Oncologic surveillance following radical cystectomy: an individualized risk-based approach

Abstract

To provide an alternative surveillance approach for bladder cancer (BC) following radical cystectomy (RC) according to more accurate predictions of a patient's projected BC course. We identified 1797 patients who underwent RC for M0 BC between 1980 and 2007. Patients were stratified by pathologic stage (pT0Nx-0, pTa/CIS/1Nx-0, pT2Nx-0, pT3/4Nx-0, and pTanyN+), relapse location (urethra, upper tract, abdomen/pelvis, chest, and other), age (≤60, 61-70, 71-80, >80years) and Charlson Co-morbidity Index (CCI ≤2 and CCI ≥3). Risks of disease recurrence and non-BC death were modeled using Weibull distributions. Recommended surveillance durations were estimated when the risk of non-BC death exceeded the risk of recurrence. At a median follow-up of 10.6years (IQR 6.8,15.2), 713 patients developed recurrence. Vastly different recurrence patterns were appreciated. Specifically, among patients ≤60years with pT2Nx-0, non-BC death risk exceeded the risk of recurrence in the abdomen at 7.5years following surgery when CCI was ≥3, versus at year 10 after RC when CCI was ≤2. Meanwhile, for patients >80years with pT2Nx-0, non-BC death risk exceeded the risk of abdominal recurrence at 1year after RC, regardless of CCI. We present an alternative post-RC surveillance approach that incorporates a patient's changing risk profile with the influence of competing health factors. We believe this strategy provides more individualized recommendations than current guidelines, and may improve the benefit derived from surveillance while reducing resource misappropriation.