AUTHOR=Yu Mingxue , Li Xiangyong , Lu Yaxin , Jie Yusheng , Li Xinhua , Shi Xietong , Zhong Shaolong , Wu Yuankai , Xu Wenli , Liu Zifeng , Chong Yutian 

TITLE=Development and Validation of a Novel Risk Prediction Model Using Recursive Feature Elimination Algorithm for Acute-on-Chronic Liver Failure in Chronic Hepatitis B Patients With Severe Acute Exacerbation

JOURNAL=Frontiers in Medicine

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2021.748915

DOI=10.3389/fmed.2021.748915

ISSN=2296-858X

ABSTRACT=Background: Chronic hepatitis B (CHB) patients with severe acute exacerbation (SAE) is at progression stage of acute-on-chronic liver failure (ACLF) but uniform models for predicting ACLF occurrence are lacking. We aimed to present a risk prediction model to early identify the patients at high-risk of ACLF and predict patient’s survival.
Methods: We selected the best variable combination using a novel recursive feature elimination algorithm to develop and validate a classification regression model as well as online application on cloud server from the training cohort with a total of 342 CHB patients with SAE and two external cohorts with a sample size of 96 and 65 patients, respectively. 
Findings: An excellent prediction model called PATA model including four predictors, prothrombin time, age, total bilirubin and alanine aminotransferase could achieve AUC of 0.959 (95% CI 0.941-0.977) in the development set and AUC of 0.932 (95% CI 0.876-0.987) and 0.905 (95% CI 0.826-0.984) in two external validation cohorts, respectively. The calibration curve for risk prediction probability of ACLF showed optimal agreement between prediction by PATA model and actual observation. After predictive stratification into different risk groups, the C-index of predictive 90-days mortality was 0.720 (0.675-0.765) for PATA model, 0.549 (0.506-0.592) for the end-stage liver disease score model, and 0.648 (0.581-0.715) for Child–Turcotte–Pugh scoring system.
Interpretation: The highly-predictive risk model and easy-to-use online application can accurately predict the risk of ACLF with poor prognosis. They may facilitate risk communication and guide therapeutic options.