Awada, Hussein

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 Introduction The 2022 WHO/ICC classifications of myeloid neoplasms (MNs) incorporated genomics into the existing clinicopathologic framework, defining new (mono)genetic entities. However, this iterative approach resulted in a mix of pathologically and genomically-defined entities, reliant on arbitrary thresholds and historically-derived boundaries between entities. We hypothesised that integrating genomics with other features, de novo from first principles, might better incorporate modern molecular knowledge to define the biological continuum, evolutionary links and overlapping features across MN. Methods The discovery cohort comprised 6484 thoroughly annotated pts. Contemporary diagnoses included 1817 AML, 1695 MDS, 737 MPN, and 1101 MDS/MPN. Non-malignant pts (AA/CHIP/CCUS/ICUS) were included to explore beyond current neoplastic boundaries. 1407 pts had ≥1 follow-up sample/s. Validation was performed in an independent international cohort (n=2012). An unsupervised random forest algorithm was used to integrate status (and VAF) for 28 mutations, karyotypes, CBC features, sex and age (total parameters 144). Features were selected for potential to be independent of BM sampling. Proximity-based multidimensional scaling/PCA were used to visualize latent structure and identify variance-driving features. Results The model projected pts onto a continuous 3-dimensional space. We first applied clustering seeking 1st-order structure, revealing 4 major clusters (C1–C4). These were further subdivided into 58 distinct subclusters (SCs) by an unsupervised iterative clustering approach. All SCs were enriched for distinctive features, with diverse associations linking genomics to clinicopathological/outcome features. Notably, contemporary diagnoses were distributed widely throughout this landscape, with 9/58 SCs containing cases spanning multiple WHO/ICC MN categories. C1 displayed the most internal heterogeneity, with admixed AML, MDS and MDS/MPN pts. It naturally subdivided down to 4th-order subclustering. Ten SCs were dominated by splicing factor mutations. For example, SF3B1MT segregated into 3 SCs, alone or with high-VAF TET2/ASXL1 or DNMT3A mutations. A separate TET2/SRSF2 SC associated with elevated monocyte%, reflecting a discrete CMML-phenotype genomic entity. NPM1MT formed a distinct SC with FLT3-ITD, distinguished further by VAFs of FLT3, TET2, and NRAS. Current genomically-defined AML subtypes separated in the 4th subclustering. C2 encompassed classical MPN driver mutations across 6 SCs, including a distinct CALRMT group and others defined by variable JAK2 VAFs, co-mutation patterns and other CBC features: eg high-VAF SF3B1MT correlated with PB blast%, and concurrent TET2MT with higher Hb/monocyte counts. C3 was characterized by TP53 alterations, with 5 SCs differing by TP53MT burden, 17p lesions, karyotypic complexity and age. Two SCs, one marked by high TP53 VAF+complex karyotype, the other by ch17 lesions, were associated with the worst OS in the whole cohort. C4 contained most known non-malignant pts, dividing into 4 SCs: 2 largely comprising CCUS and AA, while 2 reflect overlaps/convergence between AA and MDS and MPN, respectively. Repurposing single-cell pseudotime analysis suggested multiple disease origins, with various progression trajectories converging on C3. Projecting them onto the model longitudinal samples moved along these derived paths, consistent with their known progression/remission changes, validating the model's biological relevance and predictive utility. Remarkably, the 58 SCs displayed distinct risks of clonal progression and survival. To maximize potential clinical utility we created a pan-MN prognostic score integrating these features. It performed robustly (c-index 0.84), matching/exceeding existing prognostic tools and validated in an independent dataset. Conclusion Integrating clinicopathologic and genomic data we developed ab initio a continuum model of MNs that transcends historical diagnostic boundaries, captures disease evolution and stratifies clinical outcomes. It may better reflect the core biology and natural histories of specific cases, placing more rational boundaries between discrete entities. Potential advantages extend to treatment selection/trial design (in the era of molecularly-targeted therapies) and personalised prognostication throughout each disease course. Expanded validation is needed to resolve complexity in some areas (notably in C2/C4) and validate/refine this prototype framework.

