Hugo C. van der Veen James P. M. Collins Paul C. Rijk

Size: px

Start display at page:

Download "Hugo C. van der Veen James P. M. Collins Paul C. Rijk"

Annabel Melton
6 years ago
Views:

1 DOI /s ARTHROSCOPY AND SPORTS MEDICINE Value of magnetic resonance arthrography in post-traumatic anterior shoulder instability prior to arthroscopy: a prospective evaluation of MRA versus arthroscopy Hugo C. van der Veen James P. M. Collins Paul C. Rijk Received: 11 December 2010 Ó Springer-Verlag 2011 Abstract Purpose This prospective study was designed to evaluate the value of magnetic resonance arthrography (MRA) after traumatic anterior shoulder instability prior to arthroscopy. Methods Patients included had two or more shoulder dislocations, at least the first being traumatic. MRA images were scored for Hill Sachs lesions, superior labral anterior posterior (SLAP) lesions, rotator cuff tears, glenohumeral ligament (GHL) lesions and Bankart lesions. Consequently, a standardized shoulder arthroscopy was performed. Five surgeons were involved in the study, initially blinded to the MRA results. MRA and arthroscopic findings were compared. Interobserver agreement was calculated by using Cohen s Kappa coefficients (j). Results Eighteen patients (13 male, 5 female) were included (mean age 26.1 years). Hill Sachs lesions demonstrated fair agreement (j = 0.33) whereas for SLAP lesions moderate agreement was calculated (j = 0.43). On MRA, four partial thickness rotator cuff lesions were seen, not being stated by arthroscopy. GHL lesions were described on MRA in 15 patients; only two patients turned out to have GHL lesions at arthroscopy. Two arthroscopically diagnosed Bankart lesions which needed surgical treatment were not detected by MRA (moderate agreement, j = 0.47). H. C. van der Veen (&) P. C. Rijk Department of Orthopaedic Surgery, Medical Center Leeuwarden, P.O. Box 888, 8901 BR Leeuwarden, The Netherlands hcvanderveen@hotmail.com J. P. M. Collins Department of Radiology, Medical Center Leeuwarden, Leeuwarden, The Netherlands Conclusions In patients with post-traumatic anterior glenohumeral instability MRA shows many lesions that can not be confirmed by arthroscopy and therefore do not have therapeutical consequences. On the other hand some labral lesions which do need surgical treatment are not detected on MRA. At least from this study, it can be concluded that MRA has limited value prior to the arthroscopic treatment of post-traumatic shoulder instability. Keywords Post-traumatic glenohumeral instability Magnetic resonance arthrography Arthroscopy Introduction Glenohumeral instability due to a traumatic dislocation of the shoulder joint is a relatively common injury. Incidences vary between 0.08 and 0.24 per 1,000 person-years, being higher in young and active people [16]. In 95 percent of the cases the humeral head is dislocated anteriorly, often coinciding with intra-articular lesions [9]. Structures involved are the labrum, the rotator cuff, the posterolateral aspect of the humeral head (Hill Sachs lesion) and the glenohumeral ligaments (GHL). Lesions of the anterior inferior labrum, called Bankart lesions, often lead to anterior shoulder instability. The superior labral anterior posterior (SLAP) lesions are also clinically relevant, as these lesions can cause an unstable glenoid insertion of the biceps tendon, which could lead to impingement or even cuff lesions [25]. In our diagnostic work-up in patients with traumatic shoulder instability, often magnetic resonance arthrography (MRA) was performed prior to arthroscopy. We designed a prospective study to evaluate the value of MRA after traumatic anterior shoulder instability prior to arthroscopy of the shoulder.

