# paper.pdf ## Metadata - PDFFormatVersion=1.7 - Language=en - IsLinearized=true - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - CreationDate=D:20260428192750+05'30' - Subject=BLOOD ADVANCES, 10 (2026) 3054-3058. doi:10.1182/bloodadvances.2025019262 - Author=Mikkael A. Sekeres∗ - Creator=Elsevier - Producer=PDFlib+PDI 10.0.2p2 (C++/Win64) - Custom.EPSprocessor=PStill version 1.84.42 - Custom.ElsevierWebPDFSpecifications=7.0.1 - Custom.robots=noindex - Custom.doi=10.1182/bloodadvances.2025019262 - ModDate=D:20260428193229+05'30' - Title=Exposure to Agent Orange and association with myelodysplastic syndromes - dc:format=application/pdf - dc:identifier=doi:10.1182/bloodadvances.2025019262 - dc:title=Exposure to Agent Orange and association with myelodysplastic syndromes - dc:creator=Mikkael A. Sekeres∗,Amy E. DeZern∗,Michael Otterstatter,Eric Padron,Tareq Al Baghdadi,James M. Foran,Rami S. Komrokji,Gregory A. Abel,Wael Saber,Steven D. Gore,Cecilia Lee,Rafael Bejar,Jane J. Liu,H. Joachim Deeg,Seth Sherman,R. Coleman Lindsley,Matthew J. Walter,Nancy Gillis - dc:description=BLOOD ADVANCES, 10 (2026) 3054-3058. doi:10.1182/bloodadvances.2025019262 - dc:publisher=American Society of Hematology - prism:aggregationtype=journal - prism:publicationname=BLOOD ADVANCES - prism:copyright=Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0), permitting only noncommercial, nonderivative use with attribution. - prism:issn=2473-9529 - prism:volume=10 - prism:number=9 - prism:coverdisplaydate=12 May 2026 - prism:coverdate=2026-05-12 - prism:pagerange=3054-3058 - prism:startingpage=3054 - prism:endingpage=3058 - prism:doi=10.1182/bloodadvances.2025019262 - prism:url=https://doi.org/10.1182/bloodadvances.2025019262 - jav:journal_article_version=VoR - pdfx:elsevierwebpdfspecifications=7.0.1 - pdfx:doi=10.1182/bloodadvances.2025019262 - pdfx:robots=noindex - pdfx:epsprocessor=PStill version 1.84.42 - xap:creatortool=Elsevier - xap:createdate=2026-04-28T19:27:50+05:30 - xap:modifydate=2026-04-28T19:32:29+05:30 - xap:metadatadate=2026-04-28T19:32:29+05:30 - xaprights:marked=True - tdm:reservation=1 - tdm:policy=https://www.elsevier.com/tdm/tdmrep-policy.json - pdf:producer=PDFlib+PDI 10.0.2p2 (C++/Win64) - xapmm:documentid=uuid:59f6b1ee-6ce7-4997-9f7d-20246708165f - xapmm:instanceid=uuid:4f77e5ea-f6c4-4d5c-a1ea-19fa2a3a0ebf - ali:license_ref=https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode ## Contents ### Page 1 ![Figure]() TO THE EDITOR: Exposure to Agent Orange and association with myelodysplasticsyndromes Mikkael A. Sekeres,1,* Amy E. DeZern,2,* Michael Otterstatter,3 Eric Padron,4 Tareq Al Baghdadi,5 James M. Foran,6 Rami S. Komrokji,4Gregory A. Abel,7 Wael Saber,8 Steven D. Gore,9 Cecilia Lee,10 Rafael Bejar,11 Jane J. Liu,12,13 H. Joachim Deeg,14 Seth Sherman,3R. Coleman Lindsley,7 Matthew J. Walter,15 and Nancy Gillis4 - 1 Division of Hematology, Sylvester Cancer Center, University of Miami, Miami, FL;2 Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD; - 3 The Emmes Company, Rockville, MD;4 Department of Malignant Hematology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL;5 Trinity Health, Ann Arbor,MI;6 Division of Hematology & Medical Oncology, Mayo Clinic, Jacksonville, FL;7 Division of Population Sciences, Dana-Farber Cancer Institute, Boston, MA;8 Division ofHematology/Oncology, Department of Medicine, Medical College of Wisconsin, Milwaukee, WI;9 Cancer Therapy Evaluation Program, National Cancer Institute, Bethesda,MD;10 Division of Cancer Prevention, National Cancer