Comparison and development of advanced machine learning tools to predict nonalcoholic fatty liver disease: An extended study

AI-Based Platelet-Independent Noninvasive Test for Liver Fibrosis in MASLD Patients

Background and Aim

Noninvasive tests (NITs), such as platelet-based indices and ultrasound/MRI elastography, are widely used to assess liver fibrosis in metabolic dysfunction-associated steatotic liver disease (MASLD). However, platelet counts are not routinely included in Japanese health check-ups, limiting their utility in large-scale screenings. Additionally, elastography, while effective, is costly and less accessible in routine practice. Most existing AI-based models incorporate these markers, restricting their applicability. This study aimed to develop a simple yet accurate AI model for liver fibrosis staging using only routine demographic and biochemical markers.

Methods

This retrospective study analyzed biopsy-proven data from 463 Japanese MASLD patients. Patients were randomly assigned to training (N = 370, 80%) and test (N = 93, 20%) cohorts. The AI model incorporated age, sex, BMI, diabetes, hypertension, hyperlipidemia, and routine blood markers (AST, ALT, γ-GTP, HbA1c, glucose, triglycerides, cholesterol).

Results

The Support Vector Machine model demonstrated high diagnostic performance, with an area under the curve (AUC) of 0.886 for detecting significant fibrosis (≥ F2). The AUCs for advanced fibrosis (≥ F3) and cirrhosis (F4) were 0.882 and 0.916, respectively. Compared to FIB-4, APRI, and FAST score (0.80-0.96), SVM achieved comparable accuracy while eliminating the need for platelet count or elastography.

Conclusion

This AI model accurately assesses liver fibrosis in MASLD patients without requiring platelet count or elastography. Its simplicity, cost-effectiveness, and strong diagnostic performance make it well-suited for large-scale health screenings and routine clinical use.

Steatotic liver index: An interpretable predictor of steatotic liver disease using machine learning with an enhanced shrinkage method

AIM

While the Fatty Liver Index (FLI) has been the most prominent among interpretable predictors for steatotic liver disease (SLD), we aimed to prepare a novel diagnostic/prognostic index better than FLI for SLD using a non-black-box and modified parsimonious machine learning method.

METHODS

We included individuals who participated in an annual health checkup in Tokyo, Japan, between January 2008 and December 2018. In the training set (randomly selected 80% of the sample), we developed a novel interpretable model, Steatotic Liver Index (SLI), using a modified method of least absolute shrinkage and selection operator regression focusing on parsimony and interpretability using as few variables as FLI, and confirming its superiority to FLI using the test set (the remaining 20%). The predictive performance of the constructed index was assessed for the diagnosis, development, and remission of SLD.

RESULTS

Among ultrasound data of 92 968 participants at the first health checkup, 20 380 (21.9%) had SLD. Using a modified method of least absolute shrinkage and selection operator regression, SLI was constructed with four variables: body mass index, waist circumference, alanine aminotransferase, and triglycerides. The C-statistic of SLI for SLD diagnosis was superior to that of FLI (0.909 vs. 0.892, p < 0.001). In participants without SLD, SLI was more accurate than FLI in predicting SLD development, whereas among those with SLD, SLI showed better accuracy in predicting SLD remission compared with FLI.

CONCLUSIONS

We developed SLI, a novel interpretable and parsimonious index for diagnosing SLD, which demonstrates superior predictive capability compared with FLI. Further studies are necessary to validate the diagnostic ability outside Japan.

Development and validation of machine learning models for MASLD: based on multiple potential screening indicators

Background

Multifaceted factors play a crucial role in the prevention and treatment of metabolic dysfunction-associated steatotic liver disease (MASLD). This study aimed to utilize multifaceted indicators to construct MASLD risk prediction machine learning models and explore the core factors within these models.

Methods

MASLD risk prediction models were constructed based on seven machine learning algorithms using all variables, insulin-related variables, demographic characteristics variables, and other indicators, respectively. Subsequently, the partial dependence plot(PDP) method and SHapley Additive exPlanations (SHAP) were utilized to explain the roles of important variables in the model to filter out the optimal indicators for constructing the MASLD risk model.

Results

Ranking the feature importance of the Random Forest (RF) model and eXtreme Gradient Boosting (XGBoost) model constructed using all variables found that both homeostasis model assessment of insulin resistance (HOMA-IR) and triglyceride glucose-waist circumference (TyG-WC) were the first and second most important variables. The MASLD risk prediction model constructed using the variables with top 10 importance was superior to the previous model. The PDP and SHAP methods were further utilized to screen the best indicators (including HOMA-IR, TyG-WC, age, aspartate aminotransferase (AST), and ethnicity) for constructing the model, and the mean area under the curve value of the models was 0.960.

