Patients Are Generally Supportive of Artificial Intelligence in Breast Imaging: A Multisite Survey of Breast Imaging Patients.
Authors
Affiliations (10)
Affiliations (10)
- Department of Radiology, Duke University, Durham, NC, USA.
- Department of Radiology, Cornell University, New York, NY, USA.
- Department of Radiology, Northwestern University, Chicago, IL, USA.
- Department of Radiology, University of Pennsylvania, Philadelphia, PA, USA.
- Department of Radiology, Johns Hopkins University, Baltimore, MD, USA.
- Department of Radiology, University of Texas MD Anderson Cancer Center, Houston, TX, USA.
- Department of Radiology, Bronson Healthcare, Kalamazoo, MI, USA.
- Department of Radiology, University of South Carolina School of Medicine, Greenville, SC, USA.
- Duke School of Medicine, Duke University, Durham, NC, USA.
- Department of Radiology, Fred Hutchinson Cancer Center, University of Washington, WA, USA.
Abstract
To understand the perspective of patients undergoing breast imaging on the use of artificial intelligence (AI) in breast cancer screening. A 36-item survey was administered to breast imaging patients at 6 academic and 2 private practice groups in the United States. The survey included questions regarding demographics, breast imaging history, and electronic health literacy. Respondents were asked Likert scale questions on the role of AI in breast cancer screening, the role of AI as an independent or complementary reader, and concerns regarding AI in breast imaging. The survey yielded 3532 responses, a response rate of 69.9% (3532/5053). The median age was 55.9 years (SD, 12.3 years), and most respondents were White (73.0%, 2679/3532). Respondents indicated support for the role of AI to identify suspicious findings (70.6%, 2492/3532), triage findings for review (69.5%, 2382/3532), calculate breast density (73.2%, 2588/3532), and estimate breast cancer risk (61.9%, 2186/3532). Significantly higher support was noted among patients who were White, had more education, and had greater health literacy (all P <.05). There was strong agreement that it was necessary for radiologists to also review each examination (67.3%, 376/3532). Respondents were uncertain about whether AI (41.2%, 1456/3532) or radiologists (31.8%, 1124/3532) were responsible for errors. There was concern that AI will limit communication between patients and radiologists (75.7%, 2673/3532). Breast imaging patients have an overall favorable view of AI in breast cancer screening, with variable support by demographics. Education and outreach efforts should target perceived challenges to AI adoption to improve patient acceptance.