PROGRESSION OF BONE MARROW LESION VOLUME IS ASSOCIATED WITH AN INCREASED RISK OF RADIOGRAPHIC AND SYMTOMATIC KNEE OSTEOARTHRITIS: A PROSPECTIVE ANALYSIS OF KNEE MRIS FROM OSTEOARTHRITIS INITIATIVE COHORT

INTRODUCTION Bone marrow lesions (BMLs) are a risk factor for incident knee OA and deep-learning (DL) methods can help in automated segmentation and risk prediction. OBJECTIVE To develop and validate a DL model for quantifying tibiofemoral BML volume from MRIs in knees without radiographic OA and assess the association between longitudinal changes and knee OA incidence. METHODS The DL model segmented tibiofemoral joint into 10 subregions (akin to MRI Osteoarthritis Knee Score (MOAKS) system) and measured BML volume in each subregion. Baseline and 4th-year follow-up MRIs from 4700 participants (9400 knees) of the OAI cohort were analyzed. Knees without OA at baseline (KLG<2) were categorized into three groups based on 4-year BML volume changes: BML-free, regressing BML, and progressive BML. Over a 9-year period, the risk of radiographic and symptomatic knee OA incidence was compared among these groups. RESULTS We included 3869 non-OA knees from 2430 participants (age mean ± SD: 59.5±9.0, female/male:1.3). At the 4th-year follow-up, 2216 remained BML-free, 1106 showed an increase, and 547 showed a decrease in BML volume. Knees with progressive BML had a higher risk of radiographic knee OA incidence compared to BML-free (HR:3.01, P<0.001) and regressing BML (HR:2.00, P<0.001) knees. They also had a higher risk for symptomatic OA incidence compared to BML-free knees (HR:1.25, P:0.001). Larger volume changes in BML progression were associated with a higher risk of knee OA incidence (radiographic HR:1.95, symptomatic HR:1.70, P-values<0.001). In all subchondral plates, especially the medial femur and tibia, BML progression was associated with a higher risk of developing both radiographic and symptomatic knee OA compared to BML-free plates. CONCLUSION Progressive BMLs, according to the subregion and volume changes extent, are associated with an increased risk of OA incidence compared to BML-free or regressing BML knees, emphasizing the importance of monitoring BML volume changes in evaluating early interventions to prevent OA incidence.