Newsbytes

Create PDF

Eye-tracking technology: tomorrow’s tool to train pathology residents?

May 2020—Imagine the potential educational benefits of pathology residents being able to see the precise path that the eyes of experienced pathologists take as they scan a whole slide image.

Preliminary research has suggested that showing residents visual representations of a pathologist’s eye-tracking movement overlaid over a whole slide image can impact how they learn, says Sharon E. Fox, MD, PhD, a pathologist at the Southeast Louisiana Veterans Healthcare System and associate director of research and development, Department of Pathology, Louisiana State University Health Sciences Center. “If [residents] are given instruction that includes some of those videos showing eye tracking and where the eyes of their instructor are moving on the image, they tend to develop a pattern that is similar to the instructor,” adds Dr. Fox, who gave a presentation on eye tracking at the Digital Pathology Association’s Pathology Visions 2019 conference last fall.

Dr. Fox

Among early findings about the technology, summarized by Dr. Fox and Beverly Faulkner-Jones, MD, PhD, in Frontline Learning Research (2017;5[3]:43–54), were that eye-tracking “fixations”—the fractions of a second when a viewer’s eyes fixate on points in an image—show that pathology trainees and expert pathologists exhibit significantly different gaze patterns. “Trainees may actually get the answer or diagnosis correct, but typically they are using far more fixations,” explains Dr. Fox. “They may spend proportionally the same amount of time on the different parts of tissue, but they are spending far more time [overall] and far more fixations than an expert would need to.” Therefore, she adds, eye-tracking technology ideally will become an instructional aid for residents and practicing pathologists who want to hone their skills.

“It seems a lot of pathologists would be interested in seeing how [the experts] are approaching images and would find it useful that they could be shown that in an electronic format,” she says.

Dr. Faulkner-Jones, dermatopathologist and podiatric pathologist at Quantum Pathology, Waltham, Mass., and Dr. Fox’s collaborator on the 2017 eye-tracking article, has also found residents to be eager for visual guidance. Therefore, “we need to have objective measures of how we interact visually with an image,” and eye tracking could be one of those measures, adds Dr. Faulkner-Jones, who, for several years, taught dermatopathology to pathology and dermatology residents using digital pathology.

When teaching, “I would pan through a digital image, talk about what I was looking at, and then zoom in on the areas and features in the tissue that I thought were the most important diagnostically,” she says. “Residents expressed how useful it was to have a walkthrough of the image.”

Dr. Faulkner-Jones

In addition to viewing eye tracking as a tool to train residents, Dr. Faulkner-Jones considers the technology “a tool to work out how best to present tissue for teaching and diagnosis, [such as] working out the ideal number of tissue pieces and their orientation and distribution on a slide.”

Eye-tracking technology should also be used to inform artificial intelligence, says Dr. Fox, who is collaborating with members of Tulane University’s bioengineering and computer science departments to explore this application. Digital pathology image files are big and take a long time for computer algorithms to scan, she explains. Therefore, using eye-tracking technology to map the areas pathologists fixate on when they look at an image could “teach” computers to efficiently scan only those areas of the slide relevant to the pathologist.

“The question is, how is human vision and computer vision the same and different, and can human vision be used to reduce the number and types of images that go into an AI algorithm to make it faster and more efficient?” she says. Answering these questions could lead to eye-tracking technology being used as widely in pathology as in radiology, or even marketing and gaming.

Getting to this point, however, will require “quite a bit of background and training” on the technology, says Dr. Fox, who uses a device from the Swedish technology company Tobii to track eye movements. The device, which sits at the bottom of the pathologist’s computer screen, uses infrared light to record a reflection off the cornea of the eye. The viewer is told to look at certain spots on the screen before the session begins, and the eye tracker is calibrated to that person’s eyes and the screen being viewed, she explains.

The post-processing step, in particular, is complex, Dr. Fox says. Eye-tracking data, she explains, are a separate file overlaid on the image being viewed. Therefore, the user must have basic training in the technology to understand if or how the eye-tracking data file can capture a visual field.

Consequently, incorporating eye tracking into the traditional pathology workflow has been challenging, says Dr. Fox, noting that it is more difficult to use the technology to create accurate visual field images when eye movements occur through a microscope. However, she adds that, “We are already seeing increased adoption of digital pathology technology as both pathology practice and education have rapidly moved to remote diagnostic formats as a result of the COVID-19 pandemic.”

To this end, Dr. Fox, through her collaboration with Tulane University bioengineers and computer scientists, is using new three-dimensional tissue imaging techniques to better understand the disease process affecting the lungs of those with severe COVID-19. “Three-dimensional structures open up a whole new range of questions about how pathologists learn to render diagnoses from novel image formats,” she says. “And we are interested in using eye tracking to understand how three-dimensional analyses may differ from our traditional two-dimensional slide analyses.”

“Digitizing images,” concludes Dr. Faulkner-Jones, “is going to make the whole application of eye tracking easier.”

—Renee Caruthers

CDC introduces COVID-19 case reporting app

The Centers for Disease Control and Prevention announced plans to introduce an app this month to accelerate the electronic case reporting of COVID-19 data to public health agencies.

Health care institutions that do not have the capability to automatically send COVID-19 case reports from their EHRs will be able to send information through the app, called eCR Now. The app is designed to deliver data on COVID-19 to public health agencies in real time, increasing case management and contact tracing while easing the reporting burden on health care providers.