L-18:Automated Image Analysis Technology for Quantifying the “Connectivity” of Liver Fibrosis Lesions

HOME›事業紹介›徳島大学シーズ›The Tokushima University Seeds Collection›L-18:Automated Image Analysis Technology for Quantifying the “Connectivity” of Liver Fibrosis Lesions

│Return to list│Previous page│Next page│

<L-18>:Life Science, Pharmaceutical Sciences sdgs-03 
 

Automated Image Analysis Technology for Quantifying the “Connectivity” of Liver Fibrosis Lesions
― Toward the Establishment of a Diagnostic Support System for Pathology ―

Mayuko Shimizu Associate Professor
Graduate School of Biomedical Sciences, Medical Field, Medical Science, Pathological Sciences, Department of Pathology and Laboratory Medicine
Keywords Liver fibrosis, MASH, MASLD, pathological diagnosis
Laboratory URL https://plaza.umin.ac.jp/pathol2/
Research Overview
● Limited Quantitative Capability of Conventional Histopathological Assessment of Liver Fibrosis
Chronic hepatitis induces liver fibrosis, which can ultimately lead to liver cancer and liver failure; therefore, accurate assessment of liver fibrosis is essential. However, conventional histopathological evaluation methods are based on morphological observation of selected regions where lesions are prominent, relying on localized, visual inspection at high magnification. As a result, these evaluations depend heavily on the experience and subjective judgment of pathologists. This places a significant psychological burden on pathologists who bear responsibility for diagnosis, highlighting the importance of objective support for their decision-making.

● Quantifying Fibrotic Lesions in Digital Pathology Images Using Geometric Connectivity
Using whole-slide digital images of pathological specimens, we established a method to evaluate the “connectivity” of fibrotic lesions by focusing not on individual pixels but on geometric feature descriptors of tissue structure. This “connectivity” index enables assessment of liver fibrosis with performance comparable to conventional staging by pathologists, while offering the advantage of continuous numerical quantification. Furthermore, because fibrosis evaluation can be performed automatically, this method is expected to be highly useful as a diagnostic support tool for pathologists.
 
Potential Applications and Commercialization Prospects

● Application to a Pathological Diagnostic Support System for Liver Fibrosis
This method enables automated processing of liver fibrosis images from pathological specimens, providing highly quantitative and objective indicators that can support the diagnostic work of pathologists. Currently, methods capable of evaluating liver fibrosis as continuous numerical values include blood-based fibrosis biomarkers and elastography-based liver stiffness measurements; however, these methods are insufficient for discriminating early-stage or mild fibrosis. In contrast, this technique directly evaluates fibrosis within liver tissue itself and can quantify fibrosis continuously across stages from mild to severe, allowing for more accurate assessment of disease pathology. In drug discovery, this method is expected to enable precise evaluation of therapeutic effects, even in cases of subtle fibrosis improvement.

slide1.JPG
slide2.JPG
slide3.JPG
Patents ● Japanese Patent Publication No. 2026-060461 “Computer Program and Control Device”

Last updated date:October 15, 2026

 

Simple Inquiry Technical consultation Academic guidance system Joint research/commissioned research

 

 

PAGE TOP