Related Experiment Video
Updated: Aug 24, 2025
05:04
Author Spotlight: Evaluating Clinicians' Adoption of Ultrasound-Guided Vascular Cannulation Through Simulation Training
Published on: August 9, 2024
1.1K
Artificial Intelligence-Powered Clinical Decision Support and Simulation Platform for Radiology Trainee Education
Chintan Shah1, Karapet Davtyan2, Ilya Nasrallah3
1Department of Radiology, Imaging Institute, Cleveland Clinic, 9500 Euclid Ave, Mail code S3, Cleveland, OH, USA. shahc2@ccf.org.
Journal of Digital Imaging
|October 24, 2022
Summary
AI-powered clinical decision support (CDS) software enhanced the educational value of radiology simulation cases without increasing interpretation time. This technology offers a promising avenue for improving trainee education in medical imaging.
Area of Science:
- Medical Education Technology
- Radiology Training
- Artificial Intelligence in Healthcare
Background:
- Traditional radiology education relies on established methods.
- Technological advancements offer opportunities to innovate educational approaches.
- Computer-generated feedback and simulation cases are emerging tools.
Purpose of the Study:
- To investigate the utility of AI-powered Bayesian inference-based clinical decision support (CDS) software.
- To evaluate the impact of real-time automated feedback on radiology trainees during brain MRI interpretation.
- To compare the educational value and confidence levels between clinical and simulation-based learning scenarios with and without CDS.
Main Methods:
- Ten radiology trainees interpreted 75 brain MRI examinations across 25 reading sessions.
- Scenarios included clinical cases (with and without CDS) and simulation cases (with CDS).
- Trainees provided input on imaging features and differential diagnoses; CDS provided inferred diagnoses for review with an attending neuroradiologist. Confidence and educational value were assessed via surveys; timing was also recorded.
Main Results:
- Simulation cases with CDS demonstrated a statistically significant higher educational value compared to clinical cases without CDS (p < 0.05).
- Trainees reported slightly lower confidence in findings and diagnoses for simulation cases with CDS.
- No significant differences in confidence or educational value were found between clinical cases with and without CDS. Overall interpretation time did not differ across scenarios.
Conclusions:
- AI-powered CDS feedback integrated into simulation cases can enhance the educational value of interpreting imaging studies.
- This approach may improve trainee learning without extending workstation time.
- Further research is warranted to explore innovations in radiology trainee education, particularly relevant for remote work and asynchronous review models.
