TOPLINE:

Medical devices using AI were rarely indicated for children and were limited to a few specialties. Pediatric devices faced longer regulatory review times and had more clinical trial registrations than adult devices.

METHODOLOGY:

• Researchers conducted a cross-sectional study to characterize pediatric AI medical devices reviewed by the FDA.
• They retrieved 952 entries from the FDA database for AI-enabled devices spanning 1995 to 2024; of these, 948 had statements of indications for use in children and teens.
• Researchers used automated text processing with manual validation to extract indications for use, regulatory pathways, clinical specialties, sponsors, and FDA review timelines for each device.
• The median time required for FDA review and the rates of trial registration were assessed.

TAKEAWAY:

• Of 952 entries for AI devices, only 42 (4.4%) were labeled for pediatric patients. The first such device was cleared in 2015, and five devices were exclusively for pediatrics between 2020 and 2024.
• Age was omitted in 59.6% of statements for indication of use. When present, formats of labeling varied widely.
• Pediatric labels appeared in 8 of 18 clinical areas, with 42.9% related to radiology and 31% related to neurology. By contrast, 10 clinical areas had no pediatric devices.
• Pediatric devices had longer FDA review times (median, 162 vs 134 days; P = .049) and more often listed registered clinical trials (14.3% vs 2.2%) than non-pediatric devices.

IN PRACTICE:

“When devices are marketed for all ages, the FDA should require or strongly encourage pediatric performance reporting or, at a minimum, a strong justification of why it is not necessary,” the researchers of the study wrote.

“It may be a cliché, but it remains true that pediatric patients are not little adults, and algorithms trained on adult data may not perform reliably in pediatric populations. This concern is especially pressing in radiology, where more than 75% of FDA-authorized AI devices are used,” experts wrote in an accompanying invited commentary.

SOURCE:

The study was led by Grzegorz Zapotoczny, PhD, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago. It was published online on March 20 in JAMA Network Open.

LIMITATIONS:

The study used public FDA databases that were incomplete. Age ranges and trial details were reported inconsistently. Errors may have been introduced while extracting variables.

DISCLOSURES:

The study received support from the Health Resources and Services Administration through cooperative agreements to the Pediatric Pandemic Network. Additional support came from the FDA. One author reported receiving personal fees from Glooko, Dexcom, and Sanofi.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.