Citations

Abstract Nearly 300,000 cancer patients are treated with fluoropyrimidine chemotherapy in the U.S. annually. Approximately one-third of patients experience severe and life-threatening toxicities, which can result in death. Deleterious genetic variants in DPYD, which encodes the rate-limiting enzyme of fluoropyrimidine catabolism (dihydropyrimidine dehydrogenase, DPD), have been correlated with severe toxicity in clinical studies. However, these variants explain only 10-30% of toxicity cases. Phenotypic measures of DPD function are attractive biomarkers of fluoropyrimidine toxicity risk with the potential to identify unknown causes of DPD deficiency. In this study, we compared two measures of DPD enzyme activity, blood plasma dihydrouracil to uracil ratio (UH2:U) and direct measurement of DPD activity in peripheral blood mononuclear cells (PBMCs) using 204 subjects with comprehensive DPYD genotype data. DPD PBMC measurements showed higher overall inter-individual variation compared to UH2:U ratios. Low correlation between UH2:U and PBMC measurements were noted for matched samples taken from the same subject. As expected, DPD metabolites (U and UH2) correlated with UH2:U ratio; however, no correlation was noted between U or UH2 levels and the PBMC DPD activity. Furthermore, when data were analyzed by genotype, no correlations were noted between UH2:U and PBMC DPD activity. We observed the expected decrease in UH2:U ratio for carriers of the four clinically tested DPYD variants and identified two novel variants associated with a significantly decreased UH2:U ratio. To further investigate the impact of rare genotypes and previously identified DPYD haplotypes on the UH2:U ratio, additional cohorts of 1994 and 500 subjects, enriched for specific genotypes, were utilized. Our results suggest that phenotyping may have limited validity as a single test when predicting DPD deficiency at an individual level and call into question the often-reported belief that PBMC DPD activity represents the "gold standard" for identifying DPD deficiency. Additional studies are needed to investigate correlations between phenotypic measures and 5-FU toxicity in patients, potentially as secondary biomarkers in conjunction with genetic tests. Citation Format: Brianna Bembenek, Kelly Bouchonville, Carlo Largiadèr, Steven M. Offer. Comparative analysis of phenotypic markers of DPD activity as biomarkers of 5-FU toxicity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3135.

Abstract Dihydropyrimidine dehydrogenase (DPD, encoded by the DPYD gene) is the rate-limiting enzyme in the catabolism of the widely used chemotherapeutic drug 5-fluorouracil (5-FU). Genetic variations in DPYD are predictive biomarkers of 5-FU toxicity. One such biomarker, rs75017182 (the causal allele linked to the “HapB3” haplotype), has been suggested to promote alternative splicing of DPYD, leading to expression of an inactive proteoform. However, carriers of rs75017182 exhibit only moderately increased risk of 5-FU toxicity and modestly decreased ex vivo DPD activity compared to carriers of another well-studied splice variant, rs3918290 (*2A). We hypothesized that, unlike *2A, the rs75017182 variant does not cause obligate alternative splicing but instead promotes partial alternative splicing. To test this, we generated dual-fluorescent minigene reporter constructs containing either the G (wildtype) or C (variant) allele. In our minigene system, alternative splicing was elevated in cells transfected with C-containing vectors via flow cytometry, though levels were lower than those observed in a parallel *2A model. To investigate contributions of other variants in LD with rs75017182, we expanded the reporter to include additional relevant genomic regions. The analyses were complemented by the study of in vivo splicing in cells engineered to harbor various combinations of variants and in lymphoblastoid lines carrying specific genotype combinations. Overall, our findings demonstrate that rs75017182-C promotes only partial alternative splicing, eliciting a modest decrease in DPD activity and explaining its comparatively weak association with severe 5-FU toxicity. Additional variants in LD have a minimal impact on splicing and 5-FU sensitivity. Citation Format: Lauryn Allyn Hahn, Ryan Schwartz, Hannah Marie Krause, Lulu Jiang, Brianna Bembenek, Kelly Bouchonville, Steven M. Offer. Partial alternative splicing in carriers of DPYD rs75017182 explains modest association with severe 5-FU toxicity to 5-FU [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3137.

