Statistical shape models based 2D/3D registration methods for knee orthopaedic surgery

In orthopaedic surgery, and in case of Total Knee Arthroplasty (TKA) for osteoarthritic subjects, planning of the intervention is of fundamental importance. The insertion of the prosthetic components is extremely delicate because the deformations induced by the osteophytes must be resected and at the same time the mechanical axis of the leg must be corrected. The size of the implant and the plane of resection are estimated through a standing X-ray projection that covers the whole leg. This view gives a hint on the direction of the mechanical axis that must be restored, but cannot evaluate how the kinematics of the knee is influenced by the insertion. The possibility to check the pre-operative kinematic of the knee under weight bearing conditions, in order to evaluate tension of the ligaments and the distance between the bones, would be of great importance for the success of the intervention, giving to the surgeon the possibility to check the bone’s motion before entering in the surgery room. This evaluation is currently intraoperatively performed by the surgeon, who performs passive movements of the joint to check the correct placement of the prosthetic components, although the conditions are not similar to the real stress applied during everyday life, as weight and muscles strengths are missing. The use of fluoroscopic sequences is common in clinics to rapidly evaluate the knee kinematics. These fast and low dose X-ray images allow an accurate visualization of the bone movements without the invasiveness typical of other methodologies. Knee motion analysis is cur-