Objective:
To evaluate the adoption of Predetermined Change Control Plans (PCCPs) among FDA-cleared radiology AI/ML devices and assess the transparency of public documentation.
Approach:
- Study Design: A systematic scoping review was conducted following PRISMA guidelines, registered in PROSPERO, evaluating FDA-cleared radiology AI/ML devices from 2015 to 2025.
- Data Sources: FDA 510(k), de novo, premarket approval, and breakthrough device databases were accessed on April 1, 2026, with manual verification of PCCP devices against regulatory summaries.
- Assessment Method: PCCP documentation was evaluated using an 8-point rubric covering specificity, data and validation transparency, real-world evaluation, and postmarket monitoring.
Key Findings:
- Out of 1,394 FDA-listed AI/ML submissions, 1,080 (78%) were in radiology, with 1,068 (99%) cleared through the 510(k) pathway.
- Among 870 unique radiology devices, 130 (15%) underwent sequential 510(k) clearances for significant modifications.
- The mean interval between sequential submissions decreased from 25 months prior to 2021 to 13 months post-2021.
- Of 37 radiology devices cleared with PCCPs, 34 (92%) were AI/ML devices, with 22 cleared in 2025.
- Public documentation had a mean score of 5 on the 8-point rubric, with only 3 devices describing postmarket surveillance.
- Mean documentation scores increased from 4 in 2024 to 5 in 2025.
Interpretation:
The analysis indicates an increase in the adoption of PCCPs among radiology AI devices, but public documentation often lacks comprehensive details on postmarket performance monitoring.
Limitations:
- Analysis limited by publicly available data, with internal safety assessments and minor software changes inaccessible.
- Database inconsistencies and challenges in deduplicating devices due to corporate rebranding.
- The transparency rubric used was not designed as a comprehensive quality evaluation.
- The recency of final FDA guidance may limit the analysis.
Conclusion:
The study highlights the need for improved transparency in public documentation regarding postmarket monitoring of radiology AI devices.
Sources:
This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.
