Predictors Of Fatty Liver Disease Research Articles

Comparison of Machine Learning Models and the Fatty Liver Index in Predicting Lean Fatty Liver

The reported prevalence of non-alcoholic fatty liver disease in studies of lean individuals ranges from 7.6% to 19.3%. The aim of the study was to develop machine-learning models for the prediction of fatty liver disease in lean individuals. The present retrospective study included 12,191 lean subjects with a body mass index < 23 kg/m2 who had undergone a health checkup from January 2009 to January 2019. Participants were divided into a training (70%, 8533 subjects) and a testing group (30%, 3568 subjects). A total of 27 clinical features were analyzed, except for medical history and history of alcohol or tobacco consumption. Among the 12,191 lean individuals included in the present study, 741 (6.1%) had fatty liver. The machine learning model comprising a two-class neural network using 10 features had the highest area under the receiver operating characteristic curve (AUROC) value (0.885) among all other algorithms. When applied to the testing group, we found the two-class neural network exhibited a slightly higher AUROC value for predicting fatty liver (0.868, 0.841-0.894) compared to the fatty liver index (FLI; 0.852, 0.824-0.81). In conclusion, the two-class neural network had greater predictive value for fatty liver than the FLI in lean individuals.

Application of machine learning in predicting non-alcoholic fatty liver disease using anthropometric and body composition indices

Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease, which can progress from simple steatosis to advanced cirrhosis and hepatocellular carcinoma. Clinical diagnosis of NAFLD is crucial in the early stages of the disease. The main aim of this study was to apply machine learning (ML) methods to identify significant classifiers of NAFLD using body composition and anthropometric variables. A cross-sectional study was carried out among 513 individuals aged 13 years old or above in Iran. Anthropometric and body composition measurements were performed manually using body composition analyzer InBody 270. Hepatic steatosis and fibrosis were determined using a Fibroscan. ML methods including k-Nearest Neighbor (kNN), Support Vector Machine (SVM), Radial Basis Function (RBF) SVM, Gaussian Process (GP), Random Forest (RF), Neural Network (NN), Adaboost and Naïve Bayes were examined for model performance and to identify anthropometric and body composition predictors of fatty liver disease. RF generated the most accurate model for fatty liver (presence of any stage), steatosis stages and fibrosis stages with 82%, 52% and 57% accuracy, respectively. Abdomen circumference, waist circumference, chest circumference, trunk fat and body mass index were among the most important variables contributing to fatty liver disease. ML-based prediction of NAFLD using anthropometric and body composition data can assist clinicians in decision making. ML-based systems provide opportunities for NAFLD screening and early diagnosis, especially in population-level and remote areas.

Prediction of Fatty Liver Disease in a Chinese Population Using Machine-Learning Algorithms.

Fatty liver disease (FLD) is an important risk factor for liver cancer and cardiovascular disease and can lead to significant social and economic burden. However, there is currently no nationwide epidemiological survey for FLD in China, making early FLD screening crucial for the Chinese population. Unfortunately, liver biopsy and abdominal ultrasound, the preferred methods for FLD diagnosis, are not practical for primary medical institutions. Therefore, the aim of this study was to develop machine learning (ML) models for screening individuals at high risk of FLD, and to provide a new perspective on early FLD diagnosis. This study included a total of 30,574 individuals between the ages of 18 and 70 who completed abdominal ultrasound and the related clinical examinations. Among them, 3474 individuals were diagnosed with FLD by abdominal ultrasound. We used 11 indicators to build eight classification models to predict FLD. The model prediction ability was evaluated by the area under the curve, sensitivity, specificity, positive predictive value, negative predictive value, and kappa value. Feature importance analysis was assessed by Shapley value or root mean square error loss after permutations. Among the eight ML models, the prediction accuracy of the extreme gradient boosting (XGBoost) model was highest at 89.77%. By feature importance analysis, we found that the body mass index, triglyceride, and alanine aminotransferase play important roles in FLD prediction. XGBoost improves the efficiency and cost of large-scale FLD screening.

