Abstract
INTRODUCTION: Immune check point inhibitors have recently entered the armamentarium of advanced hepatocellular carcinoma (HCC) treatment. Among them, the combination of atezolizumab plus bevacizumab has pushed it a step forward; however, a number of patients still present primary non-responses without any biomarker to predict responses to different options. Here, we aimed to identify a putative baseline biomarker to predict the response to atezolizumab-bevacizumab, by investigating whether baseline PD1+ and PD-L1+ peripheral granulocyte percentages might offer a non-invasive, cheap, and easily feasible assay. METHODS: A prospective Italian cohort of 34 patients treated by atezolizumab-bevacizumab was tested to assay the baseline percentage of peripheral granulocytes and their PD1 and PD-L1 expression. The neutrophil to lymphocyte ratio (NLR) was also considered, and all data were compared with the clinical course of patients. RESULTS: A low-baseline PD1+ peripheral granulocyte percentage turned out to predict responder patients (mean ±SD of PD1+ granulocyte percentage in responders versus non-responders: 9.9 ± 9.1 vs. 29.2 ± 17.6; student's t-test, p < 0.01). In line, patients identified by a low PD1+ granulocyte percentage displayed a longer TTP (log-rank test, p < 0.0001). A lower granulocyte percentage on total white blood cells, irrespective of PD1 or PD-L1 expression, is also associated with responses to atezolizumab-bevacizumab (log-rank test, p < 0.05). No predictive value was observed for either the PD-L1+ granulocyte percentage or NLR. CONCLUSIONS: A low-baseline PD1+ peripheral granulocyte percentage is associated with responses to atezolizumab-bevacizumab treatment in advanced HCC. These findings encourage evaluating this minimally invasive, cheap, and easy test in further independent cohorts and outlining the relevance of innate immunity in the response to immune-checkpoint inhibitors.