External validation of the EORTC and NCCN bladder cancer recurrence and progression risk calculators in a U.S. community-based health system

Abstract

ObjectivesTo externally validate the European Organization for the Research and Treatment of Cancer (EORTC) risk calculator and National Comprehensive Cancer Network (NCCN) guidelines in a contemporary population of U.S. non–muscle-invasive bladder cancer (NMIBC) patients treated in a community-based setting and compare our findings to those from another U.S. health system. Materials and methodsWe identified 1,491 NMIBC patients with a median follow-up of 2.1 years (recurrence) and 4.1 years (progression). We calculated NCCN risk groupings and EORTC prognostic index for recurrence and progression. We followed Royston and Altman's guidelines for the external validation of prognostic calculators. ResultsFor predicting recurrence using the EORTC framework, Harrell's C (a measure of discrimination) was smaller in our sample (0.66) than in the European Association of Urology sample (0.61), whereas for progression, Harrell's C was larger in our sample (0.78 vs. 0.75). The EORTC calculator overestimated progression risk in the highest stratum for our sample; calibration and discrimination were adequate for all groups except the highest risk group. For NCCN risk groupings, Harrell's C was 0.54 for recurrence and 0.62 for progression, suggesting poor to fair discrimination in our sample. The NCCN framework had slightly better performance for predicting progression vs. recurrence. ConclusionsExisting NMIBC risk-stratification frameworks have acceptable accuracy to predict outcomes. However, further innovation in NMIBC care will require predictive tools with more granularity to reflect the differential risks of subgroups of NMIBC recurrence, prior treatment histories, and other prognostic variables.