[Moving towards a personalized oncology: The contribution of genomic techniques and artificial intelligence in the use of circulating tumor biomarkers]. / En marche vers une oncologie personnalisée : l'apport des techniques génomiques et de l'intelligence artificielle dans l'usage des biomarqueurs tumoraux circulants.

Technological advances, in particular the development of high-throughput sequencing, have led to the emergence of a new generation of molecular biomarkers for tumors. These new tools have profoundly changed therapeutic management in oncology, with increasingly precise molecular characterization of tumors leading to increasingly personalized therapeutic targeting. Detection of circulating tumor cells and/or circulating tumor DNA in blood samples -so-called 'liquid biopsies'- can now provide a genetic snapshot of the patient's tumor through an alternative and less invasive procedure than biopsy of the tumor tissue itself. This procedure for characterizing and monitoring the disease in real time facilitates the search for possible relapses, the emergence of resistance, or emergence of a new therapeutic target. In the long term, it might also provide a means of early detection of cancer. These new approaches require the treatment of ever-increasing amounts of clinical data, notably, with the goal of calculating composite clinical-biological predictive scores. The use of artificial intelligence will be unavoidable in this domain, but it raises ethical questions and implications for the health-care system that will have to be addressed.

[Clinical use and evolution of circulating biomarkers in the era of personalized oncology: From protein markers to bioclinical scores]. / Utilisation clinique et évolution des biomarqueurs circulants à l'ère de l'oncologie personnalisée : des marqueurs protéiques aux scores clinicobiologiques.

In oncology, the identification of targets that correlate with a type of cancer has led to a profound change in the notion of "tumor markers". Technological advances, in particular the development of high-throughput sequencing, have led to the emergence of a new generation of molecular biomarkers for tumors. Despite their limited utility for screening and diagnosis, conventional tumor markers remain interesting for evaluation of prognoses, the choice and optimization of treatments, as well as for monitoring the effectiveness of those treatments. In this article, we revisit the conventional serum markers that are enjoying a 'come back' thanks to the development of high-performance scores based on biological, cytological, clinical, or radiological criteria.