Rotator cuff fatty infiltration: are coronal images more helpful for characterization than sagittal images?
While the Goutallier system was first implemented using axial images on CT, it has been validated using sagittal oblique images on MRI, which has become the most common imaging modality for evaluation of the rotator cuff musculature. (7) Typically, the most lateral sagittal image where the scapular spine is in contact with the scapular body is used for evaluation. (7) The torn rotator cuff tendon, as well as the muscle, tend to retract medially with increased chronicity. (8,9) Given this retraction, evaluating the musculature using MR sagittal images may not give an accurate representation of the musculature and consequently the degree of fatty infiltration. Using standard imaging parameters, coronal oblique images tend to include more of the muscle bellies of the rotator cuff than sagittal oblique images. Thus, in the setting of a rotator cuff tears with retraction of the torn tendon fibers, the coronal images would likely include more of the muscle than sagittal images and allow for better characterization of the degree of fatty infiltration.
At our institution, the degree of fatty infiltration has the greatest impact on the management for patients with large or massive tears of the rotator cuff tendons. It is for these patients where the decision of surgical versus conservative treatment is the most difficult. This group typically is the most symptomatic and tends to have a greater degree of fatty infiltration than patients with smaller tendon tears. Patients with massive rotator cuff tears also tend to have high rates of healing failure after surgery, up to 41%, and inferior clinical outcomes.10
The purpose of this study is to see if there is a discrepancy between the characterization of fatty infiltration in the setting of a large or massive rotator cuff tear on sagittal and coronal images using the Goutallier classification. We hypothesize that there will be discrepancies in terms of fatty infiltration characterization between the coronal and sagittal images.
Materials and Methods
Institutional Review Board approval was obtained, and informed consent was waived for this retrospective Health Insurance Portability and Accountability Act-compliant study.
A retrospective review of patients undergoing MR examinations of the shoulder at our institution over the past 12 months was performed using the search terms "massive rotator cuff tear." Inclusion criteria were patients with rotator cuff tears of the supraspinatus and infraspinatus tendons measuring greater than or equal to 3 cm in anterior-posterior (AP) dimension and examinations, including both coronal and sagittal images. Exclusion criteria were tears less than 3 cm in AP dimension, edema within the rotator cuff musculature on the fat suppressed-fluid weighted sequences, and lack of coronal and sagittal images in the patient's study. Thirty patients met these criteria and were included in our study group. A control group of 20 randomly selected patients with MR shoulder studies performed at our institution without evidence of rotator cuff tearing or muscle edema was used for comparison.
The imaging studies were performed on either a 1.5T or 3T MR scanner (Siemens Medical) with the following sequences: oblique coronal Turbo Spin Echo (TSE) proton density (3 mm, TR range/TE range, 2100-2500/25-35) and fat suppressed T2W (3 mm, TR range/TE range 35004000/55-60); oblique sagittal T1W (3 mm, TR range/TE range 450-600/12-15) and fat suppressed T2W (3 mm, TR range/TE range, 3500-4000/55-65); and axial fat suppressed proton density (3 mm, TR range/TE range, 2100-2500/2537). The field of view was 140 mm, and the matrix was 256 x 256 for all the studies. Dedicated 16 channel shoulder array coils were used for imaging.
Each study was reviewed to record tear size in anteriorposterior dimension and assess the degree of retraction in transverse dimension. Retraction was assessed by measuring the distance between the medially retracted torn tendon and the lateral margin of the greater tuberosity. Following this, representative sagittal T1-weighted and coronal proton density images of the supraspinatus and infraspinatus muscles from each study and control group subject were selected, separated, and placed in random order. Three musculoskeletal radiologists reviewed the images blindly and independently. The degree of fatty infiltration in each muscle was quantified based on a revised version of the Goutallier 5-point classification as follows (Fig.1):
a. Grade 0/1: no or some fatty streaks,
b. Grade 2: less fat than muscle,
c. Grade 3: fat equal to muscle, and
d. Grade 4: fat greater than muscle.
Statistical analysis included logistic regression for correlated data to assess the likelihood of discrepant findings for a single reader evaluation of coronal and sagittal images of a given patient on separate occasions. Specifically, generalized estimating equations (GEE) based on a binary logistic regression model were used. The dependent variable for each reader was the binary assessment of concordance or discordance between that reader's coronal and sagittal classifications of fatty atrophy in each location (supraspinatus or infraspinatus) for each patient. Output included an estimate and 95% confidence interval (CI) for the percentage of times the coronal and sagittal images would yield discrepant evaluations of fatty atrophy and a p-value for the comparison of cases and controls in terms of this percentage. The estimate and CI was calculated for the study and control groups and for the sample as a whole. The GEE analysis was also used to identify factors predictive of discrepancy between sagittal and coronal scores, in particular, to test whether the likelihood of discrepancy is impacted by the AP size of the tendon tear or the amount of medial retraction of the torn tendon.
