Abstract
RUNX1 mutations (RUNX1mut) are recurrent alterations in chronic myelomonocytic leukemia (CMML) which promote myeloid bias and confer increased risk of progression to acute myeloid leukemia (AML). However, the genomic landscape, clonal hierarchy and clinicopathologic features of RUNX1 mutations and how these influence disease phenotype and clinical outcomes in CMML have not been fully characterized.
We retrospectively analyzed a cohort of 370 patients (pts) with newly diagnosed CMML with bone marrow targeted next generation sequencing with a sensitivity of ~2% variant allelic frequency (VAF) at the time of first diagnosis. RUNX1 coverage included exons (codons) 2-7 (1-261), 8-9 (296-437), and 9 (456-474). VAFs were used to evaluate clonal relationships within each sample.
Among 370 pts, 74 (20%) had RUNX1mut at CMML diagnosis. The median age was 71 years (range, 24–94), and 266 (72%) were male. Compared to pts with wild-type (wt) RUNX1, those with RUNX1mut had significantly lower hemoglobin levels (10.6 g/dL vs. 11.3 g/dL, p=0.02) and lower platelet counts (70×10⁹/L vs. 128×10⁹/L, p<0.001). RUNX1mut CMML was more frequently classified as CMML-2 per WHO-5th criteria (28% vs. 18%, p=0.04). No other baseline characteristics, including myeloproliferative subtype (43% vs. 46%, p=0.64), differed significantly between the two groups.
We identified 89 pts with RUNX1mut CMML, 9 pts harboring ≥2 RUNX1mut. Of these, 38 (43%) were missense, 37 (42%) frameshift, 11 (12%) nonsense, and 3 (3%) in-frame mutations; 16 pts (22%) had mutations in the C-terminal region. The median VAF was 38.2% (range, 1.6–86.2). Frequent co-mutations in included ASXL1 (66%), SRSF2 (51%), TET2 (47%), NRAS (22%), CBL (15%), and KRAS (14%). After correcting for multiple comparisons, ASXL1 was the only mutation significantly enriched in pts with RUNX1mut (OR 2.60, p=0.02), although most STAG2, FLT3-TKD and BCOR mutations occurred in RUNX1mut CMML. No significant differences were seen in the incidence of TET2 or SRSF2 mutations, although multihit TET2 alterations were less common in RUNX1mut group (33% vs. 44%). ASXL1 mutations were more frequently subclonal or co-dominant events to a RUNX1mut, with RUNX1mut rarely appearing as a subclonal event to ASXL1.
Among 364 pts with available treatment data and a median follow-up of 23 months (95% CI, 20-28), fewer pts with RUNX1mutremained untreated (18% vs. 32%, p=0.01). The majority of treated pts (72% in RUNX1mut group vs. 59% in RUNX1wt group) received hypomethylating agents (HMAs), either as monotherapy or in combination. Among pts treated with HMAs, the overall response rate (ORR) did not differ significantly by RUNX1 status (54% vs. 42%, p=0.29), including in those receiving HMAs combined therapy (82% vs. 66%).
pts with RUNX1mut had significantly shorter leukemia-free survival (LFS) (27.7 months vs. 51.8 months, p=0.01) and OS (25 months vs. 66 months, p<0.001) compared to pts with wt RUNX1 . RUNX1mut persisted at AML in 6/10 evaluable pts and emerged at AML in 3/26 RUNX1wt CMML pts. Given the frequent co-occurrence of RUNX1 with ASXL1 mutations, we analyzed the prognostic impact of this co-mutation. Pts with RUNX1mut/ASXL1mut had significantly worse LFS (median 26 months) compared to RUNX1wt/ASXL1wt (65.7 months, p=0.005), RUNX1wt/ASXL1mut (43.4 months, p=0.04), and RUNX1mut/ASXL1wt (36.2 months, p=0.16). OS was also inferior in the double-mutated group (median of 22.6 months), compared to 53.2 months for RUNX1wt/ASXL1mut (p=0.001), 46.1 months for RUNX1mut/ASXL1wt (p=0.19), and 67.2 months for RUNX1wt/ASXL1wt (p<0.001). When censoring at allogeneic stem cell transplantation, LFS was statistically shorter in RUNX1mut/ASXL1mut vs. RUNX1mut/ASXL1wt (22.7 vs. 36.2 months, p=0.04). The adverse prognostic impact of RUNX1/ASXL1 co-mutation persisted in the subset of pts with TET2mut, with median OS of 83.6 months, 39.7 months, 54.8 months, and 25.1 months in RUNX1wt/ASXL1wt,RUNX1wt/ASXL1mut, RUNX1mut/ASXL1wt, RUNX1mut/ASXL1mut, respectively.Conclusions:RUNX1mut in CMML is associated with high-risk clinical features. These mutations frequently co-occur with ASXL1mutations, and pts harboring both mutations have a particularly poor prognosis, even in the presence of TET2 mutations, suggesting a synergistic effect in disease progression. However, pts with RUNX1mut in the absence of ASXL1mut have outcomes that are not inferior to those with RUNX1wt.