Conclusions

HOMA-IR and TyG-WC are core factors in predicting MASLD risk. Ultimately, our study constructed the optimal MASLD risk prediction model using HOMA-IR, TyG-WC, age, AST, and ethnicity.

Comparison of deep learning schemes in grading non-alcoholic fatty liver disease using B-mode ultrasound hepatorenal window images with liver biopsy as the gold standard

BACKGROUND/INTRODUCTION

To evaluate the performance of pre-trained deep learning schemes (DLS) in hepatic steatosis (HS) grading of Non-Alcoholic Fatty Liver Disease (NAFLD) patients, using as input B-mode US images containing right kidney (RK) cortex and liver parenchyma (LP) areas indicated by an expert radiologist.

METHODS

A total of 112 consecutively enrolled, biopsy-validated NAFLD patients underwent a regular abdominal B-mode US examination. For each patient, a radiologist obtained a B-mode US image containing RK cortex and LP and marked a point between the RK and LP, around which a window was automatically cropped. The cropped image dataset was augmented using up-sampling, and the augmented and non-augmented datasets were sorted by HS grade. Each dataset was split into training (70%) and testing (30%), and fed separately as input to InceptionV3, MobileNetV2, ResNet50, DenseNet201, and NASNetMobile pre-trained DLS. A receiver operating characteristic (ROC) analysis of hepatorenal index (HRI) measurements by the radiologist from the same cropped images was used for comparison with the performance of the DLS.

RESULTS

With the test data, the DLS reached 89.15 %-93.75 % accuracy when comparing HS grades S0-S1 vs. S2-S3 and 79.69 %-91.21 % accuracy for S0 vs. S1 vs. S2 vs. S3 with augmentation, and 80.45-82.73 % accuracy when comparing S0-S1 vs. S2-S3 and 59.54 %-63.64 % accuracy for S0 vs. S1 vs. S2 vs. S3 without augmentation. The performance of radiologists' HRI measurement after ROC analysis was 82 %, 91.56 %, and 96.19 % for thresholds of S ≥ S1, S ≥ S2, and S = S3, respectively.

CONCLUSION

All networks achieved high performance in HS assessment. DenseNet201 with the use of augmented data seems to be the most efficient supplementary tool for NAFLD diagnosis and grading.

The power of deep learning in simplifying feature selection for hepatocellular carcinoma: a review

BACKGROUND

Hepatocellular Carcinoma (HCC) is a highly aggressive, prevalent, and deadly type of liver cancer. With the advent of deep learning techniques, significant advancements have been made in simplifying and optimizing the feature selection process.

OBJECTIVE

Our scoping review presents an overview of the various deep learning models and algorithms utilized to address feature selection for HCC. The paper highlights the strengths and limitations of each approach, along with their potential applications in clinical practice. Additionally, it discusses the benefits of using deep learning to identify relevant features and their impact on the accuracy and efficiency of diagnosis, prognosis, and treatment of HCC.

DESIGN

The review encompasses a comprehensive analysis of the research conducted in the past few years, focusing on the methodologies, datasets, and evaluation metrics adopted by different studies. The paper aims to identify the key trends and advancements in the field, shedding light on the promising areas for future research and development.

RESULTS

The findings of this review indicate that deep learning techniques have shown promising results in simplifying feature selection for HCC. By leveraging large-scale datasets and advanced neural network architectures, these methods have demonstrated improved accuracy and robustness in identifying predictive features.

CONCLUSIONS

We analyze published studies to reveal the state-of-the-art HCC prediction and showcase how deep learning can boost accuracy and decrease false positives. But we also acknowledge the challenges that remain in translating this potential into clinical reality.

Novel blood and tissue-based mitochondrial D-loop mutations detected in an Iranian NAFLD patient cohort

Non-alcoholic fatty liver disease (NAFLD) is an increasingly prevalent chronic liver disease characterized by an elusive etiology. In its advanced stages, this condition can pose life-threatening implications. Mitochondrial dysfunction due to its impact on hepatic lipid homeostasis, cytokine release, ROS production, and cell death, contributes to the pathogenesis of NAFLD. Previous research reveals a direct link between NAFLD genetic predictors and mitochondrial dysfunction. The emphasis on the D-loop stems from its association with impaired mtDNA replication, underscoring its crucial role in NAFLD progression. We included 38 Iranian NAFLD patients (comprising 16 patients with non-alcoholic fatty liver [NAFL] and 22 patients with non-alcoholic steatohepatitis [NASH]), with matched blood and liver tissue samples collected from each to compare variations in the mitochondrial D-loop sequence within samples. The mitochondrial DNA (mtDNA) D-loop region was amplified using PCR, and variations were identified through sequencing. The resultant sequences were compared with the reference sequence of human mtDNA available in the MITOMAP Database for comparative analysis. In this study, 97 somatic mutations in the mtDNA D-loop region were identified in NAFLD patients. Our study revealed significant difference between the NAFLD patients and control group in 13 detected mutations (P ≤ 0.05). Novel mutations were discovered in hepatic tissues, while mutation 16220-16221ins C was found in both tissues and blood. A significant difference was found in the distribution of D310 and mt514-mt523 (CA)n repeat variations between NAFLD patients and the control group (P < 0.001). C to T and T to C transitions were the prevalent substitution among patients. Identification of the 16220-16221ins C mutation in both blood and tissue samples from NAFLD patients holds substantial promise as a potential diagnostic marker. However, further research is imperative to corroborate these findings.