The specific sample sizes of 30 patients for the study group and 20 patients for the control group were selected so that the study would: 1. have 80% power at the 5% significance level to detect a case versus control difference in the likelihood of discrepancy provided discrepant findings are at least two times more prevalent among patients than among controls; 2. estimate the percentage of times sagittal and coronal images will yield discrepant findings for study group patients with a precision of 10 percentage points.
The study group included 15 men and 15 women, with a mean age of 65 (range: 45 to 87 years). The control group included 10 men and 10 women, with a mean age of 60 (range: 45 to 76 years). The mean size of the rotator cuff tears was 4.3 cm (range: 3.2 to 6.3 cm) while the mean tendon retraction was 4.3 cm (range: 2.1 to 7.3 cm).
Statistically significant discrepancies in the assessment of fatty infiltration between the sagittal images and the coronal images occurred 62% of the time (111/180) in the study group and 33% of the time (40/120) in the control group (p < 0.001). In the study group, each reader demonstrated a statistically significant tendency to assign a higher degree of fatty infiltration to each muscle when grading the sagittal images than when grading the coronal images (p < 0.04); a difference that was not present in the control group (p > 0.1) (Figs. 2 and 3, Table 1). There was no significant difference between the individual muscles (supraspinatus or infraspinatus) or among readers in terms of the overall likelihood of discrepancy between coronal and sagittal evaluations in either group (p = 0.669, p = 0.833, respectively).
In the study group, 33% of the time (37% supraspinatus, 29% infraspinatus), there was a Goutallier grade greater than or equal to 3 assigned on the sagittal image, but less than 3 on the corresponding coronal image. The opposite (less than 3 assigned on the sagittal image and greater than or equal to 3 on the coronal image) occurred 1% of the time, p < 0.001. Neither of these types of differences occurred in the control group. There was no statistically significant association between the likelihood of fatty infiltration discrepancy between sagittal and coronal images and AP tendon tear size or amount of tendon retraction (p = 0.5 to 0.9).
Fatty infiltration is an important consequence of rotator cuff tendon tearing; it is associated with the degree of shoulder function as well as the success of tendon repair. (2-6) The pathophysiology linking fatty infiltration with rotator cuff tears is unclear with several theories posited, including impingement of the suprascapular nerve by the torn, retracted tendon or a secondary phenomenon related to shortened, atrophied muscle cells. (11-13)
The most commonly used classification system for rotator cuff muscle fatty infiltration is the one proposed by Goutallier. (2) This scale was adapted to MRI by Fuchs, who found good to excellent interobserver reproducibility, albeit with an unsatisfactory correlation to CT. (7) Stages 0 and 1 have been categorized as non-significant, as they are rarely related to the tearing of the rotator cuff tendons. (5) There is some debate as to what stage of fatty infiltration is significant, i.e. at what stage does the risk for repair failure and decreased function occur. Studies have demonstrated this increased risk at or greater than stage 2.2, (5) At our institution, fatty infiltration greater than stage 2 is considered significant.
In our study, we found that the grade of fatty infiltration assigned to the supraspinatus and infraspinatus muscles on the sagittal images differed from the coronal image grades in 62% of the evaluations of our study group. Each reader demonstrated a tendency to assign a higher degree of fatty infiltration on the sagittal images when compared to the coronal image estimations. One-third of the time, the reader assigned a fatty infiltration grade of greater than 2 based on the sagittal images that was downgraded to 2 or less on the coronal images. In the clinical setting, this could mean a change in management decision. Based on the sagittal images, non-operative methods of treatment would be considered for this group of patients because of the lower expectations of shoulder function and repair success assigned to this degree of fatty infiltration. However, the lower degree of fatty infiltration estimated by the coronal images may favor surgical management.
In our review of the study group cases, we observed a relatively consistent degree of medial retraction of the tendons, usually to the level of the glenohumeral joint, and the adjacent muscle. With this degree of retraction, the reader is assigning a stage of fatty infiltration on the sagittal images based on the appearance of the musculotendinous junction relative to the surrounding supraspinatus fossa fat instead of the ratio of intramuscular fatty infiltration to the surrounding muscle. On the other hand, a greater proportion of the retracted muscles was consistently seen on the corresponding coronal images. We believe that the coronal images will typically provide greater coverage of the retracted supraspinatus and infraspinatus muscle bellies in the setting of a large or massive rotator cuff tear and thus may allow for a more accurate characterization of the degree of fatty infiltration.
