Venetoclax (ven)+azacitidine (aza) is the standard of care for newly-diagnosed acute myeloid leukemia (AML) patients who are not candidates for intensive chemotherapy (IC). Because prognostic factors for ven/aza and IC differ, an AML patient fit for IC may derive more benefit from ven/aza. We therefore designed a trial for younger, newly-diagnosed AML patients with non-favorable risk disease to receive ven/aza regardless of “fitness” for IC. We aimed to understand toxicity and efficacy in this population, and retrospectively compared outcomes to matched IC patients. Newly-diagnosed non-favorable risk patients ≤60 were enrolled and received ven, dose escalated to 600mg/dailyx28 days, with aza 75mg/m2x7 days on a 28-day cycle. Subjects were encouraged to move expeditiously to allogeneic stem cell transplant (ASCT) in first remission. Thirty-six subjects enrolled. Median age was 49 (22-59). Grade ≥3 neutropenia(42%), anemia(33%), thrombocytopenia(53%) and febrile neutropenia(36%) were common. The overall response rate (ORR) was 25/36 (69%) with 19 (53%) complete remissions; 68% of responders achieved MRD-negativity. Most subjects (53%) bridged to ASCT, and the majority of non-responders were successfully salvaged with IC. The median progressionfree- survival (PFS) and overall survival (OS) have not been reached (median follow-up 2.9 years). Compared to IC matched controls, the ORR, ASCT rate and PFS were significantly improved (69% vs 44% [p=0.0495], 53% vs 28% [p-0.0290] and not reached vs 60.8 months [p=0.007]). Hospital days, transfusions and infectious complications were significantly reduced for ven/aza subjects. Ven/aza is feasible for newly-diagnosed, younger, non-favorable risk AML patients, and appears at least as effective as IC.

OBJECTIVES:We discuss challenges using computational modeling approaches for personalized prediction in clinical practice to predict treatment response for rare diseases treated by novel therapies using clinical oncology as an example context. Several challenges are discussed, including data scarcity, data sparsity, and difficulties in establishing interdisciplinary teams. Machine learning (ML), mechanistic modeling (MM), and hybrid modeling (HM) are discussed in the context of these challenges. MATERIALS AND METHODS:We present an HM approach, combining ML and MM techniques for improved personalized model estimation in the context of chimeric antigen receptor T-cell therapy for aggressive lymphoma. RESULTS:The HM approach improved the root mean squared error by 61.27±23.21% compared to using MM alone (MM: 2.36*105∓1.68*105and HM: 9.57*104∓8.37*104, where the units are in cells), computed from 13 patients included in this study. DISCUSSION:By exploiting the complementary strengths of ML and MM approaches, the developed HM method addresses common limitations such as data scarcity and sparsity in medical settings, especially common for rare diseases. CONCLUSION:The HM techniques are likely required to overcome data scarcity and sparsity issues in broad medical settings. Developing these techniques requires dedicated interdisciplinary teams.

While robust tools exist for the analysis of single-cell datasets in both Python and R, interoperability is limited, and analysis tools generally only accept one object class. Considerable programming expertise is required to integrate tools across package ecosystems into a comprehensive analysis, due to their differing languages and internal data structures. This complicates validation of results and leads to inconsistent visualizations between analysis suites. Conversion between object formats is the most common solution, but this is difficult and error-prone due to the rapid pace of development of the analysis suites and their underlying data structures. To address this, we created SCUBA (Single-Cell Unified Backend API), an R package that implements a unified data access API for all common R and Python single-cell object formats. SCUBA extends the data access approach from the widely used Seurat package to SingleCellExperiment and anndata objects. SCUBA also implements new data-specific access functions for all supported object types. Performance scales well across all SCUBA-supported formats. In addition to performance, SCUBA offers several advantages over object conversion for the visualization and further analysis of pre-processed single-cell data. First, SCUBA extracts only data required for the operation at hand, leaving the original object unmodified. This process is simpler, less error prone, and less memory intensive than object conversion, which operates on the entire dataset. Second, code written with SCUBA can use any supported object class as input, with simple and consistent syntax across object formats. This allows a single analysis script or package (like our interactive single-cell browser, scExploreR) to work seamlessly with multiple object types, reducing the complexity of the code and improving both readability and reproducibility. Adoption of SCUBA will ultimately improve collaboration and reproducible research in single-cell analysis by lowering the barriers between package ecosystems.