Discriminating insulin resistance in middle-aged nondiabetic women using machine learning approaches

Objective

We employed machine learning algorithms to discriminate insulin resistance (IR) in middle-aged nondiabetic women.

Methods

The data was from the National Health and Nutrition Examination Survey (2007-2018). The study subjects were 2084 nondiabetic women aged 45-64. The analysis included 48 predictors. We randomly divided the data into training (n = 1667) and testing (n = 417) datasets. Four machine learning techniques were employed to discriminate IR: extreme gradient boosting (XGBoosting), random forest (RF), gradient boosting machine (GBM), and decision tree (DT). The area under the curve (AUC) of receiver operating characteristic (ROC), accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and F1 score were compared as performance metrics to select the optimal technique.

Results

The XGBoosting algorithm achieved a relatively high AUC of 0.93 in the training dataset and 0.86 in the testing dataset to discriminate IR using 48 predictors and was followed by the RF, GBM, and DT models. After selecting the top five predictors to build models, the XGBoost algorithm with the AUC of 0.90 (training dataset) and 0.86 (testing dataset) remained the optimal prediction model. The SHapley Additive exPlanations (SHAP) values revealed the associations between the five predictors and IR, namely BMI (strongly positive impact on IR), fasting glucose (strongly positive), HDL-C (medium negative), triglycerides (medium positive), and glycohemoglobin (medium positive). The threshold values for identifying IR were 29 kg/m2, 100 mg/dL, 54.5 mg/dL, 89 mg/dL, and 5.6% for BMI, glucose, HDL-C, triglycerides, and glycohemoglobin, respectively.

Conclusion

The XGBoosting algorithm demonstrated superior performance metrics for discriminating IR in middle-aged nondiabetic women, with BMI, glucose, HDL-C, glycohemoglobin, and triglycerides as the top five predictors.

Artificial Intelligence in Liver Diseases: Recent Advances

Liver diseases cause a significant burden on public health worldwide. In spite of great advances during recent years, there are still many challenges in the diagnosis and treatment of liver diseases. During recent years, artificial intelligence (AI) has been widely used for the diagnosis, risk stratification, and prognostic prediction of various diseases based on clinical datasets and medical images. Accumulative studies have shown its performance for diagnosing patients with nonalcoholic fatty liver disease and liver fibrosis and assessing their severity, and for predicting treatment response and recurrence of hepatocellular carcinoma, outcomes of liver transplantation recipients, and risk of drug-induced liver injury. Herein, we aim to comprehensively summarize the current evidence regarding diagnostic, prognostic, and/or therapeutic role of AI in these common liver diseases.

A data-driven approach to decode metabolic dysfunction-associated steatotic liver disease

Metabolic dysfunction-associated steatotic liver disease (MASLD), defined by the presence of liver steatosis together with at least one out of five cardiometabolic factors, is the most common cause of chronic liver disease worldwide, affecting around one in three people. Yet the clinical presentation of MASLD and the risk of progression to cirrhosis and adverse clinical outcomes is highly variable. It, therefore, represents both a global public health threat and a precision medicine challenge. Artificial intelligence (AI) is being investigated in MASLD to develop reproducible, quantitative, and automated methods to enhance patient stratification and to discover new biomarkers and therapeutic targets in MASLD. This review details the different applications of AI and machine learning algorithms in MASLD, particularly in analyzing electronic health record, digital pathology, and imaging data. Additionally, it also describes how specific MASLD consortia are leveraging multimodal data sources to spark research breakthroughs in the field. Using a new national-level 'data commons' (SteatoSITE) as an exemplar, the opportunities, as well as the technical challenges of large-scale databases in MASLD research, are highlighted.

Identifying metabolic dysfunction-associated steatotic liver disease in patients with hypertension and pre-hypertension: An interpretable machine learning approach

Objective

Metabolic dysfunction-associated steatotic liver disease (MASLD) is one of the most prevalent liver diseases and is associated with pre-hypertension and hypertension. Our research aims to develop interpretable machine learning (ML) models to accurately identify MASLD in hypertensive and pre-hypertensive populations.

