EP4.10 Accuracy of Semi-Automated 3D CT Hip Measurement for Femoroacetabular Impingement: Validation Against Surgeon and Engineering Reference Standards

Abstract

Abstract Introduction Accurate quantification of hip morphology is essential for the diagnosis and management of femoroacetabular impingement (FAI). The purpose of this study was to evaluate the measurement accuracy of CT based three-dimensional analyses (HipMap FAI Analysis) against clinically and engineering derived reference standards. Methods Measurement accuracy was assessed using 36 bilateral hip CT scans representing cam only, pincer only, and mixed FAI morphologies based on standard radiographic measurements (patient age 39 ± 11 years, 57% female/43% male, 56% right-sided affected hips/44% left-sided). Reference values for lateral center edge angle, Tönnis angle, acetabular version, femoral torsion, and femoral neck–shaft angle were obtained by ten independent high-volume hip surgeons measurements on standardized images, with a mean of three unique measurements per case. Reference values for alpha angle were based on three dimensional sphere fitting method and obtained through external engineering experts (Medical Metrics, Inc., Houston, TX, USA). Acetabular coverage accuracy was assessed by quantifying the impact of adjusting the three-dimensional surface by up to 1 mm. For all parameters, mean signed and absolute differences between HipMap and reference measurements were calculated. Signed differences indicate direction (over- or underestimation), while absolute differences reflect the magnitude of the error only. Results The differences between HipMap and the reference values were in a similar range as between surgeon measurements. The lowest mean absolute difference between HipMap and surgeon/reference measurements was observed for acetabular coverage (1%) and the largest mean absolute difference for femoral torsion (5°). Generally, the mean absolute differences had a greater magnitude than the mean signed differences, indicating that the reference measurements were not consistently smaller or larger, but equally distributed around the HipMap measurement. In comparison, the absolute difference between the surgeon rater and surgeon mean ranged between 3° ± 4° for Tönnis angle and 0.4° ± 0.3° for acetabular version. Mean signed differences between surgeon/reference and HipMap measurements were -1° for center edge angle, 2° for Tönnis angle, 0° for acetabular version, 0% for acetabular coverage, -1° for alpha angle, 3° for femoral torsion, and 2° for neck shaft angle. Similarly, the average signed difference between the surgeon rater and surgeon mean ranged between 0° and 2°. Conclusion These findings demonstrate that HipMap FAI Analysis provides accurate hip morphometric measurements across a range of clinically relevant parameters. The combination of surgeon based validation and external three dimensional engineering reference standards supports the robustness of the methodology. Clinically, reliable automated morphometry may reduce measurement variability, support consistent preoperative planning, and facilitate objective assessment of complex three- dimensional hip anatomy in patients with FAI.