Label-free characterization of pathological changes in the portal area of liver fibrosis tissue using multiphoton imaging and quantitative image analysis

Abstract Liver fibrosis plays a crucial role in the progression of liver diseases and serves as a pivotal stage leading to the development of liver cirrhosis and cancer. It typically initiates from portal area with various pathological characteristics. In this article, we employed multiphoton microscopy (MPM) to characterize the pathological changes in the portal areas of liver fibrosis tissues, and subsequently, we used our developed image analysis method to extract eight collagen morphological features from MPM images and also combined a deep learning method with a cell nuclear feature extraction algorithm to perform automatic nuclei segmentation and quantitative analysis in the H&E-stained histopathology images of portal areas. Our results demonstrate that MPM can effectively identify various pathological features in portal areas, and there are significant differences in four collagen features (collagen proportionate area, number, length and width) between normal and abnormal portal areas and in four nuclear features (mean axes distance ratio, disorder of distance to 3, 5 and 7 nearest neighbors) between normal portal area, bile duct hyperplasia and periductal fibrosis. Therefore, a combination of MPM and image-based quantitative analysis may be considered as a rapid and effective means to monitor histopathological changes in portal area and offer new insights into liver fibrosis.