Methods

The dataset for 4722 hypertensive and pre-hypertensive patients is from subjects in the NAGALA study. Six ML models, including the decision tree, K-nearest neighbor, gradient boosting, naive Bayes, support vector machine, and random forest (RF) models, were used in this study. The optimal model was constructed according to the performances of models evaluated by K-fold cross-validation (k = 5), the area under the receiver operating characteristic curve (AUC), average precision (AP), accuracy, sensitivity, specificity, and F1. Shapley additive explanation (SHAP) values were employed for both global and local interpretation of the model results.

Results

The prevalence of MASLD in hypertensive and pre-hypertensive patients was 44.3% (362 cases) and 28.3% (1107 cases), respectively. The RF model outperformed the other five models with an AUC of 0.889, AP of 0.800, accuracy of 0.819, sensitivity of 0.816, specificity of 0.821, and F1 of 0.729. According to the SHAP analysis, the top five important features were alanine aminotransferase, body mass index, waist circumference, high-density lipoprotein cholesterol, and total cholesterol. Further analysis of the feature selection in the RF model revealed that incorporating all features leads to optimal model performance.

Conclusions

ML algorithms, especially RF algorithm, improve the accuracy of MASLD identification, and the global and local interpretation of the RF model results enables us to intuitively understand how various features affect the chances of MASLD in patients with hypertension and pre-hypertension.

Non-alcoholic fatty liver disease risk prediction model and health management strategies for older Chinese adults: a cross-sectional study

BACKGROUND

Non-alcoholic fatty liver disease (NAFLD) is a common chronic liver condition that affects a quarter of the global adult population. To date, only a few NAFLD risk prediction models have been developed for Chinese older adults aged ≥ 60 years. This study presented the development of a risk prediction model for NAFLD in Chinese individuals aged ≥ 60 years and proposed personalised health interventions based on key risk factors to reduce NAFLD incidence among the population.

METHODS

A cross-sectional survey was carried out among 9,041 community residents in Shanghai. Three NAFLD risk prediction models (I, II, and III) were constructed using multivariate logistic regression analysis based on the least absolute shrinkage and selection operator regression analysis, and random forest model to select individual characteristics, respectively. To determine the optimal model, the three models' discrimination, calibration, clinical application, and prediction capability were evaluated using the receiver operating characteristic (ROC) curve, calibration plot, decision curve analysis, and net reclassification index (NRI), respectively. To evaluate the optimal model's effectiveness, the previously published NAFLD risk prediction models (Hepatic steatosis index [HSI] and ZJU index) were evaluated using the following five indicators: accuracy, precision, recall, F1-score, and balanced accuracy. A dynamic nomogram was constructed for the optimal model, and a Bayesian network model for predicting NAFLD risk in older adults was visually displayed using Netica software.

RESULTS

The area under the ROC curve of Models I, II, and III in the training dataset was 0.810, 0.826, and 0.825, respectively, and that of the testing data was 0.777, 0.797, and 0.790, respectively. No significant difference was found in the accuracy or NRI between the models; therefore, Model III with the fewest variables was determined as the optimal model. Compared with the HSI and ZJU index, Model III had the highest accuracy (0.716), precision (0.808), recall (0.605), F1 score (0.692), and balanced accuracy (0.723). The risk threshold for Model III was 20%-80%. Model III included body mass index, alanine aminotransferase level, triglyceride level, and lymphocyte count.

CONCLUSIONS

A dynamic nomogram and Bayesian network model were developed to identify NAFLD risk in older Chinese adults, providing personalized health management strategies and reducing NAFLD incidence.

Effects of Melatonin on Liver of D-Galactose-Induced Aged Mouse Model

Melatonin, a hormone secreted by the pineal gland of vertebrates, regulates sleep, blood pressure, and circadian and seasonal rhythms, and acts as an antioxidant and anti-inflammatory agent. We investigated the protective effects of melatonin against markers of D-galactose (D-Gal)-induced hepatocellular aging, including liver inflammation, hepatocyte structural damage, and non-alcoholic fatty liver. Mice were divided into four groups: phosphate-buffered saline (PBS, control), D-Gal (200 mg/kg/day), melatonin (20 mg/kg), and D-Gal (200 mg/kg) and melatonin (20 mg) cotreatment. The treatments were administered once daily for eight consecutive weeks. Melatonin treatment alleviated D-Gal-induced hepatocyte impairment. The AST level was significantly increased in the D-Gal-treated groups compared to that in the control group, while the ALT level was decreased compared to the melatonin and D-Gal cotreated group. Inflammatory genes, such as IL1-β, NF-κB, IL-6, TNFα, and iNOS, were significantly increased in the D-Gal aging model, whereas the expression levels of these genes were low in the D-Gal and melatonin cotreated group. Interestingly, the expression levels of hepatic steatosis-related genes, such as LXRα, C/EBPα, PPARα, ACC, ACOX1, and CPT-1, were markedly decreased in the D-Gal and melatonin cotreated group. These results suggest that melatonin suppresses hepatic steatosis and inflammation in a mouse model of D-Gal-induced aging.

