Temporal Evolution of MRI Findings After Arthroscopic Rotator Cuff Repair

Transcription

1 Musculoskeletal Imaging Original Research Crim et al. MRI of Arthroscopic Rotator Cuff Repair Musculoskeletal Imaging Original Research Julia Crim 1,2 Robert Burks 2 Betty J. Manaster 1 Christopher Hanrahan 1 Man Hung 2 Patrick Greis 2 Crim J, Burks R, Manaster BJ, Hanrahan C, Hung M, Greis P Keywords: rotator cuff MRI, rotator cuff repair, rotator cuff tear DOI: /AJR Received February 12, 2010; accepted after revision May 11, J. Crim is affiliated with Amirsys as an independent contractor. R. Burks is a consultant to Arthrex and is a board member of the Arthroscopy Association of North America. B. J. Manaster is a salaried part-time employee of Amirsys. 1 Department of Radiology, University of Utah, 30 N 1900 E, #1A71, Salt Lake City, UT Address correspondence to J. Crim (julia.crim@hsc.utah.edu). 2 Department of Orthopedics, University of Utah, Salt Lake City, UT. AJR 2010; 195: X/10/ American Roentgen Ray Society Temporal Evolution of MRI Findings After Arthroscopic Rotator Cuff Repair OBJECTIVE. The purpose of this article is to assess the changes occurring over time in the MRI appearance of repaired rotator cuff tendons and to correlate MRI appearance with clinical outcomes. SUBJECTS AND METHODS. MRI examinations were performed on 40 patients with full-thickness rotator cuff tears preoperatively and at 6 weeks, 3 months, and 12 months after arthroscopic repair. Preoperative scans were assessed for size of tear. Postoperative scans were evaluated for size of footprint, tendon thickness, signal intensity of the repaired tendon, and presence of full-thickness tear. Footprint and tendon thickness were graded from 1 to 4 according to percentage of normal. Tendon signal intensity was graded from 1 to 4 on the basis of the length of abnormal tendon. A composite score of footprint, tendon thickness, and tendon signal intensity was used to compare overall tendon appearance relative to the intact tendon. Rasch analysis was used to transform ordinal scale data into interval scale data. Using interval scale data, MRI findings were correlated to shoulder strength and the Constant- Murley score of clinical outcome. RESULTS. Four recurrent tendon tears occurred during the first postoperative year. Tendons appeared most disorganized compared with native tendon 3 months after surgery. Twenty-four of 36 intact tendons showed a decreased tendon score between 6 weeks and 3 months. There was considerable variability in tendon appearance among patients. There was no correlation between MRI appearance and clinical outcome score. CONCLUSION. MRI appearance of the repaired tendon changes over time but does not correlate with function or predict clinical outcomes at 1 year after surgery. M RI of the surgically repaired rotator cuff is sometimes requested for patients with persistent pain. Therefore, it is important to understand normal variation in the MRI appearance of the postoperative tendon. Increased signal intensity in the rotator cuff is commonly present and may be related to inflammation, tendinopathy, or granulation tissue [1]. As a result, postoperative images can be difficult to interpret. Several previous imaging studies have reported an abnormal MRI appearance of the repaired tendon [1 4]. Spielmann et al. [3] found that only 10% of tendons evaluated in 15 asymptomatic patients had a normal appearance on MRI. Zanetti et al. [4] reported residual defects in the MRI appearance of 21% of tendons, with an average follow-up of 39 months. Serial ultrasound evaluation of the repaired rotator cuff has revealed thinning of the tendon and increased postoperative vascularity of the peritendinous region that decreases by 6 months postoperatively [5]. Unsuccessful attempts have been made to correlate functional outcomes with the imaging appearance of a repaired tendon [6 8]. Other studies of the postoperative rotator cuff have shown that a recurrent tear results in decreased abduction strength but is not related to functional outcome or patient satisfaction [2, 9, 10]. This study documents the evolution of postoperative MRI appearance of the rotator cuff over the first year after surgery and attempts to correlate the MRI findings with functional outcome. Subjects and Methods The study protocol was approved by the institutional review board and was HIPAA compliant. All patients gave written informed consent. AJR:195, December

