August 2026 Session: AI in Radiology, Enabling Safe and Effective Clinical Adoption-AI in Radiology: Enabling Safe and Effective Clinical Adoption

Video Transcription

Video Summary

The webinar explored how the American College of Radiology (ACR) is supporting safe, effective adoption of AI in radiology. ACR treats AI as another clinical modality and has developed several initiatives to help practices evaluate and monitor it.<br /><br />AI Central is a free inventory of nearly 500 FDA-authorized radiology AI products, with model cards and comparison tools. Its error calculator illustrates how results from enriched test datasets can differ from real-world performance, particularly as disease prevalence changes. Speakers also emphasized that clinicians, patients, and juries should understand that AI can produce both false positives and false negatives.<br /><br />ARCH-AI is a recognition program that helps facilities assess whether they have the governance, training, and processes needed to use AI responsibly. Participating sites share experiences through a learning network. Assess-AI provides ongoing performance monitoring using de-identified clinical data, allowing facilities to investigate discordant results and identify issues such as scanner or protocol changes. The ACR is also developing an accreditation program.<br /><br />The speakers discussed evaluating foundation models through clinician-led, multi-institution challenges and encouraged other specialties to share resources and lessons. Overall, safe adoption requires clear understanding of intended use, reliable monitoring, and collaboration as the technology and evidence continue to evolve.

Keywords

American College of Radiology

AI in radiology

AI Central

ARCH-AI

Assess-AI

FDA-authorized AI products

performance monitoring

foundation models