Predicting the 5-Year Risk of Nonalcoholic Fatty Liver Disease Using Machine Learning Models: Prospective Cohort Study

BACKGROUND

Nonalcoholic fatty liver disease (NAFLD) has emerged as a worldwide public health issue. Identifying and targeting populations at a heightened risk of developing NAFLD over a 5-year period can help reduce and delay adverse hepatic prognostic events.

OBJECTIVE

This study aimed to investigate the 5-year incidence of NAFLD in the Chinese population. It also aimed to establish and validate a machine learning model for predicting the 5-year NAFLD risk.

METHODS

The study population was derived from a 5-year prospective cohort study. A total of 6196 individuals without NAFLD who underwent health checkups in 2010 at Zhenhai Lianhua Hospital in Ningbo, China, were enrolled in this study. Extreme gradient boosting (XGBoost)-recursive feature elimination, combined with the least absolute shrinkage and selection operator (LASSO), was used to screen for characteristic predictors. A total of 6 machine learning models, namely logistic regression, decision tree, support vector machine, random forest, categorical boosting, and XGBoost, were utilized in the construction of a 5-year risk model for NAFLD. Hyperparameter optimization of the predictive model was performed in the training set, and a further evaluation of the model performance was carried out in the internal and external validation sets.

RESULTS

The 5-year incidence of NAFLD was 18.64% (n=1155) in the study population. We screened 11 predictors for risk prediction model construction. After the hyperparameter optimization, CatBoost demonstrated the best prediction performance in the training set, with an area under the receiver operating characteristic (AUROC) curve of 0.810 (95% CI 0.768-0.852). Logistic regression showed the best prediction performance in the internal and external validation sets, with AUROC curves of 0.778 (95% CI 0.759-0.794) and 0.806 (95% CI 0.788-0.821), respectively. The development of web-based calculators has enhanced the clinical feasibility of the risk prediction model.

CONCLUSIONS

Developing and validating machine learning models can aid in predicting which populations are at the highest risk of developing NAFLD over a 5-year period, thereby helping delay and reduce the occurrence of adverse liver prognostic events.

Machine learning models predict liver steatosis but not liver fibrosis in a prospective cohort study

INTRODUCTION

Screening for liver fibrosis continues to rely on laboratory panels and non-invasive tests such as FIB-4-score and transient elastography. In this study, we evaluated the potential of machine learning (ML) methods to predict liver steatosis on abdominal ultrasound and liver fibrosis, namely the intermediate-high risk of advanced fibrosis, in individuals participating in a screening program for colorectal cancer.

METHODS

We performed ultrasound on 5834 patients admitted between 2006 and 2020, and transient elastography on a subset of 1240 patients. Steatosis on ultrasound was diagnosed if liver areas showed a significantly increased echogenicity compared to the renal parenchyma. Liver fibrosis was defined as a liver stiffness measurement ≥8 kPa in transient elastography. We evaluated the performance of three algorithms, namely Extreme Gradient Boosting, Feed-Forward neural network and Logistic Regression, deriving the models using data from patients admitted from January 2007 up to January 2016 and prospectively evaluating on the data of patients admitted from January 2016 up to March 2020. We also performed a performance comparison with the standard clinical test based on Fibrosis-4 Index (FIB-4).

RESULTS

The mean age was 58±9 years with 3036 males (52%). Modelling laboratory parameters, clinical parameters, and data on eight food types/dietary patterns, we achieved high performance in predicting liver steatosis on ultrasound with AUC of 0.87 (95% CI [0.87-0.87]), and moderate performance in predicting liver fibrosis with AUC of 0.75 (95% CI [0.74-0.75]) using XGBoost machine learning algorithm. Patient-reported variables did not significantly improve predictive performance. Gender-specific analyses showed significantly higher performance in males with AUC of 0.74 (95% CI [0.73-0.74]) in comparison to female patients with AUC of 0.66 (95% CI [0.65-0.66]) in prediction of liver fibrosis. This difference was significantly smaller in prediction of steatosis with AUC of 0.85 (95% CI [0.83-0.87]) in female patients, in comparison to male patients with AUC of 0.82 (95% CI [0.80-0.84]).

CONCLUSION

ML based on point-prevalence laboratory and clinical information predicts liver steatosis with high accuracy and liver fibrosis with moderate accuracy. The observed gender differences suggest the need to develop gender-specific models.

