Personalized surveillance for hepatocellular carcinoma in cirrhosis - using machine learning adapted to HCV status

The role of artificial intelligence in the management of liver diseases

Universal neonatal hepatitis B virus (HBV) vaccination and the advent of direct-acting antivirals (DAA) against hepatitis C virus (HCV) have reshaped the epidemiology of chronic liver diseases. However, some aspects of the management of chronic liver diseases remain unresolved. Nucleotide analogs can achieve sustained HBV DNA suppression but rarely lead to a functional cure. Despite the high efficacy of DAAs, successful antiviral therapy does not eliminate the risk of hepatocellular carcinoma (HCC), highlighted the need for cost-effective identification of high-risk populations for HCC surveillance and tailored HCC treatment strategies for these populations. The accessibility of high-throughput genomic data has accelerated the development of precision medicine, and the emergence of artificial intelligence (AI) has led to a new era of precision medicine. AI can learn from complex, non-linear data and identify hidden patterns within real-world datasets. The combination of AI and multi-omics approaches can facilitate disease diagnosis, biomarker discovery, and the prediction of treatment efficacy and prognosis. AI algorithms have been implemented in various aspects, including non-invasive tests, predictive models, image diagnosis, and the interpretation of histopathology findings. AI can support clinicians in decision-making, alleviate clinical burdens, and curtail healthcare expenses. In this review, we introduce the fundamental concepts of machine learning and review the role of AI in the management of chronic liver diseases.

Hepatocellular Cancer Surveillance in Patients with Advanced Chronic Liver Disease

BACKGROUND

Patients with advanced chronic liver disease (ACLD) are at high risk of developing hepatocellular carcinoma (HCC). Therefore, biannual surveillance is recommended. This large-scale multicenter study aimed to stratify the risk of HCC development in ACLD.

METHODS

From 3016 patients with ACLD screened in 17 European and Chinese centers, 2340 patients with liver stiffness measurement (LSM) determined using different techniques (two-dimensional shear-wave elastography [2D-SWE], transient elastography, and point shear-wave elastography) and with different disease severities were included. Cox regression was used to explore risk factors for HCC. We used these data to create an algorithm, named PLEASE, but referred to in this manuscript as "the algorithm"; the algorithm was validated in internal and two external cohorts across elastography techniques.

RESULTS

HCC developed in 127 (5.4%) patients during follow-up. LSM by 2D-SWE (hazard ratio: 2.28) was found to be associated with developing HCC, alongside age, sex, etiology, and platelet count (C-index: 0.8428). We thus established the algorithm with applicable cutoffs, assigning a maximum of six points: platelet count less than 150×109/l, LSM greater than or equal to 15 kPa, age greater than or equal to 50 years, male sex, controlled/uncontrolled viral hepatitis, or presence of steatotic liver diseases. Within 2 years, with a median follow-up of 13.7 months, patients in the high-risk group (≥4 points) had an HCC incidence of 15.6% (95% confidence interval [CI], 12.1% to 18.7%) compared with the low-risk group, at 1.7% (95% CI, 0.9% to 2.5%).

CONCLUSIONS

Our algorithm stratified patients into two groups: those at higher risk of developing HCC and those at lower risk. Our data provide equipoise to test the prospective utility of the algorithm with respect to clinical decisions about screening patients with ACLD for incident HCC. (Funded by the German Research Foundation and others; ClinicalTrials.gov number, NCT03389152.).

Artificial intelligence-based evaluation of prognosis in cirrhosis

Cirrhosis represents a significant global health challenge, characterized by high morbidity and mortality rates that severely impact human health. Timely and precise prognostic assessments of liver cirrhosis are crucial for improving patient outcomes and reducing mortality rates as they enable physicians to identify high-risk patients and implement early interventions. This paper features a thorough literature review on the prognostic assessment of liver cirrhosis, aiming to summarize and delineate the present status and constraints associated with the application of traditional prognostic tools in clinical settings. Among these tools, the Child-Pugh and Model for End-Stage Liver Disease (MELD) scoring systems are predominantly utilized. However, their accuracy varies significantly. These systems are generally suitable for broad assessments but lack condition-specific applicability and fail to capture the risks associated with dynamic changes in patient conditions. Future research in this field is poised for deep exploration into the integration of artificial intelligence (AI) with routine clinical and multi-omics data in patients with cirrhosis. The goal is to transition from static, unimodal assessment models to dynamic, multimodal frameworks. Such advancements will not only improve the precision of prognostic tools but also facilitate personalized medicine approaches, potentially revolutionizing clinical outcomes.

EASL position paper on clinical follow-up after HCV cure

Following the advent of direct-acting antivirals (DAAs), hepatitis C virus (HCV) infection can be cured in almost all infected patients. This has led to a number of clinical questions regarding the optimal management of the millions of patients cured of HCV. This position statement provides specific guidance on the appropriate follow-up after a sustained virological response in patients without advanced fibrosis, those with compensated advanced chronic liver disease, and those with decompensated cirrhosis. Guidance on hepatocellular carcinoma risk assessment and the management of extrahepatic manifestations of HCV is also provided. Finally, guidance is provided on the monitoring and treatment of reinfection in at-risk patients. The recommendations are based on the best available evidence and are intended to help healthcare professionals involved in the management of patients after treatment for HCV.

Surveillance for Hepatocellular Carcinoma in Patients with Successfully Treated Viral Disease of the Liver: A Systematic Review

Background

Surveillance for hepatocellular carcinoma (HCC) has been proven to increase the proportion of tumors detected at early stages and the chance of receiving curative therapies, reducing mortality by about 30%.

Summary

Current recommendations consist of a semi-annual abdominal ultrasound with or without serum alpha-fetoprotein measurement in patients with cirrhosis and specific subgroups of populations with chronic viral hepatitis. Antiviral therapies, such as nucleot(s)ide analogs that efficiently suppress the replication of hepatitis B virus (HBV) and direct-acting antiviral drugs able to eliminate the hepatitis C virus (HCV) in >90% of patients, have radically changed the outcomes of viral liver disease and decreased, but not eliminated, the risk of HCC in both cirrhotic and non-cirrhotic patients. HCC risk is a key starting point for implementing a cost-effective surveillance and should also guide the decision-making process concerning its modality. As the global number of effectively treated viral patients continues to rise, there is a pressing need to identify those for whom the benefit-to-harm ratio of surveillance is favorable and to determine how to conduct cost-effective screening on such patients.

Key Messages

This article addresses this topic and attempts to determine which patients should continue HCC surveillance after HBV suppression or HCV eradication, based on cost-effectiveness principles and the fact that HCC risk declines over time. We also formulate a proposal for a surveillance algorithm that switches the use of surveillance for HCC from the "one-size-fits-all" approach to individualized programs based on oncologic risk (precision surveillance).

Ensemble learning enhances the precision of preliminary detection of primary hepatocellular carcinoma based on serological and demographic indices

Primary hepatocellular carcinoma (PHC) is associated with high rates of morbidity and malignancy in China and throughout the world. In clinical practice, a combination of ultrasound and alpha-fetoprotein (AFP) measurement is frequently employed for initial screening. However, the accuracy of this approach often falls short of the desired standard. Consequently, this study aimed to investigate the enhancement of precision of preliminary detection of PHC by ensemble learning techniques. To achieve this, 712 patients with PHC and 1887 healthy controls were enrolled for the assessment of four ensemble learning methods, namely, Random Forest (RF), LightGBM, Xgboost, and Catboost. A total of eleven characteristics, comprising nine serological indices and two demographic indices, were selected from the participants for use in detecting PHC. The findings identified an optimal feature subset consisting of eight features, namely AFP, albumin (ALB), alanine aminotransferase (ALT), platelets (PLT), age, alkaline phosphatase (ALP), hemoglobin (Hb), and body mass index (BMI), that achieved the highest classification accuracy of 96.62%. This emphasizes the importance of the collective use of these features in PHC diagnosis. In conclusion, the results provide evidence that the integration of serological and demographic indices together with ensemble learning models, can contribute to the precision of preliminary diagnosis of PHC.

Risk prediction of cholangitis after stent implantation based on machine learning

The risk of cholangitis after ERCP implantation in malignant obstructive jaundice patients remains unknown. To develop models based on artificial intelligence methods to predict cholangitis risk more accurately, according to patients after stent implantation in patients' MOJ clinical data. This retrospective study included 218 patients with MOJ undergoing ERCP surgery. A total of 27 clinical variables were collected as input variables. Seven models (including univariate analysis and six machine learning models) were trained and tested for classified prediction. The model' performance was measured by AUROC. The RFT model demonstrated excellent performances with accuracies up to 0.86 and AUROC up to 0.87. Feature selection in RF and SHAP was similar, and the choice of the best variable subset produced a high performance with an AUROC up to 0.89. We have developed a hybrid machine learning model with better predictive performance than traditional LR prediction models, as well as other machine learning models for cholangitis based on simple clinical data. The model can assist doctors in clinical diagnosis, adopt reasonable treatment plans, and improve the survival rate of patients.

