Bazinet, Alexandre

PURPOSE Chronic myelomonocytic leukemia (CMML) is a rare myeloid neoplasm characterized by clinical heterogeneity and is associated with poor outcomes. To date, limited molecular information has been incorporated into disease classification and risk stratification. We aimed to integrate genomic features into the clinical decision-making process for CMML. PATIENTS AND METHODS We analyzed a retrospective cohort of 3013 patients with CMML (training set) and a prospective population of 516 patients (validation set). Using an innovative framework for multimodal data analysis, we developed molecular-based disease taxonomy and prognostication. RESULTS Unsupervised clustering identified nine entities with distinct genomic features and outcomes ( P < .001), including splicing machinery, transcription factors, signal transduction and tyrosine kinase pathways aberrations, and high-risk molecular signatures. Notably, 15% of patients showed molecular/clinical overlap with other myeloid neoplasms. We integrated molecular and clinical information to build the international CMML Prognostic Scoring System (iCPSS), incorporating mutations in nine genes together with hematologic parameters and cytogenetic abnormalities. The iCPSS identified five groups with distinct probability of overall and leukemia-free survival in both training and validation cohorts ( P < .001), outperforming existing prognostic models. Importantly, 55% of patients were reassigned to higher or lower risk groups by the iCPSS. Decision analysis demonstrated that iCPSS could refine the optimal timing of allogeneic transplantation at the individual level; compared with conventional prognostic tools, iCPSS-based decision modeling changed transplantation strategy in 31% of cases, resulting in a significant gain-in-life expectancy for eligible patient population ( P < .001). A federated learning platform was implemented to enable continuous, privacy-preserving model update across multiple centers. CONCLUSION Molecular information improves CMML classification and prognostication, supports more effective clinical decision making, and potentially refines the design of clinical trials.

Abstract Recent updates to monocyte count thresholds recognize oligomonocytic chronic myelomonocytic leukemia (OM-CMML) as an early form of CMML. However, the clinical validity of these changes remains uncertain without incorporating biological and genomic factors. In this study, we analyzed a cohort of 911 patients (249 with OM-CMML, 359 with overt CMML, and 303 with myelodysplastic syndromes) using unsupervised clustering to evaluate the role of genomic determinants in refining CMML diagnosis. Our findings show that CMML molecular signatures (biallelic TET2 mutations or SRSF2-TET2 comutations) are linked to a distinct transcriptome, monocytic bias, classical monocytosis, and a higher risk of progression to overt CMML in OM-CMML cases. We developed a weighted genomic model and diagnostic workflow showing that combining genomic signatures with bone marrow monocyte frequencies in OM-CMML more accurately predicts progression to overt CMML. These findings support integrating genomic determinants and our clinic-ready diagnostic workflow into the CMML diagnostic framework to improve accuracy. Significance: Through comprehensive clinical and genomic profiling of a large patient cohort, alongside immunophenotypic and transcriptional cellular analyses, this study provides evidence that incorporating genomic determinants into the diagnostic criteria for OM-CMML improves diagnostic accuracy and refines the identification of early-stage CMML, thereby preventing misclassification.