2 Methods Patients included had two or more shoulder dislocations, at least the first being traumatic. Exclusion criteria were previous shoulder surgery, bony Bankart lesions on X-rays, generalized joint laxity and skeletal immaturity. MRA was performed in all included patients, using an ultrasound guided anterior approach to introduce the contrast (gadolinium, Artirem Ò, Guerbet) into the shoulder joint. A 2-channel flex m coil was used in all cases (Philips intera 1.5 T). The MRI protocol consisted of 3d T1 TFE (turbo fast echo) sequence with fat suppression (SPIR). Reconstructions in three planes (axial, oblique coronal and saggital) were reconstructed from the data with a slice thickness of 1 mm, gap 0, matrix , field of view , and reconstruction voxel size mm. A T2 SE (spin echo)- weighted oblique coronal sequence with fat suppression was also performed with a slice thickness of 3 mm, gap 0.3 mm, matrix , voxel size mm and a field of view All examinations were performed and reviewed by one dedicated musculoskeletal radiologist (JC). The MR images were read and scored according to a fixed order of anatomical structures of the glenohumeral joint for the existence of (1) Hill Sachs lesion, (2) SLAP lesions, (3) rotator cuff tears, (4) GHL lesions and (5) Bankart lesions. SLAP lesions were graded as four types, according to the Snyder-classification [23]. Consequently, a shoulder arthroscopy was performed. The surgeon was initially blinded to the MRA outcome. Patients were all positioned in lateral decubitus position, using a mid-anterior and posterior portal. The bursal side of the rotator cuff was not routinely checked. Five surgeons were involved in performing the procedures. According to a standardized protocol, the same list of anatomical structures was checked as was used by the radiologist (Table 1). Before proceeding with a capsulolabral reconstruction, MRA findings and observations during arthroscopy were compared. Interobserver agreement for detection of the several lesions was calculated by using Cohen s Kappa (j) coefficients [13]. The interpretation is as follows: 0 = poor, = slight, = fair, = moderate, = substantial and = almost perfect. In cases where j could not be defined, overall agreement was calculated. Results Eighteen patients (13 male, 5 female) with anterior glenohumeral instability meeting the inclusion criteria were treated in our hospital between 2005 and No patient Table 1 Overview of intra-articular findings on MRA and at arthroscopy Pathology MRA (number of patients) Arthroscopy (number of patients) Interobserver agreement (j) Hill Sachs Positive (fair) Negative 5 9 Small (\1 cm) 4 2 Large ([1 cm) 9 7 SLAP Normal (moderate) Cuff Normal PT 4 0 GHL Normal 3 16 Superior 5 1 Middle 9 0 Inferior 5 1 Bankart Positive (moderate) Negative 4 3 SLAP Superior Labral Anterior Posterior, PT partial thickness, GHL glenohumeral ligament had more than six dislocations. The average age was 26.1 years (range years). The average time between MRA and arthroscopy was 13.1 weeks (range 5 65 weeks). No subluxations or dislocations occurred during the interval between MRA and arthroscopy. The findings on MRA and at arthroscopy for the different anatomical intra-articular structures are presented in Table 1. No chondral lesions were found on MRA and at arthroscopy. Thirteen Hill Sachs lesions were described on MRA, whereas only nine were observed during arthroscopy (j = 0.33, fair agreement). The incidence of Hill Sachs lesions was 50% (9/18). SLAP lesions showed moderate agreement (j = 0.43). Seven of ten type 2 SLAP lesions on MR could be confirmed at arthroscopy. Two of nine arthroscopically diagnosed type 2 SLAP lesions were not seen on MRA (Fig. 1). On the other hand MRA described two cases of type 2 lesions which could not be confirmed by arthroscopy. The type 1 SLAP lesion observed on MRA was not demonstrated at arthroscopy, whereas the SLAP lesion, scored as type 1 at arthroscopy was graded as type 2 on MRA. The incidence of SLAP lesions was 56% (10/18).

arthroscopically (overall agreement 0.78). Glenohumeral ligament lesions were described on MRA in 15 patients, whereas only two patients turned out to have GHL lesions at arthroscopy.