The clinical and financial burden of nonhepatocellular carcinoma focal lesions detected during the surveillance of patients with cirrhosis

BACKGROUND AND AIMS

HCC surveillance is challenged by the detection of hepatic focal lesions (HFLs) of other types. This study aimed to describe the incidence, characteristics, outcomes, and costs of non-HCC HFL detected during surveillance.

APPROACH AND RESULTS

We retrospectively analyzed nonstandardized workup performed in French patients included in HCC surveillance programs recruited in 57 French tertiary centers (ANRS CirVir and CIRRAL cohorts, HCC 2000 trial). The overall cost of workup was evaluated, with an estimation of an average cost per patient for the entire population and per lesion detected. A total of 3295 patients were followed up for 59.8 months, 391 (11.9%) patients developed HCCs (5-year incidence: 12.1%), and 633 (19.2%) developed non-HCC HFLs (5-year incidence: 21.8%). Characterization of non-HCC HFL required a median additional of 0.7 exams per year. A total of 11.8% of non-HCC HFLs were not confirmed on recall procedures, and 19.6% of non-HCC HFLs remained undetermined. A definite diagnosis of benign liver lesions was made in 65.1%, and malignant tumors were diagnosed in 3.5%. The survival of patients with benign or undetermined non-HCC HFL was similar to that of patients who never developed any HFL (5-year survival 92% vs. 88%, p = 0.07). The average cost of the diagnostic workup was 1087€ for non-HCC HFL and €1572 for HCC.

CONCLUSIONS

Non-HCC HFLs are frequently detected in patients with cirrhosis, and do not impact prognosis, but trigger substantial costs. This burden must be considered in cost-effectiveness analyses of future personalized surveillance strategies.

Study protocol for FASTRAK: a randomised controlled trial evaluating the cost impact and effectiveness of FAST-MRI for HCC suRveillance in pAtients with high risK of liver cancer

INTRODUCTION

The surveillance of hepatocellular carcinoma (HCC) using semi-annual liver ultrasound (US) is justified in patients with cirrhosis. In this context, US has a low sensitivity (<30%) for the detection of HCC at the very early stage (ie, Barcelona clinic liver cancer (BCLC) 0, uninodular tumour <2 cm). The sensitivity of abbreviated liver MRI (AMRI) is reported to exceed 80%, but its use is hampered by costs and availability. Our hypothesis is that AMRI used as a screening examination in patients at high risk of HCC (>3% per year) could increase the rates of patients with a tumour detected at an early stage accessible to curative-intent treatment, and demonstrate its cost-effectiveness in this population.

METHODS AND ANALYSIS

The FASTRAK trial is a multicentre, randomised controlled trial with two parallel arms, aiming for superiority and conducted on patients at high risk for HCC (yearly HCC incidence >3%). Randomisation will be conducted on an individual basis with a centralised approach and stratification by centre. After inclusion in the trial, each patient will be randomly assigned to the experimental group (semi-annual US and AMRI) or the control group (semi-annual US alone). The main objective is to assess the cost/quality-adjusted life year and cost/patient detected with a BCLC 0 HCC in both arms. A total of 944 patients will be recruited in 37 tertiary French centres during a 36-month period and will be followed-up during 36 months.

ETHICS AND DISSEMINATION

The FASTRAK trial received ethical approval on 4 April 2022. Results will be disseminated via publication in peer-reviewed journals as well as presentation at international conferences.

TRIAL REGISTRATION NUMBER

Clinical trial number (ClinicaTrials.gov) NCT05095714.

Predictive models for hepatocellular carcinoma development after sustained virological response in advanced hepatitis C

BACKGROUND & AIMS

Life-long hepatocellular carcinoma (HCC) surveillance is recommended after sustained virological response (SVR) in patients with advanced hepatitis C. Since the identification of patients who could be safely discontinued for surveillance is essential, we aimed to identify subsets of patients with low-risk HCC.

METHODS

491 patients with advanced and compensated fibrosis (≥F3) were prospectively followed after achieving SVR with interferon-free therapies. Clinical-biological parameters and liver stiffness measurement (LSM) were performed before starting treatment (ST) and at SVR, and HCC surveillance was carried out.

RESULTS

During a median follow-up of 49.8 months, 29 (5.9%) patients developed HCC [incidence rate: 1.6/100 patient-years (PYs)]. Two predictive models based on LSM (Model-A) or FIB-4 score (Model-B) were proposed. Only SVR parameters were included in the models, because they showed a higher accuracy for predicting HCC than ST measurements. Variables independently associated with HCC were LSM (HR, 1.03; 95% CI, 1.01-1.05), age (HR, 1.04; 95% CI, 1.01-1.08) and albumin levels (HR, 0.90; 95% CI, 0.84-0.97) in Model-A, and FIB-4 (HR, 1.22; 95% CI, 1.08-1.37) and albumin (HR, 0.90; 95% CI, 0.84-0.97) in model-B. Both models allow HCC risk stratification, identifying low-risk groups with an HCC incidence rate of 0.16/100 and 0.25/100 PYs, respectively. An overall increased hazard of HCC was observed over time.

CONCLUSION

Simple models based on non-invasive markers of liver fibrosis, LSM or FIB-4, together with age and albumin levels at SVR permit to identify subsets of patients with HCC risk clearly <1%/year, for whom HCC surveillance might not be cost-effective.

Accurate prediction of HCC risk after SVR in patients with hepatitis C cirrhosis based on longitudinal data

BACKGROUND

Most existing predictive models of hepatocellular carcinoma (HCC) risk after sustained virologic response (SVR) are built on data collected at baseline and therefore have limited accuracy. The current study aimed to construct an accurate predictive model incorporating longitudinal data using a novel modeling strategy. The predictive performance of the longitudinal model was also compared with a baseline model.

METHODS

A total of 400 patients with HCV-related cirrhosis who achieved SVR with direct-acting antivirals (DAA) were enrolled in the study. Patients were randomly divided into a training set (70%) and a validation set (30%). Informative features were extracted from the longitudinal variables and then put into the random survival forest (RSF) to develop the longitudinal model. A baseline model including the same variables was built for comparison.

RESULTS

During a median follow-up time of approximately 5 years, 25 patients (8.9%) in the training set and 11 patients (9.2%) in the validation set developed HCC. The areas under the receiver-operating characteristics curves (AUROC) for the longitudinal model were 0.9507 (0.8838-0.9997), 0.8767 (0.6972,0.9918), and 0.8307 (0.6941,0.9993) for 1-, 2- and 3-year risk prediction, respectively. The brier scores of the longitudinal model were also relatively low for the 1-, 2- and 3-year risk prediction (0.0283, 0.0561, and 0.0501, respectively). In contrast, the baseline model only achieved mediocre AUROCs of around 0.6 (0.6113, 0.6213, and 0.6480, respectively).

CONCLUSIONS

Our longitudinal model yielded accurate predictions of HCC risk in patients with HCV-relate cirrhosis, outperforming the baseline model. Our model can provide patients with valuable prognosis information and guide the intensity of surveillance in clinical practice.

Statistical perspectives on using hepatocellular carcinoma risk models to inform surveillance decisions

More than 50,000 people are diagnosed with hepatocellular carcinoma (HCC) every year in Europe. Many cases are known to specialist liver centres years before they present with HCC. Despite this, HCC is usually detected at a late stage, when prognosis is very poor. For more than two decades, clinical guidelines have recommended uniform surveillance for all patients with cirrhosis. However, studies continue to show that this broad-based approach is inefficient and poorly implemented in practice. A "personalised" approach, where the surveillance regimen is customised to the needs of the patient, is gaining growing support in the clinical community. The cornerstone of personalised surveillance is the HCC risk model - a mathematical equation predicting a patient's individualised probability of developing HCC within a specific time window. However, although numerous risk models have now been published, few are being used in routine care to inform HCC surveillance decisions. In this article, we discuss methodological issues stymieing the use of HCC risk models in routine practice - highlighting biases, evidence gaps and misconceptions that future research must address.