Institute, Rockville, MD;11 Division of Hematology and Cellular Therapy, University of California San Diego, San Diego,CA;12 Illinois Cancer Care, Peoria, IL;13 Heartland NCI Community Oncology Research Program, Decatur, IL;14 Clinical Research Division, University of Washington andFred Hutchinson Cancer Center, Seattle, WA; and15 Department of Medicine, Division of Oncology, Washington University, St. Louis, MO Myelodysplastic syndromes/neoplasms (MDS) are hematopoietic disorders characterized by cytopenias, a predilection for evolution to acute myeloid leukemia (AML), and truncated survival.1 MDSdevelops from the acquisition of somatic mutations occurring over years or decades.2-4 Although somemutations develop due to random stochastic events, others are introduced by exposure to extrinsicagents that promote genomic toxicity or instability, including chemotherapy, radiation therapy, ordiscrete environmental toxins.5 Agent Orange (AO) was a phenoxy herbicide mixture used as a defoliant, particularly during theVietnam War (1962-1971). Often, it was handled directly by military staff and was diluted with fuel topromote aerosolization. Additionally, AO was contaminated with dioxins, carcinogens associated withincreased risk of prostate cancer, bladder cancer, soft tissue sarcomas, and lymphoid hematologicmalignancies.6-8 AO has historically been suspected but not yet definitively linked as a causative agentin MDS development.9 The National Heart, Lung, and Blood Institute National MDS Natural History Study (identifier: NCT02775383) is a prospective cohort study conducted across 161 community andacademic sites in the United States from June 2016 to July 2024 that established a data and biospecimen repository for MDS and related or precursor disorders.10 Patients were enrolled whensuspected of having MDS and underwent centrally confirmed clinical, genopathologic, and detailedepidemiologic assessments, including assessment for AO exposure and other relevant environmentaland clinical exposures. Molecular mutations were identified using a custom xGen Predesigned GeneCapture Pool (Integrated DNA Technologies) that spans a 96-gene target region with 3608 probes,and cytogenetic abnormalities were identified using metaphase karyotyping. Approximately half werediagnosed with MDS or related myeloid/precursor disorders and half were not, making this an idealcohort to explore variables associated with AO exposure and the association between AO exposureand an MDS diagnosis, especially with substantial recruitment from birth cohorts eligible to serve inVietnam. Submitted 1 December 2025; accepted 3 February 2026; prepublished online onBlood Advances First Edition 24 February 2026. *M.A.S. and A.E.D. are joint first authors. Presented in part at the 67th annual meeting of the American Society of Hematology,Orlando, FL, 8 December 2025. The analyses presented in this publication are based on the use of clinical study datathat have been deposited to BioLINCC (). Targeted 3054 sequencing data presented in this publication can be downloaded from dbGaP underphs002714.v3.p2. The full-text version of this article contains a data supplement. Licensed under, permitting only noncommercial, nonderivativeuse with attribution. 