Patient Perspectives Favor AI as Support Tool for Radiologists in Breast Cancer Screening

A comprehensive study exploring patient attitudes towards artificial intelligence (AI) in breast cancer screening has revealed cautious but generally supportive views. Conducted among a diverse patient population, the research highlights that while many patients are open to AI assisting radiologists, their level of trust is influenced by factors such as medical history, educational background, and racial demographics.

Published in Radiology: Imaging Cancer, the study surveyed 518 patients who underwent mammography screenings. Most respondents expressed support for AI functioning as a secondary review tool alongside human radiologists, with approximately 71% favoring AI as a second reader. Despite this, less than 5% were comfortable with AI independently interpreting their scans. Participants expressed concerns related to data privacy, transparency of AI algorithms, and potential bias.

The study emphasizes that achieving successful integration of AI in medical imaging depends heavily on patient trust and acceptance. Dr. Basak E. Dogan, a senior researcher involved in the study, pointed out that understanding patient perspectives is vital for effective implementation of AI tools. She noted, "If patients are skeptical or hesitant about AI, it could impact their willingness to participate in screening programs and follow recommended health practices."

To gauge patient opinions, researchers developed a 29-question survey, offering it to patients visiting their institution for mammography over a seven-month period in 2023. The survey collected demographic data and explored prior experiences with breast cancer, such as personal or familial history, which significantly influenced perceptions.

Results showed that individuals with higher educational attainment or greater self-reported knowledge about AI were twice as likely to accept AI involvement. Conversely, Hispanic and non-Hispanic Black participants reported more concerns about bias and data privacy, leading to lower acceptance levels within these groups.

Personal medical history also played a role: those with close relatives diagnosed with breast cancer were more inclined to request additional reviews when abnormalities were detected, regardless of whether AI or radiologists identified the issue. Interestingly, participants with a history of abnormal mammograms were more likely to pursue further diagnostics if AI flagged an abnormality, especially when conflicting with radiologist assessments.

Dr. Dogan highlighted the importance of personalized strategies when integrating AI into breast imaging, emphasizing ongoing dialogue with patients to understand their evolving views and concerns. Her team advocates for including patient perspectives to ensure AI technology enhances, rather than hinders, patient care, fostering trust and promoting adherence to screening recommendations.

This study underscores the cautious optimism among patients regarding AI in medical imaging and stresses the need for continued efforts to build transparency and trust to maximize the benefits of AI-assisted healthcare.

Source: https://medicalxpress.com/news/2025-04-patients-ai-radiologist-backup-screening.html