AbbVie symposium — AI demystified for oncology clinicians
Presentation delivered at a symposium organised by AbbVie, bringing together a group of clinicians specialising in oncology. The objective: demystify artificial intelligence for practitioners confronting clinical research challenges daily, starting from their real-world practice.
Pedagogical presentation of what AI is beyond media hype: foundational principles (statistical learning, neural networks, language models), structural limitations, and requirements for reliable deployment in clinical settings. Emphasis on the distinction between correlation and causation, between benchmark performance and real clinical utility.
Overview of concrete AI applications across the oncology clinical research pathway: imaging-assisted diagnosis, patient prognostic stratification, treatment arm optimisation, and computational biomarker identification. Each use case is presented with its validity conditions and limitations.
Identification of haemato-oncology specificities compared to solid tumour oncology in AI applications: smaller cohorts, complex sequential therapeutic trajectories, heterogeneity of haematological malignancies. Presentation of the convergence between traditional clinical research and predictive modelling — how digital twins can complement randomised trials without replacing them.