An AI algorithm selected drug formulations during an active clinical trial and hit the study's preset pharmacokinetic target within three dosing periods, Quotient Sciences reported Wednesday. The interim results suggest machine learning could cut the time and cost of developing modified-release drugs, a process that conventionally takes months or years of iterative lab work and human testing.
How the algorithm learned during the trial
The algorithm entered the study trained only on in vitro drug release data from earlier laboratory screening. After each dosing period, researchers retrained it on tablet dissolution results and pharmacokinetic data from healthy participants. The model then selected the next formulation composition and dose. "The interim data show that the algorithm learned that relationship quickly and accurately, reaching our preset target within three dosing periods," said Andrew Lewis, Ph.D., Chief Scientific Officer at Quotient Sciences.
Quotient Sciences set the boundaries the algorithm worked within, including a dose cap on the first prototype. A safety committee approved every composition before manufacturing and dosing, maintaining human oversight throughout the trial. The study used a generic drug with an established safety record, chosen to test the algorithm rather than as a development candidate.
From lab screening to human testing
The clinical work builds on earlier laboratory results where the same algorithm mapped a formulation design space after screening one-third fewer formulations than conventional methods. That finding led to the hypothesis now under investigation: a model that learns the relationship between tablet composition and in vitro release could also learn how composition affects pharmacokinetics in humans.
"Predicting how a modified-release tablet will behave in humans is difficult," Lewis said. The algorithm's rapid learning in the clinic supports the idea that less clinical testing may be needed to develop modified-release formulations for other molecules, though dosing in the current trial continues and full data will be reported toward the end of this year.
Where it fits in the development toolkit
The AI-enhanced approach extends Quotient Sciences' existing Translational Pharmaceutics platform, which integrates drug product development, manufacturing, and clinical testing. The solution supports model-informed drug development by generating a digital twin that links formulation composition to in vitro performance and human pharmacokinetics. Quotient Sciences is encouraging drug developers with early-stage programs to contact its scientific team about applying the technology.
Why this matters for healthcare and research professionals
For scientists and clinicians working in drug development, the interim data point to a practical reduction in the number of clinical dosing periods required to reach a target pharmacokinetic profile. Fewer iterations mean shorter timelines and lower costs per program. The human-in-the-loop design also provides a template for deploying machine learning in regulated clinical environments without removing expert judgment from safety decisions.
Your membership also unlocks:
