Identification of Osteoarthritis Kinematic Phenotypes Using Cluster Analysis on Knee Kinesiography Data

Previous studies highlight that identifying phenotypes is crucial for developing effective treatments for knee osteoarthritis (OA). This study aims to identify kinematic phenotypes in a knee OA population and characterize them by patient biomechanical markers which are functional parameters used in clinical settings. Knee kinematics are measured using the KneeKG (Knee Kinesiography) system, a technology that objectively assesses 3D knee kinematics. Kinematic data were categorized into homogeneous groups using a clustering process with a discriminant model called Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH). We identified five distinct phenotypes, exhibiting significant statistical differences (p < 0.05) in 3D kinematics, and linked these phenotypes to biomechanical markers measurable in clinical settings.