Abstract
Objectives:
To compare the overall agreement in the magnitude of shoulder labral injuries between preoperative MRI and assessment at the time of diagnostic arthroscopy.
Methods:
A retrospective chart review was performed of patients undergoing primary arthroscopic labral repair procedures between February 2018 and February 2023. Exclusion criteria included prior ipsilateral shoulder surgery, MRI study performed at outside health system, and incomplete documentation of labral tear location. Age, gender, and dates of injury, imaging, and surgery were recorded. Radiology and operative reports were reviewed to record the location and extent of the labral tear using conventional clock-face coordinates. Tear locations were further classified as either anterior (2-6 o’clock), posterior (6-10 o’clock), or superior (10-2 o’clock). MRI studies were performed by 1.5-Tesla scanner and all radiology reports were performed by fellowship-trained musculoskeletal radiologists at a single health system. Subgroup analysis of labral tear size on radiology interpretation was performed between MRI arthrogram (MRA) and conventional MRI studies.
Results:
A total of 138 subjects (mean age 25.8 years, 73% male) met criteria for inclusion in this study. Of the included subjects, 105 underwent MRI arthrogram and 33 underwent conventional MRI studies. The average time from initial injury to MRI was 153.4 days and the average time from MRI to surgical intervention was 66.8 days. The average total recorded labral tear size in clock-face coordinates was significantly greater on diagnostic arthroscopy compared with the radiologist MRI interpretation (4.4 vs 3.6, p<0.01). The average labral tear size in the posterior region was also significantly greater on diagnostic arthroscopy (2.4 vs 1.6, p<0.01). The average labral tear size in the anterior (2.5 vs 2.2, p=0.08) and superior (1.6 vs 1.5, p=0.29) regions was greater on diagnostic arthroscopy but did not reach statistical significance. The difference in total and posterior tear size remained significant in both the arthrogram and conventional MRI subgroups. Subjects undergoing MRI arthrogram had a significantly greater reported size of anterior labral tears in clock-face coordinates compared with conventional MRI (2.7 vs 2.0, p=0.03), but did not reach significance for the posterior (p=0.40), superior (p=0.31), and the overall global labrum (p=0.29).
Conclusions:
MRI significantly underestimate the overall size of labral tears in setting of shoulder instability, particularly in the posterior labrum. Surgeons should account for potential underestimation of true labral injury during preoperative planning for surgical management of shoulder instability.