Limitations of Parenchymal Volume Analysis for Estimating Split Renal Function and New Baseline Glomerular Filtration Rate After Radical Nephrectomy.

ABSTRACT

PURPOSE: Accurately predicting new baseline glomerular filtration rate (NBGFR) after radical nephrectomy (RN) can improve counseling about RN vs partial nephrectomy. Split renal function (SRF)-based models are optimal, and differential parenchymal volume analysis (PVA) is more accurate than nuclear renal scans (NRS) for this purpose. However, there are minimal data regarding the limitations of PVA. Our objective was to identify patient-/tumor-related factors associated with PVA inaccuracy. MATERIALS AND METHODS: Five hundred and ninety-eight RN patients (2006-2021) with preoperative CT/MRI were retrospectively analyzed, with 235 also having NRS. Our SRF-based model to predict NBGFR was 1.25 × (GlobalGFRPre-RN × SRFContralateral), where GFR indicates glomerular filtration rate, with SRF determined by PVA or NRS, and with 1.25 representing the median renal functional compensation in adults. Accuracy of predicted NBGFR within 15% of observed was evaluated in various patient/tumor cohorts using multivariable logistic regression analysis. RESULTS: PVA and NRS accuracy were 73%/52% overall, and 71%/52% in patients with both studies (n = 235, P < .001), respectively. PVA inaccuracy independently associated with pyelonephritis, hydronephrosis, renal vein thrombosis, and infiltrative features (all P < .03). Ipsilateral hydronephrosis and renal vein thrombosis associated with PVA underprediction, while contralateral hydronephrosis and increased age associated with PVA overprediction (all P < .01). NRS inaccuracy was more common and did not associate with any of these conditions. Even among cohorts where PVA inaccuracy was observed (22% of our patients), there was no significant difference in the accuracies of NRS- and PVA-based predictions. CONCLUSIONS: PVA was more accurate for predicting NBGFR after RN than NRS. Inaccuracy of PVA correlated with factors that distort the parenchymal volume/function relationship or alter renal functional compensation. NRS inaccuracy was more common and unpredictable, likely reflecting the inherent inaccuracy of NRS. Awareness of cohorts where PVA is less accurate can help guide clinical decision-making.