GeneralAI / InformaticsResearch

FDA Predetermined Change Control Plans for Radiology AI Surge in 2025, but Postmarket Monitoring Triggers Remain Undisclosed

Radiology. Artificial intelligencetoday

Nearly one-third of FDA-cleared radiology AI devices involve Predetermined Change Control Plans, with 65% cleared in 2025 alone. However, a new analysis finds public summaries consistently omit performance monitoring metrics and drift-trigger thresholds needed for postmarket sur…

  • Among all FDA-cleared AI/ML devices with a PCCP, 34 of 37 (91.9%) were for radiology, and 77.5% of all FDA-listed AI/ML submissions were radiological.
  • Discrepancies between the FDA's public database and individual regulatory summaries required manual adjudication for 27% (9/34) of radiology PCCP devices.
  • Documentation completeness scored a mean of 5 out of 8, with continuous performance monitoring and predefined retraining triggers absent from all public summaries.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.