ذكاء اصطناعي يُغني عن قارئ ثانٍ.. ويكشف أورام الثدي مبكراً

A Dutch study published in Lancet Digital Health and the Journal of the National Cancer Institute, as part of the national breast cancer screening program, showed that reliance on an artificial intelligence system could replace the second radiologist in reading mammogram images without compromising accuracy, while reducing workload and saving millions of euros annually. A team from Radboud University led the analysis of 42,000 images from the Utrecht region, with a follow-up period of about 4 and a half years, using an algorithm developed by ScreenPoint Medical. The researchers concluded that combining human readers with the algorithm achieves earlier detection of a greater number of clinically significant tumors, which directly impacts survival rates and treatment planning. Although the design of screening programs varies between countries, the practical implication is clear; a significant portion of routine effort can be transferred to intelligent systems, while keeping the final decision with the physician in a supervisory model that reduces errors and increases consistency. The result aligns with a broader trend shown in previous literature that deep models enhance human reader performance and limit variability among experiences.