LiverScreen project: study protocol for screening for liver fibrosis in the general population in European countries

NAIF: A novel artificial intelligence-based tool for accurate diagnosis of stage F3/F4 liver fibrosis in the general adult population, validated with three external datasets

OBJECTIVE

The purpose of this study was to determine the effectiveness of a new AI-based tool called NAIF (NAFLD-AI-Fibrosis) in identifying individuals from the general population with advanced liver fibrosis (stage F3/F4). We compared NAIF's performance to two existing risk score calculators, aspartate aminotransferase-to-platelet ratio index (APRI) and fibrosis-4 (Fib4).

METHODS

To set up the algorithm for diagnosing severe liver fibrosis (defined as Fibroscan® values E ≥ 9.7 KPa), we used 19 blood biochemistry parameters and two demographic parameters in a group of 5,962 individuals from the NHANES population (2017-2020 pre-pandemic, public database). We then assessed the algorithm's performance by comparing its accuracy, precision, sensitivity, specificity, and F1 score values to those of APRI and Fib4 scoring systems.

RESULTS

In a kept-out sub dataset of the NHANES population, NAIF achieved a predictive precision of 72 %, a sensitivity of 61 %, and a specificity of 77 % in correctly identifying adults (aged 18-79 years) with severe liver fibrosis. Additionally, NAIF performed well when tested with two external datasets of Italian patients with a Fibroscan® score E ≥ 9.7 kPa, and with an external dataset of patients with diagnosis of severe liver fibrosis through biopsy.

CONCLUSIONS

The results of our study suggest that NAIF, using routinely available parameters, outperforms in sensitivity existing scoring methods (Fib4 and APRI) in diagnosing severe liver fibrosis, even when tested with external validation datasets. NAIF uses routinely available parameters, making it a promising tool for identifying individuals with advanced liver fibrosis from the general population. Word count abstract: 236.

Diffusion-Weighted MRI of the Liver in Patients With Chronic Liver Disease: A Comparative Study Between Different Fitting Approaches and Diffusion Models

BACKGROUND

Diffusion-weighted imaging (DWI) has been considered for chronic liver disease (CLD) characterization. Grading of liver fibrosis is important for disease management.

PURPOSE

To investigate the relationship between DWI's parameters and CLD-related features (particularly regarding fibrosis assessment).

STUDY TYPE

Retrospective.

SUBJECTS

Eighty-five patients with CLD (age: 47.9 ± 15.5, 42.4% females).

FIELD STRENGTH/SEQUENCE

3-T, spin echo-echo planar imaging (SE-EPI) with 12 b-values (0-800 s/mm2 ).

ASSESSMENT

Several models statistical models, stretched exponential model, and intravoxel incoherent motion were simulated. The corresponding parameters (Ds , σ, DDC, α, f, D, D*) were estimated on simulation and in vivo data using the nonlinear least squares (NLS), segmented NLS, and Bayesian methods. The fitting accuracy was analyzed on simulated Rician noised DWI. In vivo, the parameters were averaged from five central slices entire liver to compare correlations with histological features (inflammation, fibrosis, and steatosis). Then, the differences between mild (F0-F2) or severe (F3-F6) groups were compared respecting to statistics and classification. A total of 75.3% of patients used to build various classifiers (stratified split strategy and 10-folders cross-validation) and the remaining for testing.

STATISTICAL TESTS

Mean squared error, mean average percentage error, spearman correlation, Mann-Whitney U-test, receiver operating characteristic (ROC) curve, area under ROC curve (AUC), sensitivity, specificity, accuracy, precision. A P-value <0.05 was considered statistically significant.

RESULTS

In simulation, the Bayesian method provided the most accurate parameters. In vivo, the highest negative significant correlation (Ds , steatosis: r = -0.46, D*, fibrosis: r = -0.24) and significant differences (Ds , σ, D*, f) were observed for Bayesian fitted parameters. Fibrosis classification was performed with an AUC of 0.92 (0.91 sensitivity and 0.70 specificity) with the aforementioned diffusion parameters based on the decision tree method.