Abstract Introduction Blast transformation (BT) in CMML occurs in 15-30% of patients (pts) with a median overall survival (OS) &lt;6 months (mo). While clonal evolution can be observed, biological mechanisms of BT remain unknown for many pts. Using paired samples from chronic phase (CP) and at BT, this study defines the cytogenetic and molecular landscape of BT-CMML. Methods BT-CMML pts from Mayo Clinic (n=153, 31%), MD Anderson (n=195, 40%), and 7 centers of the Groupe Francophone des Myélodysplasies (n=145, 29%) were included after IRB approval. ICC 2022 criteria defined CMML and AML. Pt data were collected at initial diagnosis or referral at CP and again at BT. Clonal dominance was inferred from a natural log-fold difference of mutation variant allele fractions (VAF). OS was censored at allogeneic stem cell transplant (alloSCT). Single-cell multiomic analyses used the Tapestri platform (Mission Bio). Statistics used R. Results Among 1,883 cases of CMML with a median follow up of 96 mo, 493 (26%) underwent BT. The median age at CP and BT was 69 and 71 years, respectively. BT-CMML pts were predominantly male (69%) and had proliferative CMML (53%) or CMML-1 (69%) in CP; 39% and 33% were considered high-risk by CPSS-Mol and BLAST-Mol criteria, respectively. The median time to BT was 16 mo (interquartile range 7-35 mo) and was not different in males vs females (p=0.60). Compared to their CP parameters (p&lt;0.001 each), pts at BT had lower median hemoglobin (10.8 vs 9.2 g/dL) and platelets (100 vs 44 x109/L) but higher leukocytes (13.9 vs 16.3 x109/L), peripheral blasts (0% vs 14%), and bone marrow blasts (6% vs 32%). Blasts exhibited mostly monocytic (57%) or immature myeloid (40%) differentiation. Abnormal karyotypes (AK) were more prevalent at BT (25% vs 62%, p&lt;0.001). Among 318 pts with paired karyotype data, 221 (69%) had a normal karyotype (NK) in CP, of which 123 (56%) remained normal at BT and 98 (44%) acquired new abnormalities. Conversely, 97 (31%) pts had an AK in CP, of which 67 (69%) acquired abnormalities or increased clonal metaphases (6%), 23 (24%) remained stable, and 7 (7%) reverted to a NK at BT. Overall, 46% acquired and 8% lost a clonal cytogenetic abnormality, while 46% remained unchanged. Numeric (37%) and structural (30%) changes were most common. Complex (12% vs 3%) and monosomal (5% vs 2%) karyotypes were more frequent at BT than in CP. Only 6 (2%) pts acquired an AML-defining rearrangement at BT. Mutational data were available for 248 paired cases. The number of somatic mutations was higher at BT (median 3 vs 4, p&lt;0.001). Mutations in RUNX1 (27% vs 39%), NRAS (23% vs 33%), KRAS (11% vs 17%), FLT3 (6% vs 15%), NPM1 (0% vs 11%), IDH2 (5% vs 11%), and CEBPA (3% vs 10%) were more frequent at BT (all p&lt;0.05). At BT, the average VAF of the dominant clone increased in 32%, decreased in 35%, and remained unchanged (delta ≤5%) in 33%. The most frequent dominant clones both in CP and at BT had mutations in TET2 (8% vs 4%) or co-dominant mutations involving SRSF2-TET2 (7% vs 3%). Mutational clonal evolution was observed in 72%, with mutations gained in 45% and lost from the dominant clone in 19%. The most gained mutations at BT were in NRAS (15%) and RUNX1 (13%). Single cell proteogenomic analyses in 4 selected CP and BT pts shed light on the impact of NRAS and RUNX1 mutations on clonal hierarchies and differentiation. Among 196 fully evaluable cases, a new cytogenetic and/or molecular aberration was uniquely gained in 106 (54%) cases and lost in 9 (5%), with 40 (20%) both gaining and losing a genetic alteration. Forty-one (21%) patients had no identifiable genetic change between CP and BT, suggesting driver alterations outside of that detected by routine analyses. Median OS was 31 mo from CMML diagnosis and 5.6 mo from BT. Compared to supportive care, treatment in CP did not prolong median time to BT (14 vs 17 mo, p=0.91) or OS (35 vs 29 mo, p=0.05), regardless of intensity. After BT, median OS was 1.2 mo with supportive care (n=83), 8.7 mo with a hypomethylating agent (HMA, n=39), 10 mo with HMA plus venetoclax (n=37), 7.8 mo with chemotherapy (n=91), and 31 mo in pts who underwent alloSCT (n=57). Conclusions In this large BT-CMML cohort, cytogenetic (46% gained, 8% lost) or mutational (45% gained, 19% lost) driver events were identified in 79% of pts, with 21% having no identifiable changes by conventional testing. Whole genome/transcriptome analyses are underway in this subgroup to identify novel drivers of BT.