12 MAY 2026 • VOLUME 10, NUMBER 9 ### Page 2 | Characteristic | All (N = 2115) | Non-AO exposed (n = 1985) | AO exposed (n = 130) | P value | | - | - | - | - | - | | Sex, n (%) | <.001 | | Female | 799 (38) | 794 (40) | 5 (4) | | Male | 1316 (62) | 1191 (60) | 125 (96) | | Age group (y), n (%) | <.001 | | <60 | 334 (16) | 333 (17) | 1 (1) | | 60-69 | 491 (23) | 473 (24) | 18 (14) | | 70-79 | 831 (39) | 734 (37) | 97 (75) | | >80 | 459 (22) | 445 (22) | 14 (11) | Race, n (%) .066 | White | 1901 (90) | 1788 (90) | 113 (87) | | - | - | - | - | | Asian | 35 (2) | 34 (2) | 1 (1) | | Black or African American | 108 (5) | 95 (5) | 13 (10) | | Other | 22 (1) | 22 (1) | 00 (0) | | Unknown/not reported | 49 (2) | 46 (2) | 3 (2) | Ethnicity, n (%) .704 | Not Hispanic or Latino | 1971 (93) | 1849 (93) | 122 (94) | | - | - | - | - | | Hispanic or Latino | 74 (3) | 71 (4) | 3 (2) | | Unknown/not reported | 70 (3) | 65 (3) | 5 (4) | Disease group, n (%) .022 | ICUS | 114 (5) | 109 (5) | 5 (4) | | - | - | - | - | | Other at-risk | 156 (7) | 149 (8) | 7 (5) | | CCUS | 302 (14) | 276 (14) | 26 (20) | | MDS | 510 (24) | 466 (23) | 44 (34) | | MPN overlap | 103 (5) | 96 (5) | 7 (5) | | AML<30% | 29 (1) | 28 (1) | 1 (1) | | Nonmyeloid | 901 (43) | 861 (43) | 40 (31) | IPSS-R, n (% of MDS only) .603 | Low/very low | 240 (47) | 220 (47) | 20 (45) | | - | - | - | - | | Intermediate | 81 (16) | 75 (16) | 6 (14) | | High/very high | 90 (18) | 82 (18) | 8 (18) | | Not calculated | 99 (19) | 89 (19) | 10 (23) | Cytogenetic score, n (%) .063 ![Figure]() | Very good/good | 981 (46) | 928 (47) | 53 (41) | | - | - | - | - | | Intermediate | 55 (3) | 49 (2) | 6 (5) | | Poor/very poor | 89 (4) | 79 (4) | 10 (8) | | Missing | 990 (47) | 929 (47) | 61 (47) | MDS subtype, n (% of MDS only) .373 | MDS-EB1 | 74 (15) | 67 (15) | 7 (17) | | - | - | - | - | | MDS-EB2 | 81 (17) | 76 (17) | 5 (12) | | MDS isolated del(5q) | 120 (25) | 112 (25) | 8 (19) | | MDS-MLD | 79 (16) | 70 (16) | 9 (21) | | MDS-RSMLD | 45 (9) | 42 (9) | 3 (7) | | MDS-SLD | 30 (6) | 26 (6) | 4 (10) | | MDS-RSSLD | 38 (8) | 32 (7) | 6 (14) | | MDS-U | 21 (4) | 21 (5) | 00 (0) | ICUS, Idiopathic Cytopenia of Undetermined Significance; IPSS-R, Revised International Prognostic Scoring System; IQR, interquartile range; MDS-EB1, 5% to 9% blasts, MDS withexcess blasts-1; MDS-EB2, 10% to 19% blasts, MDS with excess blasts-2; MDS-MLD, MDS with multilineage dysplasia; MDS-RSMLD, MDS with multilineage dysplasia and ringsideroblasts; MDS-RSSLD, MDS with single lineage dysplasia and ring sideroblasts; MDS-SLD, MDS with single-lineage dysplasia; MDS-U, MDS unclassifiable; MPN, myeloproliferativeneoplasm. 12 MAY 2026 • VOLUME 10, NUMBER 9 RESEARCH LETTER 3055 ### Page 3 Characteristic All (N = 2115) Non-AO exposed (n = 1985) AO exposed (n = 130) P value Laboratory values, median (IQR) Hemoglobin 10.9 (9.3-12.7) 10.9 (9.3-12.7) 11.0 (9.3-12.7) .920 Platelets 143.0 (89.0-237.5) 144.0 (90.0-239.0) 125.5 (72.8-207.8) .070 Absolute neutrophil count 2.3 (1.3-4.0) 2.4 (1.3-4.0) 2.2 (1.2-3.9) .920 ICUS, Idiopathic Cytopenia of Undetermined Significance; IPSS-R, Revised International Prognostic Scoring System; IQR, interquartile range; MDS-EB1, 5% to 9% blasts, MDS withexcess blasts-1; MDS-EB2, 10% to 19% blasts, MDS with excess blasts-2; MDS-MLD, MDS with multilineage dysplasia; MDS-RSMLD, MDS with multilineage dysplasia and ring sideroblasts;MDS-RSSLD, MDS with single lineage dysplasia and ring sideroblasts; MDS-SLD, MDS with single-lineage dysplasia; MDS-U, MDS unclassifiable; MPN, myeloproliferative neoplasm. The Fisher exact and Kruskal-Wallis tests assessed bivariateassociations between AO exposure and demographic, diagnostic,clinical, genetic, military service, and smoking variables in thisretrospective cohort study. To identify correlates of AO exposurein the full study