Abstract Background: Higher-risk myelodysplastic syndromes/neoplasms (HR-MDS) and secondary acute myeloid leukemia (sAML) share overlapping molecular features but both remain clinically diverse with poor prognosis. Machine learning (ML)-based genomic clustering offers an opportunity to redefine disease subgroups beyond traditional clinical prognostic schemes or pathologic classifications. Prognostic scoring systems emphasize outcomes while pathologic classifications morphology resulting in inclusion of patients (pts) with different genetic background into same outcome or morphologic subgroup. We compared our ML-unsupervised clustering (UC) model named molecular nosology1 with the taxonomy model2 to evaluate differences in cluster architecture and clinical relevance in HR-MDS and sAML. Methods We analyzed HR-MDS (≥10% blasts) and sAML pts with available molecular data at diagnosis from a well-annotated cohort. Our molecular nosology model used UC via a binary latent-factor model to generate molecular clusters (MCs). The taxonomy model assigned pts to predefined genomic classes. We compared MC size, genomic enrichment, and overall survival (OS) across diagnostic categories. Concordance was assessed using Adjusted Rand Index (ARI) and Normalized Mutual Information (NMI). Purity and entropy were calculated to assess how concentrated (purity) or dispersed (entropy) each taxonomy group was across MCs. Results Among 898 pts, 367 (41%) had HR-MDS and 531 (59%) had sAML. Median age was 71 (IQR 65–77) years. Compared to sAML, HR-MDS pts had similar age (71 vs. 72, p=0.556) but higher hemoglobin (9.9 vs. 9.3, p&lt;0.001) and platelet counts (91 vs. 50, p&lt;0.001). Normal karyotype was more common in HR-MDS (55% vs. 42%, p&lt;0.001). TET2 (24%), ASXL1 (24%), RUNX1 (23%), and SRSF2 (22%) were the most frequently mutated genes. Using our nosology model, we identified 14 MCs.1 The most frequent were MC2 (normal karyotype) at 22% and MC13 (TP53-complex) at 17%, followed by MC1 (+8/ASXL1/TET2/RUNX1, 12%), MC6 (SRSF2/RAS-pathway, 9%), and MC9 and MC14 (each 8%). Less common clusters included MC12 (5%) and MC3–MC5 (2–3%). For comparison, the taxonomy model included 18 genomic groups: TP53-complex (20%), mNOS (13%), bi-TET2, IDH-STAG2, and no event (each 10%), CCUS-like, SETBP1/-7, and AML-like (each 6%), SRSF2 (4%), del(5q), BCOR/L1, EZH2-ASXL1, and SF3B1 (each 3%), and five rare subgroups (DDX41, der(1;7), U2AF1157, U2AF134, ZRSR2; all &lt;1%). Our MCs aligned closely with several taxonomy groups: 82% of TP53-complex cases were assigned to MC13, 99% of no event cases mapped to MC2, and 97% of del(5q) cases mapped to MC8. Bi-TET2 cases were distributed across MC6 and MC12 (both TET2-enriched); SETBP1/-7 to MC11; and EZH2-ASXL1 cases to MC1 and MC12. In contrast, mNOS, AML-like, and CCUS-like groups were genomically heterogeneous and spanned multiple MCs. ARI and NMI values were 0.391 and 0.458, indicating moderate concordance. Purity was highest in no event (0.989), del(5q) (0.967), and TP53-complex (0.823), and lowest in IDH-STAG2 (0.247), CCUS-like (0.241), and AML-like (0.347), all with high entropy (&gt;2.5), reflecting genomic diversity not captured by taxonomy groups alone. Median OS for the entire cohort was 17 mo (95%CI: 14–20). HR-MDS pts had longer OS vs. sAML (26 vs. 12 mo, p&lt;0.001). OS did not differ significantly between HR-MDS and sAML within most taxonomy groups or MCs, except for MC3 (NA vs. 11 mo, p=0.014), MC6 (111 vs. 15 mo, p&lt;0.001), and MC9 (18 vs. 9 mo, p=0.027) for the nosology model and SF3B1 (129 vs. 23 mo, p=0.033), del(5q) (33 vs. 6 mo, p=0.022), bi-TET2 (25 vs. 8 mo, p=0.019), mNOS (34 vs. 9 mo, p&lt;0.001) for taxonomy. Conclusions Some molecular taxonomy groups showed strong concordance with our nosology-defined clusters, which served as the gold standard for genomically distinct subsets (e.g., TP53-complex, del(5q), and no event), irrespective of diagnosis or blast percentage which may present the stage of the disease. In contrast, taxonomy-defined categories such as CCUS-like, AML-like, and mNOS were more genetically diverse and less aligned with specific clusters in the nosology model. Median OS within most clusters was not significantly influenced by blast count, underscoring the potential of molecular-driven clustering to reveal underlying disease biology not captured by morphology and to refine classification in MDS/AML and sAML with few exceptions (SF3B1 or biTET2). Ref 1.Kewan, Durmaz et al.Nat Commun 2023 2.Bernard, et al.Blood 2024

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.

