Abstract
Weight-bearing cone beam computed tomography (WB-CBCT, or simply WBCT), which captures high-resolution 3D images in a natural standing position, has gained increasing interest in recent years. This study examines the potential of WBCT as an alternative to multidetector computed tomography (MDCT) for 3D bone modelling. We generated 3D knee joint models from manually annotated WBCT and MDCT scans, performed rigid registration of these models, and assessed their similarity by evaluating the mean difference, standard deviation, and confidence intervals of the aligned models. The mean differences were computed as the average surface distances between corresponding WBCT and MDCT 3D bone models after rigid registration, providing a quantitative measure of their geometric similarity. Validation was conducted using both patient and cadaver scans to assess WBCT's clinical applicability under realistic conditions and its technical reliability with controlled samples. Our findings reveal an average absolute difference of less than 0.35 mm for patient scans and 0.30 mm for cadaveric scans between WBCT and MDCT. The patella demonstrated the smallest mean difference (-0.20 mm to 0.10 mm) and standard deviation (0.28 mm to 0.55 mm) across all scans. These results confirm the comparability of WBCT to MDCT for 3D bone modelling, highlighting WBCT's capacity to deliver appropriate image quality for the clinical assessment of bone joints.