Clinical outcomes and influence of mutation clonal dominance in oligomonocytic and classical chronic myelomonocytic leukemia

Abstract

Chronic myelomonocytic leukemia (CMML) is a myelodysplastic/myeloproliferative neoplasm characterized by peripheral blood (PB) monocytosis, defined as an absolute monocyte count (AMC) ≥1 × 109/L with ≥10% monocytes, along with myelodysplastic and myeloproliferative bone marrow features.1 Prognosis of patients with CMML is heterogeneous,2 with certain clinical, disease phenotype, and genomic features being associated with high risk of progression to acute myeloid leukemia (AML) and short overall survival. Although the current World Health Organization (WHO) definition requires presence of persistent absolute monocytosis for the diagnosis of CMML,1 several groups have proposed that oligomonocytic CMML (O-CMML), defined by presence of clinical and pathological features of CMML in the presence of an AMC of 0.5-0.9 × 109/L3 and ≥ 10% monocytes, should be recognized as a new entity. Although current available data support that the clinical, morphological, immunophenotypical, and molecular features of O-CMML are overall similar to that of patients with classical CMML,3, 4 there are scarce data on the clonal architecture, optimal therapeutic management, and survival outcomes of these patients. In order to evaluate further if O-CMML should be recognized as a new entity, we evaluated all previously untreated patients who met proposed diagnostic criteria of O-CMML treated at the University of Texas MD Anderson Cancer Center (MDACC) from 2000 to 2020, and compared their clinical and genomic features with a cohort of 271 patients with classical CMML and 86 patients with MDS and ≥ 10% monocytes without otherwise meeting criteria for O-CMML. O-CMML cases were variably classified as myeloid neoplasms, but met all of the WHO criteria for CMML except for PB AMC of 0.5-0.9 × 109/L; the diagnosis of 0-CMML was defined using the criteria proposed by Geyer et al3 Whole bone marrow DNA was subject to 81 gene-targeted next-generation sequencing (NGS) analysis in a subset of patients (Supplemental Methods). Thirty patients met criteria for O-CMML. Patient characteristics are detailed in Supplemental Table S1 in Appendix S1. Compared to classical CMML, there were no significant differences in cytogenetic abnormalities based on CMML-specific prognostic scoring system (CPSS).5 Patients with O-CMML had significantly lower WBC (4.0 × 109/L vs 11.7 × 109/L, P < .001) and absolute neutrophil counts (ANC) (1.65 × 109/L vs 6.20 × 109/L, P < .001) and lower frequency of bone marrow erythroid dysplasia (40% vs 59%, P = .046). Compared to MDS, patients with O-CMML were younger (median age 66 vs 71 years, P = .035) and had lower Hgb (9.7 vs 11.4 g/dL, P < .001). Targeted NGS was available in 10 (33%) patients with O-CMML. Frequency of identified mutations and their variant allele frequencies (VAF) are shown in Figure 1A,B. Similar to CMML, the highest frequencies of mutations were noted in TET2, SRSF2, RAS pathway genes, ASXL1, and RUNX1 genes. Frequencies of mutations in ASXL1 (50% vs 49.1%, P = .958), SRSF2 (60% vs 39.4%, P = .198), TET2 (70% vs 49.1%, P = .199), and RUNX1 (20% vs 19.4%, P = .965) were similar in O-CMML and CMML, with RAS pathway mutations (NRAS, KRAS, CBL, NF1, SETBP1, PTPN11) being more frequent in CMML compared to O-CMML (51.4% vs 20%, P = .053). The median number of mutations was 4 (range 1-12), and 4 (range 0-8) in O-CMML and CMML (P = .578), respectively. Compared to MDS patients, O-CMML had higher frequency of ASXL1 (50% vs 20%, P = .046), SRSF2 (60% vs 14%, P = .03), and TET2 (70% vs 27%, P = .009) mutations. No significant differences in median VAFs for ASXL1, SRSF2, TET2, RUNX1 and RAS pathway mutations were observed between O-CMML and CMML, or between O-CMML and MDS (Figure 1B). TET2/SRSF2 and TET2/SRSF2/ASXL1 co-mutations were observed in similar frequencies among O-CMML and CMML (40% vs 23%, P = .260; 20% vs 12%, P = .360) but were more frequent in O-CMML compared to MDS (40% vs 6%, P = .006; 20% vs 4%, P = .083). In order to determine the likely clonal dominance of identified mutations, VAF estimates were used to evaluate clonal relationships using Pearson goodness-of-fit tests. Clones with the highest