Advances and challenges in thyroid cancer: The interplay of genetic modulators, targeted therapies, and AI-driven approaches

Thyroid cancer continues to exhibit a rising incidence globally, predominantly affecting women. Despite stable mortality rates, the unique characteristics of thyroid carcinoma warrant a distinct approach. Differentiated thyroid cancer, comprising most cases, is effectively managed through standard treatments such as thyroidectomy and radioiodine therapy. However, rarer variants, including anaplastic thyroid carcinoma, necessitate specialized interventions, often employing targeted therapies. Although these drugs focus on symptom management, they are not curative. This review delves into the fundamental modulators of thyroid cancers, encompassing genetic, epigenetic, and non-coding RNA factors while exploring their intricate interplay and influence. Epigenetic modifications directly affect the expression of causal genes, while long non-coding RNAs impact the function and expression of micro-RNAs, culminating in tumorigenesis. Additionally, this article provides a concise overview of the advantages and disadvantages associated with pharmacological and non-pharmacological therapeutic interventions in thyroid cancer. Furthermore, with technological advancements, integrating modern software and computing into healthcare and medical practices has become increasingly prevalent. Artificial intelligence and machine learning techniques hold the potential to predict treatment outcomes, analyze data, and develop personalized therapeutic approaches catering to patient specificity. In thyroid cancer, cutting-edge machine learning and deep learning technologies analyze factors such as ultrasonography results for tumor textures and biopsy samples from fine needle aspirations, paving the way for a more accurate and effective therapeutic landscape in the near future.

Machine learning prediction models for clinical management of blood-borne viral infections: a systematic review of current applications and future impact

BACKGROUND

Machine learning (ML) prediction models to support clinical management of blood-borne viral infections are becoming increasingly abundant in medical literature, with a number of competing models being developed for the same outcome or target population. However, evidence on the quality of these ML prediction models are limited.

OBJECTIVE

This study aimed to evaluate the development and quality of reporting of ML prediction models that could facilitate timely clinical management of blood-borne viral infections.

METHODS

We conducted narrative evidence synthesis following the synthesis without meta-analysis guidelines. We searched PubMed and Cochrane Central Register of Controlled Trials for all studies applying ML models for predicting clinical outcomes associated with hepatitis B virus (HBV), human immunodeficiency virus (HIV), or hepatitis C virus (HCV).

RESULTS

We found 33 unique ML prediction models aiming to support clinical decision making. Overall, 12 (36.4%) focused on HBV, 10 (30.3%) on HCV, 10 on HIV (30.3%) and two (6.1%) on co-infection. Among these, six (18.2%) addressed the diagnosis of infection, 16 (48.5%) the prognosis of infection, eight (24.2%) the prediction of treatment response, two (6.1%) progression through a cascade of care, and one (3.03%) focused on the choice of antiretroviral therapy (ART). Nineteen prediction models (57.6%) were developed using data from high-income countries. Evaluation of prediction models was limited to measures of performance. Detailed information on software code accessibility was often missing. Independent validation on new datasets and/or in other institutions was rarely done.

CONCLUSION

Promising approaches for ML prediction models in blood-borne viral infections were identified, but the lack of robust validation, interpretability/explainability, and poor quality of reporting hampered their clinical relevance. Our findings highlight important considerations that can inform standard reporting guidelines for ML prediction models in the future (e.g., TRIPOD-AI), and provides critical data to inform robust evaluation of the models.

Perioperative predictors of outcome of hepatectomy for HBV-related hepatocellular carcinoma

Hepatitis B virus (HBV) is identified as a major risk factor for hepatocellular carcinoma (HCC), resulting in so-called hepatitis B virus-related hepatocellular carcinoma (HBV-related HCC). Hepatectomy for HCC is acknowledged as an efficient treatment strategy, especially for early HCC. Furthermore, patients with advanced HCC can still obtain survival benefits through surgical treatment combined with neoadjuvant therapy, adjuvant therapy, transcatheter arterial chemoembolization, and radiofrequency ablation. Therefore, preoperative and postoperative predictors of HBV-related HCC have crucial indicative functions for the follow-up treatment of patients with feasible hepatectomy. This review covers a variety of research results on preoperative and postoperative predictors of hepatectomy for HBV-related HCC over the past decade and in previous landmark studies. The relevant contents of Hepatitis C virus-related HCC, non-HBV non-HCV HCC, and the artificial intelligence application in this field are briefly addressed in the extended content. Through the integration of this review, a large number of preoperative and postoperative factors can predict the prognosis of HBV-related HCC, while most of the predictors have no standardized thresholds. According to the characteristics, detection methods, and application of predictors, the predictors can be divided into the following categories: 1. serological and hematological predictors, 2. genetic, pathological predictors, 3. imaging predictors, 4. other predictors, 5. analysis models and indexes. Similar results appear in HCV-related HCC, non-HBV non-HCV HCC. Predictions based on AI and big biological data are actively being applied. A reasonable prediction model should be established based on the economic, health, and other levels in specific countries and regions.

International Liver Cancer Association (ILCA) white paper on hepatocellular carcinoma risk stratification and surveillance

Major research efforts in liver cancer have been devoted to increasing the efficacy and effectiveness of surveillance for hepatocellular carcinoma (HCC). As with other cancers, surveillance programmes aim to detect tumours at an early stage, facilitate curative-intent treatment, and reduce cancer-related mortality. HCC surveillance is supported by a large randomised-controlled trial in patients with chronic HBV infection and several cohort studies in cirrhosis; however, effectiveness in clinical practice is limited by several barriers, including inadequate risk stratification, underuse of surveillance, and suboptimal accuracy of screening tests. There are several proposed strategies to address these limitations, including risk stratification algorithms and biomarkers to better identity at-risk individuals, interventions to increase surveillance, and emerging imaging- and blood-based surveillance tests with improved sensitivity and specificity for early HCC detection. Beyond clinical validation, data are needed to establish clinical utility, i.e. increased early tumour detection and reduced HCC-related mortality. If successful, these data could facilitate a precision screening paradigm in which surveillance strategies are tailored to individual HCC risk to maximise overall surveillance value. However, practical and logistical considerations must be considered when designing and implementing these validation efforts. To address these issues, ILCA (the International Liver Cancer Association) adjourned a single topic workshop on HCC risk stratification and surveillance in June 2022. Herein, we present a white paper on these topics, including the status of the field, ongoing research efforts, and barriers to the translation of emerging strategies.

Necessity for surveillance for hepatocellualr carcinoma in older patients with chronic hepatitis C who achieved sustained virological response

Background and Aim

Hepatocellular carcinoma (HCC) surveillance in low-risk patients (annual incidence <1.5%) is not recommended per the American Association for the Study of Liver Diseases guidelines. Because patients with chronic hepatitis C with non-advanced fibrosis who have achieved sustained virological response (SVR) have a low risk of HCC, HCC surveillance is not recommended for them. However, aging is a risk factor for HCC; threfore, the necessity for HCC surveillance in older patients with non-advanced fibrosis needs to be verified.

Methods

This multicenter, prospective study enrolled 4993 patients with SVR (1998 patients with advanced fibrosis and 2995 patients with non-advanced fibrosis). The HCC incidence was examined with particular attention to age.

Results

The 3-year incidence of HCC in patients with advanced and non-advanced fibrosis was 9.2% (95% CI: 7.8-10.9) and 2.9% (95% CI: 2.1-3.7), respectively. HCC incidence was significantly higher in patients with advanced fibrosis (P < 0.001). HCC incidence stratified by age and sex was investigated in patients with non-advanced fibrosis. The HCC incidence in the 18-49, 50s, 60s, 70s, and ≥80 age groups were 0.26, 1.3, 1.8, 1.7, and 2.9 per 100 person-years in men, and 0.00, 0.32, 0.58, 0.49, and 0.57 per 100 person-years in women, respectively.

Conclusions

Male patients with non-advanced fibrosis aged ≥60 years have a higher risk of developing HCC and, thus, require HCC surveillance.

The role of artificial intelligence in hepatology research and practice

PURPOSE OF REVIEW

The use of artificial intelligence (AI) in examining large data sets has recently gained considerable attention to evaluate disease epidemiology, management approaches, and disease outcomes. The purpose of this review is to summarize the current role of AI in contemporary hepatology practice.

RECENT FINDINGS

AI was found to be diagnostically valuable in the evaluation of liver fibrosis, detection of cirrhosis, differentiation between compensated and decompensated cirrhosis, evaluation of portal hypertension, detection and differentiation of particular liver masses, preoperative evaluation of hepatocellular carcinoma as well as response to treatment and estimation of graft survival in patients undergoing liver transplantation. AI additionally holds great promise in examination of structured electronic health records data as well as in examination of clinical text (using various natural language processing approaches). Despite its contributions, AI has several limitations, including the quality of existing data, small cohorts with possible sampling bias and the lack of well validated easily reproducible models.

