Tmall Unveils MAOSS AI Model for Fatty Liver Disease Screening
Alibaba DAMO Academy has announced the joint development of an AI model for fatty liver screening, named MAOSS, in collaboration with several institutions including Shengjing Hospital of China Medical University and Nanjing University's Gulou Hospital. The study's findings were published in the prestigious international journal, Nature Communications, in February of this year.
Fatty liver disease affects over 30% of the population. Early symptoms are often mild and easily overlooked, allowing the condition to progress to liver fibrosis or cirrhosis. Traditional screening methods like B-ultrasound have limited sensitivity, while more specialized examinations are expensive. This frequently results in missed diagnoses for high-risk patients in clinical settings.

Key breakthroughs and advantages of the MAOSS model:
Deep analysis of non-contrast CT scans: DAMO Academy employed "non-contrast CT + AI" technology, enabling the AI to automatically extract high-dimensional features such as liver texture and density. This marks the first time liver steatosis and fibrosis staging can be assessed simultaneously using only a standard non-contrast CT scan.
Diagnostic accuracy exceeding physicians: In multi-center validation trials, the MAOSS model achieved an area under the curve (AUC) of 0.904-0.917 for liver steatosis staging, significantly higher than the average radiologist score of 0.709.
Doubled detection rate for high-risk cases: For the critical window to prevent cirrhosis (stage 2 fibrosis), the model identified 52.4% of high-risk patients, compared to a mere 16.6% detected through traditional clinical pathways—more than doubling the detection rate.
Early warning for cirrhosis risk: Follow-up data indicates that patients flagged as high-risk by MAOSS had a 45.5% probability of developing cirrhosis within two years, a rate far exceeding that of the low-risk group.
Experts from DAMO Academy noted that the model can leverage existing non-contrast CT data from routine physical exams and outpatient visits. This facilitates "front-end prevention" in chronic liver disease management without incurring additional patient costs. In the future, primary care hospitals could use this AI technology to provide high-risk warnings during standard health screenings, enabling earlier detection and potential reversal of the condition.
Related article
AI Experts Deployed: Large Models Take Over Factories, Industrial Manufacturing Enters New Evolution
On the front lines of biological fermentation, architectural design, and even wastewater treatment, a new kind of "employee" is quietly reshaping traditional manufacturing. These aren't workers covered in sweat—they're industrial time-series control
Google Photos brings Clueless's iconic closet to life with AI
Google Photos announced a new AI-powered feature on Wednesday that will soon turn photos of your clothes into a digital closet, letting you create fresh outfit combinations and even virtually try them on. The concept clearly draws inspiration from Ch
Red Fruit Short Drama Accused of Using AI to Steal Ordinary People’s Faces; No Official Response
The short video industry is currently facing a controversy involving AI-related infringement. Red Fruit Short Drama’s production "The Peach Hairpin" has been accused of using AI to "steal faces"—taking the likenesses of ordinary people without permis
Related Special Topic Recommendations
Comments (0)
0/500
Alibaba DAMO Academy has announced the joint development of an AI model for fatty liver screening, named MAOSS, in collaboration with several institutions including Shengjing Hospital of China Medical University and Nanjing University's Gulou Hospital. The study's findings were published in the prestigious international journal, Nature Communications, in February of this year.
Fatty liver disease affects over 30% of the population. Early symptoms are often mild and easily overlooked, allowing the condition to progress to liver fibrosis or cirrhosis. Traditional screening methods like B-ultrasound have limited sensitivity, while more specialized examinations are expensive. This frequently results in missed diagnoses for high-risk patients in clinical settings.

Key breakthroughs and advantages of the MAOSS model:
Deep analysis of non-contrast CT scans: DAMO Academy employed "non-contrast CT + AI" technology, enabling the AI to automatically extract high-dimensional features such as liver texture and density. This marks the first time liver steatosis and fibrosis staging can be assessed simultaneously using only a standard non-contrast CT scan.
Diagnostic accuracy exceeding physicians: In multi-center validation trials, the MAOSS model achieved an area under the curve (AUC) of 0.904-0.917 for liver steatosis staging, significantly higher than the average radiologist score of 0.709.
Doubled detection rate for high-risk cases: For the critical window to prevent cirrhosis (stage 2 fibrosis), the model identified 52.4% of high-risk patients, compared to a mere 16.6% detected through traditional clinical pathways—more than doubling the detection rate.
Early warning for cirrhosis risk: Follow-up data indicates that patients flagged as high-risk by MAOSS had a 45.5% probability of developing cirrhosis within two years, a rate far exceeding that of the low-risk group.
Experts from DAMO Academy noted that the model can leverage existing non-contrast CT data from routine physical exams and outpatient visits. This facilitates "front-end prevention" in chronic liver disease management without incurring additional patient costs. In the future, primary care hospitals could use this AI technology to provide high-risk warnings during standard health screenings, enabling earlier detection and potential reversal of the condition.
AI Experts Deployed: Large Models Take Over Factories, Industrial Manufacturing Enters New Evolution
On the front lines of biological fermentation, architectural design, and even wastewater treatment, a new kind of "employee" is quietly reshaping traditional manufacturing. These aren't workers covered in sweat—they're industrial time-series control
Google Photos brings Clueless's iconic closet to life with AI
Google Photos announced a new AI-powered feature on Wednesday that will soon turn photos of your clothes into a digital closet, letting you create fresh outfit combinations and even virtually try them on. The concept clearly draws inspiration from Ch
Red Fruit Short Drama Accused of Using AI to Steal Ordinary People’s Faces; No Official Response
The short video industry is currently facing a controversy involving AI-related infringement. Red Fruit Short Drama’s production "The Peach Hairpin" has been accused of using AI to "steal faces"—taking the likenesses of ordinary people without permis





Home






