Effect of three-dimensional virtual modeling on surgical outcomes in robot-assisted partial nephrectomy for localized...

Three-dimensional virtual modeling (3DVM) is increasingly used in robot-assisted partial nephrectomy (RAPN), yet its true clinical utility remains uncertain because prior studies rarely accounted for surgeon experience or relied on unmatched cohorts. This study aimed to evaluate the association of 3DVM with Trifecta achievement and perioperative performance after accounting for tumor complexity and surgeon experience.

We retrospectively analyzed 80 RAPN cases (27 with 3DVM). Propensity-score matching (1:1) was performed using demographic and anatomical variables, while surgeon-related confounding was addressed using mixed-effects regression with surgeon included as a random effect. The primary endpoint was Trifecta (negative margin, warm ischemia time ≤25 min, and no 30-day complications). Mixed-effects logistic regression was used to evaluate independent predictors.

Propensity score matching was used to balance patient and tumor characteristics between groups, while surgeon-related effects were further addressed using mixed-effects regression models. In the matched cohort, 3DVM use resulted in significantly shorter warm ischemia time (20.3 ± 9.3 vs 28.5 ± 7.1 min; p = 0.001) and more frequent arterial-only clamping (81.5% vs 18.5%; p < 0.001). Trifecta was achieved more often with 3DVM (74.1% vs 37.0%; p = 0.014). In mixed-effects logistic regression analyses accounting for surgeon-level random effects and continuous surgeon experience, the association between 3D virtual modeling and trifecta achievement was not statistically significant (OR 2.25, 95% CI 0.44-11.45, p = 0.33).

These findings suggest that while 3D virtual modeling is associated with improved perioperative outcomes in matched comparisons, its independent effect on trifecta achievement remains uncertain after accounting for surgeon-level heterogeneity.