Hypomethylating agent and venetoclax in patients with chronic myelomonocytic leukemia: Is the combination indeed better?

To the Editor: We read with great interest the paper by Saliba et al.1 published recently in your journal. In this single-institution study, authors describe the characteristics and outcomes of 20 patients with chronic myelomonocytic leukemia (CMML, n = 6) and CMML with blast transformation (n = 14), who received treatment with venetoclax (Ven)-based regimens, including 18 patients on a combination of hypomethylating agent and Ven (HMA-Ven). The overall response rate (ORR) was 50% in patients with CMML and 43% in blast transformation, with a complete remission (CR) rate of 0% and 14%, respectively. However, the median overall survival (mOS) of the study cohort was only 8.1 months. Overall, these findings question the potential benefit of the addition of Ven to HMA in the treatment of CMML and post-CMML secondary acute myeloid leukemia (sAML). Hypomethylating agents are commonly used in CMML, particularly in high-risk disease. The ORR with HMA in CMML is only about 40%–50%, with a mOS of 17 months. However, decitabine failed to improve event-free survival (EFS) and OS compared to hydroxyurea in the randomized phase III DACOTA trial in patients with previously untreated proliferative CMML.2 Venetoclax, a BCL-2 inhibitor, in combination with HMA, significantly improved outcomes compared to HMA alone in elderly patients with AML ineligible for intensive chemotherapy. Preliminary results from a phase I/II study of azacitidine and Ven in patients with high-risk MDS and CMML3 suggested safety and clinical activity of the combination, with a high ORR (93%) and rapid onset of response within 1–2 cycles of treatment. However, the duration of response was short (approximately 6 months), and mOS in the CMML cohort (n = 5) was only 12.5 months, which is not improved from HMA alone in historical cohorts. In a retrospective study, 32 (60%) of 53 patients with CMML and post-CMML AML with myelodysplasia-related changes (AML-MRC)4 received HMA-Ven. Although ORR was high at 67% and 81% in the two groups (CMML and AML-MRC), true CR was rare (4%) in patients with CMML, similar to Saliba et al. The mOS was not reached for treatment-naive CMML, 9 months for relapsed or refractory (R/R) CMML, and 12.1 months for AML-MRC. This study did not report the outcomes in HMA-Ven cohort separately. Hence, the role of the HMA-Ven in patients with CMML in real-world settings remains unclear, with the two single-institution studies published to date showing significant differences in outcome estimates.1, 4 We retrospectively reviewed the clinical and genomic data on all patients with CMML and post-CMML sAML5 treated with HMA-Ven at H. Lee Moffitt Cancer Center from January 2018 till July 2021. Treatment response was assessed with the International Working Group criteria for CMML and European Leukemia Network 2017 criteria for sAML. Categorical variables were summarized in numbers and percentages, and their associations were assessed with the Fisher's exact test or chi-squared test. Standard descriptive statistics (median, range, standard deviation) were used for the continuous variables. The OS was estimated with the Kaplan–Meier statistics, and the comparison between groups in univariate analysis was done with log-rank test. Multivariate analysis was conducted with Cox proportional hazards model. All statistical analyses were performed in SPSS Statistics (v.26). Our study included 26 patients; with equal numbers of CMML and post-CMML sAML (n = 13 for each) at HMA-Ven initiation (Figure 1A). Median age was 74 years in both cohorts. Majority (84%) of patients were White, and there was slight male predominance (54%). We observed no significant differences in the extent of cytopenias or monocytosis between patients with CMML and sAML. Median bone marrow blast count was expectedly higher in sAML group (49% vs. 10%; p = <.001). About half the patients (n = 7) in the CMML group had proliferative CMML at HMA-Ven initiation. Fifteen (58%) patients received HMA-Ven for R/R disease and 50% had prior HMA failure, with no significant baseline differences between the two groups. On the assessment of recurrent mutations present at the time of HMA-Ven initiation (n = 23), ASXL1 was the most frequent (61%), followed by TET2 (35%) and SRSF2 (35%), with three (13%) patients (two with CMML) harboring TP53 mutation. There was no significant difference in incidence of mutations between CMML and sAML. Azacitidine was used in 69% of patients, whereas the rest received decitabine as the HMA in the combination. Median number of HMA-Ven cycles in our cohort was 3 (range 1–16), with no significant difference between groups. In patients receiving HMA-Ven for CMML, ORR was 62%, with most of the responses being marrow CR (n = 6; 46%) and only one (8%) patient experiencing a CR (Figure 1B). In the post-CMML sAML group, the ORR was 54%, with