IVIM联合动态增强MRI在非肿块强化腺病与乳腺癌中的鉴别诊断价值

Differential diagnostic value of IVIM combining with dynamic enhanced MRI in non-mass enhancement adenosis and breast cancer

目的探讨体素内不相干运动(intravoxel incoherent motion,IVIM)在非肿块强化(non-mass enhancement,NME)腺病与乳腺癌中的鉴别诊断价值。材料与方法回顾性分析2020年1月至2022年3月乳腺动态对比增强MRI(dynamic contrast-enhanced MRI,DCE-MRI)表现为NME且术后经病理证实为乳腺腺病和原发性乳腺癌的106例患者病例资料,分为乳腺腺病(55例)和乳腺癌(51例)两组,分别测量两组的IVIM定量参数[标准表观扩散系数(standard apparent diffusion coefficient,SADC)、真实扩散系数(slow ADC,D)、伪扩散系数(fast ADC,D^(*))和灌注分数(perfusion fraction,f)]并分析病灶的位置、DCE影像特征[纤维腺体组织量(fibroglandular tissue,FGT)、乳腺实质背景强化(background parenchymal enhancement,BPE)、分布特点、内部强化特征和时间-信号强度曲线(time-signal intensity curve,TIC)类型]。采用单因素分析法分别分析两组数据正态性,符合正态分布的数据选择独立样本t检验,不符合正态分布的数据选用非参数秩和检验,比较两组间差异并进行诊断效能评价,通过多因素逻辑回归分析筛选出最佳参数并进行联合诊断分析,通过受试者工作特征(receiver operating characteristic,ROC)曲线分析获得最佳阈值、敏感度和特异度。结果乳腺癌的f值低于腺病[(0.52±0.14)vs.(0.58±0.11),P=0.028],D^(*)值显著高于腺病[(7.85±7.60)×10^(-3)mm^(2)/s vs.(3.67±1.61)×10^(-3)mm^(2)/s,P<0.001],两者用于鉴别诊断的曲线下面积(area under the curve,AUC)和阈值分别为0.608、0.730和0.73、4.10×10^(-3)mm^(2)/s。乳腺癌病灶内部强化特征以不均匀强化为主(54.9%),TIC类型以平台型为主(70.6%);腺病内部强化特征以不均匀强化为主(65.5%),TIC类型以缓升型为主(72.7%),两者用于鉴别诊断的AUC分别为0.709和0.810。D^(*)值与TIC类型是预测NME乳腺癌的独立风险因素(P=0.003、0.006),两者用于鉴别诊断的AUC分别为0.730和0.810,而两者联合诊断的效能明显提高(AUC为0.860,敏感度为72.5%,特异度为85.5%)。将单纯腺病(n=39)与硬化性腺病(n=16)进一步进行比较,硬化性腺病的SADC、D、f值均高于单纯腺病(P=0.076、0.176、0.181),D^(*)值低于单纯腺病(P=0.331);单纯腺病主要表现为均匀强化或不均匀强化(占比分别为43.6%和51.3%),而硬化性腺病全部为不均匀强化,两者间差异具有统计学意义(P=0.001),内部强化特征鉴别单纯腺病与硬化性腺病的AUC为0.692,敏感度为100.0%。结论IVIM联合DCE-MRI有助于提高MRI对于NME腺病和乳腺癌的鉴别诊断效能;TIC类型和D^(*)值是预测NME乳腺癌的独立风险因素,NME病变内部的强化特点有助于单纯腺病与硬化性腺病的鉴别。

Objective:To investigate the differential diagnostic value of IVIM in non-mass enhancement(NME)adenosis and breast cancer.Materials and Methods:Retrospective analysis of 106 cases with NME manifestation on dynamic contrast-enhanced(DCE)MRI were performed between January 2020 and March 2022.The cases were divided into two groups according to pathological results:adenosis(n=55)and breast cancer(n=51),the intravoxel incoherent motion(IVIM)quantitative parameters[standard apparent diffusion coefficient(SADC),slow ADC(D),fast ADC(D^(*))and perfusion fraction(f)]were measured,and the lesion location and DCE characteristics[fibroglandular tissue(FGT),background parenchymal enhancement(BPE),distribution characteristics,internal enhancement characteristics,and time-signal intensity curve(TIC)type]were compared and analyzed,univariate analysis was used to analyze the normality of the two groups of data separately,the independent sample t-test was selected for the normal distribution data,and the nonparametric rank sum test was selected for the non-normal distribution,the statistical differences between the two groups were compared and the diagnostic efficacy was evaluated,the best parameters were screened out by multivariate logistic regression analysis for joint diagnosis,and the threshold,sensitivity and specificity were obtained by the receiver operating characteristic(ROC)curve analysis.Results:Breast cancer had f was lower than adenosis[(0.52±0.14)vs.(0.58±0.11),P=0.028;threshold was 0.73],and D^(*)was significantly higher than adenosis[(7.85±7.60)×10^(-3)mm^(2)/s vs.(3.67±1.61)×10^(-3)mm^(2)/s,P<0.001;threshold was 4.10×10^(-3)mm^(2)/s],with an area under the curve(AUC)for disease diagnosis of 0.608 and 0.730,respectively.Within breast cancer lesions were heterogeneous(54.9%),TIC(70.6%),adenosis(65.5%),and TIC type I(72.7%),both AUC for disease diagnosis were 0.709 and 0.810,respectively.D^(*)and TIC types were independent risk factors for NME breast cancer(P=0.003 and 0.006,respectively),the AUC of both for disease diagnosis was 0.730 and 0.810,respectively,and the combination was significantly better(AUC=0.860,sensitivity was72.5%,specificity was 85.5%).Further comparing pure adenosis(n=39)with sclerosing adenosis(n=16),the SADC,D,and f values were higher than those of sclerosing adenosis(P=0.076,0.176,0.181,respectively),D^(*)value lower than adenosis(P=0.331);the internal enhancement characteristics of pure adenosis were dominated by homogeneous enhancement and heterogeneous enhancement(43.6%and 51.3%,respectively),while sclerosing adenosis cases were all heterogeneous enhancement,accounting for 100.0%,the difference was statistically significant(P=0.001),diagnostic performance with an AUC of 0.692,sensitivity was 100.0%.Conclusions:Combining IVIM with DCE helps to improve the differential efficacy of MRI for the diagnosis of adenosis and breast cancer with NME manifestation.The TIC type and D^(*)values are independent risk factors for predicting breast cancer with NME manifestation,and the internal enhancement features contribute to the ability of differentiation between the pure adenosis and sclerosing adenosis.