Artificial intelligence may be better than humans at seeing lung cancer on X-rays

Artificial intelligence may be better than humans at seeing lung cancer on X-rays

Seoul, South Korea – Artificial intelligence (AI) may outperform humans in screening for lung cancer, one of the deadliest forms of the disease. One high-performance machine learning tool showed increased sensitivity, meaning it was less likely to miss cases that required immediate intervention.

The tool also improved radiologists’ performance in detecting tumors on chest X-rays and promoted greater human acceptance of AI technology. Recent studies have highlighted how useful AI is for doctors when it comes to medical diagnosis.

“The relatively large sample size in this study likely enhanced readers’ confidence in AI suggestions,” says Dr. Chang Min Park of Seoul National University, lead author of the study, in a media release. “We think the issue of human trust in AI is what we observed in vulnerability in this study: humans are more vulnerable to AI when using high-performance diagnostic AI.”

The study showed that the successful identification of cancerous lung nodules by chest X-ray increased significantly to 94% when using computer data. These abnormal growths usually develop after an infection, but in rare cases, they can be a sign of cancer.

Chest radiographs obtained as part of a health examination in a 71-year-old patient demonstrate the reader's amenability to high-resolution artificial intelligence (AI).
(A-C) Chest radiographs obtained as part of a health examination in a 71-year-old male patient demonstrate reader aptitude for high-resolution diagnostic artificial intelligence (AI). In the first session without AI, a thoracic radiologist with 16 years of experience read the chest radiograph as a plain radiograph (A). High diagnostic accuracy The AI ​​observed the probability of lung cancer in the radiograph with a confidence interval of 89% as shown in the nodule localization map (B) (as the color changes from blue to red, the probability of nodules increases). When the AI ​​suggestion was presented in the second reading session, the radiologist changed the decision and explained lung cancer in the area overlapping the right hemidiaphragm (box annotation) (c). (D,E) Contrast-enhanced CT of the chest shows a 6.8-cm pulmonary mass (arrow) with air bronchogram in the right lower lobe in the axial (D) and coronal (E) planes. Pathologically, this mass was proven to be an invasive mucinous adenocarcinoma. Therefore, the reader’s decision was incorrect in the first session but was correct in the second session after following the AI’s suggestion. (Credit: Radiological Society of North America)

In the study, 20 chest radiologists with 18 years of experience and 10 students analyzed 120 chest X-rays without the help of artificial intelligence. Half of these X-rays were from male lung cancer patients in South Korea, and the remaining half served as controls. In a follow-up session, each group re-evaluated the X-rays, with the help of high- or low-resolution AI. Unbeknownst to readers, two different AI systems were used.

The study results underscore the importance of using high-performance diagnostic AI. However, the researchers note that definitions can vary based on the task at hand and the specific clinical context.

While an AI model that can identify every anomaly may seem flawless, its practical application in reducing workload in a mass screening scenario will be limited. “Therefore, our study suggests that clinically appropriate use of AI requires the development of high-performance AI models for specific tasks and considerations about the relevant clinical environment to which the AI ​​will be applied,” says Dr. Park.

The researchers plan to expand their work on human-AI collaboration to include other abnormalities detected in chest X-rays and CT scans. Lung cancer is often diagnosed too late because the initial symptoms are not present. Previous studies have shown the ability of artificial intelligence to predict colon disease, Alzheimer’s disease, heart attacks, and dementia.

The study is published in the journal X-rays.

Southwest News Service writer Mark Waghorn contributed to this report.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *