Characteristics and Clinical Outcomes of Patients with Myeloid Malignancies and Cohesin Mutations

Background: The cohesin protein complex is involved in a broad range of cellular functions. Cohesin mutations have been observed in a subset of patients (pts) with acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS). The clinical outcomes in these pts remain unknown.

Purpose: The primary objective was to analyze the mutational landscape of a cohesin-mutated cohort of AML and MDS and assess the impact of co-occurring mutations on survival. The secondary objective was to compare outcomes in cohesin-mutated vs. wild-type (WT) cohorts.

Methods: A retrospective cohort analysis of AML and MDS pts who had next-generation sequencing at UMass Chan Medical School from March 2017 to February 2023 was performed. Cohesin mutations that were tested for were STAG2, SMC3, and SMC1A.

Results: We identified a total of 83 pts (cohesin-mutated, n=36 [STAG2, n=28; SMC1A, n=7; SMC3, n=3]; cohesin WT, n=47). We observed co-occurrence of two cohesin mutations in two pts. Within the cohesin-mutated cohort, 17 (47%) pts had AML (6 de novo and 11 secondary), and 19 (53%) had MDS. The incidence of cohesin mutations did not differ significantly by disease type. When compared with cohesin WT pts, pts with cohesin mutations were significantly older (median age, 73 vs. 63 years; P = 0.009), more likely to be male (67% vs. 53%; P = .03), and had a larger proportion of adverse-risk disease cases according to the 2022 European LeukemiaNet (ELN) criteria (82% vs. 53%; P = .04). The WT group had more de novo AML cases than did the mutated group (51% vs. 17%), along with a higher percentage of blasts (64% vs. 36%). Among the 36 pts with cohesin mutations, 32 (89%) had at least one other simultaneous gene mutation. Concurrent mutations that were significantly enriched in the cohesin-mutant group were SRSF2 (33% vs. 4%; P = 0.0007), U2AF1 (14% vs. 0%; P = 0.01), RUNX1 (28% vs. 6%; P = 0.01), TET2 (25% vs. 8%; P = 0.07), and EZH2 (17% vs. 0%; P = 0.005). In contrast, FLT3-ITD mutations were more enriched in the WT group (13% vs. 0%; P = 0.07). We did not observe a significant difference in the frequency of TP53, IDH1/IDH2, or NPM1 mutations between the cohesin-mutated and WT groups.

Induction therapies in the cohesin-mutated cohort consisted of hypomethylating agents (HMA) alone (n=17 [47%]), HMA plus venetoclax (n=5 [14%]), chemotherapy (CT; n=7 [19%]), and CT plus venetoclax (n=3 [8%]). Four (11%) patients opted to receive no treatment. In comparison, more pts in the cohesin WT group received CT (51%), and fewer received HMA (15%; P = 0.002). Pts with FLT3-ITD mutations received FLT3 inhibitors. The proportion of pts who received an allogeneic hematopoietic cell transplant (HCT) and those who opted not to receive any treatment were not significantly different between the two cohorts (19% vs. 32% [P = 0.30] and 11% vs. 9% [P = 0.90], respectively).

The median follow-up times for the cohesin-mutated and WT cohorts were 13 months (mos) (IQR, 5-31 mos) and 13 mos (IQR, 5-46 mos), respectively (P = 0.47). We found no significant differences in overall survival (OS) for pts with and without cohesin mutations (median OS time, 24 mos vs. 13 mos ; P = .6).

We also analyzed factors impacting OS in the cohesin-mutated group. We found no significant differences in OS in pts with MDS versus AML (P = 0.31), by age (<70 vs. >70 years; P = 0.47), by blast percentage (P = 0.26), or by type of treatment received (P = 0.83). Pts with STAG2 mutations had better OS than did those with SMC1A or SMC3 mutations (median OS 25.9 mos and 9.9 mos, respectively; P = 0.043). The co-occurrence of SRSF2 mutation (SRSF2(+) with cohesin mutation was associated with inferior OS when compared with that in SRSF2(-) pts (median OS 12.9 mos and 43.0 mos, respectively; P =0.016). Pts who received HCT had a median follow-up of 38.5 months (IQR, 26-74 months); their median OS time was 70 mos. Our multivariable analysis showed that SRSF2(+) was associated with worse OS (HR, 2.67 [CI, 1.04-8.86]; P = 0.04). STAG2 mutations (HR, 0.56 [0.20-1.55]; P = 0.26]) and a 3-month landmark HCT (HR, 0.32 [0.09-1.15]; P = 0.08) were associated with a trend for a better OS.

Conclusions: STAG2 was the predominant mutant cohesin subunit observed in our cohort of pts with AML and MDS. Survival outcomes were similar to those in the WT pts. The co-existence of SRSF2(+) with cohesin mutations was associated with inferior survival. The outcome of HCT in pts with cohesin mutations is favorable and must be evaluated in a larger cohort.