乳腺癌动态增强磁共振成像纹理分析对新辅助化疗疗效预测与评估研究

Prediction and evaluation of the efficacy of neoadjuvant chemotherapy for breast cancer based on the texture parameters of dynamic contrast-enhanced magnetic resonance imaging

目的:探讨乳腺癌动态增强磁共振成像(dynamic contrast-enhanced magnetic resonance imaging,DCE-MRI)常用定量参数图的纹理参数对新辅助化疗(neoadjuvant chemotherapy,NAC)的疗效预测与评估价值。方法:回顾并分析山西省肿瘤医院2014年9月—2018年10月在NAC前后均行DCE-MRI且经病理学检查证实的乳腺癌患者共63例,以病理学检查结果为依据,将患者分为治疗有效组(40例)和治疗无效组(23例),使用Omni-Kinetics软件从患者NAC前及NAC 4~8个周期后的DCE-MRI的4个参数图中各提取52个纹理参数,使用独立样本t检验或Mann-Whitney U检验对两组纹理参数进行统计分析,将有差异性的参数进行单因素逻辑回归分析,选择与NAC疗效有关联性的纹理参数,绘制受试者工作特征(receiver operating characteristic curve,ROC)曲线,根据曲线下面积(area under curve,AUC)得出DCE-MRI纹理参数对乳腺癌NAC疗效的诊断效能。结果:从患者DCE-MRI常用的4个定量参数图(Ktrans、Kep、Ve、Vp)中共提取了208个纹理参数。化疗有效组和无效组患者NAC前的DCE-MRI Ktrans、Kep、Vp参数图中,分别有13、17、10个纹理参数差异有统计学意义(P<0.05),经单因素逻辑回归分析,分别有1、7、3个纹理参数与NAC疗效相关联,而在Ve图中,未发现差异有统计学意义的纹理参数(P<0.05)。但上述参数AUC均小于0.8。NAC后的纹理参数及参数变化率可以评估NAC疗效,效能较好的参数主要有4个参数图中的体积数、游程长度不均匀性、Δ体积数、Δ游程长度不均匀性以及Ktrans图中的灰度不均匀性、Δ灰度不均匀性和Kep图中的Δ体素值总和、Δ灰度不均匀性,其AUC均较高,且有较高的灵敏度和特异度。结论:基于DCE-MRI定量参数图的纹理参数中,存在可以预测及评估乳腺癌NAC疗效的参数。

Objective:To investigate the role of texture parameters obtained from the commonly used quantitative parameter maps derived from dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI)in predicting and evaluating the efficacy of neoadjuvant chemotherapy(NAC)in breast cancer.Methods:A total of 63 breast cancer patients with pathologically confirmed in the Shanxi Province Tumor Hospital from Sep.2014 to Oct.2018 were retrospectively analyzed.All these patients underwent DCEMRI before and after NAC.According to the pathological results,the patients were divided into the effective group(40 cases)and the ineffective group(23 cases).Then we used omni-kinetics software to extract 52 texture parameters from each of the four maps derived from DCE-MRI before NAC and after 4-8 cycles of treatment.The independent-sample t test or Mann-Whitney U test were used to statistical analysis between the texture parameters of the two groups.The parameters with statistical differences were analyzed by single factor logistic regression,and the texture parameters related to the efficacy of NAC were selected.The receiver operating characteristic(ROC)curves were drawn.According to the area under the curve(AUC),the diagnostic efficacy of texture parameters in DCE-MRI image for the curative effect of NAC of breast cancer was obtained.Results:In this study,a total of 208 texture parameters were extracted from four commonly used quantitative maps(Ktrans,Kep,Ve,Vp)of patients’DCE-MRI.In Ktrans,Kep,and Vp maps of DCE-MRI before NAC in the effective and ineffective groups of patients,there were 13,17,and 10 texture parameters with statistically significant differences,respectively(P<0.05).Univariate logistic regression analysis showed that there were 1,7,and 3 texture parameters associated with the efficacy of NAC.No statistically significant texture parameters were found in the Ve map(P<0.05).But the AUC of the above parameters were all less than 0.8.Post-NAC texture parameters and parameter change rates could evaluate the efficacy of NAC.The parameters with better performance mainly included volume count,run length non-uniformity,Δvolume count,Δrun length non-uniformity in the four maps.The grey level non-uniformity andΔgrey level non-uniformity in the Ktrans maps,and theΔvoxel value sum andΔgrey level non-uniformity in the Kep maps also had good evaluation performance.All of them had higher AUC,and had high sensitivity and specificity.Conclusion:Among the texture parameters based on the DCE-MRI quantitative parameter maps,there are parameters that can predict and evaluate the efficacy of NAC in breast cancer.