Rapid learning for precision oncology

Abstract

In Precision Oncology 3.0 sophisticated algorithms analyse panomic data to hypothesize the molecular pathways that drive an individual patient's tumour, and hypothesize personalized treatments, using combinations of narrowly targeted therapies At the molecular level, where Precision Oncology 3.0 operates, there are far too many combinations of driver mutations and possible treatments to be efficiently searched by current clinical trial methodologies The 'Rapid Learning Precision Oncology' paradigm considers each patient encounter as an experiment, continuously gathering and analysing all the data to inform each subsequent encounter with the same or similar patients All patient encounters can be coordinated through a 'Global Cumulative Treatment Analysis' (GCTA) methodology, which chooses treatments according to their continuously updated performance statistics The Rapid Learning approach can help to overcome some of the technical and structural barriers facing Precision Oncology 3.0, including the facilitation of the off-label uses of targeted drugs