Bony Morphological Characteristics of the Discoid Lateral Meniscus Knee: A Volumetric Magnetic Resonance Imaging-Derived Statistical Shape Modeling Study

IntroductionWe sought to characterize the three-dimensional (3D) osseous morphology of the distal femur and proximal tibia in knees with a discoid lateral meniscus (DLM), relative to knees with normal magnetic resonance imaging (MRI) findings, using landmark-based statistical shape modeling of volumetric MRI data. MethodsWe retrospectively analyzed 19 knee MRI scans from 16 patients with symptomatic DLM and 31 scans from 30 age-and sex-matched controls with normal knee MRI findings, all younger than 30 years.Bone surface models were segmented from 3.0-T MRI, and 1003 pseudo-landmarks were placed automatically.Generalized Procrustes analysis and principal component analysis were performed for the femur and tibia combined and for each bone separately.Each principal component (PC) was tested for association with DLM using Firth logistic regression adjusted for age and sex, with patient-clustered standard errors and Holm correction for multiple testing.Tibiofemoral pose and centroid size were compared between groups, and discrimination was estimated by patient-grouped cross-validation. ResultsThe mean DLM femur showed a posterior and distal deficit of the lateral femoral condyle with relative prominence of the anterolateral condyle and metaphysis, and the lateral tibial spine was positioned more laterally, widening the intercondylar eminence.In the femur-only model, a single component (PC4, 6.1% of variance) reproduced this femoral pattern and was strongly associated with DLM (odds ratio (OR) per SD 6.43, 95% confidence interval (CI) 3.02-13.73;Holm-adjusted p < 0.001).In the combined model, the first two PCs mainly reflected knee flexion and tibial rotation; PC4 (OR 3.31, 95% CI 1.15-9.53)and PC5 (OR 2.62, 95% CI 1.24-5.54),representing tibial intercondylar eminence width and lateral compartment joint space, were associated with DLM but not after Holm correction.No tibial component remained significant after correction.Knee pose and centroid size did not differ between groups.Cross-validated areas under the curve were 0.78 for the combined model and 0.80 for the femur-only model. ConclusionsIn this exploratory study, knees with DLM showed a characteristic posterolateral hypoplasia of the lateral femoral condyle with relative anterolateral prominence, which was associated with DLM independently of age, sex, knee position, and overall size.These hypothesis-generating findings warrant validation in larger cohorts.