Magnetic Resonance Imaging Predictors of Chondral Lesions in Patients With Femoroacetabular Impingement: An Analysis of 545 Cases

Abstract

A large prospective cohort was used (1) to evaluate the overall ability of magnetic resonance imaging (MRI) to detect Outerbridge grade III and IV cartilage defects found during surgery and (2) to identify the specific MRI findings most associated with these cartilage defects so that the practicing hip arthroscopist can better predict cartilage injury before surgery. All patients undergoing hip arthroscopy between February 2015 and May 2017 at 1 institution were enrolled in a prospective cohort. Intra-articular findings were documented at the time of surgery. MRI reports were retrospectively reviewed for radiologist-reported articular cartilage, osseous, or synovial abnormalities. Sensitivity and specificity of MRI findings were calculated; multivariate logistic regression analysis determined which findings were associated with high-grade chondral damage at the time of arthroscopy and used to create an online risk calculator, https://orthop.washington.edu/hiprisk/. Of 598 patients who underwent hip arthroscopy, 550 had MRI reports available for review (92%). Grade III and IV cartilage injuries were reported on arthroscopy in 70 patients (13%) of average age 33 ± 13 years. On univariate analyses, individual MRI findings were not sensitive in detection of articular cartilage injury (mean 22%; range, 1.4%-46%), but positive findings were highly specific (mean 90%,; range, 76%-99%). Multivariate analysis revealed that older age (odds ratio [OR] 1.09 [1.06-1.11], P < .001) and osseous findings such as subchondral cyst or edema (OR 4.77 [2.51-9.05], P < .001) were most predictive of grade III and IV defects (P < .001). MRI was a specific but not sensitive tool in diagnosing articular cartilage injury. Surgeons should be aware that osseous findings such as cysts or edema are highly predictive of full-thickness cartilage loss in FAI. Level III, development of diagnostic criteria (consecutive patients with consistently applied reference standard, no blinding).