AUTHORS
Jessica Maria Stempel, MD, MHS, Seongseok Yun, MD, PhD2, Jad Othman, MD3, Mael Heiblig, MD4, Michael Heuser, MD5, Rabia Shahswar6, Andy Kaempf, MS7, Daniel Pollyea, MD8, Joshua F. Zeidner, MD9, Curtis Lachowiez, MD10, Anand Ashwin Patel, MD11
AFFILIATIONS
1 Oregon Health & Science University, Portland, Oregon, United States
2 Department of Malignant Hematology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida, United States
3 Department of Medical and Molecular Genetics, King's College London, London, United Kingdom, Haematology Department, Royal North Shore Hospital, St. Leonards, NSW, Australia, University of Sydney, Sydney, Australia
4 Hematology Department, Lyon-Sud Hospital, Hospices Civils de Lyon, Pierre Bénite, France
5 Department of Hematology, Hemostasis, Oncology and Stem Cell Transplantation, Hannover Medical School, Hannover, Germany, Department of Internal Medicine IV, University Hospital Halle (Saale), Martin-Luther-University Halle-Wittenberg, Halle, Germany
6 Department of Hematology, Hemostasis, Oncology and Stem Cell Transplantation, Hannover Medical School, Hannover, Germany, Molecular Pharmacology Program, Memorial Sloan Kettering Cancer Center, New York, United States
7 Oregon Health & Science University, Portland, Oregon, United States
8 Division of Hematology, University of Colorado School of Medicine, Aurora, Colorado, United States
9 University of North Carolina, Lineberger Comprehensive Cancer Center, Chapel Hill, North Carolina, United States
10 Oregon Health & Science University, Portland, Oregon, United States
11 12 Section of Hematology-Oncology, Department of Medicine, University of Chicago, Chicago, Illinois, United States
ABSTRACT
Background
Acute myeloid leukemia (AML) arising from an antecedent MPN is associated with a median overall survival (OS) of about 8 months even in the era of novel approved therapies (Patel et al, 2024; Bewersdorf et al, 2026). HMA-Ven is a commonly utilized strategy in MPN-BP, however the pivotal VIALE-A trial of azacitidine-Ven excluded pts with prior MPN (DiNardo et al. 2020). Ven-era prognostic models derive predominantly from de novo AML and generically defined secondary AML, therefore the utility of these models in MPN-BP is unclear. We report outcomes in a multinational cohort of pts with MPN-BP treated with frontline HMA-Ven and analyze risk stratification criteria to identify pts most likely to benefit.
Methods
Pts with MPN-BP who received frontline HMA-Ven were identified from the PRISM (Prognostic Risk Integration for Survival Modeling) database (Lachowiez et al. 2026). We extracted demographics, MPN subtypes, prior therapy, treatment characteristics, and clinical outcomes. Responses were classified by European LeukemiaNet 2022 criteria with composite complete remission (cCR) comprising CR, CR with incomplete hematologic recovery, and CR with partial hematologic recovery. Median OS was estimated by Kaplan-Meier and compared across molecular prognostic risk signature (mPRS, Döhner et al. 2024) and PRISM risk groups using the log-rank test. Discrimination for OS was assessed by Harrell’s C-index.
Results
We identified 102 pts with MPN-BP. Median age was 72 years (range, 47-87). Antecedent MPN subtypes were polycythemia vera (16, 15.7%), essential thrombocythemia (14, 13.7%), myelofibrosis (25, 24.5%), MPN-unclassifiable (3, 2.9%); the subtype was not specified in 44 pts (43.1%). Prior treatment was available for 67 pts (65.7%), of whom 20 (29.8%) received a JAK inhibitor and 3 had prior HMA (4.5%).
All pts received frontline HMA-Ven (azacitidine 79, 77.5%; decitabine 23, 22.5%). The median number of cycles was 2 (interquartile range [IQR], 2-4). Eight pts (7.8%) proceeded to allogeneic stem-cell transplant (median time to transplant, 146 days; IQR, 112-238). A total of 74 pts (72.5%) received antifungal prophylaxis during induction.
JAK2 was the most prevalent driver mutation in 66 pts (64.7%) followed by CALR (12, 11.8%) and MPL (4, 3.9%); 20 pts (19.6%) were triple-negative. Complex karyotype was identified in 33 pts (32.4%). Common mutations were ASXL1 (36, 35.3%), TP53 (30, 29.4%), RUNX1 (26, 25.5%), TET2 (26, 25.5%), SRSF2 (22, 21.6%), IDH1/2 (19, 18.6%), KRAS/NRAS (17, 16.7%), and FLT3-ITD (4, 3.9%). By mPRS, 55 pts (53.9%) had favorable, 17 (16.7%) intermediate, and 30 (29.4%) adverse risk disease. By PRISM, 6 pts (5.9%) had low, 34 (33.3%) moderate, and 62 (60.8%) high risk disease. Relative to mPRS, PRISM reclassified 58 pts (56.9%) to higher risk groups: of 55 favorable risk pts, 26 were reclassified as moderate and 23 as high risk; 9 intermediate risk pts to high risk.
Best response was evaluable in 75 pts. The cCR was 50.7% (38/75). Two pts (2.7%) achieved MLFS, 10 (13.3%) partial remission, and 25 (33.3%) had no response. Median time to cCR was 57 days (IQR, 28-84). 30-day mortality was 14.7% (n=15).
After a median follow-up of 15 months, 66 deaths occurred (64.7%). Median OS for the overall cohort was 7.3 months (95% confidence interval [CI], 3.9-10.6). Higher mPRS risk category was associated with progressively shorter OS (9.9, 7.3, 2.0 months for favorable, intermediate, and adverse risk groups, respectively; log-rank p=0.002). When applying the PRISM risk stratification, median OS was not reached for low, 14.0 months (95% CI, 4.9-23.1) for moderate risk, and 4.1 months (95% CI, 2.6-5.6) for high risk (log-rank p=0.001).
OS discrimination was similar in prognostic models (C-index 0.61 for PRISM and 0.60 for mPRS, p=0.88), although PRISM identified twice as many higher risk pts (62 vs 30), including 23 designated favorable by mPRS whose median OS was only 5.7 months.
Conclusion
In the largest cohort of pts with MPN-BP uniformly treated with frontline HMA-Ven reported to date, cCR and median OS compared similarly with historical outcomes. mPRS and PRISM stratified survival with comparable discrimination. However, PRISM captured a larger high-risk group, including a set of pts categorized as favorable by mPRS whose outcomes nonetheless remained poor. Pending external validation, PRISM could meaningfully refine risk stratification in MPN-BP and identify pts most likely to benefit from frontline HMA-Ven.
DISCLOSURES
Jessica Maria Stempel, MD, MHS: No relevant conflicts of interest to declare.