SUMMARY

AI and deep learning models have extensive applicability in assessing liver disease. However, multicenter randomized controlled trials are indispensable to validate their utility.

Interpretable machine learning models for predicting in-hospital death in patients in the intensive care unit with cerebral infarction

BACKGROUND

Research on patients with cerebral infarction in the Intensive Care Unit (ICU) is still lacking. Our study aims to develop and validate multiple machine-learning (ML) models using two large ICU databases-Medical Information Mart for Intensive Care version III (MIMIC-III) and eICU Research Institute Database (eRI)-to guide clinical practice.

METHODS

We collected clinical data from patients with cerebral infarction in the MIMIC-III and eRI databases within 24 h of admission. The opinion of neurologists and the Least Absolute Shrinkage and Selection Operator regression was used to screen for relevant clinical features. Using eRI as the training set and MIMIC-III as the test set, we developed and validated six ML models. Based on the results of the model validation, we select the best model and perform the interpretability analysis on it.

RESULTS

A total of 4,338 patients were included in the study (eRI:3002, MIMIC-III:1336), resulting in a total of 18 clinical characteristics through screening. Model validation results showed that random forest (RF) was the best model, with AUC and F1 scores of 0.799 and 0.417 in internal validation and 0.733 and 0.498 in external validation, respectively; moreover, its sensitivity and recall were the highest of the six algorithms for both the internal and external validation. The explanatory analysis of the model showed that the three most important variables in the RF model were Acute Physiology Score-III, Glasgow Coma Scale score, and heart rate, and that the influence of each variable on the judgement of the model was consistent with medical knowledge.

CONCLUSION

Based on a large sample of patients and advanced algorithms, our study bridges the limitations of studies on this area. With our model, physicians can use the admission information of cerebral infarction patients in the ICU to identify high-risk groups among them who are prone to in-hospital death, so that they could be more alert to this group of patients and upgrade medical measures early to minimize the mortality of patients.

Integrating genetic variants into clinical models for hepatocellular carcinoma risk stratification in cirrhosis

BACKGROUND & AIMS

Identifying individuals at higher risk of developing hepatocellular carcinoma (HCC) is pivotal to improve the performance of surveillance strategies. Herein, we aimed to evaluate the ability of single nucleotide polymorphisms (SNPs) to refine HCC risk stratification.

METHODS

Six SNPs in PNPLA3, TM6SF2, HSD17B13, APOE, and MBOAT7 affecting lipid turnover and one variant involved in the Wnt-β-catenin pathway (WNT3A-WNT9A rs708113) were assessed in patients with alcohol-related and/or HCV-cured cirrhosis included in HCC surveillance programmes (prospective CirVir and CIRRAL cohorts). Their prognostic value for HCC occurrence was assessed using Fine-Gray models combined into a 7-SNP genetic risk score (GRS). The predictive ability of two clinical scores (a routine non-genetic model determined by multivariate analysis and the external aMAP score) with/without the GRS was evaluated by C-indices. The standardised net benefit was derived from decision curves.

RESULTS

Among 1,145 patients, 86 (7.5%) developed HCC after 43.7 months. PNPLA3 and WNT3A-WNT9A variants were independently associated with HCC occurrence. The GRS stratified the population into three groups with progressively increased 5-year HCC incidence (Group 1 [n = 627, 5.4%], Group 2 [n = 276, 10.7%], and Group 3 [n = 242, 15.3%]; p <0.001). The multivariate model identified age, male sex, diabetes, platelet count, gamma-glutamyltransferase levels, albuminemia and the GRS as independent risk factors. The clinical model performance for 5-year HCC prediction was similar to that of the aMAP score (C-Index 0.769). The addition of the GRS to both scores modestly improved their performance (C-Indices of 0.786 and 0.783, respectively). This finding was confirmed by decision curve analyses showing only fair clinical net benefit.

CONCLUSIONS

Patients with cirrhosis can be stratified into HCC risk classes by variants affecting lipid turnover and the Wnt-β-catenin pathway. The incorporation of this genetic information modestly improves the performance of clinical scores.

IMPACT AND IMPLICATIONS

The identification of patients at higher risk of developing liver cancer is pivotal to improve the performance of surveillance. Risk assessment can be achieved by combining several clinical and biological parameters used in routine practice. The addition of patients' genetic characteristics can modestly improve this prediction and will ultimately pave the way for precision medicine in patients eligible for HCC surveillance, allowing physicians to trigger personalised screening strategies.

A risk-based postresection follow-up strategy for hepatocellular carcinoma patients

BACKGROUND

The optimal intervals for follow-up after hepatocellular carcinoma (HCC) patients undergo curative liver resection (LR) remain unclear. This study aimed to establish a risk-based post-resection follow-up strategy.

METHODS

Patients that were diagnosed with HCC and received LR from three hospitals in China were included. The risk-based strategy was established based on the random survival forest model and compared with a fixed strategy both internally and externally.

RESULTS

In total, 3447 patients from three hospitals were included. The authors' strategy showed superiority in the early detection of tumor relapse compared with fixed surveillance. Under fewer total visits, risk-based strategy achieved analogous survival time compared to the total 20 times follow-ups based on fixed strategy. Twelve total visits (five, three, one, two, and one visits in years 1-5, respectively) for American Joint Committee on Cancer/International Union Against Cancer T1a stage patients, 13 total visits (five, four, one, two, and one visits in years 1-5, respectively) for T1b stage patients, 15 total visits (eight, three, three, zero, and one visits in years 1-5, respectively) for T2 stage patients, and 15 total visits (eight, four, one, one, and one visits in years 1-5, respectively) for T3 stage patients were advocated. The detailed follow-up arrangements were available to the public through an interactive website (https://sysuccfyz.shinyapps.io/RiskBasedFollowUp/).

CONCLUSION

This risk-based surveillance strategy was demonstrated to detect relapse earlier and reduce the total number of follow-ups without compromising on survival. Based on the strategy and methodology of the authors, surgeons or patients could choose more intensive or flexible schedules depending on the requirements and economic conditions.

PLAIN LANGUAGE SUMMARY

A risk-based post-resection follow-up strategy was established by random survival forest model using a larger hepatocellular carcinoma population The strategy was demonstrated to detect tumor relapse earlier and reduce the total number of follow-ups without compromising on survival Our strategy and methodology could be widely applied by other surgeons and patients.

Profiling the risk of hepatocellular carcinoma after long-term HCV eradication in patients with liver cirrhosis in the PITER cohort

BACKGROUND AND AIMS

Severe liver disease markers assessed before HCV eradication are acknowledged to usually improve after the SVR. We prospectively evaluated, in the PITER cohort, the long-term HCC risk profile based on predictors monitored after HCV eradication by direct-acting antivirals in patients with cirrhosis.

METHODS

HCC occurrence was evaluated by Kaplan-Meier analysis. Cox regression analysis identified the post-treatment variables associated with de-novo HCC; their predictive power was presented in a nomogram.

RESULTS

After the end of therapy (median follow-up:28.47 months), among 2064 SVR patients, 119 (5.8%) developed de-novo HCC. The HCC incidence was 1.90%, 4.21%, 6.47% at 12-, 24- and 36-months from end-of-therapy, respectively (incidence rate 2.45/100 person-years). Age, genotype 3, diabetes, platelets (PLT)≤120,000/µl and albumin ≤3.5g/dl levels were identified as pre-treatment HCC independent predictors. Adjusting for age, the post-treatment PLT≤120,000/µl (AdjHR 1.92; 95%CI:1.06-3.45) and albumin≤3.5g/dl (AdjHR 4.38; 95%CI 2.48-7.75) values were independently associated with HCC occurrence. Two different risk profiles were identified by combining long-term post-therapy evaluation of PLT ≤ vs. >120,000/µl and albumin ≤ vs. >3.5g/dl showing a significant different HCC incidence rate of 1.35 vs. 3.77/100 p-y, respectively.

CONCLUSIONS

The nomogram score based on age, PLT and albumin levels after SVR showed an accurate prediction capability and may support the customizing management for early HCC detection.

