There are currently more than 20 artificial intelligence (AI) applications for breast imaging approved by the Food and Drug Administration (FDA) in the United States, most of these for full-field digital mammography (FFDM). Despite the early adoption of these algorithms, there have been no prospective clinical trials conducted. Until now, that is. A group from Sweden and Norway, led by Dr Kristina Lång (seen right) from Lund University in Malmö, has just published a carefully designed and well-executed study in The Lancet Oncology.
She commented, “In our trial, we used AI to identify screening examinations with a high risk of breast cancer, which underwent double reading by radiologists. The remaining examinations were classified as low risk and were read only by one radiologist. In the screen reading, radiologists used AI as detection support, in which it highlighted suspicious findings on the images.” In this population-based trial, 80,033 women were included and randomly assigned to two groups: 40,003 women in the intervention group who were subjected to AI-supported screening, and 40,030 women in the control group who underwent standard double reading by two radiologists.
The study, which was conducted between April 2021 and July 2022, employed the AI system Transpara version 1.7.0 supplied by the Dutch company, ScreenPoint Medical. The product is currently deployed in over 30 countries, can be used with multiple FFDM vendor platforms, and the company prides itself in having the most independent peer-reviewed publications in breast AI. Transpara provided a malignancy risk score that was then used to triage screening examinations.
Lång reported, “We found that using AI resulted in the detection of 20 % more cancers compared with standard screening, without affecting false positives. A false positive in screening occurs when a woman is recalled but cleared of suspicion of cancer after workup.” In addition, the workload for radiologists – that is, the time to read images – was reduced by 44%, a significant savings. Since 100,000 women have since been enrolled in the study, the researchers will soon be able to evaluate their primary endpoint: the rate of interval cancers.
In an invited commentary, Nero Segnan and Antonio Ponti from Turin in Italy provided positive words of support for the paper (seen right, © Science Photo Library), referring to the “breathtaking results” but they also offered “a word of caution.” They concluded: “An important research question thus remains: is AI, when appropriately trained, able to capture relevant biological features – or, in other words, the natural history of the disease – such as the capacity of tumours to grow and disseminate?”