External validation of the Simplified PADUA REnal (SPARE) nephrometry system in predicting surgical outcomes after partial nephrectomy.

BACKGROUND Pentafecta is a major goal in the era of partial nephrectomy (PN). Simplified PADUA REnal (SPARE) nephrometry system was developed to evaluate the complexity of tumor. However, the predictive ability in pentafecta of SPARE system is yet to be determined. The aim of this study was to externally validate the applicability of SPARE nephrometry system in predicting pentafecta achievement after partial nephrectomy, and to examine inter-observer concordance. METHODS We retrospectively reviewed data of 207 consecutive patients who underwent PN between January 2012 and August 2018 at a tertiary referral center. We obtained SPARE, R.E.N.A.L., and PADUA scores and evaluated correlations among the nephrometries and surgical outcomes including pentafecta by Spearman test. Logistic regression analysis was used to identify independent predictors of pentafecta outcomes. We compared the nephrometries to determine the predictive ability of achieving pentafecta using receiver operating characteristic curve analysis. Fleiss' generalized kappa was used to assessed interobserver variation in the SPARE system. RESULTS Based on the SPARE system, 120, 74, and 13 patients were stratified into low-risk, intermediate-risk, and high-risk groups, respectively. Regarding the individual components of pentafecta, there were significant differences in the complication rate (p = 0.03), ischemia time (p < 0.001), and percent change of eGFR (p < 0.001) among the three risk groups. In addition, higher tumor complexity was significantly associated with a lower achievement rate of pentafecta (p = 0.01). In Spearman correlation tests, SPARE nephrometry was correlated with ischemia time (ρ:0.37, p < 0.001), operative time (ρ:0.28, p < 0.001), complication rate (ρ:0.34, p < 0.001), percent change of eGFR (ρ:0.34, p < 0.001), and progression of chronic kidney disease stage (ρ:0.17, p = 0.02). Multivariate analysis revealed that SPARE significantly affected pentafecta (OR: 0.67, p < 0.001). In ROC curve analysis, SPARE showed fair predictive ability in the achievement pentafecta (AUC: 0.71). The predictive ability of pentafecta was similar between nephrometries (SPARE vs. R.E.N.A.L., p = 0.78; SPARE vs. PADUA, p = 0.66). The interobserver concordance of SPARE was excellent (Kappa: 0.82, p = 0.03). CONCLUSIONS SPARE system was a predictive factor of surgical outcomes after PN. This refined nephrometry had similar predictive abilities for pentafecta achievement compared with R.E.N.A.L. and PADUA.