Abstract Background: Pediatric and AYA sarcomas are rare, biologically heterogeneous tumors with limited noninvasive biomarkers to guide diagnosis, personalized treatment, and disease monitoring. Standard-of-care evaluation relies on invasive biopsies that carry procedural risks and imaging modalities with limited sensitivity for detecting minimal residual disease and distinguishing tumor types. Cell-free DNA (cfDNA) methylation profiling offers a minimally invasive alternative capable of capturing tumor-derived epigenomic alterations and has consistently shown strong clinical utility in adult malignancies. However, genome-wide cfDNA methylation patterns across diverse pediatric and AYA sarcoma subtypes remain poorly defined. Methods: Plasma was collected from 93 pediatric and AYA patients (pediatric <18 years; AYA 15-39 years) representing 27 Ewing sarcoma, 17 rhabdomyosarcoma, 11 liposarcoma, 10 chordoma, 8 synovial sarcoma, 7 chondrosarcoma, 6 osteosarcoma, and 6 fibrosarcoma cases, plus 19 matched healthy donors. cfDNA was isolated from 1 mL plasma, quantified by Qubit, and assessed by Bioanalyzer. Libraries were prepared using the NEBNext EM-seq v2 workflow with PCR cycles scaled to cfDNA input amount. Genome-wide sequencing (15× target depth) was followed by alignment, deduplication, and methylation calling. MultiQC was used to assess depth, duplication rate, conversion efficiency, and CpG coverage. Global methylation summaries were computed, and methylKit was used to generate per-CpG and 1-kb tiling matrices and initiate differential methylation analyses. Results: Ewing sarcoma patients had significantly higher plasma cfDNA concentrations than age- and sex-matched healthy donors (24.4 vs 4.6 ng/mL, p < 0.0001), whereas cfDNA levels for other sarcoma subtypes overlapped with the healthy cohort. Low or undetectable cfDNA concentrations by Qubit were far more common among non-Ewing sarcoma samples (23 of 66) than Ewing sarcoma (1 of 27). All samples achieved sufficient CpG coverage for genome-wide methylation modeling. Global methylation summaries revealed no significant differences in mean β-values across sarcoma subtypes or versus healthy donors. An independent analysis of the Ewing subset identified 299 significant differentially methylated regions (DMRs), supporting the detectability of disease-specific methylation changes. Conclusions: These results demonstrate the feasibility of high-quality cfDNA methylome profiling across a large and diverse pediatric and AYA sarcoma cohort, including in plasma samples with cfDNA concentrations below the limit of detection. While global methylation levels appear broadly similar across subtypes, proof-of-concept DMR analysis in Ewing sarcoma suggests that subtype-specific epigenomic signatures could enable future noninvasive diagnostic tools for pediatric and AYA sarcomas. Citation Format: Lilli J. Greiner, Roman O. Kowalchuk, Jeff Szymanski, Pradeep S. Chauhan, Adam Amundson, Peter J. Schoettler, Wendy Allen-Rhoades, Stephanie F. Polites, Peter S. Rose, Kelly Bailey, Linda McAllister-Lucas, Anita Mahajan, Nadia Laack, Aadel Chaudhuri. Defining genome-wide cfDNA methylation landscapes in pediatric and adolescent and young adult (AYA) bone and soft-tissue sarcomas [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7885.

Abstract 5-Fluorouracil (5-FU) is a widely prescribed chemotherapy for colorectal, breast, gastric, and head and neck cancers. The drug inhibits thymidylate synthase, blocking DNA synthesis and causing cytotoxic stress that halts tumor growth. However, patients carrying deleterious variants in DPYD, the gene encoding the rate-limiting enzyme in 5-FU metabolism (dihydropyrimidine dehydrogenase, DPD), can experience fatal toxicity due to impaired drug clearance and accumulation of 5-FU metabolites. Current pharmacogenetic screening guidelines in the U.S. recommend testing for few of the >2,000 nonsynonymous variants that have been reported for DPYD, leaving patients with rare, uncharacterized mutations at risk. To interpret expanded testing, however, a means to classify DPYD variants of unknown significance is needed. To address this, we developed MPACT-DPD, a random-forest-based classifier that accurately predicts the functional impact of DPYD missense variants. Our model was trained on in-vitro activity of 156 variants and leveraged a feature set of biochemical, evolutionary, and AlphaFold3-derived structural features. We optimized hyperparameters using ten-fold stratified cross-validation and evaluated model performance with Matthews correlation coefficient (MCC) to account for moderate class imbalance (7:3 neutral to deleterious). It achieved exceptional performance, with a Matthews correlation coeffient (MCC) of 0.906 and an accuracy of 95.1% on an independent validation set (n=43). Furthermore, a SHAP (SHapley Additive exPlanations)-based interpretability analysis revealed cofactor proximity and residue conservation as the key drivers of predictions. MPACT-DPD showed superior performance at variant classification against other tools, including a DPYD gene-specific variant classifier, and has the potential to expand pre-treatment genetic screening to improve the safety of personalized 5-FU-based chemotherapy. Citation Format: Lulu Jiang, Ryan Jonathan Swartz, Lauryn Allyn Hahn, Brianna Bembenek, Hannah Marie Krause, Kelly Bouchonville, Steven M. Offer. MPACT-DPD: An interpretable machine learning classifier for predicting the functional impact of DPYD missense variants [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2712.

