Wang Yiting, Wang Tie, Wen Xin, Feng Chongchong
Department of Laboratory Medicine, Second Hospital of Jilin University, Changchun, China.
Department of Laboratory Medicine, First Hospital of Jilin University, Changchun, China.
Parasit Vectors. 2025 May 22;18(1):186. doi: 10.1186/s13071-025-06833-9.
Hepatic clonorchiasis is one of the most prevalent foodborne parasitic diseases in China and is often overlooked because the initial symptoms are not obvious. In this study, a multivariate model for the early prediction of disease onset using laboratory test data from liver-fluke-infected patients was developed and validated.
Laboratory data from 147 liver-fluke-infected patients and 151 healthy control subjects were collected. Univariate logistic regression, Spearman correlation analysis, and collinearity diagnosis were used to screen for independent factors. A multivariate model was then constructed using the backward likelihood ratio method. For external validation, an independent patient cohort from another hospital was analyzed. The discriminative performance of the combined model was compared with that of previously identified biomarkers (eosinophil count and γ-glutamyl transpeptidase).
A 12-indicator prediction model for liver fluke infection was developed using traditional logistic regression (82.31% sensitivity and 88.08% specificity). The receiver operating characteristic curve, calibration curve, and decision curve analyses revealed that the model exhibited excellent discriminative ability (area under the curve [AUC]: training = 0.928, validation = 0.808), goodness of fit, and clinical practicability. The combined model showed superior discrimination compared with individual biomarkers, including eosinophil count (AUC = 0.577) and γ-glutamyl transpeptidase (AUC = 0.620).
This study developed an early risk prediction model for liver fluke infection using routine laboratory test data. Compared with previously reported biomarkers, the model demonstrated superior diagnostic performance and showed potential as a clinical tool for identifying early stage liver fluke infection in patients.
肝吸虫病是中国最常见的食源性寄生虫病之一，由于其初期症状不明显，常被忽视。在本研究中，我们开发并验证了一个使用肝吸虫感染患者的实验室检测数据进行疾病发病早期预测的多变量模型。
收集了147例肝吸虫感染患者和151例健康对照者的实验室数据。采用单因素逻辑回归、Spearman相关性分析和共线性诊断来筛选独立因素。然后使用向后似然比法构建多变量模型。为进行外部验证，分析了来自另一家医院的独立患者队列。将联合模型的判别性能与先前确定的生物标志物（嗜酸性粒细胞计数和γ-谷氨酰转肽酶）进行比较。
使用传统逻辑回归开发了一个用于肝吸虫感染的12指标预测模型（灵敏度为82.31%，特异度为88.08%）。受试者工作特征曲线、校准曲线和决策曲线分析表明，该模型具有出色的判别能力（曲线下面积[AUC]：训练集 = 0.928，验证集 = 0.808）、拟合优度和临床实用性。与包括嗜酸性粒细胞计数（AUC = 0.577）和γ-谷氨酰转肽酶（AUC = 0.620）在内的单个生物标志物相比，联合模型显示出更好的判别能力。
本研究使用常规实验室检测数据开发了肝吸虫感染的早期风险预测模型。与先前报道的生物标志物相比，该模型具有更好的诊断性能，显示出作为识别患者早期肝吸虫感染的临床工具的潜力。
