Comparative prognostic performance of KEYNOTE-564 and AUA risk stratification in nonmetastatic clear cell renal cell...

To compare KEYNOTE-564 and American Urological Association (AUA) risk classifications in clear cell renal cell carcinoma (ccRCC), as recurrence after treatment for localized RCC remains a significant clinical challenge, and no consensus exists on how to define high-risk disease.

We conducted a retrospective multicenter study of surgically treated patients with nonmetastatic ccRCC. Patients were stratified by AUA and KEYNOTE-564 criteria, and overall survival (OS), cancer specific survival (CSS) and recurrence free survival (RFS) were assessed using Kaplan-Meier and Cox regression analyses. Concordance index (C-index) was used to assess discrimination, and Net Reclassification Improvement (NRI) to evaluate reclassification; both were estimated with 1,000 bootstrap replications.

Among 5,711 patients included, the AUA model showed superior discrimination for RFS compared to KEYNOTE-564 (C-index 0.71 vs. 0.68; P = 0.03), with higher recurrence risk in intermediate- (HR 2.38), high- (HR 2.92), and very high-risk groups (HR 3.68) vs. low-risk (all P < 0.001). For OS, both models showed identical discrimination (C-index 0.73; P = 0.67). For CSS, discrimination was similar (AUA 0.79 vs. KEYNOTE-564 0.78; P = 0.65). NRI analysis showed that the KEYNOTE model was disadvantaged for RFS (-1.71%; P = 0.04), while differences in OS (-0.68%; P = 0.19) and CSS (-1.20%; P = 0.19) were not statistically significant.

The AUA classification showed better discrimination for RFS, and improved reclassification compared to KEYNOTE-564, likely due to inclusion of adverse features such as sarcomatoid/rhabdoid differentiation and positive surgical margins. These findings suggest that broader inclusion criteria may improve risk stratification and should be prospectively validated in future adjuvant trials.