Predictive genomic biomarkers for Atezolizumab plus Bevacizumab combination immunotherapy response in liver cancer: insights from the IMbrave150 trial

초록

Background and aims: Combination immunotherapy, exemplified by atezolizumab plus bevacizumab, have become the established standard of care for individuals with inoperable hepatocellular carcinoma (HCC). Despite this advancement, the absence of clear predictive biomarkers and an understanding of the mechanisms governing response and resistance to these combined therapies pose challenges. Our aim is to evaluate whether previously defined immune signature scores (ISS) can identify HCC patients likely to derive enhanced benefit from the combination immunotherapy. Method: We utilized the previously developed ISS predictor from analysis of TCGA pan-cancer data and applied it to gene expression data from the IMbrave150 trial. This phase 3 study compared atezolizumab plus bevacizumab versus sorafenib for treatmentnaïve advanced HCC. Using an ISS cutoff of 0.5, patients were stratified into potential immunotherapy responders (high ISS) and non-responders (low ISS). The significance of this ISS-based stratification for predicting clinical outcomes was evaluated through Kaplan-Meier plots, log-rank tests, Chi-square tests, and Cox regression analyses. Multiple statistical approaches assessed the association between ISS groups and progression-free survival in both the immunotherapy combination arm as well as the sorafenib arm. Results: In the IMBrave150 trial cohort, data on gene expression and outcomes were accessible for 247 patients who received atezolizumab plus bevacizumab, and 48 patients who received sorafenib. The high ISS demonstrated a significant association with enhanced PFS in the atezolizumab plus bevacizumab group (p = 0.003), while no such association was observed in the sorafenib group (p = 0.6). Similarly, in the high ISS subgroup, the combination therapy exhibited superior OS (p.0.001) and PFS (p = 0.02) compared to sorafenib, but no discernible benefit was noted in the low ISS subgroup (both p > 0.7). The OS hazard ratio for the combination versus sorafenib was 0.26 (95% CI 0.14–0.51, p < 0.001) among high ISS patients, while it was not statistically significant in low ISS (HR 0.88, p = 0.68). Interaction testing confirmed a notable interaction between ISS subtype and treatment benefit, particularly for OS (p = 0.01). In the combination arm, the objective response rate was 45.6% in the high ISS group versus 22.8% in the low ISS group. Conclusion: Our study suggests that ISS may serve as a promising predictive biomarker for enhanced therapeutic outcomes in patients undergoing combination immunotherapy for HCC. The identification of such markers is crucial for refining patient stratification and personalized treatment approaches, thereby advancing the effectiveness of current standard-of-care regimens. Further validation studies are warranted to solidify the clinical utility of ISS and its integration into the decision-making process for HCC immunotherapy.