AI sees what doctors miss: Fatty liver disease hidden in chest x-rays

 

Fatty liver disease, caused by the accumulation of fat in the liver, is estimated to affect one in four people worldwide. If left untreated, it can lead to serious complications, such as cirrhosis and liver cancer, making it crucial to detect early and initiate treatment.

Currently, standard tests for diagnosing fatty liver disease include ultrasounds, CTs, and MRIs, which require costly specialized equipment and facilities. In contrast, chest X-rays are performed more frequently, are relatively inexpensive, and involve low radiation exposure. Although this test is primarily used to examine the condition of the lungs and heart, it also captures part of the liver, making it possible to detect signs of fatty liver disease. However, the relationship between chest X-rays and fatty liver disease has rarely been a subject of in-depth study.

Therefore, a research group led by Associate Professor Sawako Uchida-Kobayashi and Associate Professor Daiju Ueda at Osaka Metropolitan University's Graduate School of Medicine developed an AI model that can detect the presence of fatty liver disease from chest X-ray images.

In this retrospective study, a total of 6,599 chest X-ray images containing data from 4,414 patients were used to develop an AI model utilizing controlled attenuation parameter (CAP) scores. The AI model was verified to be highly accurate, with the area under the receiver operating characteristic curve (AUC) ranging from 0.82 to 0.83.

"The development of diagnostic methods using easily obtainable and inexpensive chest X-rays has the potential to improve fatty liver detection. We hope it can be put into practical use in the future," stated Professor Uchida-Kobayashi.

 

Journal Reference:

1. Daiju Ueda, Sawako Uchida-Kobayashi, Akira Yamamoto, Shannon L. Walston, Hiroyuki Motoyama, Hideki Fujii, Toshio Watanabe, Yukio Miki, Norifumi Kawada. Performance of a Chest Radiograph–based Deep Learning Model for Detecting Hepatic Steatosis. Radiology: Cardiothoracic Imaging, 2025; 7 (3) DOI: 10.1148/ryct.240402 

Courtesy:

Osaka Metropolitan University. "AI sees what doctors miss: Fatty liver disease hidden in chest x-rays." ScienceDaily. ScienceDaily, 27 June 2025. <www.sciencedaily.com/releases/2025/06/250627021845.htm>.

 

 

 

Scientists warn of bat virus just one mutation from infecting humans

 

A group of MERS-related bat viruses could mutate to infect humans. One type, HKU5, is especially concerning due to its ability to bind to ACE2 receptors in animals—and potentially humans next. Credit: Shutterstock

A group of bat viruses closely related to the deadly Middle East respiratory syndrome coronavirus (MERS-CoV) could be one small mutation away from being capable of spilling over into human populations and potentially causing the next pandemic.

A recent study published in the journal Nature Communicationsexamined an understudied group of coronaviruses known as merbecoviruses -- the same viral subgenus that includes MERS-CoV -- to better understand how they infect host cells. The research team, which included scientists at Washington State University, the California Institute of Technology and the University of North Carolina, found that while most merbecoviruses appear unlikely to pose a direct threat to people, one subgroup known as HKU5 possesses concerning traits.

"Merbecoviruses - and HKU5 viruses in particular - really hadn't been looked at much, but our study shows how these viruses infect cells," said Michael Letko, a virologist at WSU's College of Veterinary Medicine who helped to spearhead the study. "What we also found is HKU5 viruses may be only a small step away from being able to spill over into humans."

During the past two decades, scientists have cataloged the genetic sequences of thousands of viruses in wild animals, but, in most cases, little is known about whether these viruses pose a threat to humans. Letko's lab in WSU's Paul G. Allen School for Global Health focuses on closing that gap and identifying potentially dangerous viruses.

For their most recent study, Letko's team targeted merbecoviruses, which have received limited attention apart from MERS-CoV, a zoonotic coronavirus first noted in 2012 that is transmitted from dromedary camels to humans. It causes severe respiratory disease and has a mortality rate of approximately 34%.

Like other coronaviruses, merbecovirusesrely on a spike protein to bind to receptors and invade host cells. Letko's team used virus-like particles containing only the portion of the spike responsible for binding to receptors and tested their ability to infect cells in the lab. While most merbecoviruses appear unlikely to be able to infect humans, HKU5 viruses - which have been found across Asia, Europe, Africa and the Middle East - were shown to use a host receptor known as ACE2, the same used by the more well-known SARS-CoV-2 virus that causes COVID-19. One small difference: HKU5 viruses, for now, can only use the ACE2 gene in bats, but do not use the human version nearly as well.

Examining HKU5 viruses found in Asia where their natural host is the Japanese house bat (Pipistrellus abramus), the researchers demonstrated some mutations in the spike protein that may allow the viruses to bind to ACE2 receptors in other species, including humans. Researchers on another study that came out earlier this year analyzed one HKU5 virus in China that has already been documented to have jumped into minks, showing there is potential for these viruses to cross species-barriers.

"These viruses are so closely related to MERS, so we have to be concerned if they ever infect humans," Letko said. "While there's no evidence they've crossed into people yet, the potential is there -- and that makes them worth watching."

The team also used artificial intelligence to explore the viruses. WSU postdoctoral researcher Victoria Jefferson used a program called AlphaFold 3 to model how the HKU5 spike protein binds to ACE2 at the molecular level, which could help provide a better understanding of how antibodies might block the infection or how the virus could mutate.

Up until this point, such structural analysis required months of lab work and specialized equipment. With AlphaFold, Jefferson generated accurate predictions in minutes. The results matched those recently documented by a research team that used traditional approaches.

Letko noted the study and its methods could be used for future research projects and aid in the development of new vaccines and treatments.

The research was funded through a research project grant from the National Institutes of Health. Jefferson's work was supported by an NIH T32 training grant.

Journal Reference:

1. Nicholas J. Catanzaro, Ziyan Wu, Chengcheng Fan, Victoria Jefferson, Anfal Abdelgadir, Alexandra Sch&auml;fer, Boyd L. Yount, Pamela J. Bjorkman, Ralph Baric, Michael Letko. ACE2 from Pipistrellus abramus bats is a receptor for HKU5 coronaviruses. Nature Communications, 2025; 16 (1) DOI: 10.1038/s41467-025-60286-3 

Courtesy: 

Washington State University. "Scientists warn of bat virus just one mutation from infecting humans." ScienceDaily. ScienceDaily, 12 June 2025. <www.sciencedaily.com/releases/2025/06/250612081312.htm>.