What can the pupils of pathologists tell us about making a correct cancer diagnosis? A great deal of basic research has validated the value of examining the temporal dynamics of the pupil response to understand individual differences in how the same information may be processed differently by two individuals. The proposed work will extend this research to the understanding of expertise in digital pathology by examining tonic and phasic pupil responses. Based on this prior research, we believe that the pupil may be exquisitely sensitive to the perceived difficulty of an entire case, a region of interest within a case, or the diagnostic categorization of a case. This supplement will benefit from the data collection already planned by the parent R01 (RAPID-PC), which will collect eye-tracking data from 10 sites across the US. The parent R01 will examine breast pathology interpretation in residents and experienced pathologists to understand how expertise develops across the entire diagnosis process from primary diagnoses to second opinions. During the first wave of data-collection in the parent R01, we anticipate collecting data from 84 residents and 30 experienced pathologists. Each pathologist will examine 14 cases, yielding an estimated 1596 interpreted cases to analyze. Thus, this supplement proposes to analyze one of the largest eye-tracking datasets ever collected in the field of medical image perception. Although the focus of the parent grant is to understand and categorize the cause of errors committed while examining challenging cases through eye-tracking analyses, such as saccadic amplitude and time-to-first fixation for specific regions of interest, the parent grant will not examine the pupil response. Thus, the proposed project will provide a pathway to extend the use of this rich dataset to a novel line of inquiry: can the pupil provide a meaningful index to differentiate expert from novice responses to well- validated cases of known difficulty? Our specific aim is to examine the relationship between expertise and tonic and phasic changes in pupil diameter during image interpretation. If funded, Steffi Falla, a first-generation college student from a low-income and diverse background, will work with Drs. Drew and Bruny to collect pathologist data and analyze pupillometry data to assess how expertise changes the pupil's response to overall case difficulty (tonic changes) and challenges associated with particular lesions when they are fixated (phasic changes). The overarching goal is for Ms. Falla to complete this diversity supplement for two years prior to attending graduate school so that she can ultimately pursue a career in academia. In addition to providing a novel method of analyzing the evaluation of digital pathology slides, the grant will provide Ms. Falla with a formal training plan that is designed to focus on areas critical to her success in graduate school, and eventually, as an independent researcher. Ms. Falla is an extremely promising candidate who will help diversify the viewpoints and perspectives within academia. We are excited that this funding opportunity may provide us with an opportunity to benefit from her contribution to this promising new line of research.
Pathology diagnosis is the result of human visual perception and cognitive processing of complex histopathology images. This study will contribute to the understanding of expertise in digital pathology by examining pupil responses among pathologists during the medical image interpretation process. Millions of breast biopsies are performed each year, thus our study has crucial clinical implications for patients who depend on pathologists' interpretations to guide treatment recommendation.
