Artificial intelligence sees what doctors miss: oily liver disease hidden in X -rays

Currently, standard tests for diagnosis of lipopries include ultrasound, CTS, and magnetic resonance imaging, which require expensive specialized equipment and facilities. In contrast, the X -rays on the chest are performed frequently, are relatively inexpensive, and include low radiation exposure. Although this test is used primarily to study the condition of the lungs and the heart, it picks up part of the liver, which makes it possible to discover signs of fatty liver disease. However, the relationship between x -rays and fatty liver diseases rarely was rarely the subject of in -depth study.

Therefore, a research group led by the assistant professor Sawako Uchida-kobayashi and co-professor Daiju UEDA at the Faculty of Medicine at Osaka Metropolitan University as a model for the IQ can discover the presence of sebaceous liver disease from x-rays.

In this study retroactively, a total of 6,599 images of X -rays that contain data of 4,414 patients were used to develop a controllers of the controlled mantle (CAP). The artificial intelligence model has been checked to be very accurate, as the distinctive curve of the future (AUC) ranges from 0.82 to 0.83.

“The development of diagnostic methods using X -rays that can be easily obtained and inexpensive has the ability to improve the detection of oily liver. We hope that it is used in a practical way in the future,” Professor Oshida Kobici said.