Concurrent validation of OpenCap for identifying ACL re-injury risk factors during a drop jump test in a healthy cohort

Abstract

Abstract 3D motion analysis (3DMA) can help identify patients at increased risk of ACL re-injury, but traditional marker-based systems have limited clinical accessibility. OpenCap, a novel, low-cost, markerless system, aims to enhance accessibility to 3DMA. This study evaluated the concurrent validity of a modified OpenCap version using a 2-DOF knee model for kinematics, while kinetics and ground reaction forces were derived using the native 1-DOF model, compared to a marker-based system. Twenty-four healthy participants performed 240 drop jumps, with data simultaneously captured by both systems. Root mean square error (RMSE), mean absolute error (MAE), maximum error, Pearson correlation, Bland-Altman plots, and statistical parametric mapping (SPM) were used to analyze inter-system differences. RMSE exceeded 6° for frontal-plane knee kinematics with strong waveform correlations ( r  > 0.90). Transverse-plane hip moments showed normalized MAE < 1% with weak to strong negative correlations. Sagittal-plane knee moments had normalized MAE of 5.6% and strong correlations ( r  > 0.90). Vertical GRFs showed normalized MAE > 6% and strong correlations ( r  > 0.90). SPM identified significant differences across most ground contact phases, and Bland-Altman analyses showed wide agreement limits for knee moment asymmetry at initial contact. OpenCap currently cannot be recommended for ACL re-injury risk assessment but demonstrated potential for increasing 3DMA accessibility.