2 Crim et al. Inclusion and Exclusion Criteria Forty serial patients undergoing arthroscopic rotator cuff repair by one of two subspecialty orthopedic sports surgeons were enrolled in a prospective study. Both surgeons use arthroscopy as their standard method of rotator cuff repair. Inclusion criteria included a full-thickness rotator cuff seen on preoperative MRI, the ability and willingness of the patient to undergo serial MRI and clinical examinations, and the patient s willingness to comply with a standardized rotator cuff physical therapy program. Exclusion criteria included factors that tend to complicate recovery from rotator cuff surgery, including an active history of smoking, an autoimmune or rheumatologic disease, current use of steroids, previous rotator cuff surgery of the affected shoulder, a work-related injury [11], and the presence of a subscapularis tear or a large U-shaped tear according to arthroscopic assessment [12]. Surgical and Rehabilitation Technique The rotator cuff repairs were performed by two surgeons who repair almost all rotator cuff tears arthroscopically. Nonmetallic suture anchors and sutures were used for fixation, and the cuff was secured in standard fashion with mattress sutures. After surgery, patients arms were placed in an abduction sling, with passive range of motion (ROM) during the first week, active assisted ROM at 4 6 weeks, and full active ROM at 6 8 weeks. Strengthening exercises were started at weeks. MRI Patients underwent serial shoulder MRI examinations. These were performed before surgery and at 6 weeks, 3 months, and 12 months after surgery. All MRI examinations were performed on a 1.5-T scanner (Avanto, Siemens Healthcare) with a dedicated U-shaped shoulder coil (MRI Devices) with matrix. The following sequences were obtained: oblique coronal and sagittal T1 weighted (TR/TE, 500/15), oblique coronal and sagittal fast spin-echo T2 weighted with fat saturation (4,500/60; field of view, 16 cm; 3.5- cm slice thickness; echo train, 9 on coronal and 13 on sagittal; and receiver bandwidth, 130), and axial fast spin-echo proton density with fat saturation (2,500/12; field of view, 16; 3.5-cm slice thickness; echo train, 7; and bandwidth, 130). MRI examinations were evaluated by consensus by two musculoskeletal fellowship trained radiologists with 20 and 25 years of experience who had no knowledge of the patient s clinical information or surgical history. Preoperative scans were evaluated for the thickness of the rotator cuff, size of the rotator cuff tear in the anterior posterior dimension, the degree of muscle atrophy, and the extent of retraction medial to the footprint. Postoperative scans were evaluated for the size of the footprint (i.e., attachment to greater tuberosity), the average tendon thickness at the site of repair, the signal intensity of the repaired tendon, and the presence of full-thickness tear. In each case, the tendon was compared with the preoperative anatomy for that patient. The size of the repaired footprint was compared with the expected anatomic footprint and was graded on a scale of 1 to 4. Grades were quartiles based on comparison with the normal appearance of the rotator cuff. For footprint coverage, grade 1 was 1 24%, grade 2 was 25 49%, grade 3 was 50 74%, and grade 4 was % of the width of the native tendon insertion on the greater tuberosity. Tendon thickness was compared with preoperative tendon thickness adjacent to the tear, with grade 1 as 1 24%, grade 2 as 25 49%, grade 3 as 50 74%, and grade 4 as % of preoperative normal thickness. Signal intensity of the repaired tendon was evaluated on the basis of the extent of increased signal intensity on T2-weighted sagittal and coronal images. Normal signal intensity equivalent to an intact nonoperated tendon was a grade 4. Grade 3 was a tendon with increased T2 signal over less than 1 cm of the tendon. Grade 2 constituted increased T2 signal over 1 2 cm of the tendon. Grade 1 occupied more than 2 cm of the tendon. The presence of a full-thickness tear was documented. A full-thickness tear was defined as fluid signal traversing the entire tendon from inferior to superior, or discontinuity of the tendon fibers, or both. In the 36 patients without a full-thickness recurrent tear, the three rotator cuff evaluations (footprint size, tendon thickness, and signal intensity) were summed to create a rotator cuff score. A score of 12 constituted a normal tendon. Muscle atrophy was assessed using the Thomazeau scale [12], where 1 is normal