cohort, multivariable logistic regression modelswere used with exposure (yes/no) as the dependent variable, anddemographic covariates, military service, mutational burden,smoking history, cytogenetics, and previous hematologic conditions or cancers as independent variables. To assess the association of AO exposure and clinical outcomes, including time toprogression and overall survival, among MDS and clonal cytopeniaof undetermined significance (CCUS) disease groups, separateCox regression models were used with time to progression or timeto death as dependent variables, and AO exposure, age, sex,disease group, and military service as independent variables. Disease progression assessment included disease worsening (eg,progression from CCUS to MDS, or from MDS to AML), changesin disease-directed therapy or transfusion dependency, anddisease-related death. The Bonferroni-Holm multiple comparisonscorrection was used when assessing mutation prevalence, maintaining an overall value of α of 0.05 (2-tailed). Among 2115 enrolled patients, 130 (6.1%) reported AO exposure, with the majority (96%) being male (). Compared withpatients who were not exposed, patients who were exposed to AOwere older (P < .001), more likely from the 1945 to 1954 birthcohort of age to have served in Vietnam, and to have an ultimateMDS or CCUS diagnosis (54% vs 37%, P = .022). AO exposuretrended to be more common in patients reporting being Black(P = .066, 12% overall and 21% among males, compared with5.8% and 9.2%), those with poor-risk cytogenetics (P = .063), andwith lower baseline platelets (P = .070), respectively. In multiple regression analyses including the full study cohort ofpatients with or without MDS, CCUS, and related myeloid conditions (), the odds of AO exposure were significantly higherfor the 1945 to 1954 birth cohort (odds ratio [OR], 6.4; 95%confidence interval [CI], 2.7-19.3; P < .001), those with militaryservice (OR, 18.6; 95% CI, 8.0-50.7; P < .0001) and for thosewho were Black (OR, 2.9; 95% CI, 1.1-7.0; P = .021). Althoughthe crude odds of AO exposure were higher among males (OR,16.7; 95% CI, 7.5-47.2) and trended higher for CCUS (OR, 2.0;95% CI, 1.2-3.4) or MDS (OR, 2.0; 95% CI, 1.3-3.2), differenceswere abrogated in regression analyses. Controlling for these differences by sex, race, and military service,AO exposure was significantly associated with more molecularmutations (OR, 1.10; 95% CI, 1.0-1.2; P = .047), and with havinga very poor/poor cytogenetic risk score per the revised International Prognostic Scoring System among those with MDS(OR, 2.39; 95% CI, 1.04-5.14; P = .03), and younger age at MDSdiagnosis (OR, 0.96; 95% CI, 0.93-0.99; P = .008). AO exposuretrended higher for those with previous cancer diagnoses, particularly for MDS (OR, 1.83; 95% CI, 0.95-3.53; P = .07); patientswith AO exposure and MDS had higher rates of previous prostatecancer (15% vs 6%; P = .027), which was notable given itsassociation with AO exposure. Importantly, AO exposure was notassociated with cigarette smoking in the entire cohort (OR, 1.21;95% CI, 0.81-1.82; P = .363). Across the full study cohort, AO exposure was associated with del(20q) karyotypic abnormalities (6% vs 2% overall; P = .018),mutations in TET2 (26% vs 17%; P = .008), SRSF2 (14% vs 8%;P = .035), U2AF1 (9% vs 5%; P = .032), ZRSR2 (5% vs 2%;P = .031), and KRAS (3% vs 1%; P = .049); TET2 remainedsignificant after multiple comparisons adjustment. Focusing ongene pathways,11 patients who were exposed to AO were morelikely to have RNA splicing abnormalities (35% vs 