基于全病变MRI动态增强扫描直方图分析对小乳腺癌腋窝淋巴结转移的预测价值  

Predictive value of whole-volume dynamic contrast-enhanced MRI histogram analysis to axillary lymph node metastasis of small breast cancer

作　　者：马芹芹 卢星如[2,3,4] 李芷凡 刘春翠 冯雯 雷军强 MA Qinqin;LU Xingru;LI Zhifan;LIU Chuncui;FENG Wen;LEI Junqiang(First Clinical Medical College,Lanzhou University,Lanzhou 730000,China;Department of Radiology,First Hospital of Lanzhou University,Lanzhou 730000,China;Intelligent Imaging Medical Engineering Research Center of Gansu Province,Lanzhou 730000,China;Gansu Province Clinical Research Center for Radiology Imaging,Lanzhou 730000,China)

机构地区：[1]兰州大学第一临床医学院,甘肃兰州730000 [2]兰州大学第一医院放射科,甘肃兰州730000 [3]甘肃省智能影像医学工程研究中心,甘肃兰州730000 [4]甘肃省放射影像医学临床医学研究中心,甘肃兰州730000

出　　处：《中国中西医结合影像学杂志》2024年第4期424-429,共6页Chinese Imaging Journal of Integrated Traditional and Western Medicine

基　　金：甘肃省重点人才项目资助(甘组通字[2023]20号);甘肃省自然科学基金(21JR1RA079);甘肃省教育科技创新基金项目资助(2022B-025)。

摘　　要：目的:基于小乳腺癌全病变MRI动态增强扫描(DCE-MRI)直方图分析及临床特征预测腋窝淋巴结(ALN)转移。方法:回顾性分析93例小乳腺癌患者临床病理资料,其中30例发生ALN转移(转移组),63例