Hepatocellular carcinoma, the most common form of primary liver cancer, is the 3rd leading cause of cancer-related death worldwide. Today, the reference treatment is the combination of an anti-angiogenic agent (bevacizumab) with an immune checkpoint inhibitor (atezolizumab) for patients with advanced disease. However, efficacy differs from patient to patient. Identifying non-responders to this treatment is therefore essential to adapt the therapeutic strategy.
The team led by Prof. Julien Calderaro (Hôpital Henri Mondor AP-HP, Inserm, Université Paris Est Créteil) has developed an innovative artificial intelligence technique based on the analysis of digitized histological slides. This approach makes it possible to predict response to treatment in patients with advanced hepatocellular carcinoma. These results were published in Lancet Oncology in November 2023.
This retrospective multicenter study, involving CARPEM members: the CHIC platform, N. Loménie’s team (team 22) and J. Zucman-Rossi’s team (team 2), validated the technique on a cohort of over 800 patients from Europe, the United States and Asia. The AI model thus developed is positioned as a predictive biomarker of response to treatment, offering a potential application in routine clinical practice for HCC. Used as a therapeutic decision support tool, it has the potential to significantly improve the management of HCC patients. This work paves the way for potential applications in other types of cancer.
Coupled with spatial biology techniques, this approach could contribute to a better understanding of the biological mechanisms involved in treatment response.