Generating patient-specific 3D models using a pointing device

Abstract

Computer assisted mosaic arthroplasty shows promise compared to traditional treatments of cartilage degeneration in the knee. It can be more accurate and less invasive, leading to improved outcomes. Computer assisted mosaic arthroplasty requires a computerized 3D model of the patient's distal femur, normally generated from an Arthro-CT scan or MRI, but this process is associated with high cost, error, extended wait times, and radiation exposure. Our objective was to develop an alternative method to generate a 3D model intraoperatively within 1 mm error using points collected by a tracked pointing device. After generating the 3D model, we can optionally colour it according to a confidence metric. We validated our method in a user study where several orthopaedic surgeons and residents generated models of a rapid prototyped distal femur under different conditions. Our results suggest that it is possible to generate patient-specific 3D models with submillimeter accuracy using our method, and that this can be done within the time constraints of surgical procedures. We believe that the low error and minimal intraoperative time requirement may allow 3D model generation within the operating room without the need for any preoperative preparation.