Bone Metabolism Discrepancy in Type 2 Diabetes Mellitus Patients With and Without Non-Alcoholic Fatty Liver Disease

To explore the distribution of several bone metabolic indicators in type 2 diabetes patients (T2DM) with and without non-alcoholic fatty liver disease (NAFLD) and to preliminarily evaluate the relationship of bone metabolism with NAFLD in patients with T2DM. The hospitalized patients with T2DM were divided into the group of T2DM complicated with NAFLD and the group of T2DM alone according to the results of ultrasonic diagnosis. The general information and laboratory test data such as bone metabolism indexes of these patients were collected and the differences of the indexes between the 2 groups were compared. Furthermore, the independent influencing factors of NAFLD in patients with T2DM were analyzed. A total of 186 patients were included in the study. Compared with patients with T2DM only, patients with T2DM combined with NAFLD were characterized with younger age (p < 0.001), higher BMI (p = 0.016), ALT (p = 0.001), TG (p = 0.005), HOMA-IR (p = 0.005), and lower HDL-C (p = 0.031). Significant discrepancy of age (OR 1.052, p = 0.001), ALT (OR 0.964, p = 0.047), HOMA-IR (OR 0.801, p = 0.005), and T-PINP (OR 1.022, p = 0.008) was found using multivariate logistic regression model. Significant discrepancy of T-PINP was found in T2DM patients with and without NAFLD. Further studies are needed to explore whether T-PINP could be used as a predictor of fatty liver disease, osteoporosis, and other related complications in patients with T2DM.

A family history of type 2 diabetes as a predictor of fatty liver disease in diabetes-free individuals with excessive body weight

Comprehensive screening for non-alcoholic fatty liver disease (NAFLD) may help prompt clinical management of fatty liver disease. A family history, especially of diabetes, has been little studied as a predictor for NAFLD. We characterized the cross-sectional relationship between a family history of type 2 diabetes (FHT2D) and NAFLD probability in 1185 diabetes-free Apulian (Southern-Italy) subjects aged > 20 years with overweight or obesity not receiving any drug or supplementation. Clinical data and routine biochemistry were analysed. NAFLD probability was defined using the fatty liver index (FLI). A first-degree FHT2D was assessed by interviewing subjects and assigning a score of 0, 1, or 2 if none, only one, or both parents were affected by type 2 diabetes mellitus (T2DM). Our study population featured most females (70.9%, N = 840), and 48.4% (N = 574) of the sample had first-degree FHT2D. After dividing the sample by a FHT2D, we found a higher BMI, Waist Circumference (WC), and diastolic blood pressure shared by FHT2D subjects; they also showed altered key markers of glucose homeostasis, higher triglyceride levels, and worse liver function. FLI scores were significantly lower in subjects without a first-degree FHT2D. After running logistic regression models, a FHT2D was significantly associated with the NAFLD probability, even adjusting for major confounders and stratifying by age (under and over 40 years of age). A FHT2D led to an almost twofold higher probability of NAFLD, regardless of confounding factors (OR 2.17, 95% CI 1.63 to 2.89). A first-degree FHT2D acts as an independent determinant of NAFLD in excess weight phenotypes, regardless of the age group (younger or older than 40 years). A NAFLD risk assessment within multidimensional screening might be useful in excess weight subjects reporting FHT2D even in the absence of diabetes.

Machine Learning Algorithms for Predicting Fatty Liver Disease

Background: Fatty liver disease (FLD) has become a rampant condition. It is associated with a high rate of morbidity and mortality in a population. The condition is commonly referred as FLD. Early prediction of FLD would allow patients to take necessary preventive, diagnosis, and treatment. The main objective of this research is to develop a machine learning (ML) model to predict FLD that can help medics to classify individuals at high risk of FLD, make novel diagnosis, management, and prevention for FLD. Methods: Total of 3,419 subjects were recruited with 845 having been screened for FLD. Classification models were used in the detection of the disease. These models include logistic regression (LR), random forest (RF), artificial neural networks (ANNs), k-nearest neighbors (KNNs), extreme gradient boosting (XGBoost), and linear discriminant analysis (LDA). Predictive accuracy was assessed by area under curve (AUC), sensitivity, specificity, positive predictive value, and negative predictive value. Results: We demonstrated that ML models give more accurate predictions, the best accuracy reached to 0.9415 in the XGBoost model. Feature importance analysis not only confirmed some well-known FLD risk factors, but also demonstrated several novel features for predicting the risk of FLD, such as hemoglobin. Conclusion: By implementing the XGBoost model, physicians can efficiently identify FLD in general patients; this would help in prevention, early treatment, and management of FLD.

