FDA Utilizes AI/ML for Drug Therapy Decision-Making

Food and Drug Administration

WASHINGTON, D.C. — The Food and Drug Administration (FDA) recently published a groundbreaking analysis on the use of artificial intelligence and machine learning (AI/ML) in drug therapy decisions. This marks the first instance where the FDA’s Center for Drug Evaluation and Research (CDER) employed AI/ML for a regulatory decision, leveraging these technologies to identify patient populations likely to benefit from specific drug therapies.

The FDA’s increased focus on AI/ML stems from its potential to accelerate the drug development process. These technologies can predict clinical outcomes based on patients’ baseline characteristics, such as demographic information, clinical data, and genetic markers. By identifying patients more likely to have adverse outcomes or benefit from treatments, AI/ML can streamline clinical trials and improve drug efficacy demonstrations.

Background of the AI/ML Initiative

On November 8, 2022, the FDA issued an Emergency Use Authorization (EUA) for anakinra (Kineret) to treat COVID-19 in hospitalized adults with pneumonia requiring supplemental oxygen. The approval was based on the SAVEMORE trial, a randomized, double-blind, placebo-controlled study involving high-risk patients. Anakinra is the first interleukin-1 inhibitor authorized to treat COVID-19.

A key challenge during the review process was the lack of an approved commercial test for soluble urokinase plasminogen activator receptor (suPAR) in the U.S. SuPAR levels help identify patients most likely to benefit from anakinra. To address this, the CDER team turned to AI/ML to develop a scoring rule that predicted high suPAR levels based on baseline characteristics.

Developing and Validating the Scoring Rule

The FDA team used data from the SAVEMORE trial to develop the scoring rule, applying two AI/ML algorithms: elastic net regression and an artificial neural network. These models independently selected features and cut-off values for predicting suPAR levels. The final scoring rule, validated using data from the SAVE trial, identified patients likely to have elevated suPAR levels.

The scoring rule was tested against 30 baseline variables from the SAVEMORE trial, ultimately narrowing it down to eight criteria. Patients meeting at least three of these criteria were considered likely to have high suPAR levels. This innovative use of AI/ML enabled the FDA to identify patients for anakinra treatment even without a commercial suPAR test.

Findings and Implications

The CDER team found that the scoring rule was effective in predicting suPAR levels, with a low false-positive rate. Exploratory analyses showed that patients identified by the scoring rule who received anakinra had better clinical outcomes compared to those who received a placebo.

The success of this initiative highlights the FDA’s commitment to integrating advanced technologies into regulatory processes. By harnessing the predictive power of AI/ML, the FDA aims to enhance drug development efficiency and ensure timely access to effective treatments.

As AI/ML continue to evolve, their role in regulatory decision-making is expected to grow, offering new avenues for improving public health outcomes. The FDA remains dedicated to exploring these technologies to better serve patients and healthcare providers.

For the latest news on everything happening in Chester County and the surrounding area, be sure to follow MyChesCo on Google News and MSN.