A Large-Scale Multi-Centre Research On Domain Generalisation in Mammography Artificial Learning-Based Mass Detection: A Review

In 2020, breast cancer surpassed lung cancer for the first time as the most prevalent cancer globally. Nearly 30% of all malignancies in women are caused by it, and the prevalence is rising according to current trends. During x-ray mammography, the gold standard imaging method for early detection utilized in screening programs, anomalies in the breast structures can be found to be indicative of breast cancer. These abnormalities may take the form of masses, calcifications, architectural distortions, or asymmetries in the breast. Furthermore, breast cancer screening has a high chance of false positives that could lead to unnecessary biopsies as well as a high rate of false negatives or missed malignancies. A recent, substantial study that used the interpretations of 101 radiologists to evaluate the effectiveness of an artificial intelligence (AI) system came to the conclusion that, in a retrospective setting, the AI stand-alone had cancer detection accuracy comparable to that of a typical radiologist. The performance was similar in Breast cancer screening AI stand-alone solutions was studied in several therapeutic contexts and demonstrated better cancer prediction rate in comparison to a human expert double-reader strategy. Contrary to these conclusions, related research evaluated how well AI algorithms performed.