Abstract Recent updates to the monocyte count thresholds for diagnosing chronic myelomonocytic leukemia (CMML) have recognized oligomonocytic CMML (OM-CMML) as an early stage within the CMML spectrum. However, the clinical validity of these revised criteria remains uncertain without considering biological factors and might misclassify myelodysplastic syndromes (MDS) with mild monocytosis as CMML. To characterize the genomic landscape of OM-CMML, we evaluated OM-CMML (n=249), MDS (n=304) and myelodysplastic CMML (MD-CMML, n=358) patients (pts) treated at three academic institutions. Targeted next-generation sequencing was performed on bone marrow (BM) at initial diagnosis or first referral. OM-CMML was characterized by lower frequencies of biallelic TET2 (biTET2) and RAS pathway (RASp) gene mutations (BRAF, CBL, KRAS, NF1, NRAS and PTPN11) compared to MD-CMML (p&lt;0.001) but higher frequencies than MDS (p&lt;0.001). BM monocyte frequencies, assessed morphologically, were higher in OM-CMML than in MDS (p&lt;0.001) but lower than in MD-CMML (p&lt;0.001), indicating that OM-CMML is a heterogenous group including true CMML and distinct MDS subsets. Unsupervised hierarchical clustering based on genomic alterations across all cases (n=911) identified three distinct clusters: 1) TP53 alterations (predominantly multihit TP53) and complex karyotype (TP53-complex); 2) biTET2 or TET2-SRSF2 co-mutation (biTET2/SRSF2); 3) a heterogenous group lacking these features (not otherwise specifiable [NOS]). Among OM-CMML pts, 11%, 34% and 55% belonged to these clusters, respectively. Model-based unsupervised clustering further identified 9 unique clusters that could be consolidated into 6 groups: 1) TP53-complex; 2) biTET2/SRSF2; 3) SF3B1 (without biTET2 or SRSF2); 4) SRSF2 without TET2 and frequent IDH2 mutations (SRSF2-IDH2); 5) ASXL1 and U2AF1 mutations and copy number alterations (CNA-ASXL1/U2AF1); 6) no somatic mutations (No mutation). biTET2/SRSF2 group exhibited monocytic bias with higher AMC, PB and BM monocyte % even among OM-CMML cases. This bias was not attributed to higher RAS pathway mutation frequencies, as biTET2/SRSF2 cases had similar RASp mutation frequencies but lower median variant allele frequencies for RASp mutations compared to NOS cases (9% vs. 21%, p=0.029). PB monocyte partitioning using flow cytometry (FC) in a subset of OM-CMML pts (n=50) demonstrated higher frequency of classical monocytosis (&gt;94% CD14+/CD16- monocytes) in biTET2/SRSF2 vs other OM-CMML. To further determine if biTET2/SRSF2 genotype defines a distinct biological subgroup, we performed RNA-sequencing of BM CD34+ cells from healthy donors (n=15) and pts with OM-CMML (n=12), MD-CMML (n=14) and MDS (n=55). Unsupervised clustering based on gene expression profiles revealed that biTET2/SRSF2 clustered irrespective of their diagnostic group with greater dispersion being observed based on diagnosis. MDS and CMML with biTET2/SRSF2 had less differentially expressed genes compared to MDS and CMML with biTET2/SRSF2 relative to other MDS and CMML genotypes. Among 138 OM-CMML pts in whom progression to overt CMML was evaluable, 37 (27%) progressed to overt CMML of which 26 (70%) belonged to the biTET2/SRSF2 group. Competing risk analysis confirmed higher cumulative incidence and shorter time to CMML progression among biTET2/SRSF2 OM-CMML. Using ElasticNet regression we identified BM monocytosis and 14 genomic variables (including biTET2 and TET2-SRSF2) as predictors of overt CMML progression among OM-CMML. We developed a score-based model able to predict progression with high reliability (AUC 0.805, sensitivity 0.70, specificity 0.76). Model accuracy was confirmed by optimism-corrected bootstrap internal validation (AUC 0.668). Monocytic bias, BM classical monocytosis (&gt;94% CD16- monocytes) and higher overt CMML progression in biTET2/SRSF2 was confirmed in a validation cohort (28, 66 and 22 pts with OM-CMML, MDS and MD-CMML, respectively). BM monocytosis of biTET2/SRSF2 was confirmed in this cohort with morphologic and FC monocyte frequency quantification having high concordance (r=0.737, p&lt;0.001). Genomic subgroup analysis discriminated survival outcomes more reliably than diagnostic categories alone. Our findings suggest that the presence of biTET2 or TET2-SRSF2 co-mutations, combined with BM monocytosis, can more reliably identify true CMML cases within the OM-CMML spectrum. Integrating genomic determinants into the CMML diagnostic framework enhances diagnostic accuracy.

