Danish researchers recently published a paper in the journal Radiology, saying that when a commercially available artificial intelligence tool is used to analyze patients' chest radiographs, the sensitivity of identifying abnormal signs exceeds 99%, which means that artificial intelligence is expected to help improve the diagnostic efficiency of chest radiographs and reduce the workload of radiologists. X-ray chest radiographs are widely used in the examination of diseases related to heart, lung and other parts, but reading them requires rich professional knowledge and experience. Improving the automation of film reading can significantly improve the efficiency of image screening and diagnosis in medical institutions. The research team participated by the University of Copenhagen in Denmark and several local hospitals reported that they used a commercially available AI tool to analyze the chest radiographs of 1529 patients, and invited three radiologists to analyze the chest radiographs respectively, and compared the AI with the doctor's reading results. The results showed that in 1100 chest radiographs with abnormal signs confirmed by radiologists, 1090 were identified by AI tools. The overall sensitivity of artificial intelligence to identify abnormal signs is 99.1%, and the sensitivity to identify serious abnormal signs is 99.8%. The study also showed that the radiologist judged 429 chest radiographs to be normal, and artificial intelligence identified 120 of them, that is, 7.8% of all chest radiographs used in the study could be classified as normal according to the judgment of artificial intelligence, excluding the need for further diagnosis and treatment. Researchers said that in the future, depending on AI, a certain proportion of normal chest radiographs could be screened out automatically and the workload of doctors' diagnosis could be reduced, but more large-scale research was needed to verify this to ensure the safety of patients. (Xinhua News Agency)