2953 A Novel Tool to Estimate Early Insulin Resistance and Prevent Metabolic Syndrome and Non-Alcoholic Fatty Liver Disease: Retrospective Study in the Veteran Population

INTRODUCTION: Previous studies such as the Homeostasis Model Assessment (HOMA) IR1 have been validated tools to estimate Insulin resistance and a predictor of metabolic syndrome (MS) and its related co morbidities such as Non-Alcoholic Fatty liver disease (NAFLD) and Cardiovascular disease (CVD). In the current obesity epidemic, we need simple validated tools to help detect early insulin resistance and curb the growing epidemic of NAFLD. Proinsulin was reported once and in patents as a marker of pancreatic exhaustion hence we propose a novel tool which take proinsulin into account to help predict risk of development of fatty liver disease. AIM: To study proinsulin as a predictor of progression to fatty liver disease in non-diabetic veteran population. METHODS: Single center retrospective chart review of 55 patients. Charts from the gastroenterology department at Las Vegas VA clinic were analyzed and data collected including liver enzymes, insulin, proinsulin, fasting blood sugar, HbA1C, albumin and lipid profile levels (Table 1). Using the data provided a logistic regression model was used to identify strong predictors of fatty liver disease. A novel statistical model called the Random Forrest method was also used to identify predictors of fatty liver disease. RESULTS: Using the logistic regression model, proinsulin and ratio of insulin to proinsulin were found to be significant in predicting fatty liver disease. The overall accuracy was 80%, with precision and recall for both categories ranging from 66.67% to 86.49% (Figure 2). When using the Random Forest methodology, factors such as total cholesterol, ALT, pro-insulin, triglycerides, insulin, ratio IP, and glucose were found to be the most important predictors of Fatty Liver disease (Figure 3). CONCLUSION: Early detection of insulin resistance in non-diabetic patient population and can help intervene with medical and dietary or other management to prevent the development and progression of Metabolic syndrome. Our study suggests that proinsulin should be routinely used in at risk patient population for early detection of insulin resistance and prevention of metabolic syndrome and related comorbidities such as NAFLD. Hence pro insulin is a strong predictor of insulin resistance and future prospective studies can help validate proinsulin as a predictor of NAFLD.

Steatosis and steatohepatitis found in adults after death due to non-burn trauma

OBJECTIVE:With the increasing prevalence of steatosis, the number of steatotic liver grafts from deceased donors is also increasing. Thus, determining the prevalence and the population risk factors of steatosis may assist in risk stratification. The aim of this study was to evaluate the prevalence and predictors of steatosis and steatohepatitis among livers from adults who died due to non-burn trauma.METHODS:Specimens were collected from 224 adults undergoing autopsy at a regional autopsy referral center from September 2011 to April 2013. Histopathological examination was performed on six samples obtained from different lobes of each liver. The outcomes of interest were the presence of steatosis, steatohepatitis, NASH inflammation and NASH fibrosis. The main predictors were body mass index, abdominal circumference, liver weight and volume, presence of cholelithiasis, and siderosis. Our modeling strategy made use of a series of generalized linear models with a binomial family.RESULTS:Our sample had a mean age of 40 years; steatosis was diagnosed in 48.2% of cases, and steatohepatitis was diagnosed in 2.7%. The presence of a high proportion of fatty changes was more prevalent among males and older individuals, with the most affected age group being 41-60 years. When evaluating the crude odds ratio for steatosis, the factors significantly associated with an increased risk of steatosis were greater abdominal circumference, BMI, and liver weight and the presence of siderosis.CONCLUSION:Our study reinforces the role of older age, obesity and hepatomegaly as predictors of fatty liver disease. These variables should be considered in the assessment of fatty changes in the livers of potential liver donors.

Prediction of fatty liver disease using machine learning algorithms

Background and objectiveFatty liver disease (FLD) is a common clinical complication; it is associated with high morbidity and mortality. However, an early prediction of FLD patients provides an opportunity to make an appropriate strategy for prevention, early diagnosis and treatment. We aimed to develop a machine learning model to predict FLD that could assist physicians in classifying high-risk patients and make a novel diagnosis, prevent and manage FLD. MethodsWe included all patients who had an initial fatty liver screening at the New Taipei City Hospital between 1st and 31st December 2009. Classification models such as random forest (RF), Naïve Bayes (NB), artificial neural networks (ANN), and logistic regression (LR) were developed to predict FLD. The area under the receiver operating characteristic curve (ROC) was used to evaluate performances among the four models. ResultsA total of 577 patients were included in this study; of those 377 patients had fatty liver. The area under the receiver operating characteristic (AUROC) of RF, NB, ANN, and LR with 10 fold-cross validation was 0.925, 0.888, 0.895, and 0.854 respectively. Additionally, The accuracy of RF, NB, ANN, and LR 87.48, 82.65, 81.85, and 76.96%. ConclusionIn this study, we developed and compared the four classification models to predict fatty liver disease accurately. However, the random forest model showed higher performance than other classification models. Implementation of a random forest model in the clinical setting could help physicians to stratify fatty liver patients for primary prevention, surveillance, early treatment, and management.