Abstract RAS pathway (RASp) mutations induce proliferative features, and promote transformation in chronic myelomonocytic leukemia (CMML). However, the unique clonal landscape and hierarchy of distinct RASp mutations remain unexplored. To characterize the landscape, architecture, and implications of unique RASp mutations in CMML, we evaluated a cohort of 814 patients with CMML. We identified 461 RASp mutations among 342 patients (42%). N/KRAS and CBL mutations were the most common, frequently involved the P-loop or RING domains, respectively, and frequently appeared as dominant events (63% and 65%, respectively). BRAF, NF1, and PTPN11 mutations spanned throughout the gene structure, and frequently appeared as subclonal events (75%, 64%, and 59%, respectively). CBL mutations frequently occurred in codominance with SRSF2 and multihit TET2, and were enriched for KIT mutations. PTPN11 mutations more frequently co-occurred with SETBP1 and DNMT3A mutations, and were infrequently codominant with TET2 or ASXL1. RASp mutations predicted for shorter overall survival (hazard ratio [HR], 1.55; 95% confidence interval [CI], 1.15-2.07; P = .0075) and leukemia-free survival (LFS; HR, 1.67; 95% CI, 1.26-2.20; P = .0011), influenced outcomes of myelodysplastic and TET2 mutant CMML, and cooperated with IDH2 and RUNX1 mutations to induce shorter LFS. These data set the bases for refined genomic classifications of CMML, and underscore the need to develop RAS-directed therapies for patients with CMML.

We sought to develop a survival model in chronic myelomonocytic leukemia (CMML) that is primarily based on clinical variables and examine additional impact from mutations and karyotype. 457 molecularly-annotated patients were considered. Multivariable analysis identified circulating Blasts ≥2% (1 point), Leukocytes ≥13 x 109/L (1 point), and severe (2 points) or moderate (1 point) Anemia as preferred risk variables in developing a clinical risk Stratification Tool for overall survival (OS), acronymized to "BLAST": low-risk (0 points; median 63 months); intermediate-risk (1 point; median 28 months; HR 2.2, 95% CI 1.6-3.0), and high-risk (2-4 points; median 13 months; 5.4, 4.1-7.3); the corresponding 3/5 year OS rates were 68%/53%, 43%/18%, and 12%/1%. BLAST model performance (AUC 0.77/0.85 at 3/5-years) was shown to be comparable to that of the molecular CMML-specific prognostic scoring system (CMML-mol; AUC 0.73/0.75) and the international prognostic scoring system-molecular (IPSS-M; AUC 0.73/0.74). Multivariable analysis of mutations and karyotype identified PHF6MUT and TET2MUT as being "favorable" and DNMT3AMUT, U2AF1MUT, BCORMUT, SETBP1MUT, ASXL1MUT, NRASMUT, PTPN11MUT, RUNX1MUT, TP53MUT, and adverse karyotype, "unfavorable". Molecular information was subsequently encoded in a combined clinical-molecular risk model (BLAST-mol; AUC 0.80/0.86 at 3/5-years) that included the aforementioned BLAST clinical risk variables and a 3-tiered molecular risk score. BLAST and BLAST-mol were subsequently validated by two separate external cohorts. Independent risk factors for blast transformation included DNMT3AMUT, ASXL1MUT, PHF6WT, leukocytes ≥13 x 109/L, and ≥2% circulating or ≥10% bone marrow blasts. The current study proposes an easy to implement, globally applicable, and molecularly adaptive risk model for CMML.

