The AI Theatre: Keynote

AI-SC 11 - Leading by example: Assess-AI – ACR’s real-world algorithm performance comparison and monitoring

Lectures

1
Chairperson's introduction

Chairperson's introduction

03:00Yan Chen, Nottingham / UK

2
Leading by example: Assess-AI – ACR’s real-world algorithm performance comparison and monitoring

Leading by example: Assess-AI – ACR’s real-world algorithm performance comparison and monitoring

18:00Christoph Wald, Nahant / US

3
Benchmarking in the real world: what I learned

Benchmarking in the real world: what I learned

18:00Jaishree Naidoo, Surrey / UK

4
Q&A: How will we actually set up and manage surveillance?

Q&A: How will we actually set up and manage surveillance?

21:00Q&A: How will we actually set up and manage surveillance?

Sustainable Saturday
3 min
Chairperson's introduction
Yan Chen, Nottingham / United Kingdom
18 min
Leading by example: Assess-AI – ACR’s real-world algorithm performance comparison and monitoring
Christoph Wald, Nahant / United States
  1. To learn how Assess-AI provides continuous, real-world performance monitoring of AI algorithms in clinical settings, leveraging data from the ACR National Radiology Data Registry.
  2. To appreciate the value of comparing AI algorithm performance against national benchmarks and similar facilities, ensuring safe and effective AI deployment across diverse clinical environments.
  3. To understand the role of real-time data collection, feedback, and reporting in maintaining AI quality and driving improvements in clinical decision-making and outcomes.
18 min
Benchmarking in the real world: what I learned
Jaishree Naidoo, Surrey / United Kingdom
  1. To understand the importance of real-world AI benchmarking in radiology by recognizing the limitations of pre-market validations and the need for ongoing post-market surveillance to ensure safety, effectiveness, and clinical relevance.
  2. To gain insight into the updated EU AI Act and FDA guidelines relevant for post-market AI monitoring, and explore a Practical Framework for AI Benchmarking in Real-World Settings.
  3. To appreciate the role of collaboration between radiologists, developers, and policymakers in fostering trust and ethical AI deployment while adopting actionable strategies to ensure responsible, patient-centric AI integration.
21 min
Q&A: How will we actually set up and manage surveillance?

This session offers AI-generated subtitles.