There are several limitations for our study. First, there was no gold standard for the degree of muscle fatty infiltration. Fatty infiltration is difficult to assess intraoperatively and thus not regularly reported in the operative reports of our institution's surgeons. We also understand that the shoulder MR protocols used at our institution may defer from those at other institutions. We believe that our selection of sequences and field of views are representative of the current state of MR shoulder imaging. Other limitations include the small number of patients, as well the retrospective nature. Additionally, only large and massive rotator cuff tears were included in our study group. This was done for two reasons. First, larger tears are more likely to have a higher degree of medial retraction of the musculotendinous unit. Second, it is this group of patients where the grade of fatty infiltration has the greatest impact in terms of management.
In conclusion, we found a statistically significant difference in the characterization of rotator cuff muscle fatty infiltration between the sagittal and coronal planes in the setting of a large or massive tear, a difference with clinical implications at our institution. We hypothesize that this difference is a result of incomplete visualization of the medially retracted muscle and tendons on the sagittal images and more complete visualization on the coronal images. Awareness of this difference may allow for a more accurate characterization of the degree of fatty infiltration in this setting.
Munish Chitkara, M.D., Marissa Albert, M.D., Tony Wong, M.D., and Soterios Gyftopoulos, M.D., New York University, Langone Medical Center, New York, New York. John O'Donnell, M.D., North Shore-LIJ Medical Center, New York, New York. Correspondence: Soterios Gyftopoulos, M.D., NYU Langone Medical Center, Center for Biomedical Imaging, Department of Radiology, 660 First Avenue, New York, NY 10016; soterios20@ gmail.com.
None of the authors have a financial or proprietary interest in the subject matter or materials discussed, including, but not limited to, employment, consultancies, stock ownership, honoraria, and paid expert testimony.
(1.) Bjorkenheim JM. Structure and function of the rabbit's supraspinatus muscle after resection of its tendon. Acta Orthop Scand. 1989 Aug;60(4):461-3.
(2.) Goutallier D, Postel JM, Bernageau J, et al. Fatty muscle degeneration in cuff ruptures. Pre- and postoperative evaluation by CT scan. Clin Orthop Relat Res. 1994 Jul;(304):78-83.
(3.) Nakagaki K, Ozaki J, Tomita Y, Tamai S. Function of supraspinatus muscle with torn cuff Evaluated by Magnetic Resonance Imaging. Clinical Orthop Relat Res. 1995 Sep;(318):144-51.
(4.) Goutallier D, Postel JM, Gleyze P, et al. Influence of cuff muscle fatty degeneration on anatomic and functional outcomes after simple suture of full-thickness tears. J Shoulder Elbow Surg. 2003 Nov-Dec12(6):550-4.
(5.) Gladstone JN, Bishop JY, Lo IK, Flatow EL. Fatty infiltration and atrophy of the rotator cuff do not improve after rotator cuff repair and correlate with poor functional outcome. Am J Sports Med. 2007 May;35(5):719-28.
(6.) Gerber C, SchneebergerAG, Hoppeler H, Meyer DC. Correlation of atrophy and fatty infiltration on strength and integrity of rotator cuff repairs: a study in thirteen patients. J Shoulder Elbow Surg. 2007 Nov-Dec;16(6):691-6.
(7.) Fuchs B, Weishaupt D, Zanetti M, et al. Fatty degeneration of the muscles of the rotator cuff: assessment by computed tomography versus magnetic resonance imaging. J Shoulder Elbow Surg. 1999 Nov-Dec;8(6):599-605.
(8.) Meyer DC, Mazda F, Amacker NA, et al. Quantitative analysis of muscle and tendon retraction in chronic rotator cuff tears. Am J Sports Med. 2012 Mar;40(3):606-10.
(9.) Meyer DC, Weiser K, Mazda F, et al. Retraction of supraspinatus muscle and tendon as predictors of success of rotator cuff repair. Am J Sports Med. 2012 Oct;40(10):2242-7.
(10.) Miller BS, Downie BK, Kohen RB, et al. When do rotator cuff repairs fail? Serial ultrasound examination after arthroscopic repair of large and massive rotator cuff tears. Am J Sports Med. 2011 Oct;39(10):2064-70.