Abstract Introduction: Detectable circulating tumor DNA (ctDNA) is associated with inferior outcomes in Ewing sarcoma (ES); however, ctDNA clears in nearly all patients within 2 of the 6 chemotherapy cycles before radiotherapy (RT) or surgery. Given a 3-year EFS of 37% on AEWS 1221 (metastatic) and 5-year EFS of 78% on AEWS 1031 (localized), current ctDNA testing fails to predict most relapses. Extracellular vesicles (EVs) are microscopic particles studied as a biomarker in multiple cancer types. We hope to improve the sensitivity of ctDNA in ES and provide the first application of EVs in ES. Methods: cfDNA was isolated from 1mL plasma, quantified by Qubit, and sequenced as NEBNext Enzymatic Methyl-seq v2 libraries. We performed a systematic unsupervised clustering analysis of 27 ES plasma cfDNA profiles using 4 differentially methylated regions (DMR) subsets (63 strong DMRs, top 30, top 15, and 10 promoter-proximal loci), 3 clustering algorithms (k-means, Ward hierarchical, PAM), and k=2-5 clusters, testing associations with tumor burden and progression. EV analysis involved antibody detection against ES surface markers, including LINGO1, ENPP1, CDH11, and CD99. Concentrations of EVs were determined by standardized and calibrated high-resolution flow cytometry then analyzed with custom-built software to automatically determine and apply gates. Samples were de-identified, randomly coded, and run in triplicates. The senior research technologist responsible for EV analysis was blinded to patient information and clinical data. Results: EV analysis included 36 pre-RT samples from 16 metastatic and 20 localized ES patients from a prospectively maintained biobank, with 27 pre-RT samples in the cfDNA analysis and 19 healthy donor (HD) samples. 35 of 36 patients had chemotherapy before sample collection. cfDNA concentrations were markedly elevated in ES patients compared with HD (24.4 vs 4.6 ng/mL, p<0.0001), and 74% of ES samples exceeded the HD 95th-percentile detectability threshold (7.0 ng/mL). Across clustering configurations, ES cfDNA profiles consistently exhibited reproducible methylation substructure. One configuration (PAM clustering, k = 5, 63 strong DMRs) showed a modest but permutation-supported association with primary tumor size (nominal p ≈ 0.04; empirical p ≈ 0.03-0.04). ES samples had increased EV concentration compared to HD (p<0.001), with AUC 0.988 for distinguishing between ES and HD ENPP1 (p<0.0001). The highest ENPP1 EV values were in patients with large primary tumors (>8 cm vs. ≤ 8cm) and metastatic disease (p=0.066). Conclusions: cfDNA- and EV-based liquid biopsy provide encouraging signals for high ES detectability pre-RT. We also show the first study of EVs as a novel biomarker in ES, which may provide complementary information to ctDNA. Multimodal liquid biopsy in ES may identify candidate patients for systemic or local therapy escalation and facilitate personalized care. Citation Format: Roman O. Kowalchuk, Lilli J. Greiner, Yohan Kim, Rubia Noori, Jeffrey Szymanski, William S. Harmsen, Pradeep S. Chauhan, Adam C. Amundson, Andrea M. McMahon, Peter J. Schoettler, Wendy A. Allen-Rhoades, Stephanie F. Polites, Peter S. Rose, Kelly M. Bailey, Linda M. McAllister-Lucas, Anita Mahajan, Fabrice Lucien-Matteoni, Nadia N. Laack, Aadel A. Chaudhuri. Multimodal liquid biopsy of Ewing sarcoma reveals high sensitivity prior to radiotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1149.