23%; P = .004), | Term | OR | 95% CI | P value | | - | - | - | - | | Age | 0.961 | 0.934-0.990 | .008 | | Birth cohort: 1935-1944 | 1.068 | 0.396-3.391 | .903 | | Birth cohort: 1945-1954 | 6.434 | 2.656-19.276 | <.001 | | Birth cohort: 1955-1964 | 0.868 | 0.112-4.657 | .876 | | Birth cohort: >1964 | 1.452 | 0.067-11.689 | .758 | | Male sex | 1.878 | 0.614-7.026 | .299 | | Black or African American | 2.892 | 1.135-7.000 | .021 | | Other/unknown/not reported | 0.493 | 0.076-1.808 | .359 | | Military service | 18.563 | 7.991-50.681 | <.001 | | Number of mutations | 1.100 | 0.999-1.207 | .047 | | Maximum VAF | 1.661 | 0.638-4.264 | .294 | | Previous hematologic condition | 0.858 | 0.586-1.263 | .434 | | Previous cancer | 1.156 | 0.756-1.745 | .497 | | Smoking history | 1.206 | 0.810-1.820 | .363 | | Cytogenetic risk score: very poor/poor | 2.394 | 1.042-5.141 | .031 | | Cytogenetic risk score: intermediate | 1.865 | 0.645-4.719 | .212 | Reference levels for categorical predictors: birth cohort <1935, female sex, White, nomilitary service, no previous hematologic conditions, no previous cancer, no history ofsmoking, good/very good cytogenetic risk score. Due to colinearity with age, birth cohortwas dropped from the final model. VAF, variant allelic frequency. 3056 RESEARCH LETTER 12 MAY 2026 • VOLUME 10, NUMBER 9 ### Page 4 chromatin modification (19% vs 13%; P = .029), RAS signaling(9% vs 6%; P = .017), and cohesin complex pathways (7% vs 3%;P = .043); RNA splicing remained significant after multiple comparisons adjustment (supplemental Tables 1 and 3; supplementalFigure 1). Similar patterns were observed among patients with justMDS with AO exposure (supplemental Tables 2 and 4;supplemental Figure 2). Among patients with MDS and CCUS, those who were exposed toAO had similar overall survival (hazard ratio, 1.1; 95% CI, 0.7-1.6;P = .72), but higher risk of disease progression (ie, to MDS, tohigher-risk MDS, or to AML) in the first 2 years following diagnosiscompared with patients who were not exposed (hazard ratio, 1.8;95% CI, 1.2-2.8; P = .004), after adjustment for age, diagnosis,sex, and military service. AO exposure has been linked to a variety of medical conditionsand to discrete cancers, including cancers of the skin, prostate,colon, lung, and bladder, likely mediated through AO’s contamination with the carcinogen dioxin or fuel additives. To the best of our knowledge, for the first time we were able todemonstrate the impact of AO exposure on the development ofMDS in this large cohort of patients with unexplained cytopenias,57% of whom were ultimately diagnosed with MDS and relatedconditions. AO exposure was over twice as common in patientswho reported being Black, despite Black individuals comprisingjust 16% of men who served in the Vietnam War (23% of combattroops) and being less likely to develop MDS than White individuals.12 This may reflect both differential combat roles played byBlack individuals and the earlier age at MDS diagnosis for Blackindividuals than for White individuals.13 Patients included in our cohort, by design, had cytopenias that ledto a bone marrow biopsy to evaluate MDS, although many did notultimately have that diagnosis. Patients exposed to AO were morelikely to have a final MDS diagnosis and were almost 4 times morelikely to have poor/very poor risk cytogenetic abnormalities andmore molecular mutations. This is consistent with research inJapanese citizens who were similarly exposed to a genotoxicagent, the atomic bomb, at a young age;14 those exposed whodeveloped MDS