OBJECTIVE: Cartilage surface mapping is a technique that can visualize 3D cartilage thickness variation throughout a joint without a need for arbitrary regional definitions. The objective of this cross-sectional study was to utilize this technique to evaluate the cartilage thickness distribution in knee osteoarthritis patients and to analyze to what extent it depends on demographic, radiographic, and MRI structural pathology strata. METHODS: Patients of the IMI-APPROACH cohort were included, with MRIs obtained at 1.5 T or 3 T. Tibial and femoral cartilage segmentation and registration with a canonical surface were performed semi-automatically. Kellgren-Lawrence and OARSI grading were performed on knee radiographs
MOAKS scoring was performed on MRI scans. The association of demographics and radiographic and MRI scorings with cartilage thickness distribution was analyzed with general linear models using statistical parametric mapping. RESULTS: Two hundred eighty-seven patients were included. Male sex and height were positively associated with cartilage thickness particularly in the trochlea and medial femur, respectively, with differences up to 0.5 mm (male vs female), while radiographic joint space narrowing and bone marrow lesions showed region-specific negative associations (up to 0.14-0.5 mm per grade). Kellgren-Lawrence grade, MOAKS meniscal extrusion, and osteophytes showed patterns of positive and negative associations, with increasing grades showing reduced local tibiofemoral cartilage thickness, but greater thickness in the trochlea (both up to 0.2-0.3 mm per grade). CONCLUSIONS: Decreased height, female sex, and increasing tibiofemoral pathology were associated with thinner tibiofemoral cartilage. Unexpected results such as consistently thicker cartilage in the anterior femur with increasing disease or osteophytosis states provide opportunities for future research.