Abstract Introduction: Ewing sarcoma (EwS) is a FET::ETS family member-driven primary bone cancer demonstrating vast heterogeneity. Between a patient’s primary diagnosis and disease progression, tumor cell state adaptation, microenvironment changes, and the landscape of evolving therapeutic vulnerabilities remain poorly understood. To address these gaps, the Sean Karl Cohort was established in 2025 to conduct the largest single-cell transcriptomic analysis of retrospective paired tumor samples from patients with EwS. Here, we present data from the five analytic teams, and through shared SOPs now expand this cohort to Europe. Methods: Common sample processing SOPs were established to isolate cells from FFPE material from paired patient EwS samples. Single-cell (sc) RNAseq was generated using the GEM-X Flex Gene Expression protocol (10x Genomics). Alex’s Lemonade Stand Foundation (ALSF) Data Lab established a common data processing and integration pipeline with an EwS-specific cell annotation workflow to create a harmonized dataset for downstream analyses. Analytic teams were created to define tumor cell subpopulations and their therapeutic vulnerabilities and to characterize the tumor microenvironment. Data: We demonstrate the feasibility of generating high-quality scRNAseq profiles from retrospective FFPE-preserved EwS tumors. To date, 116 samples have yielded ∼800,000 cells. Integrated analysis reveals reproducible EwS tumor cell states shared across patients with regulatory network analyses nominating candidate therapeutic vulnerabilities. The scale of this international cohort enables saturation analysis for rare cell populations. Paired and longitudinal samples further allow correlation of emergent cell states with therapy resistance and metastatic progression. For example, immune-focused analyses show increases in T cell and macrophage populations in post-therapy samples. A novel, standardized EwS-specific cell annotation workflow has been developed to harmonize analyses and findings. These data will be openly shared with the community through the ALSF Single-cell Pediatric Cancer Atlas Portal. Conclusion: Large-scale international collaborative sample sharing, standardized processing pipelines, and harmonized EwS-specific annotations now enable deep single-cell analyses to characterize tumor heterogeneity in this rare pediatric and adolescent cancer. An in vivo expansion of the Sean Karl Cohort is underway to assess conservation of EwS tumor cell states and regulatory vulnerabilities identified in human tumors pre- and post-therapy in preclinical models, further supporting translation of these findings toward therapeutic strategies. Citation Format: Abbe Pannucci, Elina Mukherjee, Jessica D. Daley, Shireen Sita Ganapathi, Elissa Boguslawski, Lea Surrey, Lauren Gutstein, Patrick Azar, Emily Stockfisch, Azfar Neyaz, Ivy John, Jennifer Picarsic, Yutaro Tanaka, Byron Butaney, Riaz Gillani, Brian D. Crompton, Katherine A. Janeway, Jaclyln Taroni, Jessica Davis, Damon Reed, Adam Shlien, Theodore W. Laetsch, Rajen Mody, Clémence Henon, Thomas G. Grünewald, Elizabeth R. Lawlor, Filemon Dela Cruz, Patrick J. Grohar, Jovana Pavisic, Ally Hawkins, Anthony R. Cillo, Kelly M. Bailey. The Sean Karl Cohort: An international single-cell RNAseq study of paired patient Ewing sarcoma specimens [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 643.

Abstract Background Photon‐counting‐detector (PCD) CT systems offer ultra‐high spatial resolution, yet the visual spatial resolution on clinical images often constrained by large pixel size, yielding resolutions below system capabilities. While reducing pixel size and using sharp kernels enhance visual spatial resolution, it increases noise, compromising image quality. Purpose To investigate the combined effects of pixel size and reconstruction kernels on visual spatial resolution using phantom and clinical images and to develop a dedicated high‐resolution deep convolutional neural network (HR‐CNN) to better utilize the intrinsic high spatial resolution of PCD‐CT in clinical imaging. Methods The relationship between spatial resolution, reconstruction kernel, and pixel size was investigated to identify strategies for utilizing the full spatial resolution potential of PCD‐CT. To overcome the increased noise associated with high‐resolution settings, a dedicated HR‐CNN was developed to push the limit of spatial resolution in routine PCD‐CT exams. The HR‐CNN was trained using patient exams acquired with ultra‐high‐resolution (UHR) mode and reconstructed with a 150‐mm field of view (FOV), matrix size of 1024×1024 (0.15‐mm pixel size) and sharpest quantitative kernel (Qr89). The impact of FOV, kernel, and denoising on spatial resolution was studied using bar‐pattern phantoms and a pilot clinical evaluation including 5 patients with interstitial lung diseases. Two thoracic radiologists evaluated 4 different FOV/reconstruction conditions: (1) FOV‐410/Qr56‐Iterative reconstruction (IR), (2) FOV‐410/Qr89‐IR, (3) FOV‐150/Qr89‐IR, and (4) FOV‐150/Qr89‐HR‐CNN in terms of overall image quality, noise, visual spatial resolution, and overall preference. Results With a FOV of 410 mm, the Qr89 sharp kernel displayed bar‐patterns up to 14 lp/cm, not much higher than the routine lung kernel Qr56. When the FOV was reduced to 150 mm, Qr89‐IR allowed for the visualization of line pair patterns ranging from 18 to 20 lp/cm, with 20 lp/cm being moderately discernible. The application of Qr89‐HR‐CNN yielded further improvement, enabling the display of line pair patterns as high as 20–22 lp/cm. In patient cases, both radiologists consistently ranked the FOV‐150 images processed with HR‐CNN as superior across metrics including overall image quality, noise reduction, visual spatial resolution, and overall preference. The HR‐CNN reduced the noise in patients’ images by 93.0 ± 0.6% and 44.9 ± 5.3% in comparison with the original FBP and IR images, respectively. Conclusions The spatial resolution of PCD‐CT is not maximized in routine practice due to the large FOV and high noise levels at sharp kernels. The proposed HR‐CNN denoising method, along with small pixel size, may allow the high spatial resolution toward the system limit to be implemented in practice, which is beneficial in the diagnosis of many diseases, including interstitial lung disease.

