A Novel Mathematical Model to Predict the Severity of Postoperative Functional Reduction before Partial Nephrectomy: The Importance of Calculating Resected and Ischemic Volume

Purpose: Preoperatively predicting postoperative kidney function is an essential step to achieve improved renal function and prevent chronic kidney disease. We introduce a novel formula especially to calculate resected and ischemic volume before partial nephrectomy. We examined whether resected and ischemic volume would have value for predicting postoperative renal function. Materials and Methods: We performed a retrospective cohort study in 210 patients who underwent robotic partial nephrectomy between September 2006 and October 2013 at a tertiary cancer care center. Based on abdominopelvic computerized tomography and magnetic resonance imaging we calculated resected and ischemic volume by the novel mathematical formula using integral calculus. We comparatively analyzed resected and ischemic volume, and current nephrometry systems to determine the degree of association and predictability regarding the severity of the postoperative functional reduction. Results: On multivariable analysis resected and ischemic volume showed a superior association with the absolute change in estimated glomerular filtration rate/percent change in estimated glomerular filtration rate (B = 6.5, p = 0.005/B = 6.35, p = 0.009). The ROC AUC revealed accurate predictability of resected and ischemic volume on the stratified event of an absolute change in estimated glomerular filtration rate/event of percent change in estimated glomerular filtration rate compared to 3 representative nephrometry systems. The calibration plot of this model was excellent (close to the 45-degree line) within the whole range of predicted probabilities. Conclusions: We report a method of preoperatively calculating resected and ischemic volume with a novel formula. This method has superior correlation with the absolute and percent change in estimated glomerular filtration rate compared to current nephrometry systems. The predictive model achieved a strong correlation for the absolute and percent change in estimated glomerular filtration rate.