Abstract
BACKGROUND: With the growing complexity of total hip arthroplasty (THA) for high hip dislocation (HHD), artificial intelligence (AI)-assisted three-dimensional (3D) preoperative planning has emerged as a promising tool to enhance surgical accuracy. This study compared clinical outcomes of AI-assisted 3D versus conventional two-dimensional (2D) X-ray preoperative planning in such cases. METHODS: A retrospective cohort of 92 patients with Crowe type II-IV HHD who underwent THA between May 2020 and January 2023 was analyzed. Patients received either AI-assisted 3D preoperative planning (n = 49) or 2D X-ray preoperative planning (n = 43). The primary outcome was the accuracy of implant size prediction. Secondary outcomes included operative time, blood loss, leg length discrepancy (LLD), implant positioning, functional scores (Harris Hip Score [HHS], WOMAC, VAS), complications, and implant survival at 24 months. RESULTS: At 24 months, both groups demonstrated significant improvements in functional outcomes. Compared to the 2D X-ray group, the AI-3D group showed higher accuracy in implant size prediction (acetabular cup: 59.18% vs. 30.23%; femoral stem: 65.31% vs. 41.86%; both p < 0.05), a greater proportion of cups placed within the Lewinnek and Callanan safe zones (p < 0.05), shorter operative time, reduced intraoperative blood loss, and more effective correction of leg length discrepancy (all p < 0.05). No significant differences were observed in HHS, WOMAC, or VAS scores between groups at 24 months (all p > 0.05). Implant survivorship was also comparable (100% vs. 97.7%; p = 0.283), with one revision noted in the 2D X-ray group. CONCLUSION: AI-assisted 3D preoperative planning improves prosthesis selection accuracy, implant positioning, and perioperative outcomes in Crowe type II-IV HHD THA, although 2-year functional and survival outcomes were comparable to 2D X-ray preoperative planning. Considering the higher cost, radiation exposure, and workflow complexity, its broader application warrants further investigation, particularly in identifying patients who may benefit most.