AI can be a powerful tool to help identify lung nodules on chest X-rays

AI can be a powerful tool to help identify lung nodules on chest X-rays

 In a groundbreaking randomized controlled study to assess the impact of artificial intelligence (AI)-based software on real-world clinical practice, researchers found that AI significantly improved the detection of lung nodules on chest X-rays. The results of the study were published today in the journal Radiology, a journal of the Radiological Society of North America (RSNA). Pulmonary nodules, which are abnormal growths that form on the lungs, are very common and are usually formed from previous lung infections. In rare cases, it can be a sign of lung cancer. A chest X-ray is one of the common screening methods used to identify lung nodules. AI can be a powerful tool to help identify lung nodules, especially when radiologists are faced with a large number of conditions.

Pulmonary nodule detection, a fundamental discovery of lung cancer, is one of the crucial tasks of chest X-rays. Several studies have suggested that AI-based computer-assisted detection software can improve radiologist performance, but it is not widely used." Jin Moo-guo, Ph.D., Ph.D., co-author of the study, Department of Radiology at Seoul National University Hospital in Korea

To determine the actual impact of AI in clinical practice, researchers included 10,476 patients with an average age of 59, who underwent a chest X-ray at the health screening center between June 2020 and December 2021. "Since our trial was conducted in a hands-on style, almost all registered participants were included, which is a real clinical setting," said Dr. Guo. Patients completed a self-reported health questionnaire to identify basic characteristics such as age, gender, smoking status, and previous history of lung cancer. 11% of patients were current or former smokers.

Patients were randomly divided equally into two groups-; AI or non-AI. X-rays of the first group were analyzed by radiologists with the help of artificial intelligence while X-rays of the second group were interpreted without AI results. Solid nodules with a diameter of more than 8 mm or solid nodules with a solid fraction larger than 6 mm were identified as executable, which means that nodules require follow-up according to lung cancer screening standards. Pulmonary nodules were identified in 2% of patients. The analysis showed that the detection rate of executable lung nodules on chest X-rays was higher with AI (0.59%) than without AI (0.25%). There were no differences in misreferral rates between AI and unexplained AI groups.While advanced age and history of lung cancer or tuberculosis are associated with positive reports, these and other health characteristics have had no impact on the effectiveness of the AI system. This suggests that AI may be consistently working across different populations, even for those with a diseased lung or after surgery. Dr Guo said: "Our study provided strong evidence that AI can really help interpret chest radiography. This will contribute to the identification of chest diseases, especially lung cancer, more effectively at an early stage." The researchers plan to conduct a similar study using chest tomography which will also determine clinical outcomes and workflow efficiency.

CANCER RESEARCH EBOOK

Cancer Research Industry Focus eBook

Compilation of the top interviews, articles, and news in the last year.