Clear-cell renal-cell carcinomas display divergent clinical behaviours. However, the molecular genetic events driving these behaviours are unknown. We discovered that BAP1 is mutated in about 15% of clear-cell renal-cell carcinoma, and that BAP1 and PBRM1 mutations are largely mutually exclusive. The aim of this study was to investigate the clinicopathological significance of these molecular subtypes and to determine whether patients with BAP1-mutant and PBRM1-mutant tumours had different overall survival.

In this retrospective analysis, we assessed 145 patients with primary clear-cell renal-cell carcinoma and defined PBRM1 and BAP1 mutation status from the University of Texas Southwestern Medical Center (UTSW), TX, USA, between 1998 and 2011. We classified patients into those with BAP1-mutant tumours and those with tumours exclusively mutated for PBRM1 (PBRM1-mutant). We used a second independent cohort (n=327) from The Cancer Genome Atlas (TCGA) for validation. In both cohorts, more than 80% of patients had localised or locoregional disease at presentation. Overall both cohorts were similar, although the TCGA had more patients with metastatic and higher-grade disease, and more TCGA patients presented before molecularly targeted therapies became available.

The median overall survival in the UTSW cohort was significantly shorter for patients with BAP1-mutant tumours (4·6 years; 95% CI 2·1–7·2), than for patients with PBRM1-mutant tumours (10·6 years; 9·8–11·5), corresponding to a HR of 2·7 (95% CI 0·99–7·6, p=0·044). Median overall survival in the TCGA cohort was 1·9 years (95% CI 0·6–3·3) for patients with BAP1-mutant tumours and 5·4 years (4·0–6·8) for those with PBRM1-mutant tumours. A HR similar to the UTSW cohort was noted in the TCGA cohort (2·8; 95% CI 1·4–5·9; p=0·004). Patients with mutations in both BAP1 and PBRM1, although a minority (three in UTSW cohort and four in TCGA cohort), had the worst overall survival (median 2·1 years, 95% CI 0·3–3·8, for the UTSW cohort, and 0·2 years, 0·0–1·2, for the TCGA cohort).

Our findings identify mutation-defined subtypes of clear-cell renal-cell carcinoma with distinct clinical outcomes, a high-risk BAP1-mutant group and a favourable PBRM1-mutant group. These data establish the basis for a molecular genetic classification of clear-cell renal-cell carcinoma that could influence treatment decisions in the future. The existence of different molecular subtypes with disparate outcomes should be considered in the design and assessment of clinical studies.

Cancer Prevention and Research Institution of Texas and National Cancer Institute.