and 3 is severe atrophy. Tendons were evaluated after completion of all scans, and observers were aware of the sequence of the scans. A combination of all planes was used for evaluation. Clinical Evaluation Clinical functional evaluations of the shoulder were performed during the course of the study. Each shoulder was evaluated by a physical therapist who was not aware of the MRI findings. The Constant-Murley score [13] is a widely accepted outcomes measure and was used as the clinical standard to compare with the MRI findings. This score evaluates subjective measures of function, namely pain and activities of daily living. The objective portion of the score includes ROM and strength. Clinical assessments of function were performed before surgery and at 3 and 12 months after surgery using the Constant-Murley clinical outcome score. Internal strength and external rotation strength were also measured at each visit. Statistical Analysis The patient s rotator cuff thickness and signal, footprint size, and rotator cuff score at 3 months and 1 year after surgery were compared with the clinical scores. Descriptive statistics, such as means and Pearson s product moment correlation coefficients, were calculated from the data. Because the rating scales of footprint, tendon thickness, and tendon signal were ordinal, we applied Rasch analysis to transform the ordinal scale data into interval scale data before computation of means and correlations to draw meaningful comparisons. Basically, in Rasch analysis, the task is to generate raw scores by summing the items and then converting them into interval scores through logarithmic transformation. Detailed discussion of Rasch analysis is beyond the scope of this article [14]. Results The 40 patients enrolled in the study had an average age of 56.5 years. There were 17 men and 23 women. The average preoperative rotator cuff tear measured 18.5 mm from anterior to posterior. MRI Findings Recurrent tendon tears Recurrent nontraumatic full-thickness tears occurred in four patients. One occurred before the 6-week follow-up visit, and the others occurred between 3 and 12 months after surgery. Of note, the tear occurring before the 6-week follow-up visit occurred at the footprint and was in the oldest patient in the study (82 years old). The remaining three recurrent tendon tears occurred at the musculotendinous junction (Fig. 1). There was one false-positive diagnosis of rotator cuff tear at 6 weeks (Fig. 2). This was followed up clinically and showed improvement in tendon thickness over time. Tendons without recurrent tear There was marked variability in the appearance of the tendons, and the rotator cuff score summarizing the appearance of all the tendon variables studied was useful in tracking overall changes (Fig. 3) Tendon footprint improved over time in many patients, but this did not reach statistical significance (Fig. 4). Although 12 patients showed a decrease in footprint size of at least one grade between 6 weeks and 3 months, none showed a decrease between 3 and 12 months. Tendon signal intensity and thickness did not show a consistent pattern of normaliza AJR:195, December 2010

3 MRI of Arthroscopic Rotator Cuff Repair Fig year-old man with new full-thickness supraspinatus tear, as seen on T2-weighted coronal image 1 year after surgery. Supraspinatus tendon is torn (arrow) at musculotendinous junction. Footprint (arrowhead) where previous tendon tear was repaired is normal. Constant-Murley score was 44 before surgery, 51 at 3 months after surgery, and 80 at 1 year after surgery. tion (Fig. 5). Signal intensity increased by one grade in six cases, by two grades in four cases, stayed constant in 16 cases, decreased by one grade in eight cases, and decreased by two grades in two cases. The mean tendon thickness remained stable, but there was considerable variability in individual cases. It decreased by one grade in five cases, by two grades in three cases, stayed constant in 18 cases, and improved by one grade in 10 cases. Two cases showed decrease in tendon signal despite decreased tendon thickness. Atrophy as measured by the Thomazeau scale was mild, with a mean score of 1.6 before surgery and a score of 2 at 1 year after surgery. The atrophy worsened in six cases, Tendon Score 30 Normalized No change More disorganized appearance Change 6 w to 3 m Change 6 w to 1 y improved in six cases, and was unchanged in 24 cases. There was no correlation between change in muscle atrophy and change in