VAF or with VAF close to 40% were defined as dominant, and those present at VAF <20% in the presence of another dominant clone were defined as minor. Mutations in RAS pathway were more likely to appear as minor clones in patients with O-CMML compared to CMML (71% vs 50%, P < .001) (Figure 1C). Sequential NGS was available in four patients with O-CMML. Transplant was associated with mutation clearance in one patient (UPN2) (Figure S1 in Appendix S1). In one patient, transformation to AML was associated with expansion of a previously not detectable ASXL1, IDH1, RUNX1, and NRAS mutations (UPN1) (Figure S1 in Appendix S1). Therapy with decitabine was associated with clearance of CBL and FLT3 mutations as well as reduction in clonal size of an ASXL1 mutation at the time of best response, with subsequent expansion of STAG2, RUNX1, and NRAS mutations at the time of progression in another patient (UPN3) (Figure S1 in Appendix S1). Acquisition of monosomy 7 and expansion of a RUNX1 mutation was observed at the time of transformation in one patient (UPN4) (Figure S1 in Appendix S1). Among the O-CMML patients, seven (23%) were followed with observation, all of which had low/intermediate-1 risk by CPSS, and 23 (77%) received therapy with a median time from diagnosis to first treatment of 1.8 months (range 0-62 months) compared to 2.1 months (range 0-60 months) in the CMML cohort (P = .483). Therapy consisted of hypomethylating agents (HMA) in 17 (57%) patients, intensive chemotherapy in two (9%) patients, and other investigational agents in four (17%) patients. Response outcomes were evaluated following the MDS/MPN IWG response criteria.6 Patients with O-CMML had a higher complete response rate (53% vs 27%, P = .032) and significantly longer median response durations compared to those with CMML (14.4 vs 6.6 months, P = .046) (Table S1 in Appendix S1). A total of six (20%) patients underwent allogeneic stem-cell transplantation. Twelve (40%) patients progressed to overt CMML within a median of 13.5 months (range 3.8-102.8 months). Three (10%) patients experienced transformation to AML compared to 40 (15%) in the control CMML cohort (P = .480), with a trend to longer median time to transformation (33.3 vs 13.2 months, P = .490). With a median follow up of 59.1 months, the median survival for patients with O-CMML was 91.2 months (95% CI 40.2-142.1 months). Survival outcomes based on diagnosis and IPSS-R category are shown in Figure 1D. Compared to MDS, patients with O-CMML had significantly better outcomes among those with very low, low, or intermediate IPSS-R, with no differences among high or very high categories. Univariate analysis for survival evaluating the O-CMML and CMML cohorts revealed that a diagnosis of O-CMML (HR 0.39, 95% CI 0.21-0.75, P = .04), particularly when compared to MP-CMML (Figure S2 in Appendix S1), age (HR 1.06, 95% CI 1.00-1.13, P = .039), and CPSS category of intermediate-2/high (HR 2.59, 95% CI 1.72-3.89, P < .001) predicted for OS. By multivariate analysis for OS, both age (HR 1.03, HR 1.01-1.51, P = .003) and CPSS category intermediate-2/high (HR 2.63, 95% CI 1.73-3.98, P < .001), but not O-CMML (HR 0.59, 95% CI 0.31-1.14, P = .120) retained their significance. Patients with intermediate-2/high CPSS O-CMML had overlapping outcomes to those with low/intermediate-1 CMML (Figure 1E). Patients with low/intermediate-1 O-CMML had a trend to improved outcomes, and those with intermediate-2/high CMML had significantly worse survival (Figure 1F). Similar findings were observed when comparing MD-CMML with O-CMML (Figure S3 in Appendix S1). Similar to previous reports,3, 4 our data confirms that O-CMML has similar clinical and mutational features to CMML, although currently not a recognized entity by the WHO. Unlike in the study by Calvo et al4 we observed lower prevalence of dyserythropoiesis in O-CMML. Similar to their reports, we did not identify differences in cytogenetic abnormalities or mutational landscape between O-CMML and CMML, with the exception of lower frequency of RAS pathway mutations. By studying clonal relationships of identified mutations, we observed that the clonal dominance of common CMML mutations is equally represented among O-CMML and CMML, with exception of RAS pathway mutations, suggesting that