Discrepancies in FDA clearance, marketing of AI-, ML-enabled medical devices

Editor's Note

This study from New York University and NYU Langone Health, New York City, finds discrepancies between the marketing and 510(k) clearance of artificial intelligence (AI)- or machine learning (ML)-enabled medical devices, with some devices being marketed as having capabilities not approved by the Food and Drug Administration (FDA).

A total of 119 public 510(k) applications and corresponding marketing materials were included in the analysis.

The devices were taxonomized into three categories, and included:

• 15 (12.6%) devices were considered discrepant
• 8 (6.72%) devices were considered contentious
• 96 (84.03%) devices were considered adherent (ie, consistent between marketing and FDA 510(k) clearance summaries).

Most devices were from radiological approval committees (75 devices), with:

• 62 (82.6%) adherent
• 3 (4%) contentious
• 10 (13.33%) discrepant.

Radiological devices were followed by the cardiovascular device approval committee (23 devices), with

• 19 (82.61%) adherent
• 2 (8.70%) contentious
• 2 (8.70%) discrepant.

The findings show significant discrepancy in the marketing of AI-or ML-enabled medical devices compared with their FDA 510(k) summaries, the researchers say. “The aim of this study was not to suggest developers were creating and marketing unsafe or untrustworthy devices but to show the need for study on the topic and more uniform guidelines around marketing of software-heavy devices,” they note.