While robust tools exist for the analysis of single-cell datasets in both Python and R, interoperability is limited, and analysis tools generally only accept one object class. Considerable programming expertise is required to integrate tools across package ecosystems into a comprehensive analysis, due to their differing languages and internal data structures. This complicates validation of results and leads to inconsistent visualizations between analysis suites. Conversion between object formats is the most common solution, but this is difficult and error-prone due to the rapid pace of development of the analysis suites and their underlying data structures. To address this, we created SCUBA (Single-Cell Unified Backend API), an R package that implements a unified data access API for all common R and Python single-cell object formats. SCUBA extends the data access approach from the widely used Seurat package to SingleCellExperiment and anndata objects. SCUBA also implements new data-specific access functions for all supported object types. Performance scales well across all SCUBA-supported formats. In addition to performance, SCUBA offers several advantages over object conversion for the visualization and further analysis of pre-processed single-cell data. First, SCUBA extracts only data required for the operation at hand, leaving the original object unmodified. This process is simpler, less error prone, and less memory intensive than object conversion, which operates on the entire dataset. Second, code written with SCUBA can use any supported object class as input, with simple and consistent syntax across object formats. This allows a single analysis script or package (like our interactive single-cell browser, scExploreR) to work seamlessly with multiple object types, reducing the complexity of the code and improving both readability and reproducibility. Adoption of SCUBA will ultimately improve collaboration and reproducible research in single-cell analysis by lowering the barriers between package ecosystems.

List of specimens used for CITE-seq analysis

WES analysis of AML subpopulations

Clinical information of primary AML specimens

Supplementary Fig. 1: Immunophenotyping and WES analysis of primary AML specimens. Supplementary Fig. 2: Immunophenotyping of engrafted cells in PDX. Supplementary Fig. 3: WES analysis and remission duration of VEN/AZA relapsed AML patients. Supplementary Fig. 4: CITE-seq data analysis. Supplementary Fig. 5: Sorting strategies for determining m-LSC immunophenotype and expression of various LSC markers in m-LSCs. Supplementary Fig. 6: Molecular properties and targeting of m-LSCs in vitro and in vivo. Supplementary Fig. 7: M5 but not M4 patients showing significantly higher refractory rate to VEN/AZA therapy.

Signal transducer and activator of transcription 3 (STAT3) is a well-described transcription factor that mediates oxidative phosphorylation and glutamine uptake in bulk acute myeloid leukemia cells and leukemic stem cells. STAT3 has also been shown to translocate to the mitochondria in acute myeloid leukemia cells, and phosphorylation at the serine 727 (pSTAT3 S727) residue has been shown to be especially important for the mitochondrial functions of STAT3. We demonstrate that inhibition of STAT3 results in impaired mitochondrial function and decreased leukemia cell viability. We discovered a novel interaction of STAT3 with voltage-dependent anion channel 1 (VDAC1) in the mitochondria which provides a mechanism through which STAT3 modulates mitochondrial function and cell survival. Through VDAC1, STAT3 regulates calcium and oxidative phosphorylation in the mitochondria. STAT3 and VDAC1 inhibition also results in significantly reduced engraftment potential of leukemia stem cells, including primary samples resistant to venetoclax. These results implicate STAT3 as a therapeutic target in acute myeloid leukemia.

