Medical Imaging Grand Rounds - Promises and Pitfalls of Deep Learning in Radiology

Wednesday, June 17, 2020 - 12:00pm
Event Location Notes: 

ZOOM Meeting link -  https://uahs.zoom.us/j/98328455278

Meeting ID  983 2845 5278

6/17/20 @ 12:00 pm

 

 

The Department of Medical Imaging is pleased to have Joshua D. Warner, MD, PhD, presenting at our Grand Rounds on Wednesday, June 17th at 12:00 pm. This is a ZOOM Meeting and the meeting login details are below.

Dr. Warner is a PGY-4 Diagnostic Radiology resident at Banner - University Medical Center in Tucson, Arizona who completed his MD and PhD degrees at Mayo Clinic in Rochester, MN.  His PhD studies were in Biomedical Engineering and Physiology, with work completed in the Radiology Informatics Lab, and his research interests include quantitative imaging and convolutional neural nets (commonly referred to as Deep Learning or Artificial Intelligence in the lay press).  

Abstract: Headlines breathlessly tout the potential for "Artificial Intelligence" in radiology; at least daily there are new reports or claims made about some computer outperforming humans.  Despite this, very few techniques have managed to mature or change practice.  There are good reasons for this, but algorithms are opaque to most in the field as well as the media, which has created problems in the review/publication process.  This talk will discuss conceptually the approaches in use today, what questions they can (and cannot) answer, and finally review a highly relevant recent publication which illustrates the danger posed when these algorithms are applied to image reconstruction (https://www.pnas.org/content/early/2020/05/08/1907377117).  All radiologists should be aware of the dangers these techniques pose to the image reconstruction pipeline.

Event Coordinator: 
Mimi Villafanae
(520) 626-2114
Event Contact Department: 
Medical Imaging