Abandoned oncology drugs -Identifying new indications using active machine learning and a biobank of patient derived dissociated tumor cells (DTCs)

Many oncology drugs have been abandoned by pharmaceutical companies due to poor performance in phase I, II, or III clinical trials.

These drugs represent untapped potential treatments for indications outside of those initially investigated. With research and development costs for a single new drug estimated between $2–3 billion, abandoned drugs represent an opportunity to bring drugs to market for new tumor indications at a fraction of the cost (approximately $300 million), with most of the cost associated with late-stage clinical trials and the regulatory approval process (1).

Here we describe a novel approach to identifying clinically promising abandoned drugs that warrant a robust yet streamlined investigation for a new indication requiring limited investment. Marrying this approach with the ability to deploy a biomarker-based strategy for selecting targeted patient cohorts can significantly derisk the clinical trial process (2).