Metabolomic analysis of m-LSC versus non m-LSC

Background:Patients with myelodysplastic syndromes (MDS) with excess blasts (MDS-EB) have poor long-term outcomes. Our preclinical studies showed MDS-EB stem cells are dependent upon protein synthesis. We designed a phase 1/2 clinical trial to examine the safety/efficacy of the protein synthesis inhibitor omacetaxine mepesuccinate (oma) with the hypomethylating agent (HMA) azacitidine (aza) for patients with untreated MDS-EB. Methods:Enrollment occurred from September 2018 to March 2024 and the study was registered at clinicaltrials.gov (NCT03564873). The phase 1 primary endpoint was to determine the maximum tolerated dose (MTD) and the phase 2 primary endpoint was to determine the overall response rate. Aza 75 mg/m2 was administered daily and oma twice daily days 1-7. Oma was escalated in three cohorts: 0.75 mg/m2, 1.0 mg/m2 and 1.25 mg/m2, with a de-escalation cohort (0.5 mg/m2), to find the maximum tolerated dose (MTD). Responders who tolerated therapy could continue sequential cycles. Those who did not respond, progressed, had significant toxicity or proceeded to allogeneic stem cell transplantation (ASCT) discontinued. Findings:The MTD of oma was 0.5 mg/m2; dose limiting toxicities included hypoxia, respiratory failure, gastrointestinal bleed and gout. Common adverse events included thrombocytopenia, anemia, neutropenia and febrile neutropenia. Overall response rate was 13/24 (54%) with four complete remissions (CR). Ten patients were bridged to ASCT. With median follow-up time of 3.5 years, median response duration and progression-free survival were 719 and 92 days, respectively. Median overall survival was 1.5 years. Interpretation:The MTD of oma in MDS-EB has been established. Responses, including CRs, occurred rapidly. This therapeutic combination, conceived based on data that it targets the malignant stem cell population, could be further studied for patients with MDS-EB but high toxicity needs to be taken into account. Funding:HYPERLINCI, Edward P. Evans Foundation, Leukemia and Lymphoma Society Career Development Program, VA Merit, V-Foundation.

List of commonly mutated genes in AML, Clinical information of VEN/AZA relapsed AML patients, CITE-seq and flow antibodies

The treatment of blast phase chronic myeloid leukemia (bpCML) remains a challenge due, at least in part, to drug resistance of leukemia stem cells (LSC). Recent clinical evidence suggests that the BCL-2 inhibitor venetoclax in combination with ABL-targeting tyrosine kinase inhibitors can eradicate bpCML LSC. In this study, we employed preclinical models of bpCML to investigate the efficacy and underlying mechanism of LSC-targeting with combinations of venetoclax/tyrosine kinase inhibitors. Transcriptional analysis of LSC exposed to venetoclax and dasatinib revealed upregulation of genes involved in lysosomal biology, in particular lysosomal acid lipase A (LIPA), a regulator of free fatty acids. Metabolomic analysis confirmed increased levels of free fatty acids in response to treatment with venetoclax/dasatinib. Pretreatment of leukemia cells with bafilomycin, a specific lysosome inhibitor, or genetic perturbation of LIPA, resulted in increased sensitivity of leukemia cells to venetoclax/dasatinib, implicating LIPA in treatment resistance. Importantly, venetoclax/dasatinib treatment did not affect normal stem cell function, suggesting a leukemia-specific response. These results demonstrate that venetoclax/dasatinib is a LSC-selective regimen in bpCML and that disrupting LIPA and fatty acid transport enhances the response to venetoclax/ dasatinib when targeting LSC, providing a rationale for exploring lysosomal disruption as an adjunctive therapeutic strategy to prolong disease remission.

Single-cell sequencing has revolutionized biomedical research by uncovering cellular heterogeneity in disease mechanisms, with significant potential for advancing personalized medicine. However, participation in single-cell data analysis is limited by the programming experience required to access data. Several existing browsers allow the interrogation of single-cell data through a point-and-click interface accessible to non-programmers, but many of these browsers are limited in the depth of analysis that can be performed, or the flexibility of input data formats accepted. Thus, programming experience is still required for comprehensive data analysis. We developed scExploreR to address these limitations and extend the range of analysis tasks that can be performed by non-programmers. scExploreR is implemented as a packaged R Shiny app that can be run locally or easily deployed for multiple users on a server. scExploreR offers extensive customization options for plots, allowing users to generate publication quality figures. Leveraging our SCUBA package, scExploreR seamlessly handles multimodal data, providing identical plotting capabilities regardless of input format. By empowering researchers to directly explore and analyze single-cell data, scExploreR bridges communication gaps between biological and computational scientists, streamlining insight generation.

