Related Experiment Video
Updated: Aug 5, 2026
The Use of Mixed Reality in Custom-Made Revision Hip Arthroplasty: A First Case Report
Published on: August 4, 2022
Three-dimensional and artificial intelligence-assisted preoperative templating in total hip arthroplasty: a
Seyed Pouya Taghavi1, Ali Salmani1, Taner Karlidag2
1Neuromusculoskeletal Research Center, Iran University of Medical Sciences, Tehran, 1449614535, Iran.
Background:
Preoperative templating in total hip arthroplasty (THA) optimizes implant sizing and positioning accuracy. Standard practice relies on two-dimensional (2D) templating, while three-dimensional (3D) and AI-assisted planning are evolving alternatives. This study aimed to evaluate the implant size prediction accuracy and its impact on intraoperative and postoperative outcomes.
Methods:
We conducted a systematic review and meta-analysis in April 2025 across MEDLINE (PubMed), Scopus, and Web of Science to identify studies comparing advanced preoperative planning (AI-based or 3D) with 2D planning in THA. Data on femoral stem and acetabular cup sizing accuracy and key clinical outcomes were extracted. Study quality was evaluated using the Joanna Briggs Institute (JBI) critical appraisal checklist. Overall, 32 studies were identified.
Results:
Meta-analysis showed that 3D planning had significantly higher exact implant size agreement than 2D (femoral OR = 2.00, 95% CI 1.34-2.97; acetabular OR = 5.32, 95% CI 1.89-14.94), and significantly lower major deviation (femoral OR = 0.37, 95% CI 0.24-0.58; acetabular OR = 0.18, 95% CI 0.06-0.50). AI-assisted planning showed significantly higher exact implant size agreement (femoral OR = 3.27, 95% CI 2.52-4.24; acetabular OR = 3.96, 95% CI 2.95-5.32) and significantly lower major deviation (femoral OR = 0.22, 95% CI 0.16-0.31; acetabular OR = 0.29, 95% CI 0.15-0.56). Significantly reduced blood loss (MD = - 36.21, 95% CI - 46.58 to - 25.85), operative time (MD = - 15.54, 95% CI - 24.25 to - 6.83), LLD (MD = - 1.58, 95% CI - 2.22 to - 0.93), and higher abduction angle (MD: 1.64; 95% CI 0.37-2.92) were associated with AI-based planning. Postoperative Harris Hip Score was marginally higher with AI-assisted planning (MD = 0.73 points), but this difference was below the minimal clinically important threshold and not robust to sensitivity analysis.
Conclusion:
Advanced planning improved implant sizing accuracy. AI-assisted planning was additionally associated with more favorable intraoperative and postoperative outcomes; however, heterogeneity, limited long-term data, and restricted external validity due to the geographic concentration of AI-assisted studies and limited diversity of AI software platforms warrant further standardized comparative studies.