Epidemiology of liver cirrhosis and associated complications: Current knowledge and future directions

Cirrhosis causes a heavy global burden. In this review, we summarized up-to-date epidemiological features of cirrhosis and its complications. Recent epidemiological studies reported an increase in the prevalence of cirrhosis in 2017 compared to in 1990 in both men and women, with 5.2 million cases of cirrhosis and chronic liver disease occurring in 2017. Cirrhosis caused 1.48 million deaths in 2019, an increase of 8.1% compared to 2017. Disability-adjusted life-years due to cirrhosis ranked 16th among all diseases and 7th in people aged 50-74 years in 2019. The global burden of hepatitis B virus and hepatitis C virus-associated cirrhosis is decreasing, while the burden of cirrhosis due to alcohol and nonalcoholic fatty liver disease (NAFLD) is increasing rapidly. We described the current epidemiology of the major complications of cirrhosis, including ascites, variceal bleeding, hepatic encephalopathy, renal disorders, and infections. We also summarized the epidemiology of hepatocellular carcinoma in patients with cirrhosis. In the future, NAFLD-related cirrhosis will likely become more common due to the prevalence of metabolic diseases such as obesity and diabetes, and the prevalence of alcohol-induced cirrhosis is increasing. This altered epidemiology should be clinically noted, and relevant interventions should be undertaken.

Is There a Place for Somatostatin Analogues for the Systemic Treatment of Hepatocellular Carcinoma in the Immunotherapy Era?

Patients with advanced hepatocellular carcinoma (HCC) have a very limited survival rate even after the recent inclusion of kinase inhibitors or immune checkpoint inhibitors in the therapeutic armamentarium. A significant problem with the current proposed therapies is the considerable cost of treatment that may be a serious obstacle in low- and middle-income countries. Implementation of somatostatin analogues (SSAs) has the potential to overcome this obstacle, but due to some negative studies their extensive evaluation came to a halt. However, experimental evidence, both in vitro and in vivo, has revealed various mechanisms of the anti-tumor effects of these analogues, including inhibition of cancer cell proliferation and angiogenesis and induction of apoptosis. Favorable indirect effects such as inhibition of liver inflammation and fibrosis and influence on macrophage-mediated innate immunity have also been noted and are presented in this review. Furthermore, the clinical application of SSAs is both presented and compared with clinical trials of kinase and immune checkpoint inhibitors (ICIs). No direct trials have been performed to compare survival in the same cohort of patients, but the cost of treatment with SSAs is a fraction compared to the other modalities and with significantly less serious side effects. As in immunotherapy, patients with viral HCC (excluding alcoholics), as well as Barcelona stage B or C and Child A patients, are the best candidates, since they usually have a survival prospect of at least 6 months, necessary for optimum results. Reasons for treatment failures are also discussed and further research is proposed.

Evaluation of the aMAP score for hepatocellular carcinoma surveillance: a realistic opportunity to risk stratify

BACKGROUND AND AIMS

The aMAP score is a model that predicts risk of hepatocellular carcinoma (HCC) development in patients with chronic hepatitis. Its performance in a 'real world' surveillance setting has not yet been ascertained.

PATIENTS AND METHODS

We had access to a cohort of 3473 individuals enrolled in a rigorously implemented and prospectively accrued surveillance programme (patients undergoing regular ultrasound and biomarker examination between 1998 and 2021). During this period 445 had HCC detected. Of these, 77.8% had early stage disease (within Milan criteria), permitting potentially curative therapy to be implemented in nearly 70% of cases. We applied the recently developed aMAP score to classify patients according to their initial aMAP score in to low, medium and high-risk groups as proposed in the original publication. The performance of the aMAP score was assessed according to the concordance-index and calibration (i.e. agreement between observed and predicted risk). Allowance was made for competing causes of death.

RESULTS

The aMAP score achieved an overall C-index of 0.81 (95% CI: 0.79-0.82) consistent with the initial report and was unaffected by allowance for competing causes of death. Sub-group analysis showed that the results did not change significantly according to gender, or aetiology. However, aMAP discrimination was greater for younger individuals (versus older individuals), and also for individuals without cirrhosis. The HCC incidence rate was 0.98, 7.05 and 29.1 events per 1000 person-years in the low-, moderate- and high-risk aMAP groups, respectively.

CONCLUSIONS

The results from this 'real-world' cohort demonstrate that risk stratification is a realistic prospect and that identification of a subgroup of chronic liver disease patients who have a very low risk of HCC is feasible.

HCC incidence after hepatitis C cure among patients with advanced fibrosis or cirrhosis: A meta-analysis

BACKGROUND AND AIMS

HCV cure reduces but does not eliminate the risk of HCC. HCC surveillance is recommended in populations where the incidence exceeds 1.5% per year. In cirrhosis, HCC surveillance should continue after HCV cure, although it is uncertain if this should be indefinite. For patients with advanced fibrosis (F3), guidelines are inconsistent in their recommendations. We evaluated the incidence of HCC after HCV cure among patients with F3 fibrosis or cirrhosis.

APPROACH AND RESULTS

This systematic review and meta-analysis identified 44 studies (107,548 person-years of follow-up) assessing the incidence of HCC after HCV cure among patients with F3 fibrosis or cirrhosis. The incidence of HCC was 2.1 per 100 person-years (95% CI, 1.9-2.4) among patients with cirrhosis and 0.5 per 100 person-years (95% CI, 0.3-0.7) among patients with F3 fibrosis. In a meta-regression analysis among patients with cirrhosis, older age (adjusted rate ratio [aRR] per 10-year increase in mean/median age, 1.32; 95% CI, 1.00-1.73) and prior decompensation (aRR per 10% increase in the proportion of patients with prior decompensation, 1.06; 95% CI, 1.01-1.12) were associated with an increased incidence of HCC. Longer follow-up after HCV cure was associated with a decreased incidence of HCC (aRR per year increase in mean/median follow-up, 0.87; 95% CI, 0.79-0.96).

CONCLUSIONS

Among patients with cirrhosis, the incidence of HCC decreases over time after HCV cure and is lowest in patients with younger age and compensated cirrhosis. The substantially lower incidence in F3 fibrosis is below the recommended threshold for cost-effective screening. The results should encourage the development of validated predictive models that better identify at-risk individuals, especially among patients with F3 fibrosis.

"Beyond MELD" - Emerging strategies and technologies for improving mortality prediction, organ allocation and outcomes in liver transplantation

In this review article, we discuss the model for end-stage liver disease (MELD) score and its dual purpose in general and transplant hepatology. As the landscape of liver disease and transplantation has evolved considerably since the advent of the MELD score, we summarise emerging concepts, methodologies, and technologies that may improve mortality prognostication in the future. Finally, we explore how these novel concepts and technologies may be incorporated into clinical practice.

External Validation of LCR1-LCR2, a Multivariable Hepatocellular Carcinoma Risk Calculator, in a Multiethnic Cohort of Patients With Chronic Hepatitis B

Background and Aims

The liver cancer risk test (LCR1-LCR2) is a multianalyte blood test combining proteins involved in liver cell repair (apolipoprotein A1, haptoglobin), hepatocellular carcinoma (HCC) risk factors (gender, age, gamma glutamyl transpeptidase), a marker of fibrosis (alpha2-macroglobulin), and alpha-fetoprotein, a specific marker of HCC. The aim was to externally validate LCR1-LCR2 in hepatitis B.

Methods

Preincluded patients were from the Hepather cohort, a multicenter, multiethnic prospective study in 6071 patients. The coprimary study outcome was the negative predictive value of LCR1-LCR2 at 5 years for the occurrence of HCC and survival without HCC according to the predetermined LCR1-LCR2 cutoffs, adjusted for risk covariables and for chronic hepatitis B treatment and quantified using time-dependent Cox proportional hazards models. A post hoc analysis compared the number of patients needed to screen one cancer by LCR1-LCR2 and PAGE-B.

Results

A total of 3520 patients, 191 (5.4%) with cirrhosis, with at least 1 year of follow-up were included. A total of 76 HCCs occurred over a median (interquartile range) of 6.0 years (4.8-7.3) of follow-up. Among the 3367 patients with low-risk LCR1-LCR2, the 5-year negative predictive value was 99.3% (95% confidence interval = 99.0-99.6), with a significant Cox hazard ratio (6.4, 3.1-13.0; P < .001) obtained after adjustment for exposure to antivirals, age, gender, geographical origin, HBe-Ag status, alcohol consumption, and type-2 diabetes. LCR1-LCR2 outperformed PAGE-B for number of patients needed to screen mean (95% CI), 8.5 (3.2-8.1) vs 26.3 (17.5-38.5; P < .0001), respectively.

Conclusion

The performance of LCR1-LCR2 to identify patients with chronic hepatitis B at very low risk of HCC at 5 years was externally validated. ClinicalTrials.gov: NCT01953458.

