摘要
目的探讨基于表观扩散系数(ADC)图的影像组学模型评价肝纤维化分期的价值。资料与方法回顾性分析103例肝纤维化患者的ADC图和临床资料,按1∶1随机分为训练组和验证组。在训练组中,提取ADC图中的影像组学特征,使用LASSO算法进行特征降维并建立影像组学纤维化指数(RFI)模型。绘制受试者工作特征(ROC)曲线,确定有临床意义的纤维化(≥S2期)、晚期纤维化(≥S3期)和肝硬化(≥S4期)的最佳截断值,并评价其在验证组中的敏感度、特异度和准确度;比较RFI、ADC值和天冬氨酸氨基转移酶/血小板比值指数(APRI)的ROC曲线下面积(AUC)。结果RFI在验证组中诊断不同肝纤维化分期的AUC为0.81~0.86,大于ADC的0.51~0.73和APRI的0.58~0.63。使用在训练组中确定的截断值,RFI在验证组中诊断≥S2、≥S3和≥S4期肝纤维化的敏感度分别为76%、87%和65%,特异度分别为82%、86%和80%,准确度分别为77%、87%和75%。结论基于ADC图的影像组学模型可以定量评估肝纤维化分期,有望为临床提供一种无创性评价工具。
Purpose To explore the value of radiomics model based on the apparent diffusion coefficient(ADC)map in evaluating the staging of hepatic fibrosis.Materials and Methods The ADC maps and clinical data of 103 patients with hepatic fibrosis were retrospectively analyzed.All patients were randomly divided into training group and verification group at 1∶1.In the training group,the radiomics features were extracted from the ADC map,and the feature dimensionality reduction was carried out by using LASSO algorithm,and the radiomics fibrosis index(RFI)model was established.The receiver operating characteristic(ROC)curves were drawn to determine the best cut-off values of clinically significant fibrosis(≥S2),advanced fibrosis(≥S3),and liver cirrhosis(≥S4),and to evaluate their sensitivity,specificity and accuracy in the verification group.The area under curves(AUCs)of RFI,ADC and aspartate aminotransferaseto-platelet ratio index(APRI)were compared between the two groups.Results The AUC(0.81-0.86)of RFI for different stages of hepatic fibrosis was higher than that of ADC(0.51-0.73)and APRI(0.58-0.63)in the verification group.Using the cut-off values determined by the training group,the sensitivity of RFI in the verification group for the diagnosis of≥S2,≥S3 and≥S4 hepatic fibrosis was 76%,87%and 65%,respectively;the specificity was 82%,86%and 80%,respectively;the accuracy was 77%,87%and 75%,respectively.Conclusion The radiomics model based on the ADC map can quantitatively evaluate the staging of hepatic fibrosis and be expected to use as a non-invasive evaluation tool in clinical practice.
作者
张宝腾
郑悠
黎金葵
翟亚楠
张波
郭顺林
ZHANG Baoteng;ZHENG You;LI Jinkui;ZHAI Yanan;ZHANG Bo;GUO Shunlin(Department of Radiology,the First Hospital of Lanzhou University,Lanzhou 730000,China;不详)
出处
《中国医学影像学杂志》
CSCD
北大核心
2021年第4期362-366,371,共6页
Chinese Journal of Medical Imaging
基金
兰州市科技发展指导性计划项目(2019-ZD-46)
兰州大学第一医院青年基金(ldyyyn2018-12)。
关键词
肝纤维化
磁共振成像
影像组学
扩散加权成像
表观扩散系数
Hepatic fibrosis
Magnetic resonance imaging
Radiomics
Diffusion weighted imaging
Apparent diffusion coefficient
