Immunotherapy has changed the way many cancers are treated. Instead of attacking the tumour directly, these drugs release the brakes on the patient's own immune system so that it can recognise and destroy cancer cells. For some patients the effect is remarkable and long-lasting. For many others, unfortunately, it is not and today we still have no reliable way of telling the two groups apart before treatment begins.
PRIME-CT sets out to close this gap. The project reads the full "activity profile" of a tumour, namely which genes it is switching on and off, what its genetic make-up looks like, and which immune cells have managed to get inside it. Combined with artificial intelligence, the project aims to learn the patterns that separate patients who respond from those who do not. The models are built on data from patients already treated at the European Institute of Oncology and are then tested on completely independent groups of patients from other institutions, to make sure that what the computer has learned genuinely transfers to new cases rather than simply memorising one hospital's data.
Two things make the approach distinctive. First, the team is developing methods that generate realistic artificial patient data, a way of compensating for the fact that real clinical datasets are always smaller than machine learning would ideally require. Second, the project asks whether a single test, based on RNA sequencing alone, can do the job, which would make the approach far cheaper and more practical for hospitals with limited resources.
If it succeeds, PRIME-CT will provide a tool that helps oncologists and patients decide, before starting treatment, whether immunotherapy is the right choice, and will point to the biological reasons why some tumours resist it, opening the way to new strategies for overcoming that resistance.
This project is funded by the Italian Ministry of University and Research (MUR) under the Fondo Italiano per la Scienza (FIS 3) programme.
Posted on September 30, 2026