Genetic Diversity Drives Differing Morphological and Clinical Presentation and Survival in Myelodysplastic-Type CMML with a Low Absolute Monocyte Count

Background:

The World Health Organization (WHO 5^th ) and International Consensus Classification (ICC) revised the defining criteria of chronic myelomonocytic leukemia (CMML), lowering the absolute monocyte count (AMC) threshold for diagnosing CMML from 1 x10⁹/ L to 0.5 x10⁹/ L while retaining the requirement for a relative monocyte count of ≥10%. This study aimed to compare the phenotypic manifestations and genetic landscape of AMC of 0.5- <1.0 x 10⁹/L (AMC^low) CMML to cases of CMML with AMC of ≥1 × 10⁹/ L (AMC^high).

Methods

As all AMC^low cases are (by definition) myelodysplastic-type CMML (MD-CMML; WBC <13 × 10⁹ /L) we restricted comparisons to AMC^high MD-CMML. Molecular and clinical data were queried from the International Working Group for Prognosis in MDS (IWG-PM) cohort of adult patients with myeloid neoplasms (N= 3,323).

Demographics were analyzed using descriptive statistics with between-group differences assessed using the Wilcoxon rank-sum test or Fisher’s exact test. Leukemia free (LFS) and overall survival (OS) were analyzed using the log-rank method. Factor analysis of mixed data was employed for dimensionality reduction to assess the inter-relationship between categorical features (i.e., mutation present vs. absent) and quantitative variables (i.e., AMC, hemoglobin, etc.). Multivariable analysis (MVA) utilized Cox proportional hazards regression modeling.

Results

469 patients (pts) met WHO 5^th and ICC criteria for MD-CMML, including 264 pts with AMC^low and 205 pts with AMC^high. Notable clinical differences between AMC^low vs. AMC^high cases included lower median BM blasts (3% vs. 3.4%, p: 0.007), median absolute neutrophil count (2.2 vs. 3 x10⁹/L, p < 0.001), hemoglobin (9.8 vs. 10.7 g/dL, p: 0.001), relative monocytes (14% vs. 25%, p < 0.001) and a higher median platelet count (159 vs. 121 x10⁹/L, p: 0.002).

Co-mutation analysis across discretized AMC categories (AMC^low vs. AMC^high) revealed distinct associations between genotype and AMC. AMC^high cases were enriched for mutations (mut) in TET2 (60% vs. 46%, p: 0.005), SRSF2 (31% vs. 19%, p: 0.002), CBL (10% vs. 4%, p:0.008), KRAS (14% vs. 2%, p < 0.001), NRAS (9% vs. 3%, p: 0.008), and PTPN11 (5% vs. 0.3%, p < 0.001), while SF3B1 (34% vs. 15% p < 0.001) and DNMT3A mut (19% vs. 9%, p: 0.005) were more frequent in AMC^low cases. Biallelic TET2 mut (48% vs. 22%, p < 0.001) and co-occurring mut in TET2 and SRSF2 (24% vs.11%, p < 0.001) were also more frequent in AMC^high cases while del(5q) abnormalities were more common in AMC^low cases (7% vs. 3%, p: 0.022)

When evaluating the overlap of presenting clinical and molecular features using factor analysis of mixed data, biallelic TET2 and TET2/SRSF2 mut (versus wild-type TET2 and SRSF2) were associated with higher AMC, while SF3B1 mut was associated with lower hemoglobin and higher platelet counts. RAS pathway mut (N/KRAS, CBL, NF1, and PTPN11) correlated with increased BM blasts, AMC, and monocyte percentage.

In accordance with the varied genetic landscape of AMC^low vs. AMC^high, LFS was superior in AMC^low vs. AMC^high MD-CMML (median 2.87 vs. 4.03 years, p: 0.037), with a trend towards longer OS (median 3.61 vs. 4.33 years, p: 0.055). However, genetic factors influenced survival more robustly, irrespective of the AMC group. In MVA, after adjusting for baseline hematologic parameters and differing molecular aberrations, RAS pathway mut were associated with an increased risk (HR: 1.60, 95% CI 1.13-2.26, p: 0.007) and SF3B1 mut was associated with a decreased risk of death (HR: 0.72, 95% CI: 0.50-1.05, p: 0.016) while AMC^low vs. AMC^high did not retain significance (HR: 0.76, 95% CI: 0.52-1.09, p: 0.13).

Conclusions

In patients with MD-CMML, the clinical disease presentation and patient outcomes appear to be driven by underlying molecular abnormalities rather than AMC. Accordingly, with the recent reduction in the AMC count to 0.5 x10⁹/ L for the classification of CMML, a portion of newly included CMML cases share overlapping genetics and blood count parameters with myelodysplastic neoplasms (MDS) . The molecular profile of a subset of AMC^low cases overlapped with AMC^high cases, suggesting the AMC^low group is genetically heterogeneous, with some cases more similar to AMC^high CMML and others more similar to MDS. This shared molecular ontogeny correlates with the clinical features of disease presentation, suggesting that molecular-based criteria may better define AMC^low CMML and refine disease classification.