A dynamic machine learning model for prediction of NAFLD in a health checkup population: A longitudinal study

Background

Non-alcoholic fatty liver disease (NAFLD) is one of the most common liver diseases worldwide. Currently, most NAFLD prediction models are diagnostic models based on cross-sectional data, which failed to provide early identification or clarify causal relationships. We aimed to use time-series deep learning models with longitudinal health checkup records to predict the onset of NAFLD in the future, and update the model stepwise by incorporating new checkup records to achieve dynamic prediction.

Methods

10,493 participants with over 6 health checkup records from Beijing MJ Health Screening Center were included to conduct a retrospective cohort study, in which the constantly updated initial 5 checkup data were incorporated stepwise to predict the risk of NAFLD at and after their sixth health checkups. A total of 33 variables were considered, consisting of demographic characteristics, medical history, lifestyle, physical examinations, and laboratory tests. L1-penalized logistic regression (LR) was used for feature selection. The long short-term memory (LSTM) algorithm was introduced for model development, and five-fold cross-validation was conducted to tune and choose optimal hyperparameters. Both internal validation and external validation were conducted, using the 20% randomly divided holdout test dataset and previously unseen data from Shanghai MJ Health Screening Center, respectively, to evaluate model performance. The evaluation metrics included area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, Brier score, and decision curve. Bootstrap sampling was implemented to generate 95% confidence intervals of all the metrics. Finally, the Shapley additive explanations (SHAP) algorithm was applied in the holdout test dataset for model interpretability to obtain time-specific and sample-specific contributions of each feature.

Results

Among the 10,493 participants, 1662 (15.84%) were diagnosed with NAFLD at and after their sixth health checkups. The predictive performance of the deep learning model in the internal validation dataset improved over the incorporation of the checkups, with AUROC increasing from 0.729 (95% CI: 0.698,0.760) at baseline to 0.818 (95% CI: 0.798,0.844) when consecutive 5 checkups were included. The external validation dataset, containing 1728 participants, was used to verify the results, in which AUROC increased from 0.700 (95% CI: 0.657,0.740) with only the first checkups to 0.792 (95% CI: 0.758,0.825) with all five. The results of feature significance showed that body fat percentage, alanine transaminase (ALT), and uric acid owned the greatest impact on the outcome, time-specific, individual-specific and dynamic feature contributions were also produced for model interpretability.

Conclusion

A dynamic prediction model was successfully established in our study, and the prediction capability kept improving with the renewal of the latest checkup records. In addition, we identified key features associated with the onset of NAFLD, making it possible to optimize the prevention and control strategies of the disease in the general population.

A risk prediction model for type 2 diabetes mellitus complicated with retinopathy based on machine learning and its application in health management

OBJECTIVE

This study aimed to establish a risk prediction model for diabetic retinopathy (DR) in the Chinese type 2 diabetes mellitus (T2DM) population using few inspection indicators and to propose suggestions for chronic disease management.

METHODS

This multi-centered retrospective cross-sectional study was conducted among 2,385 patients with T2DM. The predictors of the training set were, respectively, screened by extreme gradient boosting (XGBoost), a random forest recursive feature elimination (RF-RFE) algorithm, a backpropagation neural network (BPNN), and a least absolute shrinkage selection operator (LASSO) model. Model I, a prediction model, was established through multivariable logistic regression analysis based on the predictors repeated ≥3 times in the four screening methods. Logistic regression Model II built on the predictive factors in the previously released DR risk study was introduced into our current study to evaluate the model's effectiveness. Nine evaluation indicators were used to compare the performance of the two prediction models, including the area under the receiver operating characteristic curve (AUROC), accuracy, precision, recall, F1 score, balanced accuracy, calibration curve, Hosmer-Lemeshow test, and Net Reclassification Index (NRI).

RESULTS

When including predictors, such as glycosylated hemoglobin A1c, disease course, postprandial blood glucose, age, systolic blood pressure, and albumin/urine creatinine ratio, multivariable logistic regression Model I demonstrated a better prediction ability than Model II. Model I revealed the highest AUROC (0.703), accuracy (0.796), precision (0.571), recall (0.035), F1 score (0.066), Hosmer-Lemeshow test (0.887), NRI (0.004), and balanced accuracy (0.514).

CONCLUSION

We have built an accurate DR risk prediction model with fewer indicators for patients with T2DM. It can be used to predict the individualized risk of DR in China effectively. In addition, the model can provide powerful auxiliary technical support for the clinical and health management of patients with diabetes comorbidities.

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.

