Eye-tracking to find missed Lung Nodules
Principal Investigator
Name
Gregory DiGirolamo
Degrees
Ph.D.
Institution
University of Massachusetts, Chan Medical School
Position Title
Associate Professor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-953
Initial CDAS Request Approval
Sep 6, 2022
Title
Eye-tracking to find missed Lung Nodules
Summary
Misses are the most common and serious error in diagnostic visual search and have been investigated extensively with eye-tracking. For experienced radiologists, misses are different than for non-experts in a significant way. Eye-tracking during diagnostic visual search shows that radiologists make recognition errors (where missed abnormalities are still looked at longer than normal tissue but below some arbitrary threshold for conscious recognition. Experts make these recognition errors across modalities and anatomy, including lung nodules in chest CTs. A critical need is characterizing these non-conscious processes in diagnostic visual search and utilizing these non-conscious processes to improve nodule detection beyond conscious detection limits.
Aims
We are trying to determine why errors are made when diagnosing from a radiological image. Specifically, we want to know whether misses occur because Radiologists don’t look at the region in the image that has an abnormality. In addition, we would like to use the eye-movement data to see if deep machine learning can use the eye-movement data to learn if an abnormality is present in the images.
Collaborators
Max Rosen, University of Massachusetts Chan Medical School
Federico Sorcini, University of Massachusetts Chan Medical School
Alexander Bankier, University of Massachusetts Chan Medical School