The search for biomarkers in systemic treatment of advanced hepatocellular carcinoma.

The combination of immune checkpoint inhibitors with anti-angiogenic agents has revolutionized the treatment landscape of advanced hepatocellular carcinoma (HCC). As more therapies become available, each effective in a, yet undefined, subgroup of patients, managing advanced HCC well has become increasingly complex. With the fast pace at which the landscape evolves and the lack of data on head to head comparisons of novel regimens, an evidence-based treatment flow chart for systemic therapies in advanced HCC is currently still missing. In a fragile patient population, where the imminent risk of liver decompensation may compromise further treatment possibilities, selecting the correct treatment upfront is of utmost important. Unfortunately, significant advancements in the search for predictive biomarkers of response to systemic therapies have been hampered by the scarcity of available tumour tissue for translation research in HCC, where pre-treatment tumour biopsies, though recommended, are not mandated by current clinical guidelines. Furthermore, biomarkers successful in other solid tumour types fail in HCC, likely because they do not accurately represent the underlying tumour biology of HCC and the tumour-microenvironment that it thrives in. This reiterates the need for the development of rational, biology-based biomarker strategies that reflect the biological mechanisms that drive response or resistance to therapy. 

The two main aims of this thesis were to i) provide a clinical decision-making algorithm for advanced HCC and ii) identify biomarkers predictive of response to systemic therapies.. 

Firstly, we provide a critical appraisal of clinical recommendations for the management of advanced HCC,posed by European and American scientific societies, supplemented by an objective evaluation of clinical benefit and safety of each treatment regimen using the ESMO-MCBS score and a network meta-analysis. Integrating our findings, we propose a therapeutic flow-chart designed to aid clinicians in their decision-making process (Chapter 3).

Next, we performed in-depth characterisation of both tumour biopsies and peripheral blood samples of aHCC patients treated with systemic therapies using single-cell immune profiling technologies (Chapter 4). This resulted in one of the largest, homogeneous single-cell cohorts that allows correlation of single-cell readouts with subsequent response to therapy in advanced HCC. Focussing on HCC patients treated with immune checkpoint inhibitors, we uncover a novel paradigm, identifying clonally-expanded, PD-1 negative CD8 effector T-cells as the main anti-tumoural effector cells, characterized by a high degree of T-cell receptor sharing with peripheral blood and present in the tumour prior to therapy. Additionally, pro-inflammatory, PD-L1 expressing CXCL10+ macrophages are positioned as essential gatekeepers in the tumour-microenvironment, responsible for effective T-cell recruitment through their release of CXCL9/10/11. 

Finally, we leveraged the single-cell resolution in conventional RNA-sequencing data by generating gene signatures that recapitulate cell types present in the tumour-microenvironment of advanced HCC (Chapter 5). Applying these gene signatures in RNA-sequencing data of pre-treatment tissue biopsies of 300 advanced HCC patients treated with atezolizumab plus bevacizumab, the current standard of care in first line treatment, we identified two distinct mechanisms of response to the combination: immune-driven responders characterized by the presence of CD8 effector T-cells and pro-inflammatory macrophages versus angiogenesis-related responders that represent immune-deserted tumours with increased VEGF-Aexpression. Integrating these findings, we generated and validated a prediction model that identifies responders to atezolizumab plus bevacizumab with an accuracy of 82-98%, outperforming previously published signatures of response to atezolizumab plus bevacizumab.