LSC gene expression signatures

Background and Significance: Acute myeloid leukemia (AML) is an aggressive, biologically heterogeneous hematopoietic stem cell neoplasm with historically dismal outcomes in patients unable to receive intensive chemotherapy. Based on the pivotal phase 3 VIALE-A study (DiNardo et. al. NEJM, 2020), the combination of venetoclax (Ven), a selective oral BCL-2 inhibitor, with azacitidine (Aza) was approved for the treatment of newly diagnosed AML in adults 75 years or older, or adults ³18 years with comorbidities precluding intensive induction chemotherapy. Overall survival (OS) in VIALE-A was 14.7 months for Ven-Aza vs 9.6 months for placebo-Aza. Among patients with a poor cytogenetic risk the median OS was 7.6 months. Cusatuzumab (Cusa) is a monoclonal antibody with high affinity to human CD70, a cell surface protein expressed on a variety of cancers with limited expression in normal tissues other than activated immune cells. Cusa blocks CD70/CD27 signaling, leading to inhibition of leukemia stem cell proliferation and reduction in leukemic blast cells and exerts direct Fc-mediated, effector functions such as enhanced antibody-dependent cellular cytotoxicity. A Phase 1/2 study (NCT03030612) and a Phase 2 study (NCT04023526) combining Cusa and Aza were conducted in patients with AML unable to tolerate intensive chemotherapy. The combination was generally well-tolerated, with most treatment-emergent adverse events consistent with those expected for hypomethylating agents and there was no obvious dose dependency for toxicities. Manageable infusion-related reactions related to Cusa were noted (Pabst et al. Haematologica, 2023). Subsequently, a single arm Phase 1b study of Cusa combined with Ven and Aza (VAC) for patients with AML was performed (NCT04150887). CR for VAC was 47.6%, CR + CRh + CRi 81.0% and MLFS 11.9% in 42 evaluable patients. Among responders, 47% were MRD negative (Roboz et al, Blood 2021, 138: 369). Overall, the study results showed promising efficacy and safety and support the further development of VAC induction for newly diagnosed patients with AML. Study Design and Methods: OV-AML-1231 is a randomized, open-label, multicenter, multinational, Phase 2 trial to evaluate the efficacy and safety of VAC compared to VA in patients with newly diagnosed AML ineligible for intensive chemotherapy (NCT06384261) that started accrual in July 2024. The planned enrollment is 120 newly diagnosed AML patients, randomized 2:1 to VAC (80 patients) and VA (40 patients). Participants will receive standard of care VA or Cusa 20 mg/kg administered by intravenous infusion on Day 3 and Day 17 of 28-day cycles in combination with VA. The trial population will be enriched for patients with adverse cytogenetic and molecular risk features. The primary end point is OS. Secondary endpoints include event-free survival, response rates including CR, CR+CRh, and rates of MRD negativity. Planned exploratory studies include defining differential responses of AML genetically subgroups to the two treatment arms. Key eligibility criteria include age of 18 years or older and a confirmed diagnosis of previously untreated AML. Eligible patients will include those 75 years of age or older or patients with at least one of the following coexisting conditions: ECOG performance status of 2 or 3, history of congestive heart failure requiring treatment or ejection fraction ≤50% or chronic stable angina, diffusion capacity of the lungs for carbon monoxide (DLCO) ≤65% or forced expiratory volume in the first second (FEV1) ≤65%. Patients with prior exposure to hypomethylating agents are excluded. This ongoing, multicenter, randomized study may inform new frontline low-intensity therapeutic options for newly diagnosed AML patients with high-risk biological features.

