This study aims to identify biomarkers for treatment response of atezolizumab plus bevacizumab (Atezo+Bev) in patients with hepatocellular carcinoma (HCC). 96 patients who received Atezo+Bev or lenvatinib as a first-line systemic therapy were enrolled as the training group after propensity score matching (PSM), and 42 patients treated with Atezo+Bev were enrolled as the validation group. 17 serum cytokines were measured by Luminex multiplex assay at the start of treatment. For further assessment of the association between cytokine levels and the tumor microenvironment (TME), immunohistochemistry (IHC) was performed on pre-treatment liver biopsy specimens. In the derivation set, multivariate analysis identified elevated IL-6 as an independent risk factor in the Atezo+Bev group (HR 5.80: p<0.01), but not in the lenvatinib group; in a subset analysis of patients with low IL-6, PFS was longer in the Atezo+Bev training group than in the lenvatinib group (p = 0.02). A validation study also showed a significantly longer prognosis in the low IL-6 group for both PFS (p = 0.0001) and OS (p = 0.03). Serum IL-6 had a positive correlation with tumor IL-6 expression (ρ = 0.56, p < 0.0001) and an inverse correlation with the CD8/CD163-positive cell count ratio (ρ = -0.4, p < 0.01). Serum IL-6 levels are thought to be involved in the suppression of tumor immunity and are useful in predicting the therapeutic effect of Atezo+Bev treatment.
Read full abstract