DATA CONCLUSION

These results indicate that Bayesian fitted parameters may provide a noninvasive evaluation of fibrosis with decision tree.

EVIDENCE LEVEL

1 TECHNICAL EFFICACY: Stage 1.

Liver stiffness as a cornerstone in heart disease risk assessment

Metabolic dysfunction-associated steatotic liver disease (MASLD) typically presents with hepatic fibrosis in advanced disease, resulting in increased liver stiffness. A subset of patients further develops liver cirrhosis and hepatocellular carcinoma. Cardiovascular disease is a common comorbidity in patients with MASLD and its prevalence is increasing in parallel. Recent evidence suggests that especially liver stiffness, whether or not existing against a background of MASLD, is associated with heart diseases. We conducted a narrative review on the role of liver stiffness in the prediction of highly prevalent heart diseases including heart failure, cardiac arrhythmias (in particular atrial fibrillation), coronary heart disease, and aortic valve sclerosis. Research papers were retrieved from major scientific databases (PubMed, Web of Science) until September 2023 using 'liver stiffness' and 'liver fibrosis' as keywords along with the latter cardiac conditions. Increased liver stiffness, determined by vibration-controlled transient elastography or hepatic fibrosis as predicted by biomarker panels, are associated with a variety of cardiovascular diseases, including heart failure, atrial fibrillation, and coronary heart disease. Elevated liver stiffness in patients with metabolic liver disease should lead to considerations of cardiac workup including N-terminal pro-B-type natriuretic peptide/B-type natriuretic peptide determination, electrocardiography, and coronary computed tomography angiography. In addition, patients with MASLD would benefit from heart disease case-finding strategies in which liver stiffness measurements can play a key role. In conclusion, increased liver stiffness should be a trigger to consider a cardiac workup in metabolically compromised patients.

Implementation of a liver health check in people with type 2 diabetes

As morbidity and mortality related to potentially preventable liver diseases are on the rise globally, early detection of liver fibrosis offers a window of opportunity to prevent disease progression. Early detection of non-alcoholic fatty liver disease allows for initiation and reinforcement of guidance on bodyweight management, risk stratification for advanced liver fibrosis, and treatment optimisation of diabetes and other metabolic complications. Identification of alcohol-related liver disease provides the opportunity to support patients with detoxification and abstinence programmes. In all patient groups, identification of cirrhosis ensures that patients are enrolled in surveillance programmes for hepatocellular carcinoma and portal hypertension. When considering early detection strategies, success can be achieved from applying ad-hoc screening for liver fibrosis in established frameworks of care. Patients with type 2 diabetes are an important group to consider case findings of advanced liver fibrosis and cirrhosis, as up to 19% have advanced fibrosis (which is ten times higher than the general population) and almost 70% have non-alcoholic fatty liver disease. Additionally, patients with type 2 diabetes with alcohol use disorders have the highest proportion of liver-related morbidity of people with type 2 diabetes generally. Patients with type 2 diabetes receive an annual diabetes review as part of their routine clinical care, in which the health of many organs are considered. Yet, liver health is seldom included in this review. This Viewpoint argues that augmenting the existing risk stratification strategy with an additional liver health check provides the opportunity to detect advanced liver fibrosis, thereby opening a window for early interventions to prevent end-stage liver disease and its complications, including hepatocellular carcinoma.

Development, validation, and prognostic evaluation of a risk score for long-term liver-related outcomes in the general population: a multicohort study

BACKGROUND

Liver cirrhosis is a major cause of death worldwide. Cirrhosis develops after a long asymptomatic period of fibrosis progression, with the diagnosis frequently occurring late, when major complications or cancer develop. Few reliable tools exist for timely identification of individuals at risk of cirrhosis to allow for early intervention. We aimed to develop a novel score to identify individuals at risk for future liver-related outcomes.