ABSTRACT Patients with newly diagnosed acute myeloid leukemia (ND‐AML) derive variable survival benefit from venetoclax + hypomethylating agent (Ven‐HMA) therapy. The primary objective in the current study was to develop genetic risk models that are predictive of survival and are applicable at the time of diagnosis and after establishing treatment response. Among 400 ND‐AML patients treated with Ven‐HMA at the Mayo Clinic, 247 (62%) achieved complete remission with (CR) or without (CRi) count recovery. Multivariable analysis–derived hazard ratios (HR), including 1.8 for European LeukemiaNet (ELN) adverse karyotype, 4.7 for KMT2Ar , 1.7 for TP53 MUT , 2.6 for KRAS MUT , and 2.1 for IDH2 WT were applied to develop an HR‐weighted risk model: low, intermediate, and high; respective median survival censored for allogeneic stem cell transplant (ASCT) (3‐year survival) were “not reached” (67%), 19.1 (33%), and 7.1 months (0%). In patients achieving CR/CRi, adverse karyotype, KMT2Ar , KRAS MUT , IDH2 WT predicted inferior survival, allowing for a complementary response‐stratified risk model. The model was externally validated and was shown to be superior to the ELN 2024 risk model (AIC 179 vs. 195 and AUC 0.77 vs. 0.69). Survival was inferior with failure to achieve CR/CRi or not receiving ASCT; 3‐year survival for high‐risk with or without ASCT was 42% versus 0% ( p < 0.01); intermediate 72% versus 43% ( p = 0.06); and low‐risk 88% versus 78% ( p = 0.53). The Mayo genetic risk models offer pre‐treatment and response‐based prognostic tools for ND‐AML treated with Ven‐HMA. The current study underscores the prognostically indispensable role of achieving CR/CRi and ASCT.

Introduction Chronic myelomonocytic leukemia (CMML) is associated with a 15-30% risk for blast transformation (BT) over 3-5 years. Post-BT survival remains dismal with a median overall survival (mOS) of ~6 months (mo), even with intensive therapies. Although genetic drivers are identified in &gt;40% of BT-CMML cases (Carr, 2021), mechanisms of BT remain to be elucidated. This multicenter study examines risk factors for BT in CMML and predictors of post-BT survival. Methods CMML patients at Mayo Clinic (n = 400, 47%) and MD Anderson Cancer Center (n = 449, 53%) were included with IRB approval. Analyses used clinical and molecular data at diagnosis or first referral in chronic phase (CP) and again at BT. Categorical variables were compared by Chi squared and continuous by Mann-Whitney U or Kruskal-Wallis tests. Multivariate analyses used Cox proportional hazards models with binarization via receiver operator curves. Kaplan-Meier analyses were censored at hematopoietic cell transplantation (HCT). P &lt; 0.05 was considered significant. Calculations used BlueSky (v10.3.1) or Prism (v10.2.3). Results Of 849 CMML patients (median age 71 years, 32% female, 90% white) included in the study, 133 (16%) underwent BT. In comparison to CP-CMML without BT, those with BT were more likely (p &lt; 0.05) to have had higher WBC counts (15.1 vs 11.4 x109/L), higher absolute monocyte counts (2.9 vs 2.2 x109/L), lower platelet counts (84 vs 102 