4DCT and derivative-based functional data clustering were used to identify distinct patterns of tibiofemoral joint rotation. The approach employed a combination of first derivatives of the rotational motion, functional principal component analysis, and k-means clustering to classify rotational patterns. The method revealed meaningful kinematic variations in healthy knees by focusing on motion trends rather than amplitude. These findings provide a novel framework for assessing pathological deviations, offering potential improvements in diagnostic and therapeutic strategies.