University Hospital Center of Orleans Orléans, France
Disclosure(s): No financial relationships with ineligible companies to disclose
Background/Purpose: Osteoarthritis is one of the most prevalent chronic diseases worldwide, with knee osteoarthritis (KOA) being a particularly painful and debilitating form. The minimum tibiofemoral joint space width (mJSW) is a key criterion for assessing KOA progression. This study aims to evaluate a fully automated method for measuring mJSW in a reproducible and reliable manner, accounting for structural changes in knee anatomy. Methods: This study included 599 patients from the Foundation for the National Institutes of Health (FNIH) cohort, in which measurements of the Medial Compartment mJSW were reported. Our method consists of two main steps. First, we use a machine learning-based tool to automatically detect the tibiofemoral contour of the knee. Second, we compute mJSW as perpendicular distance from each point on the femoral side to the line connecting the two closest points on the tibial side, Figure 1. Statistical analyses were performed to compare our measurements with those obtained by the FNIH team. Reliability was assessed using the Intraclass Correlation Coefficient (ICC) and the Bland–Altman analysis. Results: The ICC demonstrated excellent agreement (0.906, 95% CI: 0.891 – 0.919) across all patients. The Bland–Altman analysis revealed a bias of -0.202 mm (limits of agreement: -1.095 mm to +0.691 mm), Figure 2. The observed bias is systematic and does not depend on the width of the joint space. Conclusion: The proposed method enables fully automated, reliable, and reproducible measurement of mJSW. While our method slightly underestimates mJSW compared to the reference measurements from the FNIH team, it remains a robust and consistent approach for assessing KOA progression.
