RLF Score: A Simplified Three-Parameter Model Integrating R.E.N.A.L. and MAP for Predicting Pentafecta Achievement in Robot-Assisted Partial Nephrectomy

Abstract Objective Precise prediction of perioperative outcomes for robot-assisted partial nephrectomy (RAPN), a cornerstone treatment for localized renal cell carcinoma (RCC), requires optimization. While the R.E.N.A.L. nephrometry score quantifies tumor anatomy and the Mayo Adhesive Probability (MAP) score assesses perinephric adhesion risk, their predictive efficacy remains debated. This study aimed to develop and validate a novel predictive model to guide RAPN clinical decision-making to develop and validate a novel predictive model to guide RAPN clinical decision-making. Methods A retrospective analysis was conducted on 704 patients undergoing primary RAPN for unilateral tumors at Nanjing Drum Tower Hospital between 2018 and 2023. Preoperative CT/MRI images were independently evaluated by three blinded physicians to obtain R.E.N.A.L. (Radius, Exophytic/endophytic, Nearness to collecting system, Location relative to polar lines) and MAP (Fat Thickness, Fat Stranding) parameters. The primary endpoint was Pentafecta achievement, defined as Trifecta (negative surgical margin, absence of Clavien-Dindo ≥ grade III complications, ≤ 30% decline in estimated glomerular filtration rate [eGFR]) plus intraoperative estimated blood loss (EBL) ≤ 50 mL and operative time (OT) ≤ 150 min. Multivariate regression identified independent predictors used to construct the RLF scoring system. Model performance was validated using the area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA). Results Multivariate linear regression identified R (β = 23.95, P < 0.001), L (β = 9.63, P < 0.001), and FS (β = 9.57, P < 0.001) as core predictors of OT. R (β = 49.79, P < 0.001) and FS (β = 13.64, P < 0.001) were independent predictors of EBL. Multivariate logistic regression confirmed R (OR = 1.60, *P = 0.011*), L (OR = 1.48, *P = 0.014*), and FS (OR = 1.35, P < 0.001) as independent predictors of Trifecta. Furthermore, R (OR = 2.99, P < 0.001), L (OR = 2.06, P < 0.001), and FS (OR = 1.88, P < 0.001) were identified as independent predictors of Pentafecta achievement. The RLF score, integrating R, L, and FS, demonstrated significantly superior predictive efficacy for Pentafecta (AUC = 0.78, 95% CI: 0.75–0.81) compared to the R.E.N.A.L. score (AUC = 0.71, 95% CI: 0.67–0.75) and MAP score (AUC = 0.70, 95% CI: 0.66–0.74). At the optimal cutoff value of 0.587, sensitivity was 0.69, specificity was 0.75, and the negative predictive value (NPV) was 0.81. Calibration curves and the Hosmer-Lemeshow test (P = 0.760) indicated good model fit. Conclusion The RLF score, the first simplified model integrating R (tumor diameter), L (polar location), and FS (fat stranding) from R.E.N.A.L. and MAP scores, effectively stratifies perioperative risk for RAPN. Its high NPV reliably identifies low-risk patients suitable for standardized protocols, while its predictive value guides intensified preoperative preparation for high-risk cases, providing an evidence-based framework for personalized surgical management. The model's reliance on fundamental anatomical parameters suggests potential for broader applicability.