ABSTRACT
PURPOSE: The incidence and risk factors for metachronous upper tract urothelial carcinoma (UTUC) following radical cystectomy (RC) remain incompletely defined, which has limited the ability to individualize postoperative surveillance. MATERIALS AND METHODS: A retrospective review of 2 institutional registries was performed to identify patients undergoing RC for urothelial carcinoma. Multivariable Cox proportional hazard models for metachronous post-RC UTUC were developed in one institutional data set and validated in the second institutional data set. A post-RC UTUC risk score was then developed from these models. RESULTS: A total of 3,170 RC patients were included from the training cohort and 959 RC patients from the validation cohort. At a median followup after RC of 4.6 years (IQR 2.1-8.7), 167 patients were diagnosed with UTUC. On multivariable analysis in the training cohort, risk factors for metachronous UTUC were the presence of positive urothelial margin (HR 2.60, p <0.01), history of bacillus Calmette-Guérin treatment prior to RC (HR 2.20, p <0.01), carcinoma in situ at RC (HR 2.01, p <0.01) and pre-RC hydronephrosis (HR 1.48, p=0.04). These factors had similar discriminative capacity in the training and validation cohorts (C-statistic 0.71 and 0.73, respectively). A UTUC risk score was developed with these variables which stratified patients into low (0 points), intermediate (1-3 points), and high risk (4+ points) for post-RC UTUC, with respective 5-year UTUC-free survivals of 99%, 96%, 89% in the training cohort and 98%, 96%, and 91% in the validation cohort. CONCLUSIONS: We developed and validated a risk score for post-RC UTUC that may optimize UTUC surveillance protocols after RC.
