Cohesin Complex Gene Mutated Acute Myeloid Leukemia (AML) Responds Favorably to Venetoclax Based Frontline Therapy

Background: Cohesin complex gene (CCG) mutations (SMC1A, SMC3, STAG1, STAG2, RAD21) occur in around 10% of newly diagnosed (ND) AML and alter DNA damage repair pathways. Additionally, STAG2 mutation is considered an adverse risk (AR) factor in ELN 2022 risk stratification. The independent impact of these mutations, based on treatment intensity, use of venetoclax and impact of alloSCT remains poorly defined. Additionally, CCG mutated leukemia cells have high preclinical sensitivity to PARP inhibitors and splicing modulators. Defining the prognostic impact and outcomes with current therapies will help establish contemporary expectations to compare results with emerging novel therapies in CCG mutated AML. Methods: We retrospectively analyzed adult patients (pts) with ND AML treated at MDACC with baseline myeloid gene sequencing that included above-mentioned CCGs and compared outcomes of pts with and without CCG mutations. AML risk stratification was as per ELN 2017 criteria, given inclusion of STAG2 as adverse in the 2022 classification. Responses were tabulated as composite complete remission (CRc, CR+CRi) and overall response (CRc+ MLFS=responders). Intensive chemotherapy (IC) included regimens with IV cytarabine + anthracycline based induction, while low-intensity therapy (LIT) included hypomethylating agent (HMA) based regimen, low-dose cytarabine based regimens or low-intensity investigational agents. Results: From Apr 2017-Jan 2024, 1446 pts with ND non-core binding factor AML and adequate baseline sequencing data were included. 147 pts (10%) had a mutation in the CCGs. The most common CCG mutation was in STAG2 (93, 63%) followed by SMC1A (24, 16%) and RAD21 (19, 13%). The med (med) age of pts in the CCG group was 71 yrs (19-95 yrs) (118 [80%] were ≥60 yrs] versus 68 yrs (20-94 yrs) in the non-CCG group (921 [71%] ≥60 yrs), p=0.02. More pts in the CCG group had history of secondary AML than non-CCG group (76 [52%] vs. 500 [38%]), p<0.01, including treated-secondary AML (41% vs. 27%, p<0.01). Amongst evaluable pts 29/133 (22%) had ELN AR CTG (including 17 [13%] complex CTG) in the CCG group versus 570/1152 pts (49%) (398 [34%] complex CTG) in the non-CTG group, p<0.01. TP53 mutation (mut) was seen less commonly in the CCG group (17 [12%] vs. 380 [29%], p<0.01, ASXL1 (38% vs. 15%, P<0.001) and RUNX1 (24% vs. 15%, p=0.01) muts were more common in the CCG group, whereas NPM1 mut (17[11%] vs. 208 [16%], p=0.19) and FLT3-ITD (12 [9%] vs. 148 [11%], p=0.41) were similar. Amongst classifiable pts, 16 (11%) were ELN 2017 favorable risk, and 91 (62%) AR in the CCG group vs. 165(13%) and 822 (63%) resp. in the non-CCG group. Treatment was with LIT in 115 pts (78%, 84 [73%] with Ven) and 908 (70%, 621 [68%] with Ven) in CCG and non-CCG groups resp.; 18/32 pts (12%) and 220/391 (30%) pts treated with IC received Ven in the 2 groups resp. A CRc was attained in 87 pts (59%) in CCG and 777 pts (60%) in non-CCG and OR in 106 (72%) and 864 (66%) pts. At a med follow-up of 30 mos in the CCG and 34 mos for the non-CCG group, the med RFS for responders was 14.4 vs. 11.1 mos (p=0.42) and OS 13.5 vs. 10.3 mos (p=0.12); 38 (26%) and 302 pts (23%) underwent SCT resp. OS amongst pts who received LIT+Ven was 14 mos for CCG (n=84) vs 9.4 mos for non-CCG group (n=621) (p=0.04) while it was dismal in both groups at 4.8 vs. 5.5 mos with LIT alone (p=0.39). In LIT+Ven treated CCG group (n=84), 45 (54%) and 10 (7%) pts were in the Dohner et al (ASH 2022) high-benefit and low-benefit groups resp, versus 257 (41%) and 216 (35%) pts in the non-CCG group (n=621). The OS in the high-benefit group was 18.4 vs. 19.4 mos (p= 0.59), intermediate benefit group 6.8 vs. 12.1 mos (p=0.21) and low benefit group 11 vs. 4.6 mos (p=0.008) for CCG and non-CCG resp. In the IC treated pts (CCG=32, non-CCG n=391) the med RFS and OS were 30.9 vs. 57.1 mos (p=0.48) and 31.0 mos vs. 24.5 mos (p=0.99) for the 2 groups resp. Med age of pts with STAG2 mut (n=93) was 72 years (33-95 years); at a med FU of 14 mos in these pts, med OS was 13.7 mos (med OS in those treated with LIT+Ven 17 mos, LIT alone 3.2 mos, IC 18 mos). On MVA in pts treated with LIT+Ven, adjusting for SCT and Dohner et al criteria, cohesin mutations were not independently associated with hazards of death. Conclusions: Cohesin mutations were present in 10% of ND AML and did not portend an adverse prognosis. OS with venetoclax based LIT appeared favorable even when adjusting for other factors. Any negative prognostic impact of STAG2 mut with Ven based therapy needs further evaluation.