Introduction: Dose-adjusted (DA)-EPOCH (etoposide, prednisone, vincristine, doxorubicin, and cyclophosphamide) is a frequently used first line chemotherapy regimen for peripheral T-cell lymphomas (PTCLs), but relapses are common and long-term outcomes are poor. Checkpoint blockade (CPB) immunotherapy has shown modest single agent efficacy in relapsed PTCLs. In other hematologic malignancies the combination of CPB and cytotoxic chemotherapy is promising. Methods: We conducted a Phase I trial to assess safety, spectrum of immune-related toxicity, and efficacy of nivolumab (Nivo) in combination with DA-EPOCH as front-line therapy for PTCLs. Patients received Nivo (360 mg) followed by DA-EPOCH every 21 days for six cycles. Patients were allowed to receive one cycle of standard-of-care chemotherapy prior to enrollment. Results: 18 patients were enrolled: 7 PTCL-not otherwise specified (NOS), 6 nodal T-cell lymphomas with a T-follicular helper phenotype, 2 primary cutaneous gamma/delta T-cell lymphoma, 2 ALK negative anaplastic large cell lymphoma, and 1 subcutaneous panniculitis like T-cell lymphoma. Fifteen had an intermediate or high-risk International Prognostic Index. Immune related (ir) adverse events (AEs) of all grades occurred in 14 and 7 patients experienced ≥ grade 3 irAEs. Eight patients required discontinuation of Nivo due to irAEs. Of the 6 patients who received a cycle of anthracycline-based combination chemotherapy prior to enrollment, none experienced an irAE resulting in Nivo dose hold or discontinuation, compared to 8 of 12 patients whose first cycle was Nivo+DA-EPOCH. There were no hyperprogression events. Interim and end of therapy overall response rates were 94% and 89%, respectively (11 complete responses, 5 partial responses, and 2 progressive diseases). With a median follow up of 707 days, median progression free and overall survival is 434 and 714 days, respectively. Conclusions: Front-line Nivo + DA-EPOCH showed good feasibility and acceptable safety when Nivo was started after chemotherapy but was associated with frequent dose-limiting irAEs when administered synchronously. Efficacy was encouraging with lengthy responses in very high risk PTCL subtypes. Further investigation of front-line line CPB-chemotherapy combinations in PTCL is warranted using a sequential approach. The trial is registered with ClinicalTrials.gov, NCT 03586999.

Background Venetoclax (ven) with azacitidine (aza) is the standard of care for newly diagnosed acute myeloid leukemia (AML) patients unfit for intensive chemotherapy (IC) due to age or comorbidities. However, fit patients with poor-risk disease biology may not benefit from IC. Furthermore, adverse risk factors for IC are not necessarily adverse risk factors for ven/aza. Therefore, we designed a pilot study of ven/aza for newly diagnosed younger AML patients with non-favorable risk disease. Methods This is a prospective, single-arm, multi-institutional investigator-initiated trial (NCT03573024). Newly diagnosed AML patients aged 18-59 with ELN 2017 non-favorable risk disease were enrolled (initially only adverse risk patients were permitted; after five years intermediate risk patients were allowed). Of 36 subjects planned, 32 have enrolled. Stopping rules for futility based on responses were planned. Subjects were matched 1:1 to historical controls who received IC based on age and ELN risk for this analysis. Subjects received aza 75mg/m2 IV d1-7 of a 28-day cycle. Ven was escalated to 600mg on days 1-4 and continued x 28 days. Interruptions to allow count recovery between cycles, with growth factor as needed, occured. Subjects had to achieve complete remission (CR), CRi or morphologic leukemia free state (MLFS) by cycle 2 to continue on study. After at least MLFS, subjects could receive up to 3 additional cycles. Once subjects achieved MRD negativity, they could receive MRD-negative maintenance (5 days of aza and 28 days of ven at 400mg). Subjects with MRD after 4 cycles remained on 7 days of aza and 600mg ven. All subjects were encouraged to proceed to allogeneic stem cell transplantation (ASCT) as soon as they achieved response. Results Among the first 28 subjects, median follow up was 35.3 months. Median age was 48.5 years (22-59); 16/28 (57%) were female. Two of 28 (7%) had treatment related disease; 8/28 (29%) had monocytic disease. Most (23/28, 82%) had adverse risk; 5/28 (18%) had intermediate risk disease. The median number of cycles was 1 (0-4). Overall response rate (ORR: CR+CRi+MLFS), was 17/28 (61%), with 14 CR, 0 CRi and 3 MLFS. After the 27th subject the ORR did not meet the futility threshold (70%) for continuation. However, 4/8 (50%) with monocytic features were non responders compared with 7/20 (35%) without monocytic features. We therefore received approval to continue accrual to 36 subjects after subsequent subjects with monocytic disease were excluded; this was defined by clinical pathologists based on morphologic and/or immunophenotypic disease features. Most (18/28, 75%) ultimately received ASCT; 13/28 (46%) were transplanted in first remission following the study treatment. Eleven of 28 (39%) were refractory to treatment; of these 9 were salvaged with IC and 7/9 responded. Four of 18 responders relapsed prior to ASCT. Only 1/28 (4%) died within 30 days. Median duration of response and OS have not been reached. For the 28 IC matched-control patients, median follow up was 82.3 months. Their ORR was 14/28 (50%) with 8 CR, 4 CRi and 2 MLFS. Twenty received ASCT but only 9 were transplanted in first remission following initial IC therapy. Fourteen were refractory to IC; of these 12 were salvaged, with IC (N=9), ven/aza (n=1) or hypomethylating agent (N=1). Only 1/14 (7%) relapsed prior to transplant; 3/28 (11%) died within 30 days. Median duration of response was not reached and OS was 60.8 months. When comparing study subjects vs controls during the 30 day post treatment period, median days in the hospital were 7.5 (4-30) vs 30 (9-30) (p<0.0001), median units of platelets transfused were 4 versus 11 (p=0.0076), median red blood cell transfusions were 3.5 vs 9 (p=0.0068) and infectious complications occurred in 11/28 (39%) versus 26/28 (93%) (p=0.0071), respectively. Conclusions Ven/aza for younger newly diagnosed AML patients given regardless of fitness for IC in a mostly adverse risk population resulted in an ORR of 61%, with most bridging to ASCT and these subjects having generally good outcomes. Those with monocytic disease features were ultimately excluded because of low response rates; the ORR in the non-monocytic population was 65%. Efficacy and ASCT rates appear similar to matched controls who received IC, with significant decreases in hospitalization, transfusion needs and infectious complications. Molecular sub-classification and propensity score-matched analysis will be presented.