one (8%) CR and three (23%) CR with incomplete count recovery (CRi). Both patients with CR in our study were HMA-naïve. Treatment with HMA-Ven induced a response in 5 (71%) out of 7 patients with HMA-refractory CMML, whereas the response rate was lower (33%) among patients with sAML with prior HMA failure. In our cohort, four patients (15%; all with post-CMML sAML) died within 60 days of HMA-Ven initiation, with one death within 30 days. Five patients (three with CMML and two with sAML) were bridged to allogeneic hematopoietic stem cell transplant (allo-HSCT). Median duration of therapy with HMA-Ven was 2.3 months overall (CMML- 5.5 months, sAML- 2 months), 4.3 months in patients experiencing a response, 2.7 months in those bridged to allo-HSCT, and 5.5 months in responding patients not proceeding to allo-HSCT. The most common reasons for treatment discontinuation were refractory cytopenias (n = 8; 31%) and progressive disease (n = 8; 31%). Overall, 61% of evaluable patients (n = 23) developed grade 4 neutropenia, and 35% experienced grade 4 thrombocytopenia, with similar rates in frontline and R/R settings. Four patients (all with R/R disease) had febrile neutropenia in our cohort. In two patients (15%) with CMML, the sAML transformation happened while on HMA-Ven. Median EFS was 4.3 months (95% CI: 1.6–7.0) in our cohort, with no significant difference between CMML and post-CMML sAML (5.7 vs. 2.5 months; p = .194). After a median follow-up duration of 17.4 months, mOS of the entire cohort was 8.1 months (95% CI: 3.6–12.6) since HMA-Ven initiation. Patients receiving HMA-Ven for CMML had a significantly prolonged mOS compared to post-CMML sAML (15.6 vs. 4.1 months; p = .007) (Figure 1C). Disease status (CMML vs. post-CMML sAML) at HMA-Ven initiation remained an independent predictor (HR- 3.14; p = .045) for OS, after adjusting for the presence of ASXL1 mutation and allo-HSCT status on multivariate analysis. We did not observe any significant difference in survival between patients receiving HMA-Ven as 1st line vs. later lines of therapy (9.7 vs. 8.1 months; p = .586). Patients with HMA-refractory disease had similar mOS compared to their HMA-naïve counterparts (8.1 vs. 9.7 months; p = .707). The presence of RAS/MAPK pathway mutations (NRAS, KRAS, CBL, or PTPN11) at HMA-Ven initiation (n = 8) was not predictive of worse OS, although this estimate is limited by small numbers (mOS- 4.1 vs. 9.7 months; p = .250). In treatment-naïve patients (n = 11), mOS was 9.7 months (CMML vs. sAML- not reached [NR] vs. 2.5 months; p = .280). Among patients experiencing an objective response to HMA-Ven, mOS was again 9.7 months (CMML vs. s-AML- NR vs. 4.9 months; p = .960). The mOS in patients bridged to allo-HSCT (n = 5) was similar (8.1 months) to the overall study population. In summary, treatment with HMA-Ven resulted in an ORR of 58% in our cohort of 26 patients with CMML (ORR-62%) and post-CMML sAML (ORR-54%), which is comparable to the response rates reported by Saliba et al.1 Notably, ORR in our sAML cohort was significantly less compared to Montalban-Bravo et al. (ORR of 81% in AML-MRC), although our study did not include patients on other Ven-based therapies than HMA-Ven.4 We noted a reasonable ORR of 54% even in patients with prior HMA exposure. However, achievement of CR was rare (8%), with a relatively short duration of treatment (median 2.3 months). Notably, CR rate with HMA-Ven in all 3 cohorts of CMML (n = 19, 6, and 13) to date has been <10%, substantially lower than patients with myelodysplastic syndrome and AML. Mutations in ASXL1, TET2, and SRSF2 were most frequently noted in our cohort, similar to other published reports.1, 4 Median OS in our study population was 8.1 months (identical to Saliba et al.) after a longer median follow-up duration since HMA-Ven initiation (17.4 months); median EFS was very short at 4.3 months. Contrary to prior reports, we observed a statistically significant worse mOS in patients getting HMA-Ven for post-CMML sAML than for CMML. Median OS of 4.1 months in the transformed CMML group in our study is poorer than 12.1 months reported by Montalban-Bravo et al. Perhaps, dependence on other anti-apoptotic survival pathways (e.g., MCL1), as in M5 AML, may be a primary mechanism for poor outcomes in this subgroup.6 Retrospective design and small sample size are important limitations of our study. Although HMA-Ven could be considered as a bridge to allo-HSCT for patients with CMML and post-CMML sAML, our data strongly support novel therapeutic strategies for this unique patient population, with the goal of true disease modification. None to declare. No relevant disclosures. Study concept: DAS and SB; Study design: SB, NAA, RSK, and DAS; Data Acquisition: NAA, AGJ, LEA, and YZ; Data analysis and interpretation: SB and DAS; Manuscript preparation: SB; Manuscript review: OC, ATK, KLS, DMS, EP, RSK, and JEL; Study supervision: DAS. Data available on request from the authors.