tendon thickness or signal. There was no correlation between the size of the tear before surgery and the Constant-Murley score, tendon thickness, or tendon signal. Interobserver Differences Of the 160 MRI examinations performed, there was disagreement between readers regarding two MRI examinations performed 3 months after surgery. One reader considered it to be a complete tear and the other reader considered a few fibers to be intact. In these cases, it was elected to categorize these tendons as A Fig year-old man with false-positive diagnosis of recurrent tear, as seen on T2-weighted coronal images. A, At 6 weeks, footprint (arrowhead) is empty, and gap is seen between footprint and tendon. However, tendon was attached to humerus more medially (arrow), finding not appreciated at time of interpretation of study. B, At 1 year, footprint is fully covered by tendon (grade 4 coverage), and tendon is continuous (arrow). Constant- Murley score was 57 at 3 months and 73 at 1 year after surgery. Change 3 m to 1 y Fig. 3 Graph showing changes in tendon score over time. thinned but not torn. Both cases showed improvement in tendon thickness over time. Comparison of MRI Findings to Clinical Score Correlation of MRI findings and clinical outcomes at 1 year is shown in Table 1. At 3 months (not shown in table) and at 1 year, there was no significant correlation between the tendon thickness, tendon signal intensity, tendon footprint, or tendon score as MRI examination outcomes, versus the clinical outcomes of internal rotation or external rotation strength or Constant-Murley score. Muscle atrophy as measured by the Thomazeau scale also showed no correlation with the Constant-Murley score. The only statistically significant correlation was between tendon thickness at 3 months and internal rotation strength at 3 months (r = 0.373; p = 0.025). Internal rotation strength increased as supraspinatus tendon thickness decreased. Four patients had recurrent rotator cuff tears documented on MRI. The first, occurring before the 6-week visit, resulted in pain and poor shoulder function with severely limited elevation of the shoulder. The Constant-Murley score at 1 year was 12. The other three tears occurred between 3 months and 1 year. The first patient had good ROM but pain and decreased strength; the Constant-Murley score was 69. B AJR:195, December

4 Crim et al. The second patient complained of pain but had good strength and ROM, with a Constant-Murley score of 93. The last patient with a tear had a Constant-Murley score of 80 with persistent pain and decreased elevation of the shoulder. Discussion Recurrent Tear Our rate of repaired tendon failure was 10%. The published rates of arthroscopic rotator cuff repair failure vary from 3% to 54% [2, 7, 9, 15 20]. The relatively low failure rate in this study may relate to patient selection. Massive (> 4 cm) tears, which have a high A A failure rate [21], were excluded from the study because they were not thought by the surgeons to be amenable to arthroscopic repair. Tears were relatively small (mean, 18.5 mm) compared with those in some studies [6, 9, 12, 22], and smaller tears have been associated with a lower rate of recurrent tear [23]. Our patients were relatively young compared with patients in many other studies [2, 9, 10, 12, 19, 20, 22, 24], and advanced age has been implicated as a risk for recurrent tear [2, 16]. Interestingly, the tendon failures that occurred after 6 weeks occurred away from the site of surgery. Some repair failures may B B Fig year-old man with improvement in footprint over time, as seen on T2-weighted coronal images. A, At 6 weeks, supraspinatus tendon attaches to 50% of normal footprint, grade 2, with lateral portion of greater tuberosity uncovered. At 3 months, it had reduced to grade 1 (not shown). B, At 1 year, tendon attachment covers entire footprint, grade 4. Constant-Murley score was 47 at 3 months and 87 at 1 year after surgery. Fig year-old woman with tendon that remained thin and contained abnormal signal intensity, as seen on T2-weighted coronal images, despite improvement of Constant-Murley score over time. Note that thickness and signal intensity were determined according to multiple slices and are not completely represented by single slice. A, At 3 months, tendon thickness (arrow) averages grade 2, signal intensity is grade 2, and footprint is grade 2. Constant-Murley score at 3 months was 52. B, At 1 year, tendon thickness (arrow) averages grade 1, signal intensity is grade 2, and footprint is grade 3. Constant-Murley score at 1 year was 80. be due to progression of tendon disease or to impaired vascularity [5]. Expected Appearance of the Postoperative Cuff The strength of the current study is the prospective temporal evaluation of the postoperative tendon by MRI. The only other temporal study, to our knowledge, of the postoperative tendon used ultrasound and showed decreasing peritendinous vascularity over the course of 6 months [5]. Spielmann et al. [3] found that intact tendon repairs are thinner than normal tendon and contain increased T2 signal intensity. Our 1364 AJR:195, December 2010

