子宫颈癌IVIM与DCE-MRI灌注相关性研究

Investigation of correlation between IVIM and DCE-MRI on uterine cervical carcinoma

目的分析体素内不相干运动(introvoxel incoherent motion,IVIM)与磁共振动态增强扫描(dynamic contrast enhanced MRI,DCE-MRI)两种模型评估子宫颈癌血流灌注参数的相关性,并探讨两种模型参数对不同病理类型子宫颈癌的诊断价值。材料与方法搜集我院2016年4月至2017年6月初诊子宫颈癌53例,所有患者于治疗前行MRI检查,包括IVIM及DCE-MRI序列,测量IVIM的灌注参数D*值、f值,计算两者的乘积f D*值,测量DCE-MRI的参数K^(trans),K_(ep)、Ve值。采用Spearman法分析全部子宫颈癌组、子宫颈鳞癌组及子宫颈腺癌组各参数的相关性。采用独立样本t或t'检验比较各参数在子宫颈鳞癌与腺癌组的差异,并进行ROC曲线分析。结果全部子宫颈癌组,D*及f D*均与DCE-MRI各参数呈轻度至中度正相关(r值:0.357~0.672),而f与DCE-MRI各参数均无相关性。D*与Ktrans在子宫颈腺癌组呈高度正相关性(r=0.900),而在鳞癌组呈轻度至中度正相关(r=0.669)。宫颈腺癌的f、K^(trans)、K_(ep)值均高于宫颈鳞癌(P值:0.002~0.013),ROC曲线分析曲线下面积为0.788、0.749及0.722,以f值>0.29为临界值,诊断子宫颈腺癌的敏感度、特异度及准确率分别为63.64%,92.86%及86.79%。结论不同病理类型子宫颈癌的IVIM及DCE-MRI模型灌注参数均呈正相关,f、Ktrans及Kep值有助于子宫颈鳞癌和腺癌的鉴别。

Objective： This study was aimed to analyze the correlation of perfusion parameters between the IVIM and DCE-MRI models to investigate the feasibility of perfusion evaluation by IVIM on uterine cervical carcinoma （UCC）, and compare the difference of those perfusion parameters between squamous cell carcinoma （SCC） and adenocarcinoma （AdC）. Materials and Methods： Fifty-three patients with newly diagnosed uterine cervical carcinoma were collected in our hospital from April 2016 to June 2017 and performed MRI examination before treatment, including intravoxel incoherent motion （IVIM） and dynamic contrasted-enhanced MRI （DCE-MRI） sequence. Perfusion parameters including D^＊, f , and fD^＊ （f multiple by D^＊） derived from IVIM, K^trans, Kep and Ve derived from DCE-MRI were calculated, and analyzed the correlation between those two models on UCC, SCC and AdC groups respectively using Spearman＇s correlation. Independent sample t test and t＇ test were conducted to compare the difference of those perfusion parameters between SCC and AdC, and ROC was used to compare the area under the curve （AUC）, and analyze the cut-off of the most meaningful parameter and diagnostic effciency. Results： D^＊ and fD^＊ were slightly and moderately positively correlated with all the three parameters of DCE-MRI on all UCC （r： 0.357-0.672）, whereas f had no correlation with parameters of DCE-MRI. D^＊ correlated positively with K^trans highly on AdC （r=0.900） and moderately on SCC （r=0.669）. AdC had higher f, K^trans and Kep value than SCC （P： 0.002-0.013）. With the cut-off of the f value less than 0.29 to diagnosis of AdC, the sensitivity, specifcity and accuracy were 63.64%, 92.86% and 86.79%, respectively. Conclusions： Positive correlations were existed between perfusion parameters by IVIM and those of DCE-MRI on UCC, suggesting that IVIM could provide information about tumor perfusion. IVIM also was helpful to identify different pathological types of UCC.