3 Fig. 1 Coronal T2 weighed MRA image of the right shoulder in a patient were the biceps anchor was described as normal (arrow). Arthroscopy however, showed a SLAP 2 lesion On MRA, four partial thickness rotator cuff lesions were seen, all of them described as partial subscapular tendon lesions, which could not be confirmed arthroscopically (overall agreement 0.78). Glenohumeral ligament lesions were described on MRA in 15 patients, whereas only two patients turned out to have GHL lesions at arthroscopy. One arthroscopically confirmed inferior GHL lesion was described as a middle GHL lesion on MRA. The other GHL lesion at arthroscopy concerned the middle GHL. This lesion was defined as an inferior and middle GHL lesion on MRA (overall agreement 0.28). The incidence of GHL lesions was 11% (2/18). Bankart lesions were described on MRA in 14 cases (Fig. 2). At arthroscopy 15 lesions were seen, two of them being undetected by MRA (j = 0.47, moderate agreement). The incidence of Bankart lesions was 83% (15/18). Fig. 2 Coronal (a) and axial (b) T2 weighed MRA images of the left shoulder in a patient where a subtle Bankart lesion was described (arrows). At arthroscopy, no Bankart lesion could be detected Discussion MRA evaluation in patients with shoulder instability is a diagnostic modality frequently used. The application of intra-articular contrast has led to an increase in accuracy of MRA. especially in detecting labral lesions [7]. However, nowadays arthroscopy still has to be considered as the gold standard in detecting intra-articular shoulder pathology. It has been mentioned that accuracy of MRA in detecting shoulder pathology might depend on the experience of the interpreting radiologist. Reuss et al. [21] found out that community radiologists were not as accurate as musculoskeletal radiologists. In our study, MRA images were performed and reviewed by a musculoskeletally dedicated radiologist. On MRA and during arthroscopy no glenoid chondral lesions were detected. This is in agreement with Page and Bhatia who demonstrated that these lesions are uncommon in anterior glenohumeral dislocations [17]. Concerning other intra-articular lesions, Hill Sachs lesions showed only fair agreement between MRA and arthroscopy findings. It is known that MRA is not the first choice in detecting bony pathology. For visualizing a Hill Sachs lesion, CT could possibly be a more suitable diagnostic modality. The incidence of Hill Sachs lesions was 50% (9/ 18). In literature, incidences vary between 87 and 100% [3, 15, 22], based on observations during early arthroscopy after first traumatic anterior dislocations. Yiannakopoulos et al. [28] describe an incidence of 65% in acute shoulder instability and 93% in chronic shoulder instability. Compared with these data, the incidence of Hill Sachs lesions in our study is remarkable low. It has been demonstrated that MRA of the shoulder is reliable and accurate for the detection of SLAP lesions [11]. In our study, SLAP lesions, showing moderate agreement between MRA and arthroscopy, were scored according to the Snyder-classification which does not

