Introduction:

Partial-nephrectomy(PN) is generally preferred for localized renal-cell-carcinoma(RCC) due to greater preservation of GFR compared to radical-nephrectomy(RN). Accurate prediction of NBGFR after PN may be important for preoperative counseling because NBGFR can have implications for long-term survival, particularly for patients with preoperative chronic-kidney-disease.  Previous studies have shown that on average PN saves ≈80% of the function of the operated kidney, and ≈90% of the global renal function when two kidneys are present.¹ Several multivariate algorithms for predicting NBGFR after PN have been developed.^2-6 While these models demonstrate strong predictive accuracies, they are methodologically complex, and the factors identified as “significant predictors of NBGFR” vary considerably.  Estimation of some of the parameters in these algorithms can also be laborious.  Herein, we propose a conceptually simple approach that presumes that 90% of the global GFR will be saved, and compare its predictive accuracy with those of previously published, multivariate algorithms.^2-6

Methods:

345 RCC patients with bilateral kidneys who underwent PN for a unifocal renal tumor at the Cleveland Clinic were analyzed. Inclusion required the availability of preoperative and postoperative measures of renal function. NBGFR was defined as the final GFR 3-12 months post-PN. Predictive accuracies were assessed from correlation coefficients (r) for each model based on alignment between predicted and observed NBGFR. Mean-squared errors (MSE) were determined for all approaches to quantify the amount of error in predicting NBGFR, with lower values representing more accurate prediction. Reported p-values are based on comparisons between the 90% functional preservation approach and each of the multivariate models, and are derived from two-tailed tests. Our simplified approach was: Predicted NBGFR=Global GFR_Pre-PN×0.90. Figure 1 provides the references for the multivariable models and an example from one of these studies to illustrate the complexity of such algorithms.

Results:

The r value for our conceptually simple model based on 90% preservation of global renal function with each PN was 0.91 (Figure 1A). For the various multivariate models, the r values were 0.88, 0.90, 0.91, 0.89, and 0.90 (Figure 1B-F, respectively). The differences in r values across the six models were not statistically significant. The MSE value for the approach based on a presumption of 90% preservation of global renal function was 83.7 (Figure 1A). The MSE values for the multivariate models were 276 (p<0.0001), 121.9 (p=0.008), 88.7 (p=.69), 104.8 (p=0.15), and 92.5 (p=0.49) (Figure 1B-F, respectively).

Conclusion:

We propose a novel and extremely simple approach to predict NBGFR after PN based on a presumption of 90% preservation of renal function with each case that can facilitate counseling about expected postoperative renal functional outcomes in RCC patients undergoing PN. This model performs equally well or better in terms of predicting NBGFR after PN when compared with complex, multivariate approaches that were previously published. Notably, our proposed approach is conceptually simpler and more clinically implementable, as it relies primarily on the anchoring effects of preoperative GFR. In contrast, the multivariate algorithms relied on unique combinations of factors such as BMI, contact-surface area, predicted percent parenchymal mass preserved, surgical approach, presence of proteinuria, and comorbidities, among others. Further studies on the implications of the R.E.N.A.L. score on the extent of renal functional preservation are warranted to refine predictions of NBGFR on a more personalized level.

Funding: N/A