x109/L), have circulating immature myeloid cells (52% vs 37%), and be classified as myeloproliferative (MP)-CMML (58% vs 44%) or CMML-2 (33% vs 17%) at diagnosis/first referral in CP. More mutations were detected in BT-CMML than in CP-CMML (mean 3.2 vs 2.8, p = 0.002). Mutations in RAS (NRAS, KRAS, BRAF, CBL, PTPN11; 51% vs 40%, p = 0.02) and signaling (CSF3R, FLT3, JAK2, SH2B3; 15% vs 7%, p = 0.004) pathway genes were more frequent in BT-CMML. Conversely, there were fewer BT-CMML patients with the favorable ASXL1WT/TET2MT genotype compared to CP-CMML (22% vs 33%, p = 0.004). Multivariate analysis identified MP-CMML (hazard ratio [HR] 1.5, p = 0.04), CMML-2 (HR 2.4, p &lt; 0.001), and the presence of ≥4 somatic gene mutations (HR 1.8, p = 0.002) as risk factors for BT whereas the ASXL1WT/TET2MTgenotype was protective (HR 0.45, p = 0.002). The ASXL1WT/TET2MT genotype was also protective in the MP-CMML (HR 0.49, p = 0.04) and RAS-mutated (HR 0.35, p = 0.005) subgroups. Amongst BT-CMML cases, the median time to BT was 10 mo from diagnosis/referral. Compared to their CP assessment, patients in an overlapping Mayo Clinic BT-CMML cohort (n = 156, median age 69 years, 35% female) were more likely to have NPM1 (24% vs 3%, p &lt; 0.001) and IDH2 (14% vs 1%, p = 0.01) mutations at BT. Amongst cases with paired molecular evaluations (n = 44), 7 (16%) acquired new mutations, including FLT3-ITD (n = 2) and NPM1 (n = 2). Conversely, 6 (14%) lost mutations, including TET2 (n = 3), SRSF2 (n = 2), CBL, NRAS, or PTPN11 (n = 1 each). There were no differences in variant allele fractions in driver genes present at CP and BT. Abnormal karyotypes were more prevalent at BT than in CP (60% vs 34%, p &lt; 0.001), and 26 (25%) patients with a normal karyotype in CP acquired structural (65%) and/or numerical (69%) changes at BT, including 6 (23%) that acquired a complex karyotype. Acute myeloid leukemia defining translocations were seen in 4 (4%) cases: 2 with t(8;21) and 1 each with inv(3) and t(10;11)(KMT2A::TET1). After considering mutation and karyotype changes, only bone marrow blasts ≥30% (HR 1.7, p = 0.04) were independently prognostic for OS after BT (mOS 4 mo). Compared to supportive care (n = 50, mOS 1.3 mo), OS was improved by hypomethylating agents (HMA, n = 14, mOS 16 mo, HR 0.3, p &lt; 0.001) or induction chemotherapy (n = 56, mOS 8 mo, HR 0.3, p &lt; 0.001), but not by HMA + venetoclax (n = 11, mOS 6 mo, HR 0.7, p = 0.4). Allogeneic HCT (n = 28) improved mOS to 46 vs 3 mo (p &lt; 0.001). Amongst 4 patients with paired whole genome sequencing before and after BT, 3 had increased copy number alterations (including 2 with loss of heterozygosity, 2 with gains, and 3 with deletions) while the 4th acquired novel likely pathogenic somatic mutations (MUC4, BAX, WNK2). Conclusions Survival in CMML after BT remains exceptionally poor. In this multicenter cohort, MP-CMML, CMML-2, and the presence of ≥4 somatic gene mutations were identified as risk factors for BT, with the ASXL1WT/TET2MT genotype being protective. Following BT, bone marrow blast burden ≥30% was an independent adverse predictor of survival.

