PHF6 mutations in chronic myelomonocytic leukemia identify a unique subset of patients with distinct phenotype and superior prognosis

Abstract

The current study was inspired by observations from exploratory analyses of an institutional cohort with chronic myelomonocytic leukemia (CMML; N = 398) that revealed no instances of blast transformation in the seven patients with plant homeodomain finger protein 6 (PHF6) mutation (PHF6^MUT). A subsequent Mayo Clinic enterprise-wide database search identified 28 more cases with PHF6^MUT. Compared with their wild-type PHF6 counterparts (PHF6^WT; N = 391), PHF6^MUT cases (N = 35) were more likely to co-express TET2 (89% vs. 45%; p < .01), RUNX1 (29% vs. 14%; p = .03), CBL (14% vs. 2%; p < .01), and U2AF1 (17% vs. 6%; p = .04) and less likely SRSF2 (23% vs. 45%; p < .01) mutation. They were also more likely to display loss of Y chromosome (LoY; 21% vs. 2%; p < .01) and platelets <100 × 10⁹/L (83% vs. 51%; p < .01). Multivariable analysis identified PHF6^MUT (HR 0.28, 95% CI 0.15–0.50) and DNMT3A^MUT (HR 5.8, 95% CI 3.3–10.5) as the strongest molecular predictors of overall survival. The same was true for blast transformation-free survival with corresponding HR (95% CI) of 0.08 (0.01–0.6) and 9.5 (3.8–23.5). At median 20 months follow-up, blast transformation was documented in none of the 33 patients with PHF6^MUT/DNMT3A^WT but in 6 (32%) of 19 with DNMT3A^MUT and 74 (20%) of 374 with PHF6^WT/DNMT3A^WT (p < .01). The specific molecular signatures sustained their significant predictive performance in the context of the CMML-specific molecular prognostic model (CPSS-mol). PHF6^MUT identifies a unique subset of patients with CMML characterized by thrombocytopenia, higher prevalence of LoY, and superior prognosis.