Most of computer-assisted computer assisted system rely on the peri-operative acquisition of the anterior pelvic plane defined as the plane crossing the two anterior iliac spine and the symphysis. The goal of this study was to evaluate in vivo and in vitro the accuracy of the anterior pelvic plane acquisition, considered as the reference for computer-assisted total hip arthroplasty (THA).

Cup placement was performed using an imageless computer-assisted system in thirty patients during THA. Post-operatively the position of the cup was evaluated on computed tomography using a validated tridimensional software. The differences between the perioperative and postoperative angles for abduction and anteversion were compared using a two-group pair test.

On two cadavers four clinicians performed ten times the anterior pelvic plane acquisition using three Methods: percutaneously, with ultrasound and by direct bony acquisition defined as the reference. The mean error for each anterior pelvic plane acquisition method was compared using a univariate variance model for repeated measurements.

In vivo, the mean difference between the perioperative and postoperative abduction angles was 4° and not statistically significant. For anteversion, the difference was 4° and not significant in patients with BMI < 27. The difference was 11° and significant in patients with BMI > 27 (p< 0.001). In vitro, the mean errors for rotation and tilt were respectively 3.8 ° and 19.25 ° for cutaneous acquisition, 2.8° and 6.2° for ultrasound acquisition method. The errors were statistically higher with the percutaneous method (p< 0.001).

According to our results, the accuracy of the standard percutaneous acquisition method of the anterior pelvic plane in computer-assisted THA is limited. The ultrasound acquisition method may represent a reliable alternative.

Intraoperative assessment of knee kinematics should help surgeons optimizing total knee replacement. The purpose of this work was to validate information delivered by an adapted navigation system in 10 healthy cadaver knees and to investigate kinematics of 10 osteoarthritic (OA) knees in patients undergoing total knee replacement. The system displayed the magnitude of axial rotation, the position of the instantaneous centre of axial rotation and the displacements of the condyles. Successive cycles from full extension to 140° of fiexion in the same knee produced a mean external rotation of 19.7±10°, which was correlated to knee fiexion (r=0,60±0.2 in healthy knees, r=0.79±0.14 in OA knees). The center of axial rotation migrated posteriorly an average of 8.2 mm in both groups. The posterior displacements were 4.0 ±5.4 mm in healthy and 5±6.3 mm in OA knees for the medial condyle, and 20.9±9.1 mm in healthy and 20.3±10 mm in OA knees for the lateral condyle. The medial condyle lifted off beyond 110° of fiexion. Results in healthy knees were consistent with those obtained using fiuoroscopy and dynamic MRI. The kinematics of healthy and of OA knees with an intact anterior cruciate ligament did not differ significantly.