New study on radiologists' experiences of using AI to detect breast cancer
We already know that artificial intelligence can be used for triage in breast cancer screening. But what do radiologists think about using AI tools after working with it for a year? A new study published in the Health Informatics Journal is unique in exploring radiologists' experiences of using artificial intelligence as a supportive tool for breast cancer detection.
Jennifer Viberg Johansson, co-author of the study.
The study was based on qualitative semi-structured interviews with well-trained clinical radiologists with extensive experience of interpreting images from mammography examinations. The radiologists that took part in the study are actively using AI in their work.
“The study is small, and conducted in one single workplace. What makes it unique is the fact that it examines how one professional group perceives AI tools after working with them for one year. The result is not presenting hypothetical scenarios or opinions. Instead, the results reflect their experience with AI, and their views based on that,” says Jennifer Viberg Johansson, Associate Professor of Medical Ethics at Uppsala University’s Centre for Research Ethics & Bioethics.
The results of the study indicate that radiologists are comfortable with the type of output that is generated by AI. Which in turn suggests that mammography could be a springboard for AI in healthcare. However, the interviewed radiologists stated that the AI tool sometimes showed ambiguous or incorrect results, resulting in more work. The AI tool was therefore perceived primarily as complementary, as opposed to an independent tool that could replace them. The AI tool was perceived as a safety net and a way to confirm diagnostic decisions, rather than as a method that could be used to detect and correct errors, or discover new things.
According to Jennifer Viberg Johansson, radiologists' experiences from working with an AI tool in a clinical setting can be of great value in the development of new AI solutions for health care. “Including clinicians' feedback could contribute to a more effective and accepted use of AI tools in healthcare in the future,” says Jennifer Viberg Johansson.
Do you want to know more? Take a look at the study!
By Fanny Klingvall & Josepine Fernow