Recently, we were intrigued by a recent study by Montironi et al,1 in which they discovered that an inflamed subclass in hepatocellular carcinoma (HCC) patients is associated with a response to immunotherapy. The authors used a 20-gene signature to distinguish these patients and further found different immune infiltration between inflamed and non-inflamed class at the bulk level. We commend the authors for undertaking this study, which holds significant clinical implications. We also observed that Li et al 2 have validated the predictive value of inflamed class in two additional RNA-seq datasets from patients who received anti-PD1 therapy. However, the use of combination immunotherapy, which includes dual immune checkpoint inhibitors or is combined with anti-VEGF agents, has become a growing trend in HCC.3–6 Here, we first performed unsupervised clustering on the RNA-seq data from 289 patients enrolled in the GO30140 Ph1b and IMbrave150 PhIII trials who received a combination of anti-PD-L1 and anti-VEGF therapy7 (figure 1). The results indicated that the subclass (C1), which exhibited high expression of genes associated with B/plasma cells and fibroblasts, had a higher inflamed-class score and better therapeutic efficacy (figure 1B–D). The performance of inflamed-class gene signature in predicting combination therapy response showed anarea under …