A Split Renal Function-Based Approach for Predicting New Baseline Glomerular Filtration Rate After Radical Nephroureterectomy.

INTRODUCTION

Accurate prediction of new baseline glomerular filtration rate (NBGFR) after radical nephroureterectomy (RNU) is important for managing upper tract urothelial carcinoma (UTUC) because it can inform timing of systemic chemotherapy. Current prediction models do not account for split renal function (SRF) and show modest accuracy. This study evaluates the accuracy of an SRF-based model, PVA, which incorporates both parenchymal volume analysis (PVA) and degree of parenchymal enhancement.

METHODS

We reviewed patients with UTUC (n = 712) managed with RNU (2013-2023) and included patients with (1) contrast-enhanced CT < 1 year preoperatively, (2) glomerular filtration rate (GFR) < 3 months preoperatively, and (3) NBGFR 1 to 12 months postoperatively. Predicted NBGFR was 1.25 × (GFR) × (SRF), with 1.25 representing the average renal functional compensation after nephrectomy. For PVA, differential parenchymal volumes and degree of enhancement were estimated using semiautomated software. SRF-based models (based on PVA, PVA alone, or nuclear renal scans) and a non-SRF-based algorithm were compared using a 20% accuracy threshold.

RESULTS

Among patients analyzed (n = 352), the median preoperative GFR was 63 mL/min/1.73 m, 101 (29%) had moderate/severe hydronephrosis, and 42 (12%) had infiltrative renal masses (IRMs). For prediction of NBGFR after RNU, PVA demonstrated superior accuracy (84%) compared with PVA alone (accuracy = 79%, < .05), nuclear renal scans-based approach (accuracy = 73%, < .01), and non-SRF-based algorithm (accuracy = 65%, < .01). Among patients with hydronephrosis, notable improvements were observed for PVA compared with PVA alone (accuracies 88%/61%, respectively, ≤ .01). For IRMs, PVA seemed to be equivalent to other approaches for predicting NBGFR after RNU.

CONCLUSIONS

PVA incorporates both differential renal function (degree of enhancement) and parenchymal volumes and outperforms other SRF-based and non-SRF-based approaches for predicting NBGFR after RNU. These findings alleviate concerns that the prevalence of hydronephrosis and IRMs in this population reduces accuracy of SRF-based approaches. PVA can inform counseling about the timing of systemic chemotherapy in patients with high-risk UTUC.