Presentation
Initial Explorations of Applying Human Factors Principles to a BI-RADS CAD System
SessionPoster Session 1
DescriptionComputer-aided detection (CAD) systems are widely implemented to improve radiologists’ sensitivity, though issues of disuse, misuse, and poor usability and trustworthiness remain. Researchers are beginning to explore how CAD aids interact with advanced imaging techniques, such as digital breast tomosynthesis (DBT), and whether integrating the Breast Imaging Reporting and Database System (BI-RADS) supports clinical decision-making. This study compared participants’ “cancer” detection by assigning them to a contemporary CAD system, analog CAD (aCAD), or one of four BI-RADS CAD systems: original BI-RADS CAD (Sabo, 2024; oBI-RADS), oBI-RADS with aCAD, a novel BI-RADS CAD (nBI-RADS), or nBI-RADS with aCAD. The nBI-RADS CAD implementation used a more intuitive color scheme, reduced clutter, and provided holistic information for each pseudo-DBT image set. Participants determined if each set had “cancer” with and without the assistance of a CAD aid. General propensity to trust automation, perceived workload, and perceived usability and trustworthiness of the CAD system were also assessed. Results suggest CAD aids may improve sensitivity, but differences across implementations - including BI-RADS-integrated versions - were not statistically significant in performance, workload, usability, or trust. These findings support CAD use with DBT, while indicating that current BI-RADS integrations may require redesign to be effective.
Event Type
Poster
TimeTuesday, October 14th5:30pm - 6:30pm CDT
LocationRiverside East