METHODS

We derived the LiverRisk score from an international prospective cohort of individuals from six countries without known liver disease from the general population, who underwent liver fibrosis assessment by transient elastography. The score included age, sex, and six standard laboratory variables. We created four groups: minimal risk, low risk, medium risk, and high risk according to selected cutoff values of the LiverRisk score (6, 10, and 15). The model's discriminatory accuracy and calibration were externally validated in two prospective cohorts from the general population. Moreover, we ascertained the prognostic value of the score in the prediction of liver-related outcomes in participants without known liver disease with median follow-up of 12 years (UK Biobank cohort).

FINDINGS

We included 14 726 participants: 6357 (43·2%) in the derivation cohort, 4370 (29·7%) in the first external validation cohort, and 3999 (27·2%) in the second external validation cohort. The score accurately predicted liver stiffness in the development and external validation cohorts, and was superior to conventional serum biomarkers of fibrosis, as measured by area under the receiver-operating characteristics curve (AUC; 0·83 [95% CI [0·78-0·89]) versus the fibrosis-4 index (FIB-4; 0·68 [0·61-0·75] at 10 kPa). The score was effective in identifying individuals at risk of liver-related mortality, liver-related hospitalisation, and liver cancer, thereby allowing stratification to different risk groups for liver-related outcomes. The hazard ratio for liver-related mortality in the high-risk group was 471 (95% CI 347-641) compared with the minimal risk group, and the overall AUC of the score in predicting 10-year liver-related mortality was 0·90 (0·88-0·91) versus 0.84 (0·82-0·86) for FIB-4.

INTERPRETATION

The LiverRisk score, based on simple parameters, predicted liver fibrosis and future development of liver-related outcomes in the general population. The score might allow for stratification of individuals according to liver risk and thus guide preventive care.

FUNDING

European Commission under the H20/20 programme; Fondo de Investigación Sanitaria de Salud; Instituto de Salud Carlos III; Spanish Ministry of Economy, Industry, and Competitiveness; the European Regional Development Fund; and the German Ministry of Education and Research (BMBF).

Setting up of a machine learning algorithm for the identification of severe liver fibrosis profile in the general US population cohort

BACKGROUND

The progress of digital transformation in clinical practice opens the door to transforming the current clinical line for liver disease diagnosis from a late-stage diagnosis approach to an early-stage based one. Early diagnosis of liver fibrosis can prevent the progression of the disease and decrease liver-related morbidity and mortality. We developed here a machine learning (ML) algorithm containing standard parameters that can identify liver fibrosis in the general US population.

MATERIALS AND METHODS

Starting from a public database (National Health and Nutrition Examination Survey, NHANES), representative of the American population with 7265 eligible subjects (control population n = 6828, with Fibroscan values E < 9.7 KPa; target population n = 437 with Fibroscan values E ≥ 9.7 KPa), we set up an SVM algorithm able to discriminate for individuals with liver fibrosis among the general US population. The algorithm set up involved the removal of missing data and a sampling optimization step to managing the data imbalance (only ∼ 5 % of the dataset is the target population).

RESULTS

For the feature selection, we performed an unbiased analysis, starting from 33 clinical, anthropometric, and biochemical parameters regardless of their previous application as biomarkers of liver diseases. Through PCA analysis, we identified the 26 more significant features and then used them to set up a sampling method on an SVM algorithm. The best sampling technique to manage the data imbalance was found to be oversampling through the SMOTE-NC. For final model validation, we utilized a subset of 300 individuals (150 with liver fibrosis and 150 controls), subtracted from the main dataset prior to sampling. Performances were evaluated on multiple independent runs.

CONCLUSIONS

We provide proof of concept of an ML clinical decision support tool for liver fibrosis diagnosis in the general US population. Though the presented ML model represents at this stage only a prototype, in the future, it might be implemented and potentially applied to program broad screenings for liver fibrosis.