Introduction: Somatic mutations in RAS pathway genes are present in ~30% of patients with chronic myelomonocytic leukemia (CMML) and are usually associated with myeloproliferative features. Whether mutations in distinct RAS pathway genes share similar genomic context and have similar implications remains unknown. Such an understanding might inform future studies evaluating genotype-specific therapies (RAS-codon specific and pan-RAS inhibitors) or cooperative leukemogenic mechanisms. This multicenter study examines the landscape of RAS pathway mutations (RASMT) in CMML. Methods: Newly diagnosed CMML patients evaluated at MD Anderson Cancer Center (n=449, 54%) and Mayo Clinic (n=383, n=46%) were included after IRB approval. Clonal relationships were tested using Pearson goodness-of-fit. Clones with the highest variant allele frequency (VAF), or with VAF close to 40% were defined as dominant, and those present at VAF &lt;20% in the presence of another dominant clone were defined as subclonal. RASMT genes included NRAS, KRAS, BRAF, CBL, CBLC, PTPN11 and NF1. Standard statistical measures were applied. Results: Among 832 CMML patients (median age 71 years, 68% male) 387 (47%) had myeloproliferative (MP) and 445 (53%) had myelodysplastic (MD) CMML. A total of 350 (42%) patients had RASMT with higher prevalence among MP-CMML compared to MD-CMML (55% vs 31%, p&lt;0.001). Among patients with RASMT, 262 (32%), 59 (7%), 20 (2%), 6 (&lt;1%) and 3 (&lt;1%) had 1, 2, 3, 4 or 5 distinct RASMT, respectively. Importantly, 17 patients harbored a RASmt as the only detectable somatic mutation, with KRASMT being the most frequent (n=10, 59%), including 3 (18%) patients with &gt;1 RASMT. Among RASMT patients, a trend for higher hemoglobin (12.7 vs 10.8 g/dL, p=0.054) was observed in NF1MT CMML and lower WBC (14.4 vs 27.9 x109/L, p=0.067) values in CBLMT CMML. In total, 479 somatic mutations in RAS pathway genes were identified the most frequent involving NRAS (n=169, 35%) followed by CBL (n=120, 25%), KRAS (n=102, 21%), PTPN11 (n=38, 8%), NF1 (n=34, 7%), BRAF (n=11, 2%) and CBLC (n=5, 1%). Most frequently involved codons for each RAS pathway gene were as follows: G594 (n=3) for BRAF, R420 (n=22) and C404 (n=15) for CBL, G12 (n=41; G12D: 13, G12C: 3) and A146 (n=16) for KRAS, G12 (n=116; G12D: 74, G12C: 5) for NRAS and A72 (n=8) for PTPN11. CBLMT CMML had higher ASXL1 (58% vs 44%, p=0.011), BRAF (8% vs 4%, p=0.017), IDH2 (10% vs 4%, p=0.024), KIT (5% vs 1%, p=0.038), SRSF2 (54% vs 40%, p=0.004), TET2 (59% vs 47%, p=0.019) and ZRSR2 (10% vs 4%, p=0.014) co-mutation frequency. Of note, DNMT3A mutations were mutually exclusive of CBLMT CMML (0% vs 7%, p=0.007), and multihit TET2 was more common among TET2MT/CBLMT vs TET2MT/CBLWT CMML (82% vs 67%, p =0.039). CMML with NF1 mutations had higher BCORL1 (8% vs 0%, p=0.005), PTPN11 (15% vs 3%, p=0.014) and SETBP1 (23% vs 8%, p=0.022) mutational cooccurrences. NRASMT CMML was enriched for ASXL1 (53% vs 43%, p=0.042), ETNK1 (5% vs 1%, p=0.004), GATA2 (4% vs 1%, p=0.034) and STAG2 (4% vs 1 %, p=0.023) mutations but had lower JAK2 (1% vs 6%, p=0.036) and TP53 (1% vs 5%, p=0.022) mutational frequency. PTPN11 mutations were associated with higher NF1 (29% vs 7%, p=0.014), SETBP1 (27% vs 9%, p=0.01) and WT1 (6% vs &lt;1%, p=0.014) mutational cooccurrences and lower SRSF2 (23% vs 42%, p=0.05) frequency. We evaluated RAS mutant clonal dominance in 253 (72%) RASMT patients. NRAS (63% vs 38%), KRAS (63% vs 38%) and CBL (65% vs 35%) mutations were more frequently observed as dominant clones while PTPN11 (44% vs 56%), BRAF (25% vs 75%) and NF1 (35% vs 65%) mutations were more frequently subclonal events, particularly of TET2, ASXL1 or SRSF2 mutant clones. Presence of a RAS pathway mutation was associated with shorter LFS (36 vs 53 months, HR 1.5, 95% CI 1.1-2.2, p=0.020), with presence of multiple RAS mutations shortening LFS further (53, 43, 27 months for 0, 1 and &gt;1, p=0.017). PTPN11 mutations predicted for shorter OS (18 vs 39 months, HR 1.7, 95% CI 1.0-2.8, p=0.035) and LFS (14 vs 36 months, HR 1.9, 95% CI 1.2-2.9, p=0.008). Presence of &gt;1 RAS pathway mutation was associated with shorter OS in ASXL1MT/TET2MT CMML compared to those with no RAS mutations (24 vs 53 months, HR 4.2, 95% CI 1.6-10.8, p=0.004). Conclusions: RAS mutations exhibit distinct genomic features and clonal hierarchy in CMML and cooperate with other somatic mutations to drive clinical outcomes. This study forms the foundation for assessment of novel RAS pathway inhibitors in CMML.