(11.) Gerber C, Meyer DC, Frey E, et al. Neer Award 2007: reversion of structural muscle changes caused by chronic rotator cuff tears using continuous musculotendinous traction. An experimental study in sheep. J Shoulder Elbow Surg. 2009 Mar-Apr;18(2):163-71.
(12.) Matsumoto F, UhthoffHK, Trudel G, Loehr JF. Delayed tendon reattachment does not reverse atrophy and fat accumulation of the supraspinatus-an experimental study in rabbits. J Orthop Res. 2002 Mar;20(2):357-63.
(13.) Albritton MJ, Graham RD, Richards RS 2nd, Basamania CJ. An anatomic study of the effects on the suprascapular nerve due to retraction of the supraspinatus muscle after a rotator cuff tear. J Shoulder Elbow Surg. 2003 Sep-Oct;12(5):497-500.
Caption: Figure 1 Four coronal proton-density images from four patients with large or massive rotator cuff tears demonstrating varying grades of supraspinatus fatty infiltration. A, Grade 0/1--no or some fatty streaks. B, Grade 2--less fat than muscle. C, Grade 3--fat equal to muscle. D, Grade 4--fat greater than muscle.
Caption: Figure 2 A 53-year-old female with history of rotator cuff repair with massive re-tear involving the supraspinatus and infraspinatus tendons. A, Medial-most sagittal T1-weighted image from the patient's MRI demonstrating significant fat within the supraspinatus and infraspinatus fossae. All three readers assigned grade 4 fatty infiltration to both the supraspinatus and infraspinatus. B, Coronal proton-density image centered on the supraspinatus. A localizer scout line (white line) indicates the slice location of the sagittal T1weighted image depicted in A. The majority of the supraspinatus muscle belly is located medial to the localizer scout line, demonstrating minimal fatty infiltration. All three readers assigned grade 0/1 fatty infiltration to the supraspinatus. C, Coronal proton-density image centered on the infraspinatus. A localizer scout line (white line) indicates the slice location of the sagittal T1-weighted image depicted in A. The majority of the infraspinatus muscle belly is located medial to the localizer scout line, demonstrating mild fatty infiltration. Two readers assigned grade 0/1 fatty infiltration to the infraspinatus, with the third assigning grade 2.
Caption: Figure 3 A 31-year-old female with chronic shoulder pain without evidence of a rotator cuff tear. A, Medial-most sagittal T1-weighted image from the patient's MRI demonstrating no significant fatty infiltration within the supraspinatus and infraspinatus muscles. All three readers assigned grade 0/1 fatty infiltration to both the supraspinatus and infraspinatus. B, Coronal proton-density image centered on the supraspinatus. A localizer scout line (white line) indicates the slice location of the sagittal T1-weighted image depicted in A. The scout line intersects the supraspinatus muscle belly. All three readers assigned grade 0/1 fatty infiltration to the supraspinatus. C, Coronal proton-density image centered on the infraspinatus. A localizer scout line (white line) indicates the slice location of the sagittal Tl-weighted image depicted in A. The scout line intersects the infraspinatus muscle belly. All three readers assigned grade 0/1 fatty infiltration to the infraspinatus.
Table 1 The Mean and Standard Deviation (SD) of the Difference Between the Coronal and Sagittal Grades Assigned by Each Reader to the Same Muscle Reader Muscle Control Group (N = 20) Study Group (N = 30) Mean SD P Mean SD P 1 Supra 0.10 0.55 0.529 -0.70 1.21 0.007 1 Infra 0.05 0.51 0.787 -0.90 1.09 0.001 2 Supra 0.30 0.66 0.103 -0.37 0.85 0.036 2 Infra -0.20 0.62 0.225 -0.67 1.27 0.014 3 Supra 0.25 0.72 0.197 -0.63 1.13 0.008 3 Infra -0.10 0.31 0.371 -1.00 1.05 < 0.001
Please note: Illustration(s) are not available due to copyright restrictions.
|Printer friendly Cite/link Email Feedback|
|Author:||Chitkara, Munish; Albert, Marissa; Wong, Tony; O'Donnell, John; Gyftopoulos, Soterios|
|Publication:||Bulletin of the NYU Hospital for Joint Diseases|
|Date:||Apr 1, 2016|
|Previous Article:||Examination under anesthesia for posterior wall acetabular fracture: a survey of the OTA membership.|
|Next Article:||Can flexible instruments create adequate femoral tunnel lengths at 90[degrees] of knee flexion in anterior cruciate ligament reconstruction?|