Role of three-dimensional printing and artificial intelligence in the management of hepatocellular carcinoma: Challenges and opportunities

Hepatocellular carcinoma (HCC) constitutes the fifth most frequent malignancy worldwide and the third most frequent cause of cancer-related deaths. Currently, treatment selection is based on the stage of the disease. Emerging fields such as three-dimensional (3D) printing, 3D bioprinting, artificial intelligence (AI), and machine learning (ML) could lead to evidence-based, individualized management of HCC. In this review, we comprehensively report the current applications of 3D printing, 3D bioprinting, and AI/ML-based models in HCC management; we outline the significant challenges to the broad use of these novel technologies in the clinical setting with the goal of identifying means to overcome them, and finally, we discuss the opportunities that arise from these applications. Notably, regarding 3D printing and bioprinting-related challenges, we elaborate on cost and cost-effectiveness, cell sourcing, cell viability, safety, accessibility, regulation, and legal and ethical concerns. Similarly, regarding AI/ML-related challenges, we elaborate on intellectual property, liability, intrinsic biases, data protection, cybersecurity, ethical challenges, and transparency. Our findings show that AI and 3D printing applications in HCC management and healthcare, in general, are steadily expanding; thus, these technologies will be integrated into the clinical setting sooner or later. Therefore, we believe that physicians need to become familiar with these technologies and prepare to engage with them constructively.

Liver cancer risk after HCV cure in patients with advanced liver disease without non-characterized nodules

BACKGROUND & AIMS

Recognition of non-characterized liver nodules (NCLN) prior to direct-acting antivirals (DAAs) is associated with increased hepatocellular carcinoma (HCC) risk in patients with HCV. The risk of HCC has not been defined in F3/F4 patients in whom NCLN have been ruled-out before starting DAAs and at sustained virological response (SVR). This study aimed to estimate HCC incidence in this population.

METHODS

We performed a prospective study including HCV-infected patients with F3/F4 fibrosis, without a history of HCC, and who achieved SVR after DAAs. Patients were only included if they had undergone ultrasound imaging that excluded the presence of HCC/NCLN within 30 days after SVR. All patients were evaluated every 6 months until developing primary liver cancer, death or withdrawal of informed consent. HCC incidence was expressed per 100 patient-years (/100PY). Adherence to screening program was calculated every 6 months for the first 48 months.

RESULTS

A total of 185 patients (63/122, F3/F4) were included. Among those with cirrhosis, 92% were Child-Pugh A and 42.7% had clinically significant portal hypertension (CSPH). Albumin-bilirubin score was 1 in 84.9% and 2 in 15.1% of patients, respectively. The median clinical and radiologic follow-up was 52.4 months and 48 months, respectively. Ten patients developed HCC: HCC incidence was 1.46/100PY (95% CI 0.79-2.71) in the whole cohort, 2.24/100PY (95% CI 1.21-4.17) in F4 only and 3.63/100PY (95% CI 1.95-6.74) in patients with CSPH. No HCC was registered in patients with F3. Median time between SVR and HCC occurrence was 28.1 months; 12 non-primary liver cancers were also identified.

CONCLUSIONS

Patients with cirrhosis without NCLN at SVR remain at risk of HCC development. The absence of HCC in patients with F3 reinforces their marginal cancer risk, but prospective studies are needed to exclude them from screening programs.

LAY SUMMARY

Patients with HCV-related cirrhosis, without non-characterized liver nodules at sustained virologic response, remain at risk of hepatocellular carcinoma despite viral cure. However, the cancer risk after successful direct-acting antiviral treatment is marginal in patients with F3 fibrosis without non-characterized liver nodules. If confirmed in larger prospective studies, current screening recommendations may need to be revisited in this group of patients.

Competing risk of the specific mortality among Asian-American patients with prostate cancer: a surveillance, epidemiology, and end results analysis

Background

Adopted the competing-risk model to investigate the relevant factors affecting the prostate cancer (PCa)-specific mortality among Asian-American PCa patients based on the Surveillance, Epidemiology, and End Results (SEER) database.

Methods

The information of 26,293 Asian-American patients diagnosed with PCa between 2004 and 2015 were extracted from the SEER 18 database. Subjects were divided into three groups: died of PCa, died of other causes, survival based on the outcomes at the end of 155 months’ follow-up. Multivariate analysis was performed by the Fine-gray proportional model. Meanwhile, subgroup analyses were conducted risk stratification by race and age.

Results

Age ≥ 65 years [Hazard ratio (HR) = 1.509, 95% confidence interval (CI) 1.299–1.754], race (HR = 1.220, 95% CI 1.028–1.448), marital status (unmarried, single or widowed, HR = 1.264, 95% CI 1.098–1.454), tumor grade II (HR = 3.520, 95% CI 2.915–4.250), the American Joint Committee on Cancer (AJCC) stage (T3: HR = 1.597, 95% CI 1.286–1.984; T4: HR = 2.446, 95% CI 1.796–3.331; N1: HR = 1.504, 95% CI 1.176–1.924; M1: HR = 9.875, 95% CI 8.204–11.887) at diagnosis, radiotherapy (HR = 1.892, 95% CI 1.365–2.623), regional nodes positive (HR = 2.498, 95% CI 1.906–3.274) increased risk of PCa-specific mortality for Asian-American PCa patients, while surgical (HR = 0.716, 95% CI 0.586–0.874) reduced the risk.

Conclusion

The study findings showed that age, race, marital status, tumor grade (II), AJCC stages (T3, T4, N1, M1) at diagnosis, radiotherapy, regional nodes positive and surgery was associated with the specific mortality of PCa patients among Asian-Americans.

Hepatocellular Carcinoma Risk Assessment for Patients With Advanced Fibrosis After Eradication of Hepatitis C Virus

The identification of patients with advanced fibrosis who do not need any further hepatocellular carcinoma (HCC) surveillance after the eradication of hepatitis C is pivotal. In this study, we developed a simple serum-based risk model that could identify patients with low-risk HCC. This was a nationwide multicenter study involving 16 Hospitals in Japan. Patients with advanced fibrosis (1,325 in a derivation cohort and 508 in a validation cohort) who achieved sustained virological responses at 24 weeks after treatment (SVR24) were enrolled. The HCC risk model at any point after SVR24 and its change were evaluated, and subsequent HCC development was analyzed. Based on the multivariable analysis, patients fulfilling all of the factors (GAF4 criteria: gamma-glutamyl transferase < 28 IU/L, alpha-fetoprotein < 4.0 ng/mL, and Fibrosis-4 Index < 4.28) were classified as low-risk and others were classified as high-risk. When patients were stratified at the SVR24, and 1 year, and 2 years after SVR24, subsequent HCC development was significantly lower in low-risk patients (0.5-1.1 per 100 person-years in the derivation cohort and 0.9-1.1 per 100 person-years in the validation cohort) than in high-risk patients at each point. HCC risk from 1 year after SVR24 decreased in patients whose risk improved from high-risk to low-risk (HCC incidence: 0.6 per 100 person-years [hazard ratio (HR) = 0.163 in the derivation cohort] and 1.3 per 100 person-years [HR = 0.239 in the validation cohort]) than in those with sustained high risk. Conclusion: The HCC risk model based on simple serum markers at any point after SVR and its change can identify patients with advanced fibrosis who are at low HCC risk, and these patients may be able to reduce HCC surveillance.

Competing Risk Bias in Prognostic Models Predicting Hepatocellular Carcinoma Occurrence: Impact on Clinical Decision-making

Background and Aims

Existing models predicting hepatocellular carcinoma (HCC) occurrence do not account for competing risk events and, thus, may overestimate the probability of HCC. Our goal was to quantify this bias for patients with cirrhosis and cured hepatitis C.

Methods

We analyzed a nationwide cohort of patients with cirrhosis and cured hepatitis C infection from Scotland. Two HCC prognostic models were developed: (1) a Cox regression model ignoring competing risk events and (2) a Fine-Gray regression model accounting for non-HCC mortality as a competing risk. Both models included the same set of prognostic factors used by previously developed HCC prognostic models. Two predictions were calculated for each patient: first, the 3-year probability of HCC predicted by model 1 and second, the 3-year probability of HCC predicted by model 2.

Results

The study population comprised 1629 patients with cirrhosis and cured HCV, followed for 3.8 years on average. A total of 82 incident HCC events and 159 competing risk events (ie, non-HCC deaths) were observed. The mean predicted 3-year probability of HCC was 3.37% for model 1 (Cox) and 3.24% for model 2 (Fine-Gray). For most patients (76%), the difference in the 3-year probability of HCC predicted by model 1 and model 2 was minimal (ie, within 0 to ±0.3%). A total of 2.6% of patients had a large discrepancy exceeding 2%; however, these were all patients with a 3-year probability exceeding >5% in both models.