Introduction: In many cases, treatment decisions for multiple myeloma patients must be made in the absence of high quality randomized controlled clinical trials. As a result, many clinicians lean on prognostic models to drive treatment selection. The most widely used of these models incorporate a limited number of patient related factors that reflect tumor burden, cytogenetic features, or gene expression profiling. These prediction scores are derived from regression models that incorporate these variables at the time of diagnosis. They do not account for variables that change dynamically over time, do not explore the relationship between treatment selection and progression, and imperfectly predict survival outcomes. We designed a prognostic model based on an ensemble machine learning platform to predict progression in NDMM using Extreme Gradient Boosting Machine (XGBoost) combined with accelerated failure time modeling for survival analysis. Methods: We utilized a large retrospective data set containing treatment and response information for 1127 patients with newly diagnosed multiple myeloma (NDMM) treated at the Cleveland Clinic Foundation between 2000 and 2023. Following data preprocessing based on data completeness, 953 patient records were included in our training, testing, and cross validation datasets. For the purposes of our analysis, the initial dataset was randomly split into training (70%), testing (15%) and validation (15%) subsets at the patient level. We also defined the lower and upper bounds for each subset, which is critical due to the right-censored nature of survival data. Hyperparameters for XGBoost were optimized using Bayesian search, minimizing the negative log-likelihood. A random forest method was applied for the imputation of missing data (MissForest). We then trained the XGBoost model using GPU acceleration for enhanced computational efficiency. Finally, a log-rank test on predicted survival times was used to test the model's performance on patients with and without known progression. Results: Our preprocessed data set had 953 patients with a mean age at disease onset of 65 years and a slight male predominance (55%). Approximately 27% of patients harbored high cytogenetic risk disease and 3.5% of the cohort presented with extramedullary disease a diagnosis. At a median follow up of 35 months, 47% of patients had experienced disease progression or death. Induction therapy included an immunomodulatory drug in 40% of patients and proteasome inhibitors in 22%. Frontline autologous stem cell transplantation followed induction in 28% of patients. Median progression free survival in the overall data set was 44 months. These features were consistent across the randomly assigned training, testing, and validation cohorts. Following data preprocessing, 34 independent clinical and genomic variables including selection of first line treatment were assessed as model inputs each available record. Under optimized parameters, the model was trained with a maximum tree depth = 7 and a learning rate ~ 0.22 for the desired output of progression free survival (PFS). We then queried PFS for both the training and validation sets and divided patients into those known to have progressed and those without a progression event. Survival analysis was undertaken using predicted PFS values for patients known to have progressed and those who remained progression free. For the validation data set, our model successfully discerned between progressed and non-progressed cases (log-rank test statistic = 22.07, p &lt; 0.005) (Figure 1). Discussion: We present a machine learning approach based on regularized gradient boosting that accurately discerns between patients who experience progression and those who remain progression free at a median ~35 months of follow-up in a large retrospective data set in patients presenting with NDMM. Further elaboration of our model will allow for the incorporation of large amounts data to predict survival outcomes on the basis of dynamic variables such as depth of response and treatment selection. With the ability to parse outcomes among multiple myeloma patients at high resolution, future clinicians and clinical trialists may be able to overcome limitations in trial design and patient-level therapy selection for both newly diagnosed and relapsed patients.