Machine learning classifiers for screening nonalcoholic fatty liver disease in general adults

Nonalcoholic fatty liver disease (NAFLD) is one of major causes of end-stage liver disease in the coming decades, but it shows few symptoms until it develops into cirrhosis. We aim to develop classification models with machine learning to screen NAFLD patients among general adults. This study included 14,439 adults who took health examination. We developed classification models to classify subjects with or without NAFLD using decision tree, random forest (RF), extreme gradient boosting (XGBoost) and support vector machine (SVM). The classifier with SVM was showed the best performance with the highest accuracy (0.801), positive predictive value (PPV) (0.795), F1 score (0.795), Kappa score (0.508) and area under the precision-recall curve (AUPRC) (0.712), and the second top of area under receiver operating characteristic curve (AUROC) (0.850). The second-best classifier was RF model, which was showed the highest AUROC (0.852) and the second top of accuracy (0.789), PPV (0.782), F1 score (0.782), Kappa score (0.478) and AUPRC (0.708). In conclusion, the classifier with SVM is the best one to screen NAFLD in general population based on the results from physical examination and blood testing, followed by the classifier with RF. Those classifiers have a potential to screen NAFLD in general population for physician and primary care doctors, which could benefit to NAFLD patients from early diagnosis.

SWOT analysis of noninvasive tests for diagnosing NAFLD with severe fibrosis: an expert review by the JANIT Forum

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease. Nonalcoholic steatohepatitis (NASH) is an advanced form of NAFLD can progress to liver cirrhosis and hepatocellular carcinoma (HCC). Recently, the prognosis of NAFLD/NASH has been reported to be dependent on liver fibrosis degree. Liver biopsy remains the gold standard, but it has several issues that must be addressed, including its invasiveness, cost, and inter-observer diagnosis variability. To solve these issues, a variety of noninvasive tests (NITs) have been in development for the assessment of NAFLD progression, including blood biomarkers and imaging methods, although the use of NITs varies around the world. The aim of the Japan NASH NIT (JANIT) Forum organized in 2020 is to advance the development of various NITs to assess disease severity and/or response to treatment in NAFLD patients from a scientific perspective through multi-stakeholder dialogue with open innovation, including clinicians with expertise in NAFLD/NASH, companies that develop medical devices and biomarkers, and professionals in the pharmaceutical industry. In addition to conventional NITs, artificial intelligence will soon be deployed in many areas of the NAFLD landscape. To discuss the characteristics of each NIT, we conducted a SWOT (strengths, weaknesses, opportunities, and threats) analysis in this study with the 36 JANIT Forum members (16 physicians and 20 company representatives). Based on this SWOT analysis, the JANIT Forum identified currently available NITs able to accurately select NAFLD patients at high risk of NASH for HCC surveillance/therapeutic intervention and evaluate the effectiveness of therapeutic interventions.

Age patterns of nonalcoholic fatty liver disease incidence: heterogeneous associations with metabolic changes

BACKGROUND

As the nonalcoholic fatty liver disease (NAFLD) epidemic matures, understanding how metabolic changes in NAFLD development vary over the age distribution is important to guide precise prevention. We aimed to clarify metabolic trends in age-specific NAFLD incidence.

METHODS

We conducted a 4-year longitudinal retrospective cohort study enrolling 10,240 consecutive healthy individuals who received annual physical examination during 2012-2019. Baseline and dynamic changes in metabolism and hepatic steatosis determined with ultrasound were collected and analyzed stratified by age into the following groups: 20-34, 35-49, 50-64, and over 65 years.

RESULTS

Overall, 1701 incident NAFLD participants (16.6%) were identified. Adjusted Cox regression analysis showed that the baseline and increased body mass index were the main risk factors for NAFLD in people ≤ 65 years old. Baseline high-density lipoprotein (HR = 0.56; 95% CI 0.39-0.78) was a protective factor for newly onset NAFLD in the 50-to-64-year-old group, while baseline SBP (HR = 1.03; 95% CI 1.01-1.05), baseline uric acid (HR = 1.04; 95% CI 1.01-1.07), triglyceride increase (HR = 4.76; 95% CI 3.69-6.14), fasting blood glucose increase (HR = 1.32; 95% CI 1.06-1.65) were independently associated with incident NAFLD in over-65-year-old group.

CONCLUSIONS

NAFLD incidence attributable to potentially metabolic risk factors varied substantially across age groups in a cohort of Chinese people. The adoption of age targeted metabolic prevention strategies might reduce the burden of NAFLD.