mutations in these signaling genes might be responsible for the increased proliferation and monocytosis observed in classical CMML. By comparing to a cohort of MDS patients with ≥10% monocytes but otherwise no other criteria for O-CMML, we could confirm that O-CMML shares clinical and genomic features closer to CMML than to MDS. In addition, although patients with O-CMML tended to present with lower CPSS risk categories and showed improved clinical outcomes when compared to CMML, these survival differences were not significant when corrected by relevant clinical and cytogenetic features, and risk of transformation to AML was similar in both populations. Finally, therapy with HMAs was effective and associated with similar ORR with higher CR and median response duration as compared to classical CMML. Although this might be driven by small patient numbers, it may also suggest that early intervention in these patients prior to the disease becoming proliferative might be associated with improved response outcomes. We acknowledge that this study has several limitations, including its retrospective nature, small numbers of patients with O-CMML (partly as an effect of its low frequency), and absence of NGS in all included patients. However, we believe our findings suggest that the AMC threshold for the diagnosis of CMML should be revised to enable identification of O-CMML patients and facilitate their enrollment in clinical trials for CMML. These data also support the idea that clinical and mutational features should guide confirmation of the diagnosis of CMML. Koji Sasaki: This author declares an advisory role with Pfizer Japan. Elias Jabbour: This author declares research support and an advisory role with Adaptive, AbbVie, Amgen, Pfizer, Cyclacel LTD, Takeda, and Bristol Myers Squibb. Courtney DiNardo: This author declares consultancy fee for Abbvie, Agios, Celgene and honoraria from Medimmune, Daiichi Sankyo, Abbvie, Agios, Jazz, Celgene, and Syros. Koichi Takahashi: This author declares an advisory role with Symbio Pharmaceuticals. Marina Konopleva: This author declares research support and an advisory role with Eli Lilly, AbbVie, Cellectis, Amgen, F. Hoffman La-Roche, Genentech, Ascentage, Kisoji, Reata Pharmaceuticals, Ablynx, Astra Zeneca, Agios, Forty-Seven, Stemline Therapeutics, and Calithera. Tapan Kadia: This author declares research support and an advisory role with Amgen, Bioline RX, Pfizer, Jazz, Bristol Myers Squibb, Celgene, Genentech, Pharmacyclics, Takeda, and AbbVie. Farhad Ravandi: This author declares research support and an advisory role with Macrogeni, Selvita, Cyclacel LTD, Menarini Ricerche, and Xencor. Hagop Kantarjian: This author declares research support and an advisory role with Actinium, and research support from AbbVie, Agio, Amgen, Ariad, Astex, Bristol Myers Squibb, Cyclacel, Daiichi-Sankyo, Immunogen, Jazz Pharma, Novartis, and Pfizer. Guillermo Garcia-Manero: This author declares research support and an advisory role with Bristol Myers Squibb, Astex, and Helsinn, and research support from Amphivena, Novartis, AbbVie, H3 Biomedicine, Onconova, and Merck. Guillermo Montalban-Bravo and Guillermo Garcia-Manero: Concept and design; administrative support; provision of study materials and patients; data collection, analysis, interpretation; and manuscript writing and final approval. Sherry Pierce: Provision of study materials; data collection, and final approval. Rashmi Kanagal-Shamanna, Veronica Guerra, Jorge Ramos-Perez, Danielle Hammond, Paul Shilpa, Kiran Naqvi, Koji Sasaki, Elias Jabbour, Courtney DiNardo, Koichi Takahashi, Marina Konopleva, Naveen Pemmaraju, Tapan Kadia, Farhad Ravandi, Naval Daver, Gautam Borthakur, Zeev Estrov, Joseph D. Khoury, Sanam Loghavi, Carlos Bueso-Ramos, Keyur Patel, and Hagop Kantarjian: Collection and assembly of data; data analysis and interpretation; manuscript writing; and final approval of manuscript. This work was supported in part by the University of Texas MD Anderson Cancer Center Support Grant CA016672. The datasets generated during and/or analyzed during the current study are not publicly available due to patient privacy concerns but are available from the corresponding author on reasonable request. Appendix S1 Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts