Accuracy of Computer Navigation–Assisted Arthroscopic Osteochondroplasty for Cam-Type Femoroacetabular Impingement Using the Model-to-Image Registration Method

Abstract

Background: Precise osteochondroplasty is important in arthroscopic hip surgery for cam-type femoroacetabular impingement (FAI). Although computer-assisted surgery with a navigation system may enhance the accuracy of arthroscopic osteochondroplasty, few clinical studies have assessed its accuracy. Purpose: To evaluate the accuracy of arthroscopic osteochondroplasty by a computed tomography (CT)–based navigation system for cam-type FAI, using 3-dimensional (3D) reconstruction with more detail compared with previous methods. Study design: Case series; Level of evidence, 4. Methods: Twenty patients (14 men and 6 women) who underwent navigation-assisted arthroscopic surgery for cam-type FAI were included. The preoperative 3D model of the femur was constructed from each patient’s CT data, and a planned model with virtual cam resection was generated. A femoral model was reconstructed from CT data postoperatively. The 3 models for each patient were overlaid using a 3D model registration method. Then, the contours of the bone resection area of each model were compared by measuring them. To measure the deviation between planned and actual bone resections, 4 cross-sectional images of the 3 femoral models were set at one-quarter intervals from the femoral head radius. All measurements were based on clockface lines set around the femoral neck axis at 30-minute intervals. Differences between the planned and postoperative contour lines were deemed resection deviations. Results: All cam resections were performed in the anterior half of the region of interest. Therefore, only the anterior half (48 points) of the 96 points per case were analyzed. In 876 (91.3%) points of the total measurement points (960 points/20 cases), the error in resection depth was within 3 mm. Overresection was observed at 35 (3.6%) points and underresection at 49 (5.1%) points. The observed maximum deviations from the planned models were 6.3 mm overresection and –7.1 mm underresection. The alpha angles of the postoperative model at the posterior 9- to 3-o’clock position were <55° in all patients. Conclusion: Navigation-assisted arthroscopic osteochondroplasty showed favorable accuracy. Underresection was more frequent than overresection on the anterosuperior side of the femur, despite assistance of the navigation system.