Development and external validation of a preoperative nomogram for predicting pathological locally advanced disease of clinically localized upper urinary tract carcinoma

Objective To develop and validate a preoperative nomogram to predict pathological locally advanced disease (pLAD) of clinically localized upper urinary tract urothelial carcinoma (UTUC) treated with extirpative surgery. Methods In total, 1101 patients with cN0M0 UTUC (development cohort, n = 604; validation cohort, n = 497) from 2 independent academic databases were retrospectively analyzed. pLAD was defined as pT3/4 and/or pN+. Multivariate logistic regression was used to develop a nomogram. The accuracy of the nomogram was evaluated with a receiver operating characteristic curve, calibration plot, and decision curve analysis. Results The development and validation cohorts comprised 204 (33.8%) and 178 (35.8%) patients with pLAD, respectively. The multivariate analyses showed that the neutrophil-to-lymphocyte ratio (hazard ratio [HR], 2.27;P < .001), chronic kidney disease (HR, 1.56;P = .032), tumor location (HR, 1.60;P = .029), hydronephrosis (HR, 2.71;P < .001), and local invasion on imaging (HR, 8.59;P < .001) were independent predictive factors. After bootstrapping, a well-calibrated nomogram achieved discriminative accuracy of 0.77 in the development cohort. The decision curve analysis demonstrated improved risk prediction against threshold probabilities (>= 8%) of pLAD. These results were consistent in the validation cohort. Conclusion Our novel nomogram allows for more highly accurate prediction of pLAD of UTUC. This nomogram integrates standard imaging and laboratory factors that help to identify patients who will benefit from preoperative chemotherapy, extended lymph node dissection, or both.