Background. Study of myeloid neoplasms (MN) has been rapidly transformed by genome characterization. Gene mutations have been reported to influence disease phenotype and risk of disease progression, and the evaluation of mutational status can provide valuable insights to improve the decision-making process. In this international study, we evaluated the clinical impact of mutational screening in patients with CMML, which is characterized by significant clinical and genomic heterogeneity, including a substantial proportion of patients with poor outcomes and unmet needs. Currently, only limited molecular information has been incorporated into CMML classification and prognostication. Methods. We retrospectively collected 3182 CMML patients. We used MOSAIC, an AI-based framework for multimodal analysis in rare cancers (PMID: 38875514) to develop innovative, molecular-based tools for classification and prognostication. An independent validation of the results on 516 prospectively collected patients was performed. Results. Bayesian Networs and Hierarchical Dirichlet Processes were used to identify genomic associations and define a CMML molecular classification. We identified 11 clusters with distinct clinical and genomic features, including splicing machinery, signal transduction and tyrosine kinase pathways aberrations, and high-risk molecular signatures (i.e., AML-like and TP53 mutations). Up to 15% of patients showed clear genomic overlap with other MN. Importantly, extensive multi-color flow cytometry on T lymphocytes, NK and myeloid cells (BD FACSymphony A5 Cell Analyzer) revealed specific immunologic and inflammation signatures associated with molecular subgroups. We then integrated molecular and clinical information to build an international CMML Prognostic Scoring System (iCPSS). We assessed the performance of different statistical and AI-based methods; L1-penalized Cox Model resulted as the best performing method. Selected features included hematological parameters (WBC, Hb, PLT and marrow blasts), cytogenetic abnormalities, and mutations in 10 genes (ASXL1, DNMT3A, EZH2, NRAS, RUNX1, SETBP1, STAG2, TET2, TP53, U2AF1). By Bayesian thresholds optimization, we identified 5 risk classes (i.e., very-low, low, intermediate, high and very-high) with median overall survival (OS) ranging from 99 to 9 months (P&lt;.001), and median leukemia-free survival (LFS) ranging from 210 to 18 months (P&lt;.001). iCPSS provided better patient discrimination across all clinical endpoints compared to currently available prognostic tools, with a Concordance Index (CI) for OS of 0.75 versus 0.62-0.64, respectively. Notably, compared to available scoring systems, up to 40% of patients were reassigned to higher or lower risk classes by the iCPSS. External validation on an independent prospective cohort confirmed that the performance of iCPSS is superior to existing prognostic models (CI for OS 0.71 vs 0.54-0.61, respectively) To demonstrate the clinical utility of iCPSS, we focused on 753 patients who underwent allogeneic stem cell transplantation (HSCT). In multivariable analysis, iCPSS stratified the probability of OS post-HSCT (P&lt;.001) and identified groups of patients with different probabilities of disease relapse, ranging from 9% to 62% (P&lt;.001). We employed a clinical- and genomic-based decision support system (PMID: 38723212) to determine the optimal timing of HSCT. Preliminary analyses indicated that higher-risk patients according to iCPSS (i.e., intermediate, high, and very high) benefit from an immediate HSCT strategy. Lower-risk patients (i.e., very low and low) had life expectancy maximized with delayed HSCT. Modeling decision analysis using iCPSS compared to conventional scores, resulted in a change in transplantation policy for a significant proportion of patients (up to 25%). Conclusion. Molecular information significantly improves the classification of CMML patients, providing a basis for refining diagnostic boundaries with other MN and for a more rational inclusion of patients in clinical trials. The iCPSS demonstrated superior performance over currently available scoring systems and improved HSCT decision making process.