Currently, most chronic hepatitis B (CHB) patients cannot achieve a cure. While guidelines suggest that some CHB patients can discontinue nucleos(t)ide analogs (NAs) after consolidation therapy, virological recurrence remains common. Early virological recurrence typically predicts disease flare-up; however, there are limited biomarkers for assessing the risk of virological recurrence after treatment cessation.
Dr. Yong-an Ye ‘s team at the Dongzhimen Hospital of Beijing University of Chinese Medicine (Chinese Institute of Traditional Chinese Medicine for Liver Diseases) conducted a prospective randomized controlled trial (RCT) and constructed a new predictive model based on serological markers to identify the risk of early virological recurrence in Chinese CHB patients after discontinuing entecavir (ETV). The findings were presented in poster format during the 58th European Association for the Study of the Liver (EASL) International Liver Congress 2023.
Study Overview
Chronic hepatitis B virus (HBV) infection is a significant global public health concern, associated with various liver diseases such as hepatitis, cirrhosis, and hepatocellular carcinoma (HCC). The World Health Organization’s goal is to reduce HBV infection incidence by 90% and HBV-related mortality by 65% by 2030. Notably, Chinese patients account for approximately 30% of all HBV infections. Achieving this goal has important clinical and public health implications.
CHB patients often require lifelong use of nucleos(t)ide analogs (NAs) to maintain disease stability. However, there is still a risk of disease recurrence due to factors such as viral resistance. Additionally, existing guidelines lack sufficient evidence regarding NA discontinuation criteria, and traditional serological markers, including HBV DNA and HBsAg, do not accurately reflect the transcriptional activity of covalently closed circular DNA (cccDNA), making it challenging to predict the risk after treatment cessation. Safely discontinuing treatment is a crucial issue in clinical practice, and the key challenge is finding high-quality markers or techniques to predict recurrence risk and achieve safe discontinuation in eligible patients.
This study aimed to use machine learning techniques to combine new biomarkers, including HBV RNA, hepatitis B core-related antigen (HBcrAg), and large surface antigen (L-HBsAg), with traditional serological markers and liver enzymes at the end of treatment (EOT) to build a model that predicts the risk of early virological recurrence (EVR) within 24 weeks after ETV discontinuation in CHB patients.
This study included a total of 420 patients, of which 172 patients (128 males and 44 females) experienced EVR within 24 weeks after discontinuation, while 248 patients (192 males and 56 females) did not. Patients with higher HBcrAg levels had a worse prognosis and shorter overall survival (P=0.0018). LASSO regression revealed that eight factors [HBcrAg, HBV RNA, L-HBsAg, aspartate aminotransferase (AST), hepatitis B e antibody (anti-HBe), HBsAg, direct bilirubin (DBIL), and total serum protein (TP)] were important features for recurrence.
The RF model based on TI (HBsAg, anti-HBe, AST, DBIL, and TP) [AUC: 0.72, 95% confidence interval (CI): 0.58–0.85] outperformed the SVM (AUC: 0.63, 95% CI: 0.50–0.75) and ANNs (AUC: 0.63, 95% CI: 0.50–0.80) models. Furthermore, the performance of ANNs models that integrated TI and all NSBs (AUC: 0.82, 95% CI: 0.70–0.93) was superior to RF (AUC: 0.81, 95% CI: 0.70–0.92) and SVM (AUC: 0.71, 95% CI: 0.57–0.84). Additionally, the diagnostic efficiency of diagnosing EVR was higher when combining TI and all NSB indicators compared to using a single composite indicator
In conclusion,Machine learning techniques can be used to create predictive models for EVR in CHB patients after discontinuing ETV. Patients with certain EOT features, such as high levels of HBcrAg and L-HBsAg, as well as detectable serum HBV RNA, are at higher risk of EVR. These models can help identify patients suitable for ETV discontinuation with the lowest risk in clinical practice.
Researchers’ Remarks
Chronic HBV infection is the most important cause of HCC in Asia, leading healthcare professionals in high-prevalence HBV infection regions to favor conservative strategies for long-term NA treatment. Moreover, guidelines in most countries have recently lowered the threshold for initiating NA antiviral therapy in adult hepatitis B patients, requiring only that they are over 30 years old and have positive HBV DNA results. However, compared to NA initiation criteria, NA discontinuation criteria are more stringent and less conducive to clinical practice. Therefore, in clinical practice, the risk of discontinuation is still unknown compared to conventional antiviral treatment.
Although current retrospective studies suggest a higher rate of HBsAg clearance in CHB patients after discontinuation, and even small-scale attempts to discontinue NAs in cirrhotic patients, its clinical application remains controversial. The controversy arises from the lack of reliable recurrence prediction factors or methods. Furthermore, doctors and patients are more likely to focus on short-term recurrence risk associated with NA interruption, compared to health risks associated with long-term NA use. To further address this issue, our team conducted a prospective randomized controlled trial (RCT) in 19 centers in China[9] and constructed a new predictive model based on serological markers. This model is one of the few prospective randomized controlled trial predictive models in discontinuation studies and uses readily available biomarkers in clinical practice.
First Author
Associate Chief Physician at Dongzhimen Hospital, Beijing University of Chinese Medicine.
Co-Corresponding Author
Dr. Du Hongbo, Department of Gastroenterology and Infectious Diseases, Dongzhimen Hospital, Beijing University of Chinese Medicine.
Corresponding Author
Dr. Yong-an Ye , Chief Physician, Dongzhimen Hospital, Beijing University of Chinese Medicine, and Director of the Chinese Institute of Traditional Chinese Medicine for Liver Diseases.
TAG: EASL2023;Voice of China;CHB;HBV
