Comparison of Knee Articular Cartilage Defect Size Between Measurements Obtained on Preoperative MRI Versus During Arthrotomy

Abstract

Background: Treatment decisions for cartilage defects are often based on lesion size. Magnetic resonance imaging (MRI) is widely used to diagnose cartilage defects noninvasively; however, their size estimated from MRI may differ from defect sizes measured during arthrotomy, especially after debridement to healthy cartilage if undergoing autologous chondrocyte implantation. Purpose/Hypothesis: The purpose of this study was to evaluate the reliability of 2 methods to assess knee cartilage defect size on preoperative MRI and determine their accuracy in predicting postdebridement defect sizes recorded during arthrotomy. It was hypothesized that defect size would be predicted more accurately by the total area of abnormal articular cartilage rather than the area of full-thickness cartilage loss as identified on MRI. Study Design: Cohort study (diagnosis); Level of evidence, 3. Methods: This study included 64 patients (mean age, 41.8 ± 9.6 years) who underwent autologous cell therapy. Each patient received a 3-T MRI at 6.1 ± 3.0 weeks before cell implantation. Three raters, a radiologist, a surgeon, and a scientist, measured (1) the full-thickness cartilage defect area and (2) the total predicted abnormal cartilage area, identified by an abnormal signal on MRI. Interrater reliability was assessed using the intraclass correlation coefficient (ICC). Actual pre- and postdebridement defect sizes were obtained from intraoperative surgical notes. Postdebridement surgical measurements were considered the clinical reference standard and were compared with the radiologist’s MRI measurements. Results: Eighty-seven defects were assessed, located on the lateral (n = 8) and medial (n = 26) femoral condyle, trochlea (n = 17), and patella (n = 36). The interrater reliability of the cartilage defect measurements on MRI was good to excellent for the full-thickness cartilage defect area (ICC = 0.74) and the total predicted abnormal cartilage area (ICC = 0.78). The median full-thickness cartilage defect area on MRI underestimated the median postdebridement defect area by 78.3%, whereas the total predicted abnormal cartilage area measurement underestimated the postdebridement defect area by 14.3%. Conclusion: Measuring the full-thickness cartilage defect area on MRI underestimated the area to treat, whereas measuring the total abnormal area provided a better estimate of the actual defect size for treatment.