NARLabs and NCKU Launch AI Platform to Accelerate MASH Diagnosis and Drug Development
The National Laboratory Animal Center (NLAC) of the National Applied Research Laboratories (NARLabs), in collaboration with National Cheng Kung University (NCKU), launched the "AI+MASH Pathology Quantification Platform" on June 4, 2025. This platform uses AI image analysis to reduce the time for generating mouse liver pathology reports from the traditional 2 months to just 2 weeks, achieving 98% accuracy. It aims to accelerate the screening and development of drugs for Metabolic dysfunction-Associated Steatohepatitis (MASH) and boost Taiwan's biotech competitiveness.
📋 Article Processing Timeline
- 📰 Published: June 4, 2026 at 21:40
- 🔍 Collected: June 4, 2026 at 21:58 (18 min after Published)
- 🤖 AI Analyzed: June 6, 2026 at 15:29 (41h 30m after Collected)
(Central News Agency, Reporter Zhao Minya, Taipei, June 4) Metabolic dysfunction-Associated Steatohepatitis (MASH) has become the leading cause of liver cancer in Taiwan. The National Laboratory Animal Center (NLAC) of the National Applied Research Laboratories (NARLabs), in collaboration with National Cheng Kung University (NCKU), today announced the "AI+MASH Pathology Quantification Platform." This platform utilizes AI image analysis technology to shorten the time required for generating mouse liver pathology reports, enabling research teams to quickly evaluate the intervention effects of drugs on disease progression and enhance the efficiency of MASH drug screening.
The NLAC of NARLabs and the Department of Electrical Engineering at NCKU, led by Distinguished Professor Pau-Choo Chung, collaborated across disciplines to launch the "AI+MASH Pathology Quantification Platform." The platform integrates the preclinical animal testing expertise of NARLabs' NLAC, the AI image analysis technology of the NCKU team, and the high-performance computing resources of NARLabs' National Center for High-performance Computing (NCHC). This integration achieves the advantages of panoramic, objective, high-throughput, and standardized pathological analysis.
NARLabs explained that fatty liver can progress to MASH, then to fibrosis, and eventually to cirrhosis and liver cancer. If appropriate drug treatment is administered during the MASH or even early fibrosis stage, there is an opportunity to restore health. The "AI+MASH Pathology Quantification Platform" can precisely assess the degree of fatty liver, MASH, and fibrosis in experimental mice, and can be used to test the efficacy of various drugs.
NARLabs pointed out that the traditional evaluation method involves a pathologist examining liver pathology slides, selecting a small portion for semi-quantitative scoring to determine the amount of lipid droplets, the density of inflammatory foci, and the extent of fibrosis, in order to infer the pathological progression. The "AI+MASH Pathology Quantification Platform" uses panoramic scanning and automated AI analysis to examine mouse liver pathology slides.
Dr. Yu-Jia Su, a researcher at NARLabs' NLAC, stated that through the human-machine collaboration of the "AI+MASH Platform," AI automatically analyzes the entire area of the liver pathology slide. This expands from local sampling to panoramic, objective analysis, removing subjective errors and establishing absolute statistical standards. The accuracy rate of the platform's analysis matches that of professional manual interpretation by up to 98%.
Dr. Su stated that the quantitative report can be generated in just 2 weeks, after which it is confirmed by manual analysis. Compared to the traditional fully manual interpretation process, which takes at least 2 months, the AI-human collaboration reduces the time to one-quarter, significantly increasing the capacity for MASH pathological diagnosis.
NARLabs explained that the AI model can precisely quantify the accumulation of lipid droplets, the density of inflammatory foci, and the distribution ratio of steatotic hepatocytes and liver fibrosis across the entire slide. For example, a normal liver slide has only 3.07 inflammatory foci per square millimeter, while the MASH group has as many as 39.7. Research teams can use these four pathological quantifications to evaluate the intervention effect of drugs on disease progression.
NARLabs noted that the platform also provides a "Liver Panoramic View" function. Based on different models, it identifies various pathological features of MASH, marks their locations, and helps quickly locate lesions. This serves as an important indicator for assessing the severity and affected area of MASH, providing objective data for drug efficacy evaluation.