4 correspond with the severity of the SLAP lesion. In clinical practice, only type 2 and 4 SLAP lesions have therapeutic consequences, namely the requirement of refixation of the detached biceps insertion on the glenoid. Except for one type 1, all lesions at arthroscopy were SLAP 2 lesions. Type 1 and 2 lesions combined with a Bankart lesion are often associated with acute or chronic instability [25]. This is in accordance with the findings in our population with anterior instability where type 3 and 4 tears were not observed. We found an incidence for SLAP lesions of 56%. This is slightly higher than incidences reported by previous studies. Kirkley et al. [12] found an incidence of SLAP lesions in acute shoulder instability of 44%. However, Yiannakopoulos et al. [28] described a much lower incidence of 22% in patients with chronic shoulder instability (20% in acute shoulder instability). Most of their patients were military cadets with a mean age of 24 years. Patient characteristics could be an explanation for this lower incidence of SLAP lesions. No rotator cuff lesions were seen at arthroscopy. MRA suggested in four patients a partial subscapular tendon tear. An overall agreement of 0.83 seems high, but says nothing about specificity of the MRA in detecting cuff lesions, because in our group there were no cuff lesions at all arthroscopically. It must be noted here, that the bursal side of the rotator cuff was not routinely checked during surgery. Several studies have reported that the incidence of cuff lesions increases with age [1, 26, 27]. In patients older than 40 years, the incidence of a rotator cuff tear after the first traumatic anterior dislocation is 15%, whereas in patients older than 60 years the incidence climbs to 40% [10]. Our population had a mean age of 26.1 years, so that could be an explanation for the low incidence of cuff lesions. MRA is known to be suitable for detecting GHL lesions [6, 18 20]. In our series, GHL lesions were seen significantly more often on MRA than at arthroscopy. Glenohumeral ligaments serve as static stabilizers of the glenohumeral joint. It has been demonstrated that rupture of the inferior GHL is often associated with anteroinferior instability [5]. Because the inferior GHL originates from the inferior labrum and inserts on the humeral neck, it might be difficult on MRA to differentiate between GHL lesions and labral lesions [8]. However, the clinical importance of differentiating pre-operatively between these anatomical structures is questionable because in both situations a refixation of the labro-ligamentous complex has to be performed. As a disruption of the labro-ligamentous complex, Bankart lesions form an important factor in traumatic anterior instability. MRA has been proven to be an appropriate diagnostic tool in determining these lesions [2, 4, 11, 18]. Kirkley et al. [12] investigated 16 patients in a similar study format, showing perfect agreement (j = 1.0) for the detection of Bankart lesions. In our study, these lesions showed the highest inter-observer agreement of all scored parameters (j = 0.47, moderate agreement). Nevertheless, two Bankart lesions which needed surgical repair were not observed on MRA. Furthermore, the difficult distinction between inferior GHL and Bankart lesions, as previously stated, could have been interfering with the total number of Bankart lesions. The incidence of Bankart lesions was 83%. In literature, incidences are described between 72 and 97% [3, 22, 24]. To date, the reliability of MRA in describing humeral avulsions of the glenohumeral ligament (HAGL) and anterior labroligamentous periosteal sleeve avulsion (AL- PSA) lesions has to be established. Melvin et al. [14] illustrated the difficulty in distinguishing HAGL from other abnormalities of the glenohumeral ligament complex with MRI and stated that the diagnosis of HAGL should be reserved for arthroscopy. Therefore, HAGL and ALPSA lesions were not evaluated in our protocol. This study has some limitations we would like to address. First, not all operations were performed by the same surgeon. This could have interfered with our results. However, we think this influence has been overcome due to the fact that the anatomical items were systematically scored in a fixed order using a uniform scoring list by all investigators. Furthermore, the variation in length of the interval between MRA and arthroscopy (5 65 weeks) has to be addressed, as well as the fact that an interval of 65 weeks is quite large. Because none of the patients had subluxations, dislocations or other traumatic events during the period between MRA and arthroscopy we believe this did not lead to alterations in outcome of any significance. Conclusions In our study, we found a limited value of MRA in detecting clinically relevant intra-articular lesions after traumatic anterior dislocations. In patients with post-traumatic anterior glenohumeral instability MRA shows many lesions that cannot be confirmed by arthroscopy and therefore do not have therapeutic consequences. On the other hand some labral lesions which do need surgical treatment are not detected on MRA. Therefore, it can be concluded, at least from this study, that MRA has limited value prior to the arthroscopic treatment of post-traumatic shoulder instability. Acknowledgment We thank Dr. Henk Groen, epidemiologist, for the statistical analysis of the data.

