Three-dimensional virtual models accurately predict preserved parenchymal volume after robot-assisted partial nephrectomy: a prospective study.

Abstract

Introduction: Accurate prediction of postoperative renal function remains a challenge in robot-assisted partial nephrectomy (RAPN). Preserved parenchymal volume (PPV) has emerged as a key determinant of long-term renal function. Three-dimensional virtual models (3DVM) enable patient-specific anatomical reconstruction and volumetric quantification, potentially allowing preoperative estimation of PPV. However, their accuracy in predicting postoperative PPV and functional outcomes has not been prospectively validated. This study aimed to assess the accuracy of 3DVM in predicting PPV after RAPN and to explore their relationship with postoperative renal function and anatomical complexity.

Materials and methods: A prospective single-centre study included patients undergoing RAPN with preoperative 3DVM-based planning (2022-2024). Estimated preoperative PPV was compared with real postoperative PPV obtained from follow-up 3DVM. Agreement was evaluated using paired statistics and Spearman's correlation. Functional outcomes (serum creatinine, eGFR) were correlated with PPV and WIT, whilst tumour complexity was assessed by PADUA and RENAL scores.

Results: Twenty-nine patients were included (mean age 59.7 years; 72% male). Median estimated PPV was 156.7 ml and real postoperative PPV was 148.6 ml, with a mean absolute difference of 7.3 ml (~ 6%). Estimated and real PPV were strongly correlated (ρ = 0.97; p < 0.0001). PADUA score correlated negatively with postoperative eGFR (ρ ≈ -0.4; p < 0.05), whilst RENAL score and WIT did not. Median parenchymal loss was 6%, with low morbidity (17% Clavien I-II).

Conclusions: 3DVM accurately predict postoperative PPV after RAPN, showing excellent concordance with real volumetric measurements. Integration of 3D volumetry into surgical planning may provide a quantitative tool to support nephron-sparing strategies and functional preservation.