RADIANT: The application of machine learning to imaging and biological features from a clinical trial to predict outcomes to chemoradiation in locally advanced rectal cancer

Artificial intelligence approaches have the potential to affect several facets of cancer therapy and represent a gateway to the next frontier of precision cancer therapy, reconciling extraordinary amounts and different types of data into actionable decision aids.

Anonymized patient-level data from clinical research are increasingly recognised as a fundamental and valuable resource. Post-hoc analysis of trials can take a long time, but treatment concepts might have changed.

The proposal is a demonstrator pathway making use of data collected as part of a CRUK funded trial, fulfilling CRUK’s objective to learn as much as possible from every patient on a clinical trial.

In this exemplar study we will demonstrate the feasibility of combining a clinical trial dataset, including clinical covariates, treatment toxicity and efficacy outcomes (ARISTOTLE ASCO 2020) with the associated histopathological, imaging and biological data (from the S:CORT consortium) that will provide the basis for future hypothesis driven and hypothesis free analyses.