Applications of Machine Learning in Fatty Live Disease Prediction.

: Fatty liver disease (FLD) is considered the most prevalent form of chronic liver disease worldwide. The prediction of fatty liver disease is an important factor for effective treatment and reduce serious health consequences. We, therefore construct a prediction model based on machine learning algorithms. A dataset was developed with ten attributes that included 994 liver patients in which 533 patients were females and others were male. Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Logistic Regression (RF) data mining technique with 10-fold cross-validation was used in the proposed model for the prediction of fatty liver disease. The performances were evaluated with accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. In this proposed model, logistic regression technique provides a better result (Accuracy 76.30%, sensitivity 74.10%, and specificity 64.90%) among all other techniques. This study demonstrates that machine learning models particularly logistic regression model provides a higher accurate prediction for fatty liver diseases based on medical data from electronic medical. This model can be used as a valuable tool for clinical decision making.

A Comparison of the Liver Fat Score and CT Liver-to-Spleen Ratio as Predictors of Fatty Liver Disease by HIV Serostatus.

Non-alcoholic fatty liver disease (NAFLD) is common among HIV-infected (HIV+) adults. The Liver Fat Score (LFS) is a non-invasive, rapid, inexpensive diagnostic tool that uses routine clinical data and is validated against biopsy in HIV-uninfected (HIV-) persons. CT liver-to-spleen (L/S) attenuation ratio is another validated method to diagnose NAFLD. We compared NAFLD prevalence using the LFS versus L/S ratio among Multicenter AIDS Cohort Study participants to assess the LFS's performance in HIV+vs. HIV-men. In a cross-sectional analysis of men reporting<3 alcoholic drinks daily (308 HIV+, 218 HIV-), Spearman correlations determined relationships between LFS and L/S ratio by HIV serostatus. Multivariable regression determined factors associated with discordance in LFS- and L/S ratio-defined NAFLD prevalence. NAFLD prevalence by LFS and L/S ratio were 28%/15% for HIV+men and 20%/19% for HIV-men, respectively. Correlations between LFS and L/S ratio were weaker among HIV+than HIV-men, but improved with increasing BMI and exclusion of HCV-infected men. LFS and L/S ratio discordance occurred more frequently and across BMI strata among HIV+men, but predominantly at BMI<30 kg/m2 among HIV-men. In multivariate analysis, only lower total testosterone levels were significantly associated with discordance. NAFLD prevalence was similar by LFS and L/S ratio identification among HIV-men, but dissimilar and with frequent discordance between the two tests among HIV+men. As discordance may be multifactorial, biopsy data are needed to determine the best non-invasive diagnostic test for NAFLD in HIV+persons.

Fat mass and obesity-associated gene variations are related to fatty liver disease in HIV-infected patients.

Fatty liver disease (FLD) is frequently observed in HIV-infected patients. Obesity and type 2 diabetes mellitus (T2DM) are strongly associated with FLD. Because genetic variants within the fat mass and obesity-associated (FTO) gene have been associated with both pathologies, our aim was to evaluate the association of single nucleotide polymorphisms (SNPs) within the FTO, previously related to obesity or T2DM, with FLD in HIV-infected patients. FLD was defined as a value of the controlled attenuation parameter (CAP) ≥ 238 dB/m, obtained by transient elastography. Four SNPs within FTO intron 1 (rs11642841, rs8050136, rs9939609 and rs9940128) were genotyped in 421 individuals using a custom Golden Gate protocol. The results were replicated in a validation sample consisting of a further 206 HIV-infected patients. Multivariate logistic regression analyses were conducted in the entire population. Three SNPs (rs8050136, rs9939609 and rs9940128) were associated with FLD, with rs9940128 showing the strongest association. This polymorphism also showed an association with FLD in the validation sample. In total, rs9940128 was genotyped in 627 HIV-infected patients, including 267 (42.6%) FLD-diagnosed individuals. The frequency of FLD among rs9940128 AA carriers was 55.7% (63 of 113 individuals) and that in patients without this genotype was 39.7% (204 of 514 individuals) [P = 0.009; adjusted odds ratio 1.88; 95% confidence interval (CI) 1.17-3.01]. Variations within FTO may be predictors of FLD in HIV-infected patients independently of metabolic factors.

Fractal Dimension of Tc-99m DTPA GSA Estimates Pathologic Liver Injury due to Chemotherapy in Liver Cancer Patients.