Conclusion

Prognostic models that ignore competing risks do overestimate the future probability of developing HCC. However, the degree of overestimation-and the way it is patterned-means that the impact on HCC screening decisions is likely to be modest.

Machine Learning-Based Personalized Prediction of Hepatocellular Carcinoma Recurrence After Radiofrequency Ablation

Background and Aims

Radiofrequency ablation (RFA) is a widely accepted, minimally invasive treatment for hepatocellular carcinoma (HCC). This study aimed to develop a machine learning (ML) model to predict the risk of HCC recurrence after RFA treatment for individual patients.

Methods

We included a total of 1778 patients with treatment-naïve HCC who underwent RFA. The cumulative probability of overall recurrence after the initial RFA treatment was 78.9% and 88.0% at 5 and 10 years, respectively. We developed a conventional Cox proportional hazard model and 6 ML models-including the deep learning-based DeepSurv model. Model performance was evaluated using Harrel's c-index and was validated externally using the split-sample method.

Results

The gradient boosting decision tree (GBDT) model achieved the best performance with a c-index of 0.67 from external validation, and it showed a high discriminative ability in stratifying the external validation sample into 2, 3, and 4 different risk groups (P < .001 among all risk groups). The c-index of DeepSurv was 0.64. In order of significance, the tumor number, serum albumin level, and des-gamma-carboxyprothrombin level were the most important variables for the prediction of HCC recurrence in the GBDT model. Also, the current GBDT model enabled the output of a personalized cumulative recurrence prediction curve for each patient.

Conclusion

We developed a novel ML model for the personalized risk prediction of HCC recurrence after RFA treatment. The current model may lead to the personalization of effective follow-up strategies after RFA treatment according to the risk stratification of HCC recurrence.

Early hepatocellular carcinoma detection using magnetic resonance imaging is cost-effective in high-risk patients with cirrhosis

Background & Aims

Reinforced hepatocellular carcinoma (HCC) surveillance using magnetic resonance imaging (MRI) could increase early tumour detection but faces cost-effectiveness issues. In this study, we aimed to evaluate the cost-effectiveness of MRI for the detection of very early HCC (Barcelona Clinic Liver Cancer [BCLC] 0) in patients with an annual HCC risk >3%.

Methods

French patients with compensated cirrhosis included in 4 multicentre prospective cohorts were considered. A scoring system was constructed to identify patients with an annual risk >3%. Using a Markov model, the economic evaluation estimated the costs and life years (LYs) gained with MRI vs. ultrasound (US) monitoring over a 20-year period. The incremental cost-effectiveness ratio (ICER) was calculated by dividing the incremental costs by the incremental LYs.

Results

Among 2,513 patients with non-viral causes of cirrhosis (n = 840) and/or cured HCV (n = 1,489)/controlled HBV infection (n = 184), 206 cases of HCC were detected after a 37-month follow-up. When applied to training (n = 1,658) and validation (n = 855) sets, the construction of a scoring system identified 33.4% and 37.5% of patients with an annual HCC risk >3% (3-year C-Indexes 75 and 76, respectively). In patients with a 3% annual risk, the incremental LY gained with MRI was 0.4 for an additional cost of €6,134, resulting in an ICER of €15,447 per LY. Compared to US monitoring, MRI detected 5x more BCLC 0 HCC. The deterministic sensitivity analysis confirmed the impact of HCC incidence. At a willingness to pay of €50,000/LY, MRI screening had a 100% probability of being cost-effective.

Conclusions

In the era of HCV eradication/HBV control, patients with annual HCC risk >3% represent one-third of French patients with cirrhosis. MRI is cost-effective in this population and could favour early HCC detection.

Lay summary

The early identification of hepatocellular carcinoma in patients with cirrhosis is important to improve patient outcomes. Magnetic resonance imaging could increase early tumour detection but is more expensive and less accessible than ultrasound (the standard modality for surveillance). Herein, using a simple score, we identified a subgroup of patients with cirrhosis (accounting for >one-third), who were at increased risk of hepatocellular carcinoma and for whom the increased expense of magnetic resonance imaging would be justified by the potential improvement in outcomes.

A 2-Step Strategy Combining FIB-4 With Transient Elastography and Ultrasound Predicted Liver Cancer After HCV Cure

INTRODUCTION

Despite the direct-acting antiviral therapy has dramatically decreased the likelihood of having liver-related complications and extrahepatic outcomes, the risk of developing hepatocellular carcinoma (HCC) is not totally eliminated after sustained virological response (SVR). We aimed to develop an easy-to-apply strategy to be adopted in clinical practice for accurately classifying the HCC risk in hepatitis C virus patients after SVR.

METHODS

Prospective and multicenter study enrolling hepatitis C virus patients with advanced fibrosis (transient elastography [TE] > 10 kPa) or cirrhosis by ultrasound showing SVR. They were followed up until HCC, liver transplantation, death, or until October 2020, which occurred first, with a minimum follow-up period of 6 months after SVR (follow-up: 49 [interquartile range 28-59] months).

RESULTS

Patients with cirrhosis by ultrasound represented 58% (611/1,054) of the overall cohort. During the study, HCC occurrence was 5.3% (56/1,054). Multivariate analyses revealed that Fibrosis-4 (FIB-4) > 3.25 (hazard ratio [HR] 2.26 [1.08-4.73]; P = 0.030), TE (HR 1.02 [1.00-1.04]; P = 0.045) and cirrhosis by ultrasound (HR 3.15 [1.36-7.27]; P = 0.007) predicted HCC occurrence. Baseline HCC screening criteria (TE > 10 kPa or cirrhosis) identified patients at higher risk of HCC occurrence in presence of FIB-4 > 3.25 (8.8%; 44/498) vs FIB-4 < 3.25 (2.4%; 12/506), while those with only FIB > 3.25 had no HCC (0%; 0/50) (logRank 22.129; P = 0.0001). A combination of baseline FIB-4 > 3.25 and HCC screening criteria had an annual incidence >1.5 cases per 100 person-years, while the rest of the groups remained <1 case. Patients who maintained post-treatment FIB-4 > 3.25 and HCC screening criteria remained at the highest risk of HCC occurrence (13.7% [21/153] vs 4.9% [9/184]; logRank 7.396, P = 0.007).

DISCUSSION

We demonstrated that a two-step strategy combining FIB-4, TE, and ultrasound could help stratify HCC incidence risk after SVR.

Performance of models to predict hepatocellular carcinoma risk among UK patients with cirrhosis and cured HCV infection

BACKGROUND & AIMS

Hepatocellular carcinoma (HCC) prediction models can inform clinical decisions about HCC screening provided their predictions are robust. We conducted an external validation of 6 HCC prediction models for UK patients with cirrhosis and a HCV virological cure.

METHODS

Patients with cirrhosis and cured HCV were identified from the Scotland HCV clinical database (N = 2,139) and the STratified medicine to Optimise Treatment of Hepatitis C Virus (STOP-HCV) study (N = 606). We calculated patient values for 4 competing non-genetic HCC prediction models, plus 2 genetic models (for the STOP-HCV cohort only). Follow-up began at the date of sustained virological response (SVR) achievement. HCC diagnoses were identified through linkage to nation-wide cancer, hospitalisation, and mortality registries. We compared discrimination and calibration measures between prediction models.

RESULTS

Mean follow-up was 3.4-3.9 years, with 118 (Scotland) and 40 (STOP-HCV) incident HCCs observed. The age-male sex-ALBI-platelet count score (aMAP) model showed the best discrimination; for example, the Concordance index (C-index) in the Scottish cohort was 0.77 (95% CI 0.73-0.81). However, for all models, discrimination varied by cohort (being better for the Scottish cohort) and by age (being better for younger patients). In addition, genetic models performed better in patients with HCV genotype 3. The observed 3-year HCC risk was 3.3% (95% CI 2.6-4.2) and 5.1% (3.5-7.0%) in the Scottish and STOP-HCV cohorts, respectively. These were most closely matched by aMAP, in which the mean predicted 3-year risk was 3.6% and 5.0% in the Scottish and STOP-HCV cohorts, respectively.

CONCLUSIONS

aMAP was the best-performing model in terms of both discrimination and calibration and, therefore, should be used as a benchmark for rival models to surpass. This study underlines the opportunity for 'real-world' risk stratification in patients with cirrhosis and cured HCV. However, auxiliary research is needed to help translate an HCC risk prediction into an HCC-screening decision.

