false
OasisLMS
Catalog
ASGE Annual Postgraduate Course: Clinical Challeng ...
Learning from Other Disciplines: How Far Behind is ...
Learning from Other Disciplines: How Far Behind is GI?
Back to course
[Please upgrade your browser to play this video content]
Video Transcription
Video Summary
In this video, Dr. Charles Kahn discusses the challenges and considerations of applying artificial intelligence (AI) in medical imaging, with a focus on radiology. He highlights the importance of testing and validating AI systems on specific populations, as well as understanding the potential biases and limitations of these systems. Dr. Kahn also emphasizes the need for information standards in medical imaging, using the example of radiology's implementation of DICOM. He presents cases where AI systems have displayed shortcut learning, such as mistaking radiographic markers for pneumonia detection or chest tubes for pneumothorax detection. Dr. Kahn advises clinicians to thoroughly evaluate AI solutions, considering factors like intended users, potential benefits and risks, integration into clinical workflow, and maintenance. He also urges researchers to report false positives and false negatives to enhance the understanding and improvement of AI systems. The speech ends with Dr. Kahn expressing solidarity with other specialties facing similar challenges and thanking the audience for their attention. No credits were mentioned in the video.
Asset Subtitle
Charles Kahn, Jr., MD, MS
Keywords
artificial intelligence
medical imaging
radiology
testing and validating
biases and limitations
×
Please select your language
1
English