FDA Good Machine Learning Practice guiding principles — what the page says

FDA’s page, Good Machine Learning Practice for Medical Device Development: Guiding Principles, says the International Medical Device Regulators Forum (IMDRF) released a final document in January 2025 that identifies 10 guiding principles that can inform the development of Good Machine Learning Practice (GMLP). The page says that document builds on guiding principles released in October 2021 by FDA, Health Canada and the UK’s MHRA.

The page is a pointer to the IMDRF text. This article does not restate the ten principles; read them in the IMDRF document linked from the FDA page.

The PCCP guidance is a different kind of document: it makes recommendations for a marketing submission. IEC 62304:2006 is a different kind again: a medical device software life cycle standard that does not cover validation and final release of the device.

Editorial judgment: Treat the three as separate evidence layers. Principles inform development practice, IEC 62304 frames the software life cycle, and a PCCP describes planned modifications within a submission. We have not found an official FDA sentence that conforming to the guiding principles substitutes for either of the other two.