4%. The AI tool Mia was used in routine breast screening of 10,889 women in NHS Grampian. Among these, 106 women were diagnosed with cancer through routine screening, and an additional 11 cancers were diagnosed with AI support, seven of which were invasive.
AI reduced the time to notify affected women from two weeks to three days. According to the trial, AI reduced the number of women unnecessarily recalled for further tests, including unnecessary biopsies. Researchers found that AI cut the workload of healthcare staff by as much as 31%.
Using AI as a second scan reader resulted in the best combination of workload savings and increased early cancer detection without recalling more women for additional tests. Dr. ' A study on 100,000 Swedish women published in The Lancet shows that AI can make breast cancer testing better and more efficient.
When radiologists got help from AI, more cancer was detected, and more tumors were detected early. ' No more 'false' cancer cases were detected with AI, which is important to avoid unnecessary tests and operations. The workload for radiologists decreased with AI.
In Sweden, radiologists got help from AI throughout the process, which could make them overly dependent on AI. Only experienced radiologists were allowed to interpret images that AI had processed in the Swedish study. In a Norwegian study, doctors do not see AI results until after experts have made their own assessment.
AI can halve the interpretation time for radiologists. It takes a radiologist only about one minute to assess whether a mammogram shows signs of cancer or not. Even with AI, two humans are always involved in assessing the most difficult cases.
According to researchers, the main gain from AI is in better quality, not necessarily in saving radiologist resources. Yvonne Cook's aggressive breast cancer was detected by artificial intelligence after being missed by medics. Women in the UK aged 50 to 70 are invited for breast cancer screening every three years.
Mammograms are assessed by two radiologists, but some cancers are missed. 3 million Norwegian women have been through the Mammography Program, with over 30,000 breast cancer cases detected over 30 years. The long-term impact of AI on radiologist training and experience, especially if only experienced radiologists are used with AI, remains uncertain.
Additionally, the exact cost savings or resource allocation changes from implementing AI in breast cancer screening programs have not been determined. The potential risks or error rates of AI in breast cancer detection that might not be covered in current studies are also unknown.