Dr. Hsien-Ching Chin, Director of NARLabs' NLAC, emphasized that this technology not only improves the efficiency of pathological analysis but also provides crucial support for MASH drug screening, accelerating the clinical translation of treatment plans. In the future, the platform will be continuously optimized, integrating more pathological models to support global MASH drug research and development, thereby enhancing the international competitiveness of Taiwan's biotechnology industry. (Editor: Yang Kaixiang) 1150604
The NLAC of NARLabs and the Department of Electrical Engineering at NCKU, led by Distinguished Professor Pau-Choo Chung, collaborated across disciplines to launch the "AI+MASH Pathology Quantification Platform." The platform integrates the preclinical animal testing expertise of NARLabs' NLAC, the AI image analysis technology of the NCKU team, and the high-performance computing resources of NARLabs' National Center for High-performance Computing (NCHC). This integration achieves the advantages of panoramic, objective, high-throughput, and standardized pathological analysis.
NARLabs explained that fatty liver can progress to MASH, then to fibrosis, and eventually to cirrhosis and liver cancer. If appropriate drug treatment is administered during the MASH or even early fibrosis stage, there is an opportunity to restore health. The "AI+MASH Pathology Quantification Platform" can precisely assess the degree of fatty liver, MASH, and fibrosis in experimental mice, and can be used to test the efficacy of various drugs.
NARLabs pointed out that the traditional evaluation method involves a pathologist examining liver pathology slides, selecting a small portion for semi-quantitative scoring to determine the amount of lipid droplets, the density of inflammatory foci, and the extent of fibrosis, in order to infer the pathological progression. The "AI+MASH Pathology Quantification Platform" uses panoramic scanning and automated AI analysis to examine mouse liver pathology slides.
Dr. Yu-Jia Su, a researcher at NARLabs' NLAC, stated that through the human-machine collaboration of the "AI+MASH Platform," AI automatically analyzes the entire area of the liver pathology slide. This expands from local sampling to panoramic, objective analysis, removing subjective errors and establishing absolute statistical standards. The accuracy rate of the platform's analysis matches that of professional manual interpretation by up to 98%.
Dr. Su stated that the quantitative report can be generated in just 2 weeks, after which it is confirmed by manual analysis. Compared to the traditional fully manual interpretation process, which takes at least 2 months, the AI-human collaboration reduces the time to one-quarter, significantly increasing the capacity for MASH pathological diagnosis.
NARLabs explained that the AI model can precisely quantify the accumulation of lipid droplets, the density of inflammatory foci, and the distribution ratio of steatotic hepatocytes and liver fibrosis across the entire slide. For example, a normal liver slide has only 3.07 inflammatory foci per square millimeter, while the MASH group has as many as 39.7. Research teams can use these four pathological quantifications to evaluate the intervention effect of drugs on disease progression.
NARLabs noted that the platform also provides a "Liver Panoramic View" function. Based on different models, it identifies various pathological features of MASH, marks their locations, and helps quickly locate lesions. This serves as an important indicator for assessing the severity and affected area of MASH, providing objective data for drug efficacy evaluation.
Dr. Hsien-Ching Chin, Director of NARLabs' NLAC, emphasized that this technology not only improves the efficiency of pathological analysis but also provides crucial support for MASH drug screening, accelerating the clinical translation of treatment plans. In the future, the platform will be continuously optimized, integrating more pathological models to support global MASH drug research and development, thereby enhancing the international competitiveness of Taiwan's biotechnology industry. (Editor: Yang Kaixiang) 1150604
FAQ
What is the main advantage of the AI+MASH Pathology Quantification Platform?
The main advantage is reducing the time to generate pathology reports from 2 months to 2 weeks while improving accuracy to 98%.
How does this platform help in developing MASH drugs?
It precisely quantifies lipid droplets, inflammation, and fibrosis in mouse livers, allowing for rapid evaluation of drug efficacy.
Does this technology replace human pathologists?
No, it is designed for human-machine collaboration, where AI performs the initial analysis and humans confirm the results.