LAY SUMMARY

Patients with cirrhosis and cured HCV are at high risk of developing liver cancer, although the risk varies substantially from one patient to the next. Risk calculator tools can alert clinicians to patients at high risk and thereby influence decision-making. In this study, we tested the performance of 6 risk calculators in more than 2,500 patients with cirrhosis and cured HCV. We show that some risk calculators are considerably better than others. Overall, we found that the 'aMAP' calculator worked the best, but more work is needed to convert predictions into clinical decisions.

Surveillance for hepatocellular carcinoma in chronic viral hepatitis: Is it time to personalize it?

Surveillance with abdominal ultrasound with or without alpha-fetoprotein is recommended by clinical practice guidelines for patients who are considered to be at risk of developing hepatocellular carcinoma (HCC), including those with cirrhosis, advanced fibrosis and special subgroups of chronic hepatitis B (CHB). Application of the standard surveillance strategy to all patients with chronic liver disease (CLD) with or without cirrhosis imposes major sustainability and economic burdens on healthcare systems. Thus, a number of HCC risk scores were constructed, mainly from Asian cohorts, to stratify the HCC prediction in patients with CHB. Similarly, even if less than for CHB, a few scoring systems were developed for chronic hepatitis C patients or cirrhotic patients with CLD of different etiologies. Recently, a few newsworthy HCC-risk algorithms were developed for patients with cirrhosis using the combination of serologic HCC markers and clinical parameters. Overall, the HCC risk stratification appears at hand by several validated multiple score systems, but their optimal performance is obtained only in populations who show highly homogenous clinic-pathologic, epidemiologic, etiologic and therapeutic characteristics and this limitation poses a major drawback to their sustainable use in clinical practice. A better understanding of the dynamic process driving the progression from CLD to HCC derived from studies based on molecular approaches and genetics, epigenetics and liquid biopsy will enable the identification of new biomarkers to define the individual risk of HCC in the near future, with the possibility to achieve a real and cost/effective personalization of surveillance.

External validation of LCR1-LCR2, a multivariable HCC risk calculator, in patients with chronic HCV

BACKGROUND & AIMS

The Liver Cancer Risk test algorithm (LCR1-LCR2) is a multianalyte blood test combining proteins involved in liver cell repair (apolipoprotein-A1 and haptoglobin), known hepatocellular carcinoma (HCC) risk factors (sex, age, and gamma-glutamyl transferase), a marker of fibrosis (alpha2-macroglobulin) and alpha-fetoprotein (AFP), a specific marker of HCC. The aim was to externally validate the LCR1-LCR2 in patients with chronic HCV (CHC) treated or not with antivirals.

METHODS

Pre-included patients were from the Hepather cohort, a multicentre prospective study in adult patients with CHC in France. LCR1-LCR2 was assessed retrospectively in patients with the test components and AFP, available at baseline. The co-primary study outcome was the negative predictive value (NPV) of LCR1-LCR2 for the occurrence of HCC at 5 years and for survival without HCC according to the predetermined LCR1-LCR2 cut-offs. The cut-offs were adjusted for risk covariables and for the response to HCV treatment, and were quantified using time-dependent proportional hazards models.

RESULTS

In total, 4,903 patients, 1,026 (21.9%) with baseline cirrhosis, were included in the study. Patients were followed for a median of 5.7 (IQR 4.2-11.3) years. A total of 3,788/4,903 (77.3%) patients had a sustained virological response. There were 137 cases of HCC at 5 years and 214 at the end of follow-up. HCC occurred at 5 years in 24/3,755 patients with low-risk LCR1-LCR2 compared with 113/1,148 patients with high-risk LCR1-LCR2. The NPV was 99.4% (95% CI 99.1-99.6). Similar findings (hazard ratio, 10.8; 95% CI, 8.1-14.3; p <0.001) were obtained after adjustment for exposure to antivirals, age, sex, geographical origin, HCV genotype 3, alcohol consumption, and type 2 diabetes mellitus.

CONCLUSIONS

The results showed that LCR1-LCR2 can be used to successfully identify patients with HCV at very low risk of HCC at 5 years.

LAY SUMMARY

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer-related death worldwide and the fastest growing cause of cancer death in many countries. We constructed and internally validated a new multianalyte blood test to assess this Liver Cancer Risk (LCR1-LCR2). This study confirmed the performance of LCR1-LCR2 in patients with chronic HCV in the national French cohort Hepather, and its ability to identify patients at a very low risk of HCC at 5 years.

CLINICAL TRIALS REGISTRATION

The study is registered at ClinicalTrials.gov (NCT01953458).

We need stronger evidence for (or against) hepatocellular carcinoma surveillance

Current guidelines from EASL recommend that most patients with cirrhosis are offered surveillance for hepatocellular carcinoma (HCC), but fewer patients than expected actually receive it. The recommendation is based on observational studies and simulations, not randomised trials. In this opinion piece we argue that a randomised trial of HCC surveillance vs. no surveillance is necessary and feasible, and we believe that clinician and patient participation in HCC surveillance would be better if it were based on trial results demonstrating its value.

Epidemiological projections of viral-induced hepatocellular carcinoma in the perspective of WHO global hepatitis elimination

Hepatitis B is an eminent risk factor for hepatocellular carcinoma (HCC) in Southeast Asia and sub-Saharan Africa, whereas hepatitis C is a key risk factor for HCC in Western Europe and North America. Increased awareness of the global burden of viral hepatitis resulted, in May 2016, in the adoption of the first global health sector strategy on viral hepatitis by the World Health Assembly, which calls for the elimination of viral hepatitis as a public health threat by 2030. Although the incidence of liver cancer resulting from viral infections has increased since the 1990s, the implementation of public health interventions, such as hepatitis B vaccination and antiviral therapies might have reduced the global burdens of HCC. Hepatitis B immunization in infancy has been associated with a reduction in the risk of infant fulminant hepatitis, chronic liver disease, and HCC in Taiwan. Achieving viral hepatitis elimination by 2030 can be accelerated by improving the access to HCC screening programs. HCC surveillance programs in developed countries must be refined to increase an access to personalized surveillance program, whereas the limited access to surveillance and treatment of HCC in developing countries remains a significant public health issue.

Artificial intelligence in precision medicine in hepatology

The advancement of investigation tools and electronic health records (EHR) enables a paradigm shift from guideline-specific therapy toward patient-specific precision medicine. The multiparametric and large detailed information necessitates novel analyses to explore the insight of diseases and to aid the diagnosis, monitoring, and outcome prediction. Artificial intelligence (AI), machine learning, and deep learning (DL) provide various models of supervised, or unsupervised algorithms, and sophisticated neural networks to generate predictive models more precisely than conventional ones. The data, application tasks, and algorithms are three key components in AI. Various data formats are available in daily clinical practice of hepatology, including radiological imaging, EHR, liver pathology, data from wearable devices, and multi-omics measurements. The images of abdominal ultrasonography, computed tomography, and magnetic resonance imaging can be used to predict liver fibrosis, cirrhosis, non-alcoholic fatty liver disease (NAFLD), and differentiation of benign tumors from hepatocellular carcinoma (HCC). Using EHR, the AI algorithms help predict the diagnosis and outcomes of liver cirrhosis, HCC, NAFLD, portal hypertension, varices, liver transplantation, and acute liver failure. AI helps to predict severity and patterns of fibrosis, steatosis, activity of NAFLD, and survival of HCC by using pathological data. Despite of these high potentials of AI application, data preparation, collection, quality, labeling, and sampling biases of data are major concerns. The selection, evaluation, and validation of algorithms, as well as real-world application of these AI models, are also challenging. Nevertheless, AI opens the new era of precision medicine in hepatology, which will change our future practice.

Stratification of Hepatocellular Carcinoma Risk Following HCV Eradication or HBV Control

Hepatocellular carcinoma (HCC) incidence has dramatically decreased in patients infected with HCV and HBV due to the widespread use of highly effective antiviral agents. Nevertheless, a substantial proportion of patients with advanced fibrosis or cirrhosis following HCV clearance of in case of HBV control whatever the stage of fibrosis remains at risk of liver cancer development. Cancer predictors in these virus-free patients include routine parameters estimating coexisting comorbidities, persisting liver inflammation or function impairment, and results of non-invasive tests which can be easily combined into HCC risk scoring systems. The latter enables stratification according to various liver cancer incidences and allocation of patients into low, intermediate or high HCC risk probability groups. All international guidelines endorse lifelong surveillance of these patients using semi-annual ultrasound, with known sensibility issues. Refining HCC prediction in this growing population ultimately will trigger personalized management using more effective surveillance tools such as contrast-enhanced imaging techniques or circulating biomarkers while taking into account cost-effectiveness parameters.