Training Perceptual Expertise in Radiologists

Visual search is a ubiquitous task (e.g., find the < key on your keyboard) that can have dire consequences for missed targets in a variety of different fields (e.g., cancer in radiographs). Despite improvements in training and technology, experienced radiologists miss diagnostic aspects of radiographic images at a rate of 3-5% per day. Given that experienced radiologists, like experts in other specialized contexts, not only identify targets more easily in radiographic images (relative to novice and naïve subjects), but also appear to enjoy enhanced visual skills in that domain of expertise-– why do these daily errors in performance persist, and what training interventions might reduce such error? This interdisciplinary research project will comprise researchers from Philosophy, Psychology, and Radiology and will focus on understanding how to best train visual expertise, with central emphasis on practicing radiologists and trainees. Utilizing visual search tasks, we will measure the impacts of search pattern, distractions, and training schedules on perceptual performance using behavioral and eye-tracking methods of study, with the aim of improving perceptual performance. This interest generalizes to the following theoretical question: how and in what contexts can humans improve their perceptual skills? And an educational question: what training interventions optimize perceptual performance in specialized domains? Accordingly, the aims are both theoretical and practical, engaging at least three U strategic goals: development and transfer of new knowledge, community engagement and promotion of student success to improve health and quality of life.

Budget:
Student Expenses:
-1 graduate student ~0.75 year (7 mos) = $15,200
-1 undergraduate student 1 year hourly pay (10 hrs per week, 9 months) = $8K
-Equipment Expenses: $2K (new eye tracking computer)
-Participant payment (Material) Expenses: 40 radiologists at $100 per hour = $4K
-Travel Expenses: $800

Current Status

2024-01-24
Radiologists develop incredible perceptual expertise for searching for anomalies across images. The implementation of new scanning technologies now allows for searching these images in 3D; however, little perceptual training is provided to develop search processes for 3D imagery. This proposed study will implement a new training procedure in 3D using virtual reality.

Expert radiologists can rapidly and accurately identify anomalies in chest radiographs.One standard explanation for this ability appeals to an underlying mechanism that allows well-trained radiologists to quickly see the gist or “gestalt” of a radiograph. This kind of holistic
processing may be what separates novice radiologists from seasoned experts. Research suggests that the ability to holistically search the image is guided by prior knowledge and experience (Drew et al., 2013, Ivy et al., 2023). And eye tracking expert radiologists while they read images provides some insight into trained behaviors that aid search.

Little research has questioned whether 2D image search behaviors can translate to
searching the novel, three-dimensional images that are now becoming more prevalent. Studies suggest that expert radiologists rely on different mental representations (due to more training) to effectively guide search. The question is whether training protocols for either 2D radiography or volumetric imagery translate to genuine 3D images.

This project aims to develop training protocols for 3D musculoskeletal scans and compare to current 2D training practices. We will test whether identifying and classifying fractures in 3D radiography can be trained. We will additionally assess whether spatial cognitive abilities—those involving mental imagery—correlate with efficiency of training of 3D scans. Prior work suggests that mental rotation ability is correlated with success in STEM fields (Uttal & Cohen, 2012). Developing training procedures for 3D imagery could benefit from information about who has learned strategies that could aid search, so we will test for this ability in participants.

Participants in this experiment will be medical students/residents (N ≥ 40) who self-enroll in the perceptual training and will be randomized to control and experimental groups. All participants will complete the short version of the Vandenburg-Kuse (1978) mental rotation task prior to training. Experimental and control groups will receive training at the RadSimPE radiology workstation simulator. Participants will be asked to identify abnormalities on medical images of acetabular fractures. Subjects will be given several sets of cases with interleaved perceptual education in the form of VR tools. The groups will vary with respect to when training is introduced. Accuracy of search will be recorded. When in virtual reality, eye tracking data will be collected and analyzed. Surveys that assess subjective perceptions of both types of training will be given post training as in Banerjee et al. (2021). We hypothesize that subjects in the experimental group receiving perceptual education using VR will be better at identifying and classifying image abnormalities than the control group.

Appropriate cases for training have been curated and rendered into the 3D VR platform. The mental rotation survey has been created. Educational materials are being developed with anticipated enrollment of study participants in the coming months.

Banerjee, S., & Auffermann, W. F. (2021). RadSimPE—a Radiology Workstation Simulator for Perceptual Education. Journal of Digital Imaging, 34, 1059-1066.

Drew, T., Evans, K., Võ, M. L. H., Jacobson, F. L., & Wolfe, J. M. (2013). Informatics in
radiology: what can you see in a single glance and how might this guide visual search in medical images? Radiographics, 33(1), 263-274.

Ivy, S., Rohovit, T., Stefanucci, J., Stokes, D., Mills, M., & Drew, T. (2023). Visual expertise is more than meets the eye: an examination of holistic visual processing in radiologists and
architects. Journal of Medical Imaging, 10(1), 015501-015501.

Uttal, D. H., & Cohen, C. A. (2012). Spatial thinking and STEM education: When, why, and how? In Psychology of learning and motivation (Vol. 57, pp. 147-181). Academic Press.

Vandenberg, S. G., & Kuse, A. R. (1978). Mental rotations, a group test of three-dimensional spatial visualization. Perceptual and Motor Skills, 47(2), 599-604.