Dataset Information

Noninvasive Diagnosis of Nonalcoholic Steatohepatitis and Advanced Liver Fibrosis Using Machine Learning Methods: Comparative Study With Existing Quantitative Risk Scores.

ABSTRACT:

Background

Nonalcoholic steatohepatitis (NASH), advanced fibrosis, and subsequent cirrhosis and hepatocellular carcinoma are becoming the most common etiology for liver failure and liver transplantation; however, they can only be diagnosed at these potentially reversible stages with a liver biopsy, which is associated with various complications and high expenses. Knowing the difference between the more benign isolated steatosis and the more severe NASH and cirrhosis informs the physician regarding the need for more aggressive management.

Objective

We intend to explore the feasibility of using machine learning methods for noninvasive diagnosis of NASH and advanced liver fibrosis and compare machine learning methods with existing quantitative risk scores.

Methods

We conducted a retrospective analysis of clinical data from a cohort of 492 patients with biopsy-proven nonalcoholic fatty liver disease (NAFLD), NASH, or advanced fibrosis. We systematically compared 5 widely used machine learning algorithms for the prediction of NAFLD, NASH, and fibrosis using 2 variable encoding strategies. Then, we compared the machine learning methods with 3 existing quantitative scores and identified the important features for prediction using the SHapley Additive exPlanations method.

Results

The best machine learning method, gradient boosting (GB), achieved the best area under the curve scores of 0.9043, 0.8166, and 0.8360 for NAFLD, NASH, and advanced fibrosis, respectively. GB also outperformed 3 existing risk scores for fibrosis. Among the variables, alanine aminotransferase (ALT), triglyceride (TG), and BMI were the important risk factors for the prediction of NAFLD, whereas aspartate transaminase (AST), ALT, and TG were the important variables for the prediction of NASH, and AST, hyperglycemia (A_1c), and high-density lipoprotein were the important variables for predicting advanced fibrosis.

Conclusions

It is feasible to use machine learning methods for predicting NAFLD, NASH, and advanced fibrosis using routine clinical data, which potentially can be used to better identify patients who still need liver biopsy. Additionally, understanding the relative importance and differences in predictors could lead to improved understanding of the disease process as well as support for identifying novel treatment options.

SUBMITTER: Wu Y

PROVIDER: S-EPMC9210198 | biostudies-literature | 2022 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Noninvasive Diagnosis of Nonalcoholic Steatohepatitis and Advanced Liver Fibrosis Using Machine Learning Methods: Comparative Study With Existing Quantitative Risk Scores.

Wu Yonghui Y Yang Xi X Morris Heather L HL Gurka Matthew J MJ Shenkman Elizabeth A EA Cusi Kenneth K Bril Fernando F Donahoo William T WT

JMIR medical informatics 20220606 6

<h4>Background</h4>Nonalcoholic steatohepatitis (NASH), advanced fibrosis, and subsequent cirrhosis and hepatocellular carcinoma are becoming the most common etiology for liver failure and liver transplantation; however, they can only be diagnosed at these potentially reversible stages with a liver biopsy, which is associated with various complications and high expenses. Knowing the difference between the more benign isolated steatosis and the more severe NASH and cirrhosis informs the physician ...[more]