A Novel Method for Survival Prediction of Hepatocellular Carcinoma Using Feature-Selection Techniques

The World Health Organization (WHO) predicted that 10 million people would have died of cancer by 2020. According to recent studies, liver cancer is the most prevalent cancer worldwide. Hepatocellular carcinoma (HCC) is the leading cause of early-stage liver cancer. However, HCC occurs most frequently in patients with chronic liver conditions (such as cirrhosis). Therefore, it is important to predict liver cancer more explicitly by using machine learning. This study examines the survival prediction of a dataset of HCC based on three strategies. Originally, missing values are estimated using mean, mode, and k-Nearest Neighbor (k-NN). We then compare the different select features using the wrapper and embedded methods. The embedded method employs Least Absolute Shrinkage and Selection Operator (LASSO) and ridge regression in conjunction with Logistic Regression (LR). In the wrapper method, gradient boosting and random forests eliminate features recursively. Classification algorithms for predicting results include k-NN, Random Forest (RF), and Logistic Regression. The experimental results indicate that Recursive Feature Elimination with Gradient Boosting (RFE-GB) produces better results, with a 96.66% accuracy rate and a 95.66% F1-score.

Machine-Learning Algorithm for Predicting Fatty Liver Disease in a Taiwanese Population

The rising incidence of fatty liver disease (FLD) poses a health challenge, and is expected to be the leading global cause of liver-related morbidity and mortality in the near future. Early case identification is crucial for disease intervention. A retrospective cross-sectional study was performed on 31,930 Taiwanese subjects (25,544 training and 6386 testing sets) who had received health check-ups and abdominal ultrasounds in Changhua Christian Hospital from January 2009 to January 2019. Clinical and laboratory factors were included for analysis by different machine-learning algorithms. In addition, the performance of the machine-learning algorithms was compared with that of the fatty liver index (FLI). Totally, 6658/25,544 (26.1%) and 1647/6386 (25.8%) subjects had moderate-to-severe liver disease in the training and testing sets, respectively. Five machine-learning models were examined and demonstrated exemplary performance in predicting FLD. Among these models, the xgBoost model revealed the highest area under the receiver operating characteristic (AUROC) (0.882), accuracy (0.833), F1 score (0.829), sensitivity (0.833), and specificity (0.683) compared with those of neural network, logistic regression, random forest, and support vector machine-learning models. The xgBoost, neural network, and logistic regression models had a significantly higher AUROC than that of FLI. Body mass index was the most important feature to predict FLD according to the feature ranking scores. The xgBoost model had the best overall prediction ability for diagnosing FLD in our study. Machine-learning algorithms provide considerable benefits for screening candidates with FLD.

Application of artificial intelligence in non-alcoholic fatty liver disease and liver fibrosis: a systematic review and meta-analysis

BACKGROUND

The global prevalence of non-alcoholic fatty liver disease (NAFLD) continues to rise. Non-invasive diagnostic modalities including ultrasonography and clinical scoring systems have been proposed as alternatives to liver biopsy but with limited performance. Artificial intelligence (AI) is currently being integrated with conventional diagnostic methods in the hopes of performance improvements. We aimed to estimate the performance of AI-assisted systems for diagnosing NAFLD, non-alcoholic steatohepatitis (NASH), and liver fibrosis.

METHODS

A systematic review was performed to identify studies integrating AI in the diagnosis of NAFLD, NASH, and liver fibrosis. Pooled sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and summary receiver operating characteristic curves were calculated.

RESULTS

Twenty-five studies were included in the systematic review. Meta-analysis of 13 studies showed that AI significantly improved the diagnosis of NAFLD, NASH and liver fibrosis. AI-assisted ultrasonography had excellent performance for diagnosing NAFLD, with a sensitivity, specificity, PPV, NPV of 0.97 (95% confidence interval (CI): 0.91-0.99), 0.98 (95% CI: 0.89-1.00), 0.98 (95% CI: 0.93-1.00), and 0.95 (95% CI: 0.88-0.98), respectively. The performance of AI-assisted ultrasonography was better than AI-assisted clinical data sets for the identification of NAFLD, which provided a sensitivity, specificity, PPV, NPV of 0.75 (95% CI: 0.66-0.82), 0.82 (95% CI: 0.74-0.88), 0.75 (95% CI: 0.60-0.86), and 0.82 (0.74-0.87), respectively. The area under the curves were 0.98 and 0.85 for AI-assisted ultrasonography and AI-assisted clinical data sets, respectively. AI-integrated clinical data sets had a pooled sensitivity, specificity of 0.80 (95%CI: 0.75-0.85), 0.69 (95%CI: 0.53-0.82) for identifying NASH, as well as 0.99-1.00 and 0.76-1.00 for diagnosing liver fibrosis stage F1-F4, respectively.

CONCLUSION

AI-supported systems provide promising performance improvements for diagnosing NAFLD, NASH, and identifying liver fibrosis among NAFLD patients. Prospective trials with direct comparisons between AI-assisted modalities and conventional methods are warranted before real-world implementation.

PROTOCOL REGISTRATION

PROSPERO (CRD42021230391).