Chemotherapy-induced liver injury after potent chemotherapy is a considerable problem in patients undergoing liver resection. The aim of this study was to assess the relationship between the fractal dimension (FD) of Tc-99m diethylenetriaminepentaacetic acid (DTPA) galactosyl human serum albumin (GSA) and pathologic change of liver parenchyma in liver cancer patients who have undergone chemotherapy. We examined 34 patients (10 female and 24 male; mean age, 68.5 years) who underwent hepatectomy. Hepatic injury was defined as steatosis more than 30%, grade 2-3 sinusoidal dilation, and/or steatohepatitis Kleiner score ≥4. Fractal analysis was applied to all images of Tc-99m DTPA GSA using a plug-in tool on ImageJ software (NIH, Bethesda, MD). A differential box-counting method was applied, and FD was calculated as a heterogeneity parameter. Correlations between FD and clinicopathological variables were examined. FD values of patients with steatosis and steatohepatitis were significantly higher than those without (P>.001 and P>.001, respectively). There was no difference between the FD values of patients with and without sinusoidal dilatation (P=.357). Multivariate logistic regression showed FD as the only significant predictor for steatosis (P=.005; OR 36.5; 95% CI 3.0-446.3) and steatohepatitis (P=.012; OR, 29.1; 95% CI 2.1-400.1). FD of Tc-99m DTPA GSA was the significant predictor for fatty liver disease in patients who underwent chemotherapy. This new modality is able to differentiate steatohepatitis from steatosis; therefore, it may be useful for predicting chemotherapy-induced pathologic liver injury.

Neck circumference, along with other anthropometric indices, has an independent and additional contribution in predicting fatty liver disease.

Background and AimPrevious studies have indicated that neck circumference is a valuable predictor for obesity and metabolic syndrome, but little evidence is available for fatty liver disease. We examined the association of neck circumference with fatty liver disease and evaluated its predictive value in Chinese adults.MethodsThis cross-sectional study comprised 4053 participants (1617 women and 2436 men, aged 20-88) recruited from the Health Examination Center in Guangzhou, China between May 2009 and April 2010. Anthropometric measurements were taken, abdominal ultrasonography was conducted and blood biochemical parameters were measured. Covariance, logistic regression and receiver operating characteristic curve analyses were employed.ResultsThe mean neck circumference was greater in subjects with fatty liver disease than those without the disease in both women and men after adjusting for age (P<0.001). Logistic regression analysis showed that the age-adjusted ORs (95% CI) of fatty liver disease for quartile 4 (vs. quartile 1) of neck circumference were 7.70 (4.95-11.99) for women and 12.42 (9.22-16.74) for men. After further adjusting for other anthropometric indices, both individually and combined, the corresponding ORs remained significant (all P-trends<0.05) but were attenuated to 1.94-2.53 for women and 1.45-2.08 for men. An additive interaction existed between neck circumference and the other anthropometric measures (all P<0.05). A high neck circumference value was associated with a much greater prevalence of fatty liver disease in participants with both high and normal BMI, waist circumference and waist-to-hip ratio values.ConclusionsNeck circumference was an independent predictor for fatty liver disease and provided an additional contribution when applied with other anthropometric measures.

Serum Leptin as a Predictor of Fatty Liver in 7-Year-Old Korean Children

Aim: This study investigated the usefulness of serum leptin as a predictor of fatty liver disease in Korean children. Methods: One hundred and twenty-four children were recruited from urban areas in Korea. Anthropometrical parameters and clinical variables, such as the levels of serum lipids, insulin and leptin, were measured. Fatty liver disease was detected via hepatic sonogram. We also collected dietary intake data using a 24-hour dietary recall for 3 days. Overweight children were identified using the standards established by the International Obesity Task Force. Results: The prevalence of fatty liver disease was 12.1% among all children and was significantly higher in overweight than in normal children (22.5 vs. 7.1%, respectively; p < 0.05). Within normal weight children, children with fatty livers showed a 2-fold increase in serum leptin levels compared with children with healthy livers (6.2 vs. 2.9 ng/ml, respectively; p < 0.05). Within overweight children, elevated serum leptin levels were observed in children with fatty livers. Multiple logistic regression analysis revealed that serum leptin was strongly associated with fatty liver disease, independent of the body mass index. Energy and fat intake, the frequency of fruit and vegetable consumption, and the frequency of physical activity did not differ among children with or without fatty livers. Conclusion: Elevated serum leptin may be a useful parameter when screening for early fatty liver disease among children.