Background Patients with myelodysplastic syndromes (MDS) with excess blasts (MDS-EB) have poor long-term outcomes. Hypomethylating agents (HMA) are the only approved therapies; more efficacious and rationally designed treatment are needed. Previously we identified a stem cell population in MDS and reported its unique dependency on protein synthesis. This led us to design a phase 1/2 study of the protein synthesis inhibitor omacetaxine mepesuccinate (oma) with the HMA azacitidine (aza) for newly diagnosed MDS-EB. We report clinical outcomes and detailed correlative data from this single-institution trial (NCT03564873). Methods Subjects were eligible if they had MDS-EB, no prior therapy and adequate organ function. Aza 75 mg/m2 was administered d1-7. Oma was administered subcutaneously BID d1-7. Three oma dose escalation cohorts (0.75 mg/m2, 1.0 mg/m2, 1.25 mg/m2) with a de-escalation cohort (0.5 mg/m2) were planned in a 3+3 design. Responders who tolerated therapy could continue to receive serial treatment cycles. Non responders, those who progressed, had significant treatment related toxicity or proceeded to allogeneic stem cell transplantation (ASCT) came off study. Correlative endpoints used techniques detailed below. Clinical Results Nine subjects were required to complete phase 1. The median number of cycles was 2 (1-3). Two of the first 3 subjects in cohort 1 experienced DLT (grade 3 hypoxia/grade 4 respiratory failure), requiring de-escalation. In cohort 0, 1/6 had DLT (grade 4 GI bleed/grade 3 gout). The MTD was 0.5mg/m2 SC oma BID d1-7 with aza. Fifteen additional subjects enrolled at the MTD. Of the 24 total, median age was 70, median blasts were 10% and median IPSS-M score was 1.5. The most common grade >2 AEs were leukopenia (N=14), thrombocytopenia, (N=14) anemia (N=9), and febrile neutropenia (N=9); the most common SAEs were febrile neutropenia (N=9), infection/pneumonia (N=5), fever (N=3) and sepsis (N=3). There were two grade 5 events, both sepsis. The overall response rate was 15/24 (63%) with 4 complete remissions. Responses occurred after a median of 1 cycle (1-3). Hematologic improvement (HI) occurred in 10/24 (42%). The median number of cycles was 1 (1-3). Ten subjects were bridged to ASCT; of these 2 relapsed after transplant. Five responders did not bridge to transplant; 3 died of disease progression. Five of 15 responders experienced disease progression; median response duration was 621 days (19-1771) and median progression free survival was 125 days (35-1806). Fifteen subjects have died, from disease progression (N=9), ASCT related mortality (N=3), sepsis (N=2) and pulmonary hypertension (N=1); median overall survival was 438 days. Correlative Studies To gain insight into factors associated with responses, we profiled baseline marrow specimens from responders (N=7) and non-responders (N=5) at single-cell resolution using flow cytometry and Cellular Indexing of Transcriptomes and Epitopes (CITE)-seq. CITE-seq from malignant stem cell populations showed responders had increased protein synthesis signatures, consistent with our preclinical studies, as well as enrichment for NF-kB and pro-inflammatory pathways. In contrast, non-responders had elevation of several metabolic pathways associated with cellular energetics, indicating a fundamentally distinct physiological state that we hypothesize influenced therapeutic response to this regimen. Conclusions Oma/aza in newly diagnosed MDS-EB can be safely administered; higher and more rapid rates of response and HI were observed compared to expectations for HMA alone. A significant number of patients were successfully bridged to ASCT and they had generally favorable long-term outcomes. Given the rapid responses and likelihood of progression without a transplant, oma/aza might best be utilized as a bridge to ASCT for appropriate patients. Correlative studies revealed significant inter-patient heterogeneity in the baseline stem cell population demonstrating two molecular phenotypes, one of which, involving increased protein synthesis and inflammation, was responsive to oma/aza. Collectively, our data identify oma/aza as a promising stem cell-directed therapy for newly diagnosed MDS-EB.