5 References 1. Antonio GE, Griffith JF, Yu AB, Yung PS, Chan KM, Ahuja AT (2007) First-time shoulder dislocation: High prevalence of labral injury and age-related differences revealed by MR arthrography. J Magn Reson Imaging 26: Applegate GR, Hewitt M, Snyder SJ, Watson E, Kwak S, Resnick D (2004) Chronic labral tears: value of magnetic resonance arthrography in evaluating the glenoid labrum and labral bicipital complex. Arthroscopy 20: Baker CL, Uribe JW, Whitman C (1990) Arthroscopic evaluation of acute initial anterior shoulder dislocations. Am J Sports Med 18: Bencardino JT, Beltran J (2006) MR imaging of the glenohumeral ligaments. Radiol Clin North Am 44: Burkart AC, Debski RE (2002) Anatomy and function of the glenohumeral ligaments in anterior shoulder instability. Clin Orthop Relat Res 400: Chaipat L, Palmer WE (2006) Shoulder magnetic resonance imaging. Clin Sports Med 25: Chandnani VP, Yeager TD, DeBerardino T et al (1993) Glenoid labral tears: prospective evaluation with MRI imaging, MR arthrography, and CT arthrography. Am J Roentgenol 161: Dodson CC, Cordasco FA (2008) Anterior glenohumeral joint dislocations. Orthop Clin North Am 39: Goss TP (1988) Anterior glenohumeral instability. Orthopedics 11: Henry JH, Genung JA (1982) Natural history of glenohumeral dislocation revisited. Am J Sports Med 10: Jee WH, McCauley TR, Katz LD, Matheny JM, Ruwe PA, Daigneault JP (2001) Superior labral anterior posterior (SLAP) lesions of the glenoid labrum: reliability and accuracy of MR arthrography for diagnosis. Radiology 218: Kirkley A, Litchfield R, Thain L, Spouge A (2003) Agreement between magnetic resonance imaging and arthroscopic evaluation of the shoulder joint in primary anterior dislocation of the shoulder. Clin J Sport Med 13: Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33: Melvin JS, Mackenzie JD, Nacke E, Sennett BJ, Wells L (2008) MRI of HAGL lesions: four arthroscopically confirmed cases of false-positive diagnosis. AJR Am J Roentgenol 191: Norlin R (1993) Intraarticular pathology in acute, first-time anterior shoulder dislocation: an arthroscopic study. Arthroscopy 9: Owens BD, Dawson L, Burks R, Cameron KL (2009) Incidence of shoulder dislocation in the United States military: demographic considerations from a high-risk population. J Bone Joint Surg Am 91: Page R, Bhatia DN (2010) Arthroscopic repair of a chondrolabral lesion associated with anterior glenohumeral dislocation. Knee Surg Sports Traumatol Arthrosc 18: Palmer WE, Brown JH, Rosenthal DI (1994) Labral ligamentous complex of the shoulder: evaluation with MR arthrography. Radiology 190: Parmar H, Jhankaria B, Maheshwari M et al (2002) Magnetic resonance arthrography in recurrent anterior shoulder instability as compared to arthroscopy: a prospective comparative study. J Postgrad Med 48: Reijnierse M (2009) MR arthrography of the shoulder: glenohumeral ligaments. JBR-BTR 92: Reuss BL, Schwartzberg R, Zlatkin MB, Cooperman A, Dixon JR (2006) Magnetic resonance imaging accuracy for the diagnosis of superior labrum anterior-posterior lesions in the community setting: eighty-three arthroscopically confirmed cases. J Shoulder Elbow Surg 15: te Slaa RL, Brand R, Marti RK (2003) A prospective arthroscopic study of acute first-time anterior shoulder dislocation in the young: a five-year follow-up study. J Shoulder Elbow Surg 12: Snyder SJ, Karzel RP, Del Pizzo W, Ferkel RD, Friedman MJ (1990) SLAP lesions of the shoulder. Arthroscopy 6: Taylor DC, Arciero RA (1997) Pathologic changes associated with shoulder dislocations. Arthroscopic and physical examination findings in first-time, traumatic anterior dislocations. Am J Sports Med 25: Waldt S, Burkart A, Lange P, Imhoff AB, Rummeny EJ, Woertler K (2004) Diagnostic performance of MR arthrography in the assessment of superior labral anteroposterior lesions of the shoulder. Am J Roentgenol 182: Yamaguchi K, Ditsios K, Middleton WD, Hildebolt CF, Galatz LM, Teefey SA (2006) The demographic and morphological features of rotator cuff disease. A comparison of asymptomatic and symptomatic shoulders. J Bone Joint Surg Am 88: Yamamoto A, Takagishi K, Osawa T et al (2010) Prevalence and risk factors of a rotator cuff tear in the general population. J Shoulder Elbow Surg 19: Yiannakopoulos CK, Mataragas E, Antonogiannakis E (2007) A comparison of the spectrum of intra-articular lesions in acute and chronic anterior shoulder instability. Arthroscopy 23:

Sensitivity and Specificity in Detection of Labral Tears with 3.0-T MRI of the Shoulder

Sensitivity and Specificity in Detection of Labral Tears with 3.0-T MRI of the Shoulder Magee and Williams MRI for Detection of Labral Tears Musculoskeletal Imaging Clinical Observations C M E D E N T U R I C L I M G I N G JR 2006; 187:1448 1452 0361 803X/06/1876 1448 merican Roentgen Ray