Impact of three-dimensional reconstruction in preoperative planning for partial nephrectomy: A prospective study

ABSTRACT Introduction: Two-dimensional (2D) imaging has traditionally guided surgical planning for partial nephrectomy (PN). However, it has limitations in depicting complex renal anatomy. Three-dimensional (3D) CT reconstructions may enhance spatial understanding, but evidence of their added value remains limited. This study evaluates the role of 3D reconstruction in preoperative planning. Materials and Methods: This was a prospective, observational, single-center cohort study conducted between June 2021 and October 2023. Patients with cT1 renal masses planned for robotic PN willing for additional 3D reconstruction were enrolled. Preoperative predictions were made first with 2D computed tomography and then with combined 2D+3D reconstructions across seven parameters: Sinus dissection, selective clamping, margin status, warm ischemia (<30 min), pelvicalyceal system (PCS) entry, venous clamping, and vascular injury. Predictions were compared with intraoperative findings. Results: Fifty-four patients with cT1 renal masses were included in the study. The median age of the patients was 49 years, and most tumors were solitary (90.7%) and of moderate complexity (64.8%). Trifecta outcomes were achieved in 92.6%. Incorporating 3D reconstruction resulted in consistent descriptive improvements in predictive accuracy of 3.7%–9.2%, although statistical significance was not reached. McNemar’s test demonstrated statistically significant discordance between 2D+3D-based prediction and intraoperative findings for warm ischemia time (11 patients), PCS entry (21 patients), and major vascular injury (7 patients). Nevertheless, predictive accuracy and Cohen’s kappa values favored combined imaging. Conclusion: Integration of 3D reconstruction with 2D improved the accuracy of preoperative prediction of intraoperative events, thereby supporting more informed preoperative planning. Further studies with larger sample size are warranted to identify tumor complexity subsets that derive the greatest benefit and the impact on the learning curve of PN.