CT-based navigation for acetabular cup placement in supine and lateral positions: a simulated surgery study using patient-specific 3D-printed pelvic models

Purpose Accurate acetabular cup placement is critical for the success of total hip arthroplasty (THA). The aim of this study was to evaluate the accuracy and precision of a CT-based, AI-driven navigation tool for acetabular cup placement in supine and lateral patient positions, using a patient-specific 3D-printed pelvic model with an embedded acetabular cup rim as a physical measurement reference. Methods Twenty-six patient-specific pelvic models were generated from preoperative CT data. A virtual acetabular cup was added to each pelvic mesh, cropped to the rim, and 3D-printed together with the bone, yielding a physical cup reference embedded in the printed pelvis. The ground-truth cup orientation and position were defined within the planning application by overlaying a cup component template on the embedded rim. Each model was tested once in supine and once in lateral position in a paired design. Ten cases were repeated three times per position to assess intra-session precision. Outcomes were surface matching root-mean-square error (RMSE), cup anteversion, inclination, and depth errors. Analyses included two one-sided tests (TOST) for equivalence (±0.5 mm), Bland–Altman agreement, Levene's test, and intraclass correlation coefficients (ICC). Results Mean alignment-point RMSE was 0.26 mm (supine) and 0.28 mm (lateral); mean independent test-point RMSE was 0.30 mm in both positions, with TOST confirming equivalence within ±0.5 mm. Mean absolute anteversion error was 0.67° (supine) and 0.44° (lateral), with comparable inclination errors. Mean absolute cup depth error remained well below ±3 mm. Bland–Altman analyses showed narrow 95% limits of agreement between positions. All ICC(3,1) values for cup orientation exceeded 0.99 in both positions. Conclusions The evaluated navigation tool achieved sub-millimeter surface matching, small cup placement errors, and excellent intra-session precision in both supine and lateral positions. The embedded-rim phantom enabled direct measurement of cup orientation and seating depth on the printed model and is applicable to other CT-based navigation systems.