Yang Jun, Xue Ran, Wu Jing, Jia Lin, Li Juan, Yu Hongwei, Zhu Yueke, Dong Jinling, Meng Qinghua
Department of Critical Care Medicine of Liver Disease, Beijing Youan Hospital, Capital Medical University, Beijing, China.
Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Phase I Clinical Trial, Peking University Cancer Hospital & Institute, Beijing, China.
J Clin Transl Hepatol. 2022 Jun 28;10(3):458-466. doi: 10.14218/JCTH.2021.00202. Epub 2021 Oct 18.
It is challenging to predict the 90-day outcomes of patients infected with hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) via prevailing predictive models. This study aimed to develop an innovative model to enhance the analytical efficacy of 90-day mortality in HBV-ACLF.
In this study, 149 HBV-ACLF patients were evaluated by constructing a death risk prediction nomogram. Bootstrap resampling and an independent validation cohort comprising 31 patients from June 2019 to February 2020 were assessed for model confirmation.
The nomogram was constructed by entering and identifying five factors (age, total bilirubin, prothrombin activity (PTA), lymphocyte (L)%, and monocyte (M)%. Healthy refinement was achieved from the nomogram analysis, where the area under the receiver operating characteristic curve was 0.864 for the training cohort and 0.874 was achieved for the validation cohort. There was admirable concordance between the predicted and true results in the equilibrium curve. The decision curve assessment revealed the useful clinical application of the nomogram.
We constructed an innovative nomogram and validated it for the prediction of 90-day HBV-ACLF patient outcomes. This model might help develop optimized treatment protocol recommendations for HBV-ACLF patients.
通过现有的预测模型预测乙型肝炎病毒相关慢加急性肝衰竭（HBV-ACLF）患者90天的预后具有挑战性。本研究旨在开发一种创新模型，以提高HBV-ACLF患者90天死亡率的分析效能。
在本研究中，通过构建死亡风险预测列线图对149例HBV-ACLF患者进行评估。采用自助重抽样法，并对2019年6月至2020年2月期间的31例患者组成的独立验证队列进行模型验证。
通过纳入并确定五个因素（年龄、总胆红素、凝血酶原活动度（PTA）、淋巴细胞（L）%和单核细胞（M）%）构建列线图。列线图分析实现了良好的优化，训练队列的受试者工作特征曲线下面积为0.864，验证队列的该面积为0.874。在一致性曲线中，预测结果与真实结果之间具有良好的一致性。决策曲线评估显示了列线图的临床应用价值。
我们构建了一种创新的列线图，并对其预测HBV-ACLF患者90天预后的能力进行了验证。该模型可能有助于为HBV-ACLF患者制定优化的治疗方案建议。
