed to determine how this additional diagnostic information should influence therapeutic choices and modalities. REFERENCES 1. Bardet JF: Diagnosis of shoulder instability in dogs and cats: a retrospective study. J Am Anim Hosp Assoc 34:42–54, 1998 2. Sidaway BK, McLaughlin RM, Elder SH, et al: Role of the tendons of the biceps brachii and infraspinatus muscles and the medial glenohumeral ligament in the maintenance of passive shoulder joint stability in dogs. Am J Vet Res 65:1216–1222, 2004 3. Cook JL, Tomlinson JL, Fox DB, et al: Treatment of dogs diagnosed with medial shoulder instability using radiofrequency-induced thermal capsulorrhaphy. Vet Surg 34:469– 475, 2005 4. Mitchell RA, Innes JF: Lateral glenohumeral ligament rupture in three dogs. J Small Anim Pract 41:511–514, 2000 5. Laitinen OM, Flo GL: Mineralization of the supraspinatus tendon in dogs: a long term follow-up. J Am Anim Hosp Assoc 36:262–267, 2000 6. Bruce WJ, Spence S, Miller A: Teres minor myopathy as a cause of lameness in a dog. J Small Anim Pract 38:74–77, 1997 7. Stobie D, Wallace LJ, Lipowitz AJ, et al: Chronic bicipital tenosynovitis in dogs: 29 cases (1985–1992). J Am Vet Med Assoc 207:201–207, 1995 8. Bennett AR: Contracture of the infraspinatus muscle in dogs: a review of 12 cases. J Am Anim Hosp Assoc 22:481–487, 1986 9. Bardet JF: Uncommon shoulder lameness in dogs and cats, in Proceedings of the 2nd World Veterinary Orthopedic Congress, Keystone, CO, February 25–March 4, 2006 10. Cook J: Multicenter data on epidemiology, diagnostics, and treatment for shoulder disease in dogs, in Proceedings of

11. 12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

the American College of Veterinary Surgeons Symposium, Washington, DC, October 5–7, 2006 Stark DD, Bradley WG (eds). Magnetic Resonance Imaging, Vol II (ed 3). St. Louis, MO, Mosby, 1999 Schaefer SL, Forrest LJ: Magnetic resonance imaging of the canine shoulder: an anatomic study. Vet Surg 35:721–728, 2006 Van Bree H, Degryse H, Van Ryssen B, et al: Pathologic correlations with magnetic resonance images of osteochondrosis lesions in canine shoulders. J Am Vet Med Assoc 202:1099–1105, 1993 Van Bree H, Van Ryssen B, Ramon F: Magnetic resonance arthrography of the scapulohumeral joint in dogs using gadopentetate dimeglumine. Am J Vet Res 56:286–288, 1995 Fransson BA, Gavin PR, Lahmers KK: Supraspinatous tendinosis associated with biceps brachi tendon displacement in a dog. J Am Vet Med Assoc 227:1429–1433, 2005 Beale BS, Hulse DA, Schulz KS, et al: (eds). Small Animal Arthroscopy. Philadelphia, PA, Saunders, 2003, pp 24–49 Bearcroft PWP, Blanchard TK, Dixon AK, et al: An assessment of the effectiveness of magnetic resonance imaging of the shoulder: literature review. Skelet Radiol 29:673–679, 2000 Fan L, Sarkar K, Franks DJ, et al: Estimation of total collagen and types I and III collagen in canine rotator cuff tendons. Calcif Tissue Int 61:223–229, 1997 Muir P, Johnson KA: Supraspinatus and biceps brachii tendinopathy in dogs. J Small Anim Pract 35:239–243, 1994 Sahin G, Demirtas M: An overview of MR arthrography with emphasis on the current technique and applicational hints and tips. Eur J Radiol 53:416–430, 2006 Gerbig J, Schaefer SL: Magnetic resonance arthrography of the canine shoulder joint: evaluation of the diagnostic potential, in Proceedings of the Veterinary Orthopedic Society Conference, Sun Valley, ID. March 3–10, 2007. Long CD, Nyland TG: Ultrasonographic evaluation of the canine shoulder. Vet Radiol Ultrasound 40:372–379, 1999 Balich SM, Sheley RC, Brown TR, et al: MR imaging of the rotator cuff tendon: interobserver agreement and analysis of interpretive errors. Radiology 204:191–194, 1997