Objective: This study aimed to examine whether parent–child interaction therapy (PCIT), compared with other therapy modalities, was associated with differences in substance use and psychiatric outcomes in a cohort of children with attention-deficit/hyperactivity disorder (ADHD). Methods: This retrospective cohort study was conducted using medical record data from the Rochester Epidemiology Project (REP). The sample included 643 children aged ≤7 years who were diagnosed with ADHD and received either stimulant or nonstimulant medication, along with psychosocial therapy. Participants were categorized into two groups: those who documented exposure to PCIT versus other therapy modalities such as behavioral management, play therapy, cognitive behavioral therapy, parent management training, occupational therapy, speech therapy, parenting class, applied behavior analysis, and social skills training. Conditional logistic regression models were used to examine differences in baseline characteristics and outcomes between the two treatment groups. Results: Individuals in the PCIT group were more likely to have private insurance (66% vs. 57%; p = 0.016) and to fall into higher socioeconomic status quartiles ( p = 0.024) at baseline. PCIT recipients also had a higher prevalence of emotional and behavioral disturbances at baseline (40% vs. 25%; p < 0.001). The median age of the participants at follow-up was 19.6 years, indicating a median follow-up period of approximately 13 years from the initiation of treatment. At follow-up, the PCIT group demonstrated higher rates of problematic alcohol (24% vs. 17%; HR = 1.50, 95% CI: [1.02, 2.20]; p = 0.038) and cannabis use (25% vs. 18%; HR = 1.48, 95% CI: [1.03, 2.15]; p = 0.037). However, after controlling for covariates in the adjusted model, no statistically significant differences emerged between the groups across any substance use outcome. No significant differences were observed in other psychiatric diagnoses or mental health service utilization. Conclusions: Findings indicate that individuals who documented exposure to PCIT demonstrated elevated rates of alcohol and cannabis use in young adulthood, but had no other differences in psychiatric outcomes. While this aligns with prior literature reporting increased substance use risk in ADHD populations, our study cannot confirm this association relative to non-ADHD controls because such a comparison was not included. These associations likely reflect underlying differences in clinical severity and clinical comorbidities at baseline, rather than a direct effect of PCIT. Further investigation is warranted to determine the treatment effects based on baseline risk factors.