5 MRI of Arthroscopic Rotator Cuff Repair TABLE 1: Clinical Outcomes Compared With MRI Findings at 1 Year After Surgery Clinical Outcome study is in agreement with those findings and adds data showing that improvement, when it occurs, generally develops between 3 and 12 months. MRI findings in the early postoperative period should be viewed with caution. Boileau et al. [2] evaluated rotator cuff repairs on a single study performed between 6 months and 3 years after surgery and concluded that the repair should be considered failed if footprint coverage is less than 50%. Our data show that this is not a good indicator of surgical failure. The footprint coverage may appear poor at 6 weeks or 3 months after surgery, and there may be subsequent improvement of footprint coverage by the end of the first postoperative year. On the basis of our findings, we suggest that during the first postoperative year, it is not prudent to consider the tendon repair as failed according to tendon irregularity, thinning, or increased signal intensity. The only exception to this rule is a study showing unequivocal full-thickness tear (i.e., complete discontinuity of fibers and fluid signal traversing the entire tendon). The appearance of the rotator cuff after arthroscopic repair shows considerable variability in the first postoperative year and does not correlate to outcome. The tendon appearance often becomes more like normal tendon by 1 year after surgery, despite extensive signal changes and tendon thickening at 6 weeks or 3 months after surgery. MRI Appearance Does Not Correlate With Clinical Outcome In agreement with prior studies, we found no statistically significant correlation between clinical outcome scores and MRI findings [6 8]. Some studies have shown a statistically significant correlation between abduction strength and the presence of an intact tendon [2, 9], but this was not seen in our study group according to Constant-Murley scores. Limitations There are two main limitations to this study. First, the 1-year follow-up is shorter MRI Finding Footprint Thickness Signal Constant-Murley score r = 0.081; p = r = 0.075; p = r = 0.092; p = Internal rotation strength r = 0.037; p = r = 0.165; p = r = 0.147; p = External rotation strength r = 0.106; p = r = 0.060; p = r = 0.005; p = than that in some other studies [2, 4, 7, 10, 19, 21, 22, 24 26]. Second, we did not perform second-look arthroscopy to confirm the postoperative MRI findings. Instead, we relied on published criteria to assess for a full-thickness tear, namely absence of visualized tendon fibers and fluid signal extending through its entire thickness. The single false-positive tear found in this study (Fig. 1) corresponds favorably to false-positive rate reported in a study by Owen et al. [27]. Practical Applications MRI of the postoperative rotator cuff is difficult to interpret. In the first year after surgery, MRI appearance of the cuff is variable and should be interpreted with caution. Clinical evaluation of the postoperative rotator cuff should be stressed during the first year after surgery. In conclusion, the appearance of the rotator cuff after arthroscopic repair shows considerable variability during the first postoperative year and does not correlate with outcome. Acknowledgment We thank Barbara Fink, Physical Therapist, for her clinical assessment of the patients. References 1. Zlatkin MB. MRI of the postoperative shoulder. Skeletal Radiol 2002; 31: Boileau P, Brassart N, Watkinson DJ, Carles M, Hatzidakis AM, Krishnan SG. Arthroscopic repair of full-thickness tears of the supraspinatus: does the tendon really heal? J Bone Joint Surg Am 2005; 87: Spielmann AL, Forster BB, Kokan P, Hawkins RH, Janzen DL. Shoulder after rotator cuff repair: MR imaging findings in asymptomatic individuals initial experience. Radiology 1999; 213: Zanetti