利用MR动态增强纹理分析评估不同分子亚型乳腺癌异质性

Dynamic contrast-enhanced MR texture analysis for assessment of heterogeneity of different molecular subtypes of breast cancer

目的利用磁共振纹理分析评估不同分子亚型乳腺癌的异质性。方法回顾性分析353例经病理证实为乳腺浸润性导管癌患者的临床资料及MRI表现,根据免疫组化或荧光原位杂交结果分为Luminal A型、Luminal B型、人类表皮生长因子受体2(human epidermal growth factor receptor-2,HER-2)过表达型与三阴型(triple-negative breast cancer,TNBC)四型,利用第三方软件手动勾画病灶所有层面融合为三维感兴趣区并提取纹理特征(18种),采用2检验、秩和检验、单因素方差分析及LSD-t检验进行统计学分析。结果Luminal A型90例,Luminal B型115例,HER-2过表达型66例,TNBC型82例。四种亚型中发病年龄、标准差、方差、范围、偏度、峰度、均匀性、能量、熵、惯性矩、聚类萌、聚类突、总均值、最小强度、最大强度、表面积体积比的差异均有统计学意义(P<0.05)。两两比较Luminal A型与HER-2过表达型、TNBC型均有14个特征差异具有统计学意义,Luminal B型与HER-2过表达型、TNBC型分别有12、11个特征差异具有统计学意义,Luminal A型与Luminal B型、HER-2过表达型与TNBC型间均有6个特征差异具有统计学意义。结论基于动态增强MRI纹理分析证实不同亚型乳腺癌具有异质性,纹理特征能够很好地区分HR+与HR-型乳腺癌,纹理特征是潜在评估乳腺癌异质性的一种方法。

Objective To evaluate and monitor the heterogeneity of different molecular subtypes of breast cancer using magnetic resonance texture analysis.Methods We retrospectivly analysed of 353 clinical data and MRI images of patients with breast invasive ductal carcinoma.According to the results of IHC and FISH,they were divided into four molecular subtypes:Luminal A,Luminal B,HER2 over expression,and triple-negative breast cancer.Using third-party software,we manually delineazted all aspects of the lesion into a three-dimensional region of interest and extract texture features.All the features were analyzed by 2,ANOVA,Rank sum test and LSD-t test.Results Of all the patients,90 were Luminal A,115 were Luminal B,66 were HER-2 over expressed,and 82 were TNBC.Age of onset,StdDeviation,Variance,Range,skewness,kurtosis,uniformity,Energy,Entropy,Inertia,ClusterShade,ClusterProminence,sumAverage,sumVariance,Minintensity,Maxintensity,Surface Volume Ratio were significantly different among the four subtypes(P<0.05).There were 14 features in Luminal A vs.HER-2 over expression and TNBC,with significant difference.Luminal B vs.HER-2 over expression and TNBC were 12,11,with statistically significant features respectively.6 features were significantly different between Luminal A vs.Luminal B and HER-2 over expression vs.TNBC.Conclusion Heterogeneity of different subtypes of breast cancer is confirmed by Dynamic enhanced MRI texture analysis.Texture features can distinguish HR+and HR-breast cancer.Texture analysis can be used to assess breast cancer heterogeneity.