decades later similarly were more likely to havecomplex karyotypes, reflecting the serial acquisition of geneticmutations over decades following an initial inciting event, as withAO here.15 We also noted higher rates of discrete genetic mutations following AO exposure, including del(20q) and TET2 mutations, along with alterations in specific functional gene pathways. Exposure to dioxin may cause epigenetic modifications and alterations in DNA methylation profiles, events also common in MDS,16and may induce impaired DNA damage response.17 Exposure togasoline, an AO diluent, has also been linked to MDS and AMLdevelopment,18 further supporting a viable biologic rationale for theassociation of AO exposure with MDS. Another study of 29 VietnamWar-era veterans with MDS reporting AO exposure found that themajority of mutations occurred along epigenetic regulation and RNAsplicing pathways, similar to our cohort.19 Our AO-exposed MDScohort also appeared to differ slightly in the prevalence of discretegene mutations from patients who were not exposed to AO. Study limitations include lack of information on the amount orduration of AO exposure (although the US Veterans Administration considers anyone on active military duty in the country to havebeen exposed) and inability to control for subsequent occupationalexposures associated with MDS. In conclusion, in the National MDS Natural History Study cohort ofpatients with ≥1 cytopenia, AO exposure appears to be associated with younger age at diagnosis, being Black, MDS diagnosis,disease progression, and genetic complexity of MDS. The study received institutional review board approval through theNational Institutes of Health and at individual participating EasternCooperative Oncology Group-American College of RadiologyImaging Network, Southwest Oncology Group, Alliance, and NCICommunity Oncology Research Program sites and was conducted in accordance with the Declaration of Helsinki. Acknowledgments: The National MDS Natural History study hasbeen supported by US federal government contractsHHSN268201400003I and HHSN268201400002I from theNational Heart, Lung, and Blood Institute and additional fundingby the National Cancer Institute (NCI) to NCI CommunityOncology Research Program and the participating clinical centers. Contribution: M.A.S., A.E.D., M.O., and N.G. conceived anddesigned the study, analyzed data, and wrote and edited themanuscript; E.P., T.A.B., J.M.F., R.S.K., G.A.A., W.S., S.D.G.,C.L., R.B., J.J.L., H.J.D., S.S., R.C.L., and M.J.W. analyzed dataand wrote and edited the manuscript. Conflict-of-interest disclosure: The authors declare nocompeting financial interests. | ORCID profiles: M.A.S., | ; M.O., | | - | - | | ; | R.S.K., | ; | R.B., | | ; | H.J.D., | ; | S.S., | | ; | R.C.L., | ; | N.G., | . Correspondence: Mikkael A. Sekeres, Division of Hematology,Sylvester Cancer Center, University of Miami, 1120 14th St,Miami, FL 33136; email:. References - 1.. - 2.. - 3.. - 4.. - 5.. - 6. 12 MAY 2026 • VOLUME 10, NUMBER 9 RESEARCH LETTER 3057 ### Page 5 . - 7.. - 8.. - 9.. - 10.. - 11.. - 12. African American Veterans Monument. Stories of War: Vietnam War1964-1975. Accessed 9 March 2026. - 13. Surveillance, Epidemiology, and End Results (SEER) Program.SEER*StatDatabase: Incidence-SEER9RegsPublic-Use,Nov2004Sub(1973-2002), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, ReleasedApril 2005, Based on the November 2004 Submission. Accessed 27February 2026. - 14.. - 15.. - 16.. - 17.. - 18.. - 19.. 3058 RESEARCH LETTER 12 MAY 2026 • VOLUME 10, NUMBER 9