What comes after the DDT pathway?

The drug development tool pathway was added to the FDA act, however, key provisions of the 21st Century cure act are set to expire in September 2026 including what looks like funding for this program. In the process of renewing the program, it could provide a good opportunity to rethink the DDT pathway.

Recently released guidelines for artificial intelligence use in drug development could provide some clues for what we could expect. Specifically, the guidance document outlines a “risk-based credibility assessment framework” for evaluating the risk posed by a specific model along two axes: 1) decision consequences and 2) model influence. If we use this framework to evaluate COA risk, a COA used as a primary endpoint in a phase 3 trial would be deemed high risk as it’s the primary way to determine drug efficacy (influence) and drug approval has significant impact (decision consequences). As we transition to secondary and then to exploratory risk drops dramatically due to the COA output having lower consequences and influence on decision making.

The current DDT pathway takes on average 6 years for a COA to be qualified as an experimental endpoint. The time and stringency is in no way commensurate with the risk posed by an experimental endpoint (e.g. low decision consequences and low influence). What if the pathway was broken into 3 steps: qualification for exploratory, to secondary and finally primary. Each step would have a different stringency that was commensurate with the level of risk.

It will be interesting to see what comes next, maybe they’ll copy key language from the AI guidance document for a revised drug development tool pathway and amend the FDA act to incorporate the updates.