磁共振弥散成像参数预测肝细胞癌微血管模式的价值

Diagnostic value of diffusion-weighted imaging parameters for prediction of microvascular patterns in hepatocellular carcinoma

陈千娟 ¹龙莉玲 ¹李晨晖 ¹谢金桓 ¹张会婷²

扫码查看

作者信息

1. 530021 南宁,广西医科大学第一附属医院放射科
2. 430015 武汉,西门子医疗系统有限公司磁共振事业部
折叠

摘要

目的:评估术前磁共振检查中使用两种非高斯弥散模型衍生的定量参数以及常规的表观弥散系数(ADC)预测肝细胞癌(HCC)肿瘤包绕型血管(VETC)及微血管侵犯(MVI)联合表型方面的潜力.方法:前瞻性搜集105例HCC患者,所有患者均在术前两周内进行常规序列及多个b值(0～3000 s/mm2)的DWI检查.通过弥散后处理技术获得体素内非相干运动(IVIM)模型、弥散峰度成像(DKI)模型的衍生定量参数及常规的ADC值,分别由2位放射科医生测量整个病灶的所有弥散参数的平均值.由2位病理科医师联合VETC和MVI结果对HCC组织的微血管模式进行分类,将其分为3个不同的VETC/MVI(VM)组.比较不同VM分组间各个定量参数的差异,采用受试者工作特征(ROC)曲线评估差异具有统计学意义的定量参数的诊断效能,并采用DeLong检验比较各组AUC值的差异.结果:IVIM-Dstar值、DKI-K值在不同VM分组之间差异均有统计学意义(P＜0.001).VM-组的IVIM-Dstar值、DKI-K值低于VM±和VM+组,差异均有统计学意义(P＜0.05).两两比较结果显示,不同VM分组间的DKI-K值差异均有统计学意义(P均＜0.05).ROC曲线分析结果显示,IVIM-Dstar 及 DKI-K 值鉴别不同 VM 分组的 AUC 值分别为 0.756、0.863、0.630 及 0.653、0.802、0.673;IVIM-Dstar值及DKI-K值组成的联合模型鉴别不同VM分组的AUC值分别为0.769、0.896和0.702(P＜0.05).Delong检验结果表明,IVIM-Dstar、DKI-K以及两者组成的联合模型在鉴别不同VM分组的效能方面差异无统计学意义(P＞0.05).结论:磁共振非高斯弥散模型定量参数在术前预测HCC的VETC及 MVI联合表型方面具有较好的应用价值.

Abstract

Objective:To evaluate the potential of two quantitative parameters derived from two non-Gaussian diffusion models and conventional apparent diffusion coefficient(ADC)in preoperative magnetic resonance imaging(MRI)for predicting tumor encapsulated blood vessels(VETC)and mi-crovascular infiltration(MVI)in hepatocellular carcinoma(HCC).Methods:This study prospectively collected 105 HCC patients.All patients underwent routine sequence and diffusion-weighted imaging(DWI)MRI with multiple b-values(0～3000s/mm2)within two weeks before surgery.The voxel-wise non-Gaussian incoherent motion(IVIM)model,diffusion kurtosis imaging(DKI)model-derived quantitative parameters,and conventional ADC were obtained using diffusion post-processing tech-niques.The average values of all diffusion parameters of the entire lesion were measured by two radiol-ogists.The VETC and MVI results of HCC tissues were confirmed and divided into three different VETC/MVI(VM)groups by two pathologists.The quantitative parameters were compared between VM groups.The diagnostic performance of quantitative parameters with statistically significant differ-ences was then evaluated by receiver operating characteristic(ROC)curves.Finally,the differences in AUC values between groups were compared using the DeLong test.Results:There were significant differences in IVIM-Dstar and DKI-K values among the VM groups(P＜0.001).The IVIM-Dstar and DKI-K values in the VM-group were significantly lower than those in the VM±and VM+groups.In pairwise comparisons,the DKI-K values showed significant differences between each VM subgroup(P all＜0.05).The AUC values for differentiating various VM subgroups were 0.756,0.863,0.630 for IVIM-Dstar,and 0.653,0.802,0.673 for DKI-K.The combined model,utilizing IVIM-Dstar and DKI-K values,demonstrated AUC values of 0.769,0.896,and 0.702(P＜0.05).The Delong test demonstrated that there were no significant differences in diagnostic performance among the VM groups for IVIM-Dstar,DKI-K,and the combined model(P＞0.05).Conclusion:Quantitative parameters derived from non-Gaussian diffusion models have significant practical value in the preoperative prediction of VETC and MVI in HCC.

关键词

肝细胞癌/肿瘤包绕型血管/微血管侵犯/磁共振成像/非高斯弥散模型

Key words

Hepatocellular carcinoma/Tumor encapsulated blood vessels/Microvascular infil-tration/Magnetic resonance imaging/Non-Gaussian diffusion model

引用本文复制引用

基金项目

国家自然科学基金(82060310)

出版年

2024

放射学实践

华中科技大学同济医学院

放射学实践

CSTPCDCSCD北大核心

影响因子：1.08

ISSN：1000-0313

浏览量1

参考文献量34

段落导航