Quantification of Liver Fibrosis—A Comparative Study

Abstract

Liver disease has been targeted as the fifth most common cause of death worldwide and tends to steadily rise. In the last three decades, several publications focused on the quantification of liver fibrosis by means of the estimation of the collagen proportional area (CPA) in liver biopsies obtained from digital image analysis (DIA). In this paper, early and recent studies on this topic have been reviewed according to these research aims: the datasets used for the analysis, the employed image processing techniques, the obtained results, and the derived conclusions. The purpose is to identify the major strengths and “gray-areas” in the landscape of this topic.

Software-Assisted Image Analysis for Identification and Quantification of Hepatic Sinusoidal Dilatation and Centrilobular Fibrosis

“…The three main limitations of the current study are: 1) the small number of livers specimens to establish a reliable statistical correlation between quantitative and qualitative assessment; 2) not be fully automated, requiring a qualified technician to manually extraction of the sinusoidal lumen, being subjectively dependent on the visual interpretation of the observer; and 3) do not use sirius red stain, the ideal one to bind to collagen proteins 1 . Besides that, this method has two disadvantages: 1) limited information, image analysis provides a quantitative approach to liver fibrosis and cannot replace qualitative determination; 2) cost, image analysis is expensive, since there is a need to scan all slides 23 .…”

“…Hepatic fibrosis plays the most important role in the evolutionary process to cirrhosis, independent of the primary etiology. Even with advances in radiological imaging and the development of several putative serum and urinary markers of liver fibrosis, it’s evaluation in liver biopsy specimens remains the ‘gold standard’ for diagnosis, degree assessment and prognosis of liver diseases 1 . Sinusoidal dilatation, a more specific and frequent finding in venous outflow obstruction syndromes, is also very relevant for the evaluation of this condition 2 , 12 .…”

MicroRNA 200a inhibits liver fibrosis of schistosoma

“…100 mg of fresh mouse liver tissue was taken and put into 4% paraformaldehyde fixative. The fixed liver tissue was washed, dehydrated, transparent, embedded in wax, sliced (thickness of about 4 μm), and then stained by HE and Masson staining, respectively, [ 13 , 14 ]. HE staining was performed by dewaxing, ethanol elution, rinsing with running water, staining with hematoxylin staining solution for 6 min, 0.5% hydrochloric acid ethanol for 3 min, and 0.5% eosin staining solution for 1 min.…”

An Ensemble Learning Approach for Enhanced Classification of Patients With Hepatitis and Cirrhosis

“…Although more classical data sources are still used (for example, insurance claims [61]), EHRs represent now a solid base for building learning models upon, both for diagnostic purposes in chronic hepatitis C [9], [62]- [64] and in derived diseases such as liver cancer [65]. Notably, in the last few years several research groups worldwide have been actively involved in assessing and predicting the progression of chronic hepatitis C into fibrosis first and then cirrhosis, proposing several methodological alternatives to carefully stage such a progressively deteriorating condition, both by classic shallow models [61], [66]- [69] as well as by more advanced deep learning approaches such as Recurrent Neural Networks [70], marking another domain where Artificial Intelligence will definitely have an impact in the near future. If better predictive performances often characterize learning based approaches, on the other hand their higher complexity (and sometimes their difficult interpretability, too) constitutes an hindrance towards their adoption in clinical practice.…”