High prevalence of liver fibrosis among general population: a Romanian population-based study

BACKGROUND

Cirrhosis can be present undected for years prior to a symptomatic presentation. Early detection may result in improved outcomes. Data are lacking, however, regarding the yield of screening in many populations. We aimed to determined prevalence of significant liver fibrosis diagnosed by vibration-controlled transient elastography (VCTE) in apparently healthy Romanians.

METHODS

Between December 2021 and March 2022, we prospectively screened 1,027 subjects from different counties of Northeastern Romania using VCTE and B-mode ultrasonagraphy after a comprehensive medical history questionnaire. Participants with abnormal liver stiffness measurement values were further evaluated by laboratory tests to identify the severity and etiology of chronic liver disease.

RESULTS

Overall, 17.9% of subjects had liver stiffness measurments (LSM) ≥8 kpa, including 55 with LSM ≥13.0 kpa. Among these subjects, 26.1% had a history of heavy alcohol intake, 22.3% tested positive for hepatitis B and/or C infection, and 49.5% were diagnosed with NAFLD. The prevalence of elevated LSM was highest among older subjects (>60 y old) and those with diabetes. Among those with LSM ≥13 kPa and ≥9.6 kpa, FIB-4 was <2.67 in 46.9% and 87.5% respectively.

CONCLUSION

There is high prevalence of significant liver fibrosis in the Romanian general population. VCTE is a usefool tool for early detection of liver disease and appears more sensitive than FIB-4.

Update on the role of elastography in liver disease

The diagnosis of liver fibrosis and the assessment of its severity are important to provide appropriate management, to determine the prognosis or the need for surveillance. Currently, for fibrosis staging, liver stiffness measurement (LSM) with the shear wave elastography (SWE) techniques is considered a reliable substitute for liver biopsy in several clinical scenarios. Nonetheless, it should be emphasized that stiffness value is a biomarker of diffuse liver disease that must be interpreted taking into consideration anamnesis, clinical and laboratory data. In patients with diffuse liver disease, it is more clinically relevant to determine the likelihood of advanced disease rather than to obtain an exact stage of liver fibrosis using a histologic classification. In this regard, a 'rule of five' for LSMs with vibration-controlled transient elastography (VCTE) and a 'rule of four' for LSMs with the acoustic radiation force impulse (ARFI)-based techniques have been proposed. In patients with advanced chronic liver disease (CLD), the risk of liver decompensation increases with increasing liver stiffness value. SWE has been proposed as a tool to predict the risk of death or complications in patients with CLD. LSM by VCTE combined with platelets count is a validated non-invasive method for varices screening, with very good results in terms of invasive procedures being spared. ARFI-based techniques also show some promising results in this setting. LSM, alone or combined in scores or algorithms with other parameters, is used to evaluate the risk of hepatocellular carcinoma occurrence. Due to the high prevalence of CLD, screening the population at risk is of interest but further studies are needed.

Ultrasound-Based Hepatic Elastography in Non-Alcoholic Fatty Liver Disease: Focus on Patients with Type 2 Diabetes

Non-alcoholic fatty liver disease (NAFLD) is the most prevalent liver disease and is the hepatic expression of metabolic syndrome. The development of non-invasive methods for the diagnosis of hepatic steatosis and advanced fibrosis in high-risk patients, especially those with type 2 diabetes mellitus, is highly needed to replace the invasive method of liver biopsy. Elastographic methods can bring significant added value to screening and diagnostic procedures for NAFLD in patients with diabetes, thus contributing to improved NAFLD management. Pharmacological development and forthcoming therapeutic measures that address NAFLD should also be based on new, non-invasive, and reliable tools that assess NAFLD in at-risk patients and be able to properly guide treatment in individuals with both diabetes and NAFLD. This is the first review aiming to outline and discuss recent studies on ultrasound-based hepatic elastography, focusing on NAFLD assessment in patients with diabetes.