OBJECTIVE The aim of this study was to elucidate the long-term morbidity associated with untreated cerebral cavernous malformation (CCM). METHODS From a prospectively maintained database of patients with CCM, we assessed functional outcome as measured using the modified Rankin Scale (mRS) score at baseline and the last untreated follow-up evaluation. Disability was defined as an mRS score ≥ 3. Kaplan-Meier curves were used to assess the time to disability, and Cox proportional hazards models were used to determine risk factors for disability. Furthermore, the predictive performance of a previously published nomogram for morbidity in CCM was assessed. RESULTS The cohort consisted of 332 patients (58.4% female, mean age at diagnosis 44.5 years), with 19.8% having the familial form and 28.0% having a CCM located in the brainstem. Of 48 patients who presented with symptomatic hemorrhage (SH) and who did not undergo surgery or have a prospective SH, 27.1% had an mRS score ≥ 3 at diagnosis. By 1 year and 5 years, this improved to 6.2% and 4.7%, respectively. Over a mean follow-up duration of 6.6 years, 103 of 332 patients (31.0%) had one prospective SH and 48 (14.5%) had multiple prospective SHs. Patients with brainstem-located CCM had an 18.8% 5-year and 35.4% 10-year risk of disability compared with 4.1% and 7.5%, respectively, in patients with a nonbrainstem location (p < 0.0001). Multivariate analysis showed that a brainstem location, history of self-reported psychiatric disorder, and ≥ 2 SHs were predictive of disability at the last follow-up. External validation of a published nomogram demonstrated high specificity but limited sensitivity for predicting disability, with an area under the receiver operating curve of 0.687 for mRS scores ≥ 2 and 0.783 for mRS scores ≥ 3. CONCLUSIONS Most patients improved after a first hemorrhage, with the most improvement occurring in year 1. Disability increased with each SH. Disability was associated with 2 or more SHs, a brainstem location, and a self-reported psychiatric diagnosis. Ten-year disability risk with a nonbrainstem CCM location was < 8%.

Objective: The identification of reliable biomarkers in trauma patients (pts) remains a clinical challenge. Nucleosomes, which are DNA wrapped around histone protein cores, have been used as reliable markers for quantification in sepsis, acute respiratory distress syndrome, and lymphoma. We hypothesized that: 1) levels of circulating nucleosomes quantified after traumatic injury, specifically histone H3.1 and its citrullinated R8 posttranslation modification (H3R8 Cit), would be increased after trauma and 2) that levels of circulating nucleosomes would be further, increased in pts who went on to develop venous thromboembolism (VTE), as these nucleosomes may be clot-enhancing mediators. Methods: Trauma pts presenting to a Level I trauma center were evaluated for inclusion in a prospective case-cohort study, and citrated blood samples were collected within 12 hours of injury. Pts were followed for up to 90 days and VTE occurrence was confirmed via autopsy or imaging. Pts who developed incident, symptomatic VTE, and those who did not develop VTE were selected at a 1:3 ratio. The effect of trauma was assessed by comparison with healthy volunteer samples. Circulating nucleosomes were quantified using Nu.Q H3.1 and Nu.Q H3R8 Cit assays. Data are presented as median [interquartile range] or n (%), with Wilcoxon Rank-Sum or chi-squared test performed between trauma pts who developed symptomatic VTE versus those who did not, with a P value < 0.05 as statistically significant. Results: A total of 639 trauma pts were analyzed [51 years (33, 65), 71.0% male, 94.1% blunt, injury severity scores 17 (9, 27)]. Both H3.1 [359.7 (58.4, 1769.3) vs. 21.5 (15.3, 26.6) ng/mL, P < 0.001] and H3R8 Cit levels [209.4 (45.4, 665.5) vs. 22.3 (16.2, 28.9) ng/mL, P < 0.001] were greater in trauma pts than in 10 healthy volunteers. The 160 pts with VTE were then compared with 479 without VTE. VTE pts had a median time to VTE of 8.5 days, with 91 developing symptomatic deep venous thrombosis, 48 with pulmonary embolism, and 21 with both deep venous thrombosis and pulmonary embolism. No significant differences were found in age, sex, or mechanism between pts with or without VTE. However, VTE pts had higher injury severity scores [24 (14, 34), 14 (9, 24), P < 0.001] and body mass index [29.1 (24.3, 34.2), 27.6 (24.2, 32.0) kg/m 2 , P = 0.037]. A greater percentage of VTE pts underwent surgery requiring general anesthesia (53.1%, 34.0%, P < 0.001) and received blood transfusions within 24 hours of injury (51.9%, 24.0%, P < 0.001). H3.1 levels were significantly greater in trauma pts who developed VTE compared with those who did not [828.4 (112.0, 3186.5), vs. 276.7 (50.3, 1237.0) ng/mL, P < 0.001], as were H3R8 Cit levels [350.4 (77.4, 729.5), vs. 189.2 (40.8, 632.6) ng/mL, P = 0.005]. No differences in levels of H3.1 and H3R8 Cit were found between early (diagnosed day 0–7 after injury) and late (diagnosed > day 7) VTE. Conclusions: Levels of the H3.1 and H3R8 Cit nucleosomes are elevated early after traumatic injury, especially in those who developed VTE. These findings underscore the importance of understanding the pathophysiology of nucleosomes in inducing VTE and their role as biomarkers