Tufts analysis: Agentic AI tools could shorten cancer clinical development by roughly 10 weeks and save up to $5.6 million in late-stage trial operating costs. The financial benefits scale with indications—a drug with 50 active uses might see gains near $565 million. Tufts modeled a Medable monitoring agent in an undisclosed Phase 2–3 oncology program. Experts caution that non-data challenges like recruitment, consent and manufacturing still require human-led work and oversight.
Study: AI Cuts Cancer Trial Timelines by ~10 Weeks and Saves Millions in Late-Stage Costs
Illustration: Sarah Grillo/Axios. Stock: Getty Images
New research from the Tufts Center for the Study of Drug Development, shared first with Axios, finds that agentic artificial intelligence tools are beginning to deliver measurable time and cost savings in oncology clinical trials.
Why this matters: Clinical trials for cancer are often lengthy and expensive. AI-driven tools that help recruit and enroll participants, monitor outcomes, and interpret data can shorten timelines, reduce operational costs and free resources for additional studies—potentially improving the efficiency of drug development.
What the study found: Tufts modeled the impact of a clinical monitoring AI agent supplied by Medable on an unnamed oncology drug development program running Phase 2 and Phase 3 studies. The analysis estimated that AI agents could shorten clinical development by about 10 weeks and cut direct operating costs in late-stage trials by up to $5.6 million. Tufts also projected that the financial upside grows with the number of indications a drug targets—an experimental therapy with 50 active uses could have net benefits approaching $565 million.
Practical efficiencies observed: According to Tufts and Medable, deploying the agentic AI led to operational improvements such as fewer required on-site visits, faster patient enrollment, earlier data locking and more efficient monitoring workflows—changes that translate into both time savings and reduced spending.
"To our knowledge, this is the first time that [predictive] modeling based on actual use and benchmark data has been applied to quantify the net financial impact of an agentic AI solution in a drug development program," said Ken Getz, executive director of the Tufts center.
Industry outlook: Medable executives told Tufts that AI agents could become routine components of some clinical trials within three to five years, performing repetitive record-keeping and monitoring tasks much like self-driving systems automate tedious work in other industries. The tools may also help track trial population diversity and surface safety or efficacy signals earlier in development.
Limitations and caveats: AI is not a guaranteed shortcut to successful trials. Key challenges—finding eligible patients, obtaining informed consent, manufacturing and distributing therapies—remain largely human-led and may limit gains. Human verification of AI outputs will likely still be required, which could reduce some of the predicted net time savings. The study models potential impacts and uses a specific commercial monitoring agent in its scenario; real-world results will vary by program, disease area and implementation.
Research was shared first with Axios.
Help us improve.
