Allogeneic stem cell transplantation-related outcomes in chronic myelomonocytic leukemia: Data from a large academic center

Abstract

Abstract Introduction: Chronic myelomonocytic leukemia (CMML) is a rare myeloid neoplasm. The outcomes of patients (pts) with CMML treated with allogeneic stem cell transplantation (SCT) are incompletely understood. Methods: We evaluated a cohort of pts presenting with newly diagnosed CMML between 2000-2023 (n=855) to assess outcomes with SCT. Time-to-event endpoints were calculated from SCT day 0. Cumulative incidence (CI) analyses were performed using the Fine-Gray method with relapse and treatment-related mortality (TRM) as competing risks. Multivariate analysis (MVA) was performed using a Cox proportional hazards regression. Results: Among the cohort of 855 pts, 109 (13%) proceeded to SCT. Pts who underwent SCT had a median age of 63 years (range 27-78) and 68 (62%) were male. Thirty-three (30%), 11 (10%), and 51 (47%) had CMML-2, therapy-related CMML (t-CMML), and proliferative-type CMML (MP-CMML), respectively. RAS pathway (RASp) mutations were identified in 45/65 (69%) pts. Other high-risk mutations included: ASXL1 in 37/70 (53%) pts, RUNX1 in 18/70 (26%), SETBP1 in 6/44 (14%), IDH2 in 7/79 (9%), and DNMT3A in 12/74 (16%). CPSS score was Low in 14/107 (13%), Int-1 in 26 (24%), Int-2 in 52 (49%), and High in 15 (14%). CPSS-mol score was Low in 3/51 (6%), Int-1 in 8 (16%), Int-2 in 16 (31%), and High in 24 (47%). The median time from diagnosis to SCT was 9.3 months (m). Conditioning was MAC in 65 (60%) and RIC in 44 (40%) pts, with PTCy administered in 56 pts (51%). Compared to MAC, pts who received RIC were older (median age 64 vs 61 yrs, p=0.001) and had worse HCT-CI (median 3 vs 2, p=0.18). 28-day, 56-day, and 100-day mortality were 3%, 6%, and 10%, respectively. Acute graft-versus-host disease (GVHD) was diagnosed in 62/104 (60%) pts (9 [9%] pts with grade 3/4). Chronic GVHD was diagnosed in 28/94 (30%) pts. The median follow-up time from SCT day 0 was 54.7 m. At the time of SCT, 83/109 (76%) pts had CMML and 26/109 (24%) had transformed to secondary acute myeloid leukemia (sAML). The median OS was 57.7 m (2-yr 59%) for CMML and 12.0 m (2-yr 42%) for sAML (p=0.47). The median PFS was 13.8 m (2-yr 42%) for CMML and 9.4 m (2-yr 42%) for sAML (p=0.65). The CI of relapse at 2 years was 33% in CMML and 37% in sAML (p=0.29). The CI of TRM at 2 years was 24% in CMML and 20% in sAML (p=0.40). We next explored subsets in the pts with CMML (excluding sAML). OS was significantly longer in MAC vs RIC conditioning (median 81.3 vs 20.2 m; p=0.005). The cumulative incidence of relapse was similar between MAC and RIC (2-yr 37% vs 29%; p=0.55) but TRM was lower in MAC vs RIC (2-yr 12% vs 41%; p=0.013). Causes of TRM (n=14) in RIC were infections in 6/14 (43%), GVHD in 5/14 (36%), and other in 3/14 (21%). The median OS was 38 m (2-yr 55%) in the pts with RASp mutations vs not reached (NR) (2-yr 87%) in those without (p=0.069). The median OS was 60.8 m (2-yr 77%) in pts with ASXL1 mutations (n=33) vs not reached (NR) (2-yr 62%) in those without (n=25, p=0.88). The median OS was 38.0 m (2-yr 66%) in pts with RAS pathway + ASXL1 mutations (n=22) vs NR (2-yr 80%) in those without any (n=11; p=0.39). The median OS was NR (2-yr 80%) in pts with RAS pathway + RUNX1 mutations (n=10). There were no significant differences in OS by CMML type (median 57.7 m [2-yr 57%] with MD-CMML vs 43.9 m [2-yr 62%] with MP-CMML; p=0.70). The median OS was 57.7 m (2-yr 60%) in CMML-1, 43.9 m (2-yr 60%) in CMML-2, and 41.4 m (2-yr 50%) in t-CMML (p=0.89). By CPSS, the median OS was 57.7 m (2-yr 59%) in Low/Int-1 and 43.9 m (2-yr 62%) in Int-2/High (p=0.60). By CPSS-mol, the median OS was 57.7 m (2-yr 90%) in Low/Int-1 and 81.3 m (2-yr 70%) in Int-2/High (p=0.28). By MVA considering age, HCT-CI, conditioning intensity, CMML type (MD vs MP), RASp mutations, and CPSS, only RASp mutated status was significantly associated with adverse OS (HR 5.34, p=0.039) and PFS (HR 5.61, p=0.013). By landmark analysis compared to non-transplanted pts, both RASp mutated (median OS from diagnosis 63.1 m with SCT vs 30.3 m without, p=0.003) and RASp unmutated pts (median OS from diagnosis NR with SCT vs 37.3 m without, p=0.02) appeared to benefit from SCT. Conclusions: SCT provides durable disease control in a significant proportion of pts with CMML. Outcomes were inferior with RIC regimens, possibly related to pt age and comorbidities. RAS pathway mutations were associated with shorter post-SCT OS, but these patients still appear to benefit from SCT.