Abstract 5767: Pan-viral response in African Americans and European Americans with hepatocellular carcinoma

Abstract Hepatocellular carcinoma (HCC) is often diagnosed at late-stages, with African Americans (AA) being diagnosed at a later stage with worse prognosis and high mortality rates. The poor screening and diagnosis of HCC and differences between racial/ethnic groups show a disproportion in mortality rates. Examining an individual’s viral exposure history might serve as a useful biomarker of early disease that could address and, potentially, alleviate this disparity. We aim to investigate whether viral features can be used to identify HCC and individuals at high risk (HR) of HCC. Additionally, we examine the racial/ethnic differences in viral exposures and whether viral exposure history accounts for epidemiological disparity described previously. Herein we use a technique called VirScan, which conjugates Phage Immunoprecipitation sequencing (PhIP-seq) and DNA sequencing, to profile an individual’s epitope antibody repertoire upon viral exposure. We performed VirScan on serum and plasma samples from 199 AA (130 HR and 79 HCC) and 260 European Americans (EA) (210 HR and 50 HCC) participants from the NCI-UMD cohort (NCT00913757; clinicaltrials.gov). We used a machine-learning XGBoost model to classify viral features that can discriminate HCC or HR. A composite viral score was generated using the SHAP value of the most significant viral features from the model with a higher score more predictive of HCC. Secondary analyses using Wilcoxon-Rank sums tests and Cox proportional hazard analysis revealed that while race/ethnicity did not differentially predict HCC in this cohort, the viral score was significantly higher in AA compared to EA in the HR group (p = 0.027). XGBoost search identified 235 important viral features which distinguish HCC from HR. Out of the 235 viral features, 20 viruses were different between AA and EA (FDR < 0.05). We also found the viral score was positively associated with incident mortality in HCC group over the study period of 8 years (p=0.03). Thus, we demonstrate that the viral history of an individual might be useful in early detection of HCC. This approach could be used to account for the disparities we see in both incidence and mortality of HCC between AA and EA. Citation Format: Theressa Ewa, Whitney L. Do, Limin Wang, Marshonna Forgues, Xin W. Wang. Pan-viral response in African Americans and European Americans with hepatocellular carcinoma. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5767.