• Author
    Ria Dawar
  • Discovery PI

    Hannah Milch

  • Project Co-Author

  • Abstract Title

    Radiologist Perceptions of Artificial Intelligence in Screening Mammography

  • Discovery AOC Petal or Dual Degree Program

    Informatics & Data Science

  • Abstract

    Purpose: To evaluate breast radiologist perceptions of AI support tools in screening mammography

    Methods and Materials: As part of the PRISM (Pragmatic Randomized Trial of Artificial Intelligence for Screening Mammography), 119 breast radiologists from 6 academic centers across the United States were recruited to trial the AI tool Transpara (by ScreenPoint medical) in their breast mammography workflow over a 2–3-year study period. Prior to clinical implementation, they were surveyed on their attitudes toward AI in screening mammography (September to December 2025). The survey consisted of 47 Likert-scale questions, 15 multiple-choice questions, and 1 open-ended question. Responses were analyzed using descriptive statistics.

    Results: Among 118 respondents (response rate: 99%, 118/119), the vast majority (94%, 111/118) perceived newer AI cancer detection support tools as helpful, as compared to traditional Computer-Aided Detection tools (60%, 71/118). Further, most breast radiologists believed that the use of AI could improve the delivery of patient care (84%, 98/117), clinical decision making (86%, 101/118), and population health outcomes (84%, 99/118). Many felt that AI support would help reduce their level of burnout (55%, 65/118) and their cognitive workload and fatigue (69%, 82/118). However, many radiologists were concerned that AI would lead them to overcall findings (65%, 77/118), as opposed to undercalling or missing cancers (8%, 9/118). They also expressed concern about the lack of preparedness regarding the introduction of AI technology (72%, 85/118) as well as a lack of ethical framework in place (47%, 55/118). Overall, 78% (92/118) believed that AI would change their role as a healthcare professional in the future.

    Conclusion: Breast radiologists are broadly optimistic regarding the introduction of AI into screening mammography workflow and believe it will help them improve their delivery of patient care and outcomes. However, they raise concerns around ethical frameworks and institutional preparedness that only serve to highlight the need for structured integration guidelines for AI use in healthcare.  

Back