M, Jost B, Hodler J, Gerber C. MR imaging after rotator cuff repair: full-thickness defects and bursitis-like subacromial abnormalities in asymptomatic subjects. Skeletal Radiol 2000; 29: Fealy S, Adler RS, Drakos MC, et al. Patterns of vascular and anatomical response after rotator cuff repair. Am J Sports Med 2006; 34: Klepps S, Bishop J, Lin J, et al. Prospective evaluation of the effect of rotator cuff integrity on the outcome of open rotator cuff repairs. Am J Sports Med 2004; 32: Liu SH, Baker CL. Arthroscopically assisted rotator cuff repair: correlation of functional results with integrity of the cuff. Arthroscopy 1994; 10: Oh JH, Kim SH, Ji HM, Jo KH, Bin SW, Gong HS. Prognostic factors affecting anatomic outcome of rotator cuff repair and correlation with functional outcome. Arthroscopy 2009; 25: Bishop J, Klepps S, Lo IK, Bird J, Gladstone J, Flatow EL. Cuff integrity after arthroscopic versus open rotator cuff repair: a prospective study. J Shoulder Elbow Surg 2006; 15: Jost B, Pfirrmann CW, Gerber C, Switzerland Z. Clinical outcome after structural failure of rotator cuff repairs. J Bone Joint Surg Am 2000; 82: Misamore GW, Ziegler DW, Rushton JL 2nd. Repair of the rotator cuff: a comparison of results in two populations of patients. J Bone Joint Surg Am 1995; 77: Burkhart SS, Barth JR, Richards DP, Zlatkin MB, Larsen M. Arthroscopic repair of massive rotator cuff tears with stage 3 and 4 fatty degeneration. Arthroscopy 2007; 23: Pallant JF, Tennant A. An introduction to the Rasch measurement model: an example using the Hospital Anxiety and Depression Scale (HADS). Br J Clin Psychol 2007; 46: Constant CR, Murley AH. A clinical method of functional assessment of the shoulder. Clin Orthop Relat Res 1987; 214: Lafosse L, Brozska R, Toussaint B, Gobezie R. The outcome and structural integrity of arthroscopic rotator cuff repair with use of the double-row suture anchor technique. J Bone Joint Surg Am 2007; 89: Liem D, Lichtenberg S, Magosch P, Habermeyer P. Magnetic resonance imaging of arthroscopic supraspinatus tendon repair. J Bone Joint Surg Am 2007; 89: Sugaya H, Maeda K, Matsuki K, Moriishi J. Repair integrity and functional outcome after arthroscopic double-row rotator cuff repair: a prospective outcome study. J Bone Joint Surg Am 2007; 89: Magee TH, Gaenslen ES, Seitz R, Hinson GA, Wetzel LH. MR imaging of the shoulder after surgery. AJR 1997; 168: Franceschi F, Ruzzini L, Longo UG, et al. Equivalent clinical results of arthroscopic single-row and double-row suture anchor repair for rotator cuff tears: a randomized controlled trial. Am J Sports Med 2007; 35: AJR:195, December

6 Crim et al. 20. Frank JB, ElAttrache NS, Dines JS, Blackburn A, Crues J, Tibone JE. Repair site integrity after arthroscopic transosseous-equivalent suture-bridge rotator cuff repair. Am J Sports Med 2008; 36: Galatz LM, Ball CM, Teefey SA, Middleton WD, Yamaguchi K. The outcome and repair integrity of completely arthroscopically repaired large and massive rotator cuff tears. J Bone Joint Surg Am 2004; 86A: Anderson K, Boothby M, Aschenbrener D, van Holsbeeck M. Outcome and structural integrity FOR YOUR INFORMATION after arthroscopic rotator cuff repair using 2 rows of fixation: minimum 2-year follow-up. Am J Sports Med 2006; 34: Levy O, Venkateswaran B, Even T, Ravenscroft M, Copeland S. Mid-term clinical and sonographic outcome of arthroscopic repair of the rotator cuff. J Bone Joint Surg Br 2008; 90: Fuchs B, Gilbart MK, Hodler J, Gerber C. Clinical and structural results of open repair of an isolated one-tendon tear of the rotator cuff. J Bone Joint Surg Am 2006; 88: Deutsch A, Kroll DG, Hasapes J, Staewen RS, Mark your calendar for the following ARRS annual meetings: May 1 6, 2011 Hyatt Regency Chicago, Chicago, IL April 29 May 4, 2012 Vancouver Convention Center, Vancouver, BC, Canada April 14 19, 2013 Marriott Wardman Park Hotel, Washington, DC May 4 9, 2014 Manchester Grand Hyatt San Diego, San Diego, CA Pham C, Tait C. Repair integrity and clinical outcome after arthroscopic rotator cuff repair using single-row anchor fixation: a prospective study of single-tendon and two-tendon tears. J Shoulder Elbow Surg 2008; 17: Wilson F, Hinov V, Adams G. Arthroscopic repair of full-thickness tears of the rotator cuff: 2- to 14- year follow-up. Arthroscopy 2002; 18: Owen RS, Iannotti JP, Kneeland JB, Dalinka MK, Deren JA, Oleaga L. Shoulder after surgery: MR imaging with surgical validation. Radiology 1993; 186: AJR:195, December 2010