Introduction Immune checkpoint inhibitor (ICI-) based therapies represent the current standard
for advanced HCC. Unfortunately, prognostic biomarkers are still lacking. The serological
repertoire of circulating antibodies may reflect the individual capability of the
immune system and has the potential to serve as a biomarker for outcomes.
Material and Methods We included HCC patients treated with ICI-regimen at two institutions (Medical University
of Vienna and University Medical Center Hamburg-Eppendorf). Phage-display immunoprecipitation
sequencing (PhIP-Seq) was performed in plasma samples obtained prior to treatment
initiation to characterize the antibody epitope repertoires against 357,000 bacterial,
viral and mycotic proteins. Data analyses was performed utilising the XGBoost machine
learning pipeline. The 15 peptides with the highest SHapley Additive exPlanations
(SHAP) values to predict survival were combined to build our antigen score ([Fig. 1]).
Fig. 1 Comparison of overall survival between patients with a low as compared to a high
antigen score.
Results We included 73 patients (n=55, 75.3% male) with a mean age of 69.1±9.6 years. The
majority of patients had advanced stage HCC (BCLC C:n=56, 76.7%) with preserved or
slightly impaired liver function (CPS A/B: n=44 (60.3%)/n=29 (39.7%)). A low antigen
score was associated with a significantly better median overall survival (n=38, mOS:27.1
(95%CI:11.1-45.4) months vs. high antigen score: n=35, mOS:10.6 (95%CI:3.0-13.4) months;
p<0.001). These results were also confirmed upon multivariable Cox Regression analyses
with backward selection, where a high antigen score (aHR:2.79 (95%CI:1.36-5.72); p=0.005)
along Child Pugh score and AFP≥400 IU/mL were identified as independent predictors
of OS. The antigen score was also associated with radiological response. While the
disease control rate (DCR) was 76.3% in the ‘antigen-low’ group, it was only 42.9%
in the ‘antigen-high’ group (p=0.004).
Conclusion The antigen-score identifies patients with a favourable prognosis as well as a better
radiological response. External validation in a different ICI-cohort as well as in
a cohort of patients treated with multi-tyrosine-kinase inhibitors (TKIs) is required
to determine if the predictive value of the score is specific to ICI-treatment.
