Pan-microbial serological repertoire as a biomarker of immunotherapy response in hepatocellular carcinoma.

Abstract

Immune checkpoint inhibition (ICI) has become a first-line treatment strategy for advanced hepatocellular carcinoma (HCC). However, treatment efficacy remains varied, and there are no reliable biomarkers of response. We used phage immunoprecipitation sequencing technology to measure circulating viral and bacterial antibodies as a biomarker of ICI response. The XGBoost Cox-proportional hazard model identified 23 viral and bacterial strains associated with survival. An antigen score developed from the 23 reactive microbial peptides significantly predicted overall survival (HR: 81.7, [95% CI]: 9.88 to 10,690.96, p<0.0001) and a high antigen score was predictive of progressive disease (HR: 4.03, [95% CI]: 0.78 to 20.87, p=0.09). The time-dependent area under the curve was 0.8 at 1 year. Findings were validated in an independent cohort and confirmed the antigen score association with survival (HR: 2.38, [95% CI 0.98 to 5.76, p=0.055). Our results suggest a unique microbial reactivity profile may serve as a potential biomarker of ICI response in patients with HCC.