An extra pair of (digital) eyes - how AI will improve breast cancer screening 

Take our own Dr Kristina Lång, based at the Unilabs Mammography Unit at Skåne University Hospital in Malmö. She is looking into how AI can improve breast cancer screening in the world´s first randomised AI screening trial, involving 100,000 women. 

"The purpose of our trial is to assess whether AI can improve the efficacy of mammography screening by adapting single and double reading based on AI-derived cancer-risk scores, and to use AI as detection support in the screen reading. We hypothesise that the “interval cancer” and false positive rates will be reduced with AI-integrated mammography screening, along with a considerable reduction in the screen-reading workload,” she says. 

This last point is especially relevant considering the lack of breast radiologists – a shortage that is putting screening services under pressure. Fortunately, the findings of the study are promising. 

Preliminary results, in retrospective studies, are interesting, with AI identifying 20% of interval cancers at the time of screening. In addition, AI has been shown to be effective in determining screen exams that are cancer free or show a very low risk of malignancy.  

“Expectations are high, but we need to thoroughly evaluate how AI can be integrated for an efficient and safe screening programme. This has to be tested in a prospective setting in order to fully understand its impact, for example, how will AI influence radiologists’ decisions. One thing is clear: in the coming years, AI in healthcare will become unstoppable. As a physician, we must look for ways to use these new technologies to provide high-value care for our patients.”