华支睾吸虫感染背景的肝细胞癌影像组学特征及预测华支睾吸虫感染的性能评价

Imaging omics characteristics of hepatocellular carcinoma with background of Clonorchis sinensis infection and performance evaluation for predicting C.sinensis infection

目的本研究旨在探讨华支睾吸虫感染的肝细胞癌的影像组学特征,并评价其在预测华支睾吸虫感染方面的性能。方法选择在本院接受治疗的150例肝细胞癌患者,纳入时间为2021年1月至2023年12月,根据华支睾吸虫感染阳性和阴性将患者分为阳性组62例,阴性组88例,收集患者临床特征,比较华支睾吸虫感染阳性和阴性患者临床特征、训练组和验证组肝细胞癌临床特征,ROC曲线分析模型预测华支睾吸虫感染的临床价值。结果不同华支睾吸虫感染状态的肝细胞癌患者ALT水平比较存在差异(P<0.05),其他临床特征比较差异无统计学意义(P>0.05)。训练组和验证组肝细胞癌临床特征临床特征比较差异无统计学意义(P>0.05)。使用SVM和KNN分类器分别评估训练组和验证组中MRI成像组学特征模型对华支睾吸虫感染的预测性能。计算全组预测华支睾吸虫感染的MRI四个阶段和综合模型的AUC值、特异性和敏感性,全组中SVM、KNN分类器各分期预测性能相当。全组中SVM、KNN分类器综合模型预测价值较高,且具有较好的敏感性和特异性。结论本研究基于影像组学特征,建立了预测华支睾吸虫感染的模型。该模型在预测华支睾吸虫感染方面表现出较高的准确性,有助于临床医生在早期阶段识别华支睾吸虫感染,从而采取及时有效的干预措施。然而,模型预测性能的提高仍有待于进一步的研究和优化。

Objective The aim of this study is to explore the imaging omics characteristics of hepatocellular carcinoma infected with C.sinensis and evaluate its performance in predicting C.sinensis infection.Methods 150 liver cancer patients who received treatment in our hospital were selected and included from January 2021 to December 2023.Patients were divided into a positive group of 62 cases and a negative group of 88 cases based on positive and negative C.sinensis infections.Clinical characteristics of patients were collected,and the clinical characteristics of liver cancer in the training and validation groups were compared.The ROC curve analysis model was used to predict the clinical value of C.sinensis infections.Results There were differences in ALT levels among liver cancer patients with different levels of C.sinensis infection(P<0.05),while there was no statistically significant difference in other clinical features(P>0.05).There was no statistically significant difference in clinical characteristics between the training group and the validation group for hepatocellular carcinoma(P>0.05).Evaluate the predictive performance of MRI imaging omics feature models for C.sinensis infection in the training and validation groups using SVM and KNN classifiers,respectively.Calculate the AUC values,specificity,and sensitivity of the MRI four stages and comprehensive model for predicting C.sinensis infection in the entire group.The predictive performance of SVM and KNN classifiers for each stage is equivalent in the entire group.The SVM and KNN classifiers in the entire group have high predictive value and good sensitivity and specificity.Conclusion This study established a model for predicting C.sinensis infection based on imaging omics features.This model shows high accuracy in predicting C.sinensis infections,which helps clinical doctors identify C.sinensis infections in the early stages and take timely and effective intervention measures.However,further research and optimization are needed to improve the predictive performance of the model.