AI Model Detects Fatty Liver Disease Using Routine Chest X-Rays

Researchers are leveraging artificial intelligence to identify fatty liver disease through routine medical imaging, potentially transforming how a silent global epidemic is screened. A team at Osaka Metropolitan University has developed a deep learning model capable of detecting the condition using standard chest X-rays, which capture the upper portion of the liver.

The AI model achieved an Area Under the Curve (AUC) of 0.82 to 0.83. To reach this level of accuracy, the researchers utilized a retrospective study consisting of 6,599 chest X-ray images sourced from 4,414 patients. This approach allows for the detection of Metabolic dysfunction-associated steatotic liver disease (MASLD)—the updated term for fatty liver disease—without requiring specialized liver-specific scans.

The Silent Epidemic

Fatty liver disease is characterized by the accumulation of excess fat in the liver and is estimated to affect approximately 25% to 30% of the global adult population. The condition is particularly dangerous because it often remains asymptomatic in its early stages. If left undetected, simple steatosis can progress to inflammation, fibrosis, and cirrhosis, eventually leading to liver cancer.

Traditionally, diagnosing the condition requires specialized and often expensive imaging, such as MRI, CT scans, or ultrasound. Because these tools are not typically used for general population screening, many patients only discover they have the disease by accident or after the condition has reached an advanced, irreversible stage.

Implications for Public Health

Shifting the detection of MASLD to routine, inexpensive tests like chest X-rays could enable mass screening and earlier medical intervention. By identifying at-risk patients during standard health checks, clinicians may be able to prevent millions of cases from progressing to liver failure.

Professor Sawako Uchida-Kobayashi noted that the development of diagnostic methods using easily obtainable and inexpensive chest X-rays has the potential to significantly improve fatty liver detection. This shift toward accessible AI-driven screening could reduce the reliance on costly specialized imaging for initial detection.

Next Steps

While the current model shows strong predictive power, the transition from a retrospective study to clinical practice will require further validation in diverse patient populations. Observers will be watching to see if this AI tool can be integrated into standard radiology workflows to provide automated alerts for liver health during routine thoracic imaging.