• Development and Commercialization Pathways of AI Medical Devices in the United States: Implications for Safety and Regulatory Oversight

    Branden Lee, B.S., Shivam Patel, B.S., Crystal Favorito, B.S., Sara Sandri, B.S., Maria Rain Jennings, Ph.D., and Tinglong Dai, Ph.D.

    Abstract

    The landscape of U.S. Food and Drug Administration�regulated artificial intelligence�enabled medical devices (AIMDs) has expanded rapidly, with clearances and authorizations increasing, on average, from 1.4 to 146 per year (1995 through 2014 vs. 2020 through 2024, respectively). In this review of 950 AIMDs and their associated commercialization trends, private firms comprise 68.9% of the manufacturers, yet public companies produce more devices per firm on average (5.1 vs. 2.0), reflecting their larger scale and better resources. Deep learning now powers half of all new AIMDs, underscoring its growing role in medical AI. Transparency has improved, as the proportion of devices without explicit AI descriptions declined from 62.4% (2015 through 2019) to 8.9% (2020 through 2024). However, public companies exhibit a 30-fold higher recall rate than private firms, raising concerns about patient safety and regulatory oversight. This review examines disparities in development models, market positioning, and postmarket performance, highlighting the distinct commercialization strategies of public and private manufacturers. These findings contextualize how AI medical device manufacturers� characteristics relate to transparency and recall rates, highlighting commercialization aspects requiring regulatory and provider attention. (Funded by Johns Hopkins University.)