PMID: 35666557

Similar Datasets

Project description:ImportanceNonalcoholic steatohepatitis (NASH) is projected to be the leading cause of liver transplantation by 2020. Advanced fibrosis (stage F3-F4) on liver biopsy independently predicts all-cause and liver-related mortality in NASH. There are no known efficacious treatments for advanced fibrosis related to NASH. Thiazolidinedione therapy has been extensively evaluated in NASH, and new randomized clinical trials (RCTs) of its efficacy have been completed.ObjectiveTo synthesize the evidence about the association of thiazolidinedione therapy with advanced liver fibrosis in NASH.Data sourcesMEDLINE, Ovid MEDLINE In-Process, Cochrane Library, EMBASE, clinicaltrials.gov, PubMed, and Scopus databases (without language restrictions), as well as other registries and scientific meeting presentations, from database inception through August 15, 2016.Study selectionRandomized clinical trials evaluating the effect of thiazolidinedione therapy on histologic features of the liver in biopsy-proven NASH.Data extraction and synthesisTwo investigators extracted study data independently and in duplicate and rated the risk of bias using the Cochrane Risk of Bias Tool.Main outcomes and measuresThe primary outcome was a dichotomous improvement in advanced fibrosis on liver biopsy, defined as an improvement in fibrosis stage from F3-F4 to F0-F2. Secondary outcomes were at least a 1-point improvement in fibrosis of any stage and NASH resolution. This meta-analysis also evaluated adverse effects of thiazolidinedione therapy, including weight gain, lower limb edema, congestive heart failure, bone fractures, cancer, and anemia. With the use of random-effects models, dichotomous variables are presented as odds ratios (ORs) with 95% CIs, and continuous variables are presented as weighted mean differences with 95% CIs.ResultsThis study analyzed 8 RCTs (5 evaluating pioglitazone use and 3 evaluating rosiglitazone maleate use) enrolling 516 patients with biopsy-proven NASH for a duration of 6 to 24 months. Among all studies combined, thiazolidinedione therapy was associated with improved advanced fibrosis (OR, 3.15; 95% CI, 1.25-7.93; P = .01; I2 = 0%), fibrosis of any stage (OR, 1.66; 95% CI, 1.12-2.47; P = .01; I2 = 0%), and NASH resolution (OR, 3.22; 95% CI, 2.17-4.79; P < .001; I2 = 0%). Analyses restricted to RCTs enrolling patients without diabetes yielded similar results for improvement in advanced fibrosis (OR, 2.95; 95% CI, 1.04-10.90; P = .02; I2 = 0%), improvement in fibrosis of any stage (OR, 1.76; 95% CI, 1.02-3.03; P = .02; I2 = 0%), and NASH resolution (OR, 3.40; 95% CI, 1.95-5.93; P < .001; I2 = 0%). All effects were accounted for by pioglitazone use. Weight gain and lower limb edema occurred more frequently with thiazolidinedione therapy (initial body weight +2.70%; 95% CI, 1.96%-4.34%; P = .001). The small sample size of included RCTs prevented evaluation of more serious adverse effects of thiazolidinedione therapy.Conclusions and relevancePioglitazone use improves advanced fibrosis in NASH, even in patients without diabetes. Whether this finding translates to improvement in risk for clinical outcomes requires further study.

Project description:UnlabelledThe characteristics of nonalcoholic fatty liver disease (NAFLD) in elderly patients are unknown. Therefore, we aimed to examine the differences between elderly and nonelderly patients with NAFLD and to identify determinants of nonalcoholic steatohepatitis (NASH) and advanced fibrosis (bridging fibrosis or cirrhosis) in elderly patients. This is a cross-sectional analysis of adult participants who were prospectively enrolled in the NASH Clinical Research Network studies. Participants were included based on availability of the centrally reviewed liver histology data within 1 year of enrollment, resulting in 61 elderly (age ≥65 years) and 735 nonelderly (18-64 years) participants. The main outcomes were the presence of NASH and advanced fibrosis. Compared to nonelderly patients with NAFLD, elderly patients had a higher prevalence of NASH (56% versus 72%, P = 0.02), and advanced fibrosis (25% versus 44%, P = 0.002). Compared to nonelderly patients with NASH, elderly patients with NASH had higher rates of advanced fibrosis (35% versus 52%, P = 0.03), as well as other features of severe liver disease including the presence of ballooning degeneration, acidophil bodies, megamitochondria, and Mallory-Denk bodies (P ≤ 0.05 for each). In multiple logistic regression analyses, independent determinants of NASH in elderly patients included higher aspartate aminotransferase (AST) (odds ratio [OR] = 1.12, P = 0.007) and lower platelets (OR = 0.98, P = 0.02); and independent determinants of advanced fibrosis included higher AST (OR = 1.08, P = 0.007), lower alanine aminotransferase value (OR = 0.91, P = 0.002), and an increased odds of having low high-density lipoprotein (OR = 8.35, P = 0.02).ConclusionElderly patients are more likely to have NASH and advanced fibrosis than nonelderly patients with NAFLD. Liver biopsy may be considered in elderly patients and treatment should be initiated in those with NASH and advanced fibrosis.

Dataset Information

Noninvasive Diagnosis of Nonalcoholic Steatohepatitis and Advanced Liver Fibrosis Using Machine Learning Methods: Comparative Study With Existing Quantitative Risk Scores.

Background

Objective

Methods

Results

Conclusions

Publications

Noninvasive Diagnosis of Nonalcoholic Steatohepatitis and Advanced Liver Fibrosis Using Machine Learning Methods: Comparative Study With Existing Quantitative Risk Scores.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets