scholarly journals Deep learning enables pathologist-like scoring of NASH models

2019 ◽  
Vol 9 (1) ◽  
Author(s):  
Fabian Heinemann ◽  
Gerald Birk ◽  
Birgit Stierstorfer

AbstractNon-alcoholic fatty liver disease (NAFLD) and the progressive form of non-alcoholic steatohepatitis (NASH) are diseases of major importance with a high unmet medical need. Efficacy studies on novel compounds to treat NAFLD/NASH using disease models are frequently evaluated using established histological feature scores on ballooning, inflammation, steatosis and fibrosis. These features are assessed by a trained pathologist using microscopy and assigned discrete scores. We demonstrate how to automate these scores with convolutional neural networks (CNNs). Whole slide images of stained liver sections are analyzed using two different scales with four CNNs, each specialized for one of four histopathological features. A continuous value is obtained to quantify the extent of each feature, which can be used directly to provide a high resolution readout. In addition, the continuous values can be mapped to obtain the established discrete pathologist-like scores. The automated deep learning-based scores show good agreement with the trainer - a human pathologist.

2018 ◽  
Vol 107 ◽  
pp. 979-988 ◽  
Author(s):  
Erenius Toppo ◽  
S. Sylvester Darvin ◽  
S. Esakkimuthu ◽  
K. Buvanesvaragurunathan ◽  
T.P. Ajeesh Krishna ◽  
...  

2021 ◽  
pp. 153537022110092
Author(s):  
Manush Saydmohammed ◽  
Anupma Jha ◽  
Vineet Mahajan ◽  
Dillon Gavlock ◽  
Tong Ying Shun ◽  
...  

Metabolic syndrome is a complex disease that involves multiple organ systems including a critical role for the liver. Non-alcoholic fatty liver disease (NAFLD) is a key component of the metabolic syndrome and fatty liver is linked to a range of metabolic dysfunctions that occur in approximately 25% of the population. A panel of experts recently agreed that the acronym, NAFLD, did not properly characterize this heterogeneous disease given the associated metabolic abnormalities such as type 2 diabetes mellitus (T2D), obesity, and hypertension. Therefore, metabolic dysfunction-associated fatty liver disease (MAFLD) has been proposed as the new term to cover the heterogeneity identified in the NAFLD patient population. Although many rodent models of NAFLD/NASH have been developed, they do not recapitulate the full disease spectrum in patients. Therefore, a platform has evolved initially focused on human biomimetic liver microphysiology systems that integrates fluorescent protein biosensors along with other key metrics, the microphysiology systems database, and quantitative systems pharmacology. Quantitative systems pharmacology is being applied to investigate the mechanisms of NAFLD/MAFLD progression to select molecular targets for fluorescent protein biosensors, to integrate computational and experimental methods to predict drugs for repurposing, and to facilitate novel drug development. Fluorescent protein biosensors are critical components of the platform since they enable monitoring of the pathophysiology of disease progression by defining and quantifying the temporal and spatial dynamics of protein functions in the biosensor cells, and serve as minimally invasive biomarkers of the physiological state of the microphysiology system experimental disease models. Here, we summarize the progress in developing human microphysiology system disease models of NAFLD/MAFLD from several laboratories, developing fluorescent protein biosensors to monitor and to measure NAFLD/MAFLD disease progression and implementation of quantitative systems pharmacology with the goal of repurposing drugs and guiding the creation of novel therapeutics.


Author(s):  
Jeniffer Danielle M. Dutra ◽  
Quelson Coelho Lisboa ◽  
Silvia Marinho Ferolla ◽  
Carolina Martinelli M. L. Carvalho ◽  
Camila Costa M. Mendes ◽  
...  

Abstract. Some epidemiological evidence suggests an inverse correlation between non-alcoholic fatty liver disease (NAFLD) frequency and vitamin D levels. Likewise, a beneficial effect of vitamin D on diabetes mellitus (DM) and insulin resistance has been observed, but this is an unsolved issue. Thus, we aimed to investigate the prevalence of hypovitaminosis D in a NAFLD Brazilian population and its association with disease severity and presence of comorbidities. In a cross-sectional study, the clinical, biochemical and histological parameters of 139 NAFLD patients were evaluated according to two different cut-off points of serum 25-hydroxyvitamin D levels (20 ng/mL and 30 ng/mL). The mean age of the population was 56 ± 16 years, most patients were female (83%), 72% had hypertension, 88% dyslipidemia, 46% DM, 98% central obesity, and 82% metabolic syndrome. Serum vitamin D levels were < 30 ng/mL in 78% of the patients, and < 20 ng/mL in 35%. The mean vitamin D level was 24.3 ± 6.8 ng/mL. The comparison between the clinical, biochemical and histological characteristics of the patients according to the levels of vitamin D showed no significant difference. Most patients with NAFLD had hypovitaminosis D, but low vitamin D levels were not related to disease severity and the presence of comorbidities.


2008 ◽  
Vol 78 (1) ◽  
pp. 27-32 ◽  
Author(s):  
Suano de Souza ◽  
Silverio Amancio ◽  
Saccardo Sarni ◽  
Sacchi Pitta ◽  
Fernandes ◽  
...  

Objectives: To evaluate the frequency of non-alcoholic fatty liver disease, the retinol serum levels, lipid profile, and insulin resistance in overweight/obese children. To relate these biochemical variables with the risk of this disease in the population studied. Methods: The study was cross-sectional and prospective, with 46 overweight/obese school children (28 female, 18 male; mean age 8.6 years). The control group consisted of 45 children, paired by age and gender. Hepatic steatosis, evaluated by ultrasound, was classified as normal, mild, moderate, or severe. Also evaluated were serum retinol levels; thiobarbituric acid reactive substances; lipid profile; and fasting glucose and serum insulin levels, used for the calculation of the Homeostasis Model Assessment. Results: Hepatic ultrasound alterations were found in 56.5% and 48,9% of the overweight/obese and control group children, respectively. Presence of obesity was associated with high levels of triglycerides (OR = 4.6; P = 0.002). In the studied children, the risk of steatosis was related to a trend to a higher percentage of retinol inadequacy (OR = 2.8; p = 0.051); there was no association with thiobarbituric acid reactive substances, lipid profile, or insulin resistance. Conclusions: The high frequency of non-alcoholic fatty liver disease in both groups, evaluated by hepatic ultrasound, in low-socioeconomic level children, independent of nutritional condition and without significant association with insulin resistance, emphasizes that especially in developing countries, other risk factors such as micronutrient deficiencies (e.g. vitamin A) are involved.


Sign in / Sign up

Export Citation Format

Share Document