AI 3D preoperative planning for total hip arthroplasty: meta-analysis of 8 studies finds exact cup and stem sizing predicted roughly 3 to 4 times more often than with 2D templating
Source: Journal of Experimental Orthopaedics·Published: 2026
Authors: Altahtamouni SB, Salman LA, Al-Ani A, Ahmed G·DOI: 10.1002/jeo2.70427Open Access
Key figure: Forest plots (Figures 2 and 3) — Forest plots of the pooled odds ratios for acetabular cup and femoral stem size prediction, showing AI-assisted 3D planning outperforming 2D templating across all eight included studies. View in source
Bottom line: Across eight studies from Chinese centers, AI-assisted 3D planning was 3.85 times more likely to predict the acetabular cup to the exact size than 2D templating (95% CI 2.79 to 5.32, p < 0.0001) and 3.28 times more likely to predict the femoral stem to the exact size (95% CI 2.56 to 4.22, p < 0.0001). All included studies originated in China, limiting geographic generalizability.
What the study did
The authors searched PubMed, Scopus, and Embase from inception through October 2024 for studies comparing AI-assisted 3D preoperative planning against conventional 2D templating for acetabular cup and femoral stem sizing in total hip arthroplasty. Eight studies with 1,371 participants were included in the meta-analysis. A random-effects model was used given high between-study heterogeneity. Odds ratios with 95% confidence intervals were calculated for exact-size predictions and for predictions within one standard deviation. The Newcastle-Ottawa Scale assessed study quality. The analysis followed PRISMA reporting guidelines.
What they found
For the acetabular cup, AI-assisted planning predicted the exact size significantly more often than 2D templating (OR 3.85, 95% CI 2.79 to 5.32, p < 0.0001; I² = 42%) and predicted a size within one standard deviation more often as well (OR 3.49, 95% CI 1.21 to 10.13, p = 0.0212; I² = 81%). For the femoral stem, AI outperformed 2D templating for exact-size prediction (OR 3.28, 95% CI 2.56 to 4.22, p < 0.0001; I² = 0%) and for within-one-standard-deviation prediction (OR 5.35, 95% CI 3.84 to 7.45, p < 0.0001; I² = 0%). Newcastle-Ottawa quality scores ranged from 6 to 9.
Why it matters for orthopedic practice
Implant sizing is one of the most predictable determinants of stable fixation and soft-tissue balance in total hip arthroplasty. Correct preoperative sizing reduces reliance on intraoperative trial-and-error, shortens operative time, and lowers the risk of periprosthetic fracture in press-fit designs. This meta-analysis provides the first pooled effect sizes suggesting AI-assisted 3D planning is materially more accurate than conventional 2D templating, rather than equivalent. For surgeons considering a move to 3D planning workflows, the data support the investment on sizing grounds alone, before any operative-time or complication analyses.
Limitations
Every included study originated in a Chinese center, and Asian anatomy may not generalize to North American or European populations, particularly for femoral morphology. Study designs were mixed, with 38% prospective cohorts and the remainder retrospective. Heterogeneity for within-one-standard-deviation cup prediction was high (I² = 81%). The meta-analysis does not report operative time, functional outcomes, or revision rates, only sizing accuracy. No study directly compared different AI platforms against each other.
Altahtamouni SB, Salman LA, Al-Ani A, Ahmed G. The accuracy of artificial intelligence in 3D preoperative planning for total hip arthroplasty: a systematic review and meta-analysis. J Exp Orthop. 2026. doi:10.1002/jeo2.70427
Publishing AI research in orthopedics?
OSCRSJ accepts case reports and series on novel AI-assisted diagnoses and surgical planning. Free to publish in 2026.
Submit a manuscriptOSCRSJ News items are editorial summaries for educational purposes. They are not clinical recommendations, endorsements, or substitutes for the primary literature. Always consult the source paper and applicable specialty-society guidelines before changing practice.