54 PathBrowser: A pathology education tool for interactive flashcard-based self-learning

Background Flashcards are a favorite learning tool that enable repetitive exposure to a wide breadth of images representing clinical pathology specimens, which is critical for achieving diagnostic competence. While popular flashcard platforms exist, they are not tailored for pathology learning, and do not actively engage user input, allowing users to bypass the active recall process that is necessary for test-enhanced learning. Methods To address these limitations, we developed PathBrowser, an R-shiny based web browser-accessible program for interactive flashcards. Given a picture (or sets of pictures), users are challenged to enter their diagnoses, and are provided instant feedback by having their answers scored based on similarity to the correct answer. Users can also choose to show captions embedded in the metadata for each picture, which store expert-curated information from the original source. Finally, the web application integrates a pathology database, which provides additional information about the diagnostic entity represented in the picture (including definitions, clinical and prognostic correlations, and ancillary studies). The application is designed to allow users to quickly browse through large sets of questions, using keystrokes to navigate, score, and filter out correctly answered questions in subsequent rounds. Results We created a flashcard deck for common peripheral blood smear findings on PathBrowser (https://pbrowser.shinyapps.io/PathBrowser_v5/) and made it available to a group of pathology residents, who were asked to run the app and complete a short survey. Seven of 8 (87.5%) respondents agreed the web application helped them better recognize blood smear findings, and that the web application was a more effective learning tool compared to only reading from the textbook. Conclusions We developed an interactive flashcard program for pathology education, with highly favorable responses from pathology trainees. Current work aims to integrate high-quality questions and answer explanations, curated from textbooks or online sources, or generated using large language models.