Venetoclax plus azacitidine (ven/aza) is a new standard of care for adult Acute Myeloid Leukemia (AML) patients who are not candidates for intensive therapies. Risk stratification approaches have been proposed to identify patients with favorable, intermediate, and adverse therapeutic outcomes following ven/aza and other lower intensive therapies. However, most have been developed for retrospective data analyses and have limitations in their application to upfront risk stratification of newly diagnosed patients. Here, we describe an AML risk model, termed the Refined Risk Model (RRM), that is specific for ven/aza, addresses important real-world considerations and utilizes pathology features that have the potential to be available relatively quickly-and-broadly following diagnosis. The RRM was developed and internally validated using a single center cohort of 316 AML patients from the University of Colorado treated upfront with ven/aza, and then externally validated on an AML cohort from a nationwide electronic health record-derived de-identified AML database. The RRM effectively stratified patients into Adverse, Intermediate, and Favorable groups across both the internal and external cohorts; it performed well in subsets with or without allogeneic transplant recipients, demonstrated tolerance to missing data, and showed numerical performance comparable to or exceeding the existing alternatives such as the European Leukemia Network (ELN 2022) and molecular prognostic risk signature (mPRS) models. These findings suggest that the RRM may have potential application in defining the prognostic mortality risk for newly diagnosed AML patients, which may help guide clinical trial design and execution as well as other important elements of AML clinical decision support. ### Competing Interest Statement Both CAS and MB are employees of and hold equity in OncoVerity. In addition, CAS is a consultant to RefinedScience. All other authors declare no conflicts of interest. ### Funding Statement This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: This study was a retrospective analysis utilizing a limited data set. Study approval and a full HIPAA waiver of consent was granted by the Colorado Multiple Institutional Review Board (approval number 23-2059) I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes This retrospective study was approved by CU internal review board (IRB) and used a limited dataset with a waiver of consent from the CU IRB. The raw, individual patient data are protected and not available due to data privacy laws. The processed data are available at reasonable request to the corresponding author. The Flatiron Health data that supported the findings of this study were originated by and are the property of Flatiron Health, Inc., which has restrictions prohibiting the authors from making the data set publicly available. Requests for data sharing by license or by permission for the specific purpose of replicating results in this manuscript can be submitted to PublicationsDataAccess@flatiron.com.