FDA Turns to AI to Speed Up Clinical Trials
The agency aims to get drugs to patients faster and make the U.S. more competitive
- The U.S. Food and Drug Administration plans to pilot an artificial intelligence approach to accelerate clinical trials.
- The new approach uses artificial intelligence to extract study data directly from electronic health records for real-time submission.
- Drugmakers AstraZeneca and Amgen are testing real-time data reporting in clinical trials for specific cancer medicines.
The U.S. Food and Drug Administration seeks to accelerate clinical trials of new medicines by using artificial intelligence to streamline the laborious process of collecting and submitting study data.
Typically, medical centers involved in clinical trials pull study data from electronic-health records and enter them manually into a data-capture system. Then, the drug company developing the medicine reviews the data and submits them to the FDA.
The FDA this summer plans to pilot an approach that would upend this practice, which hasn’t changed much in decades. Through the program, AI would extract data directly from electronic records so they can be submitted in real-time to both the FDA and the pharmaceutical company.
If successful and adopted widely, the new approach could speed drug development and help the U.S. compete with other nations making a strong push into biotechnology, officials said.
Advances in AI and electronic-health records enable this initiative, but it is also driven by a recognition of inefficiencies in U.S. trials, said Dr. Emma Meagher, senior vice dean for clinical and translational research at Penn Medicine
“In general, we have a sense that the way we do clinical trials is dysfunctional in many ways,” she added.
After reviewing industry comments the FDA intends to launch a pilot that includes five to nine companies, said Jeremy Walsh, the FDA’s chief AI officer.
“The whole purpose here is to accelerate cures,” Walsh said.
As a prelude to the pilot, drugmakers AstraZeneca and Amgen agreed to test real-time data reporting in clinical trials of specific cancer medicines.
The University of Pennsylvania and the University of Texas MD Anderson Cancer Center are participating in the AstraZeneca study of patients with treatment-naïve mantle cell lymphoma. Site selection for the Amgen trial, focused on patients with limited-stage small cell lung carcinoma, is under way.
Startup Paradigm Health is providing AI capabilities for this proof-of-concept work.
Large language models enable unstructured as well as structured data to be extracted and quality-controlled in real-time in a way that wasn’t possible before, said Paradigm founder and Chief Executive Kent Thoelke.
In addition to enabling data to be reviewed quickly, the real-time reporting could make clinical trials more accessible to people who live far from academic medical centers, where trials have typically occurred, said Dr. Paul Burton, senior vice president and chief medical officer of Amgen.
Drugmakers have relied on medical centers with long histories of running trials and entering data into databases, he said. By enabling data to be pulled directly from electronic records, studies can more easily be conducted at any center that has these records, he said.
“This program has the potential to be transformational on so many levels in how we do clinical research,” Burton added.
Today, drugmakers don’t see study data in real-time. That can be especially difficult for biotech startups, which have limited resources.
A real-time system would help them shut failing programs down faster or fundraise after getting an early read on a drug that looks promising in a clinical trial, said Krish Ramadurai, a partner with AIX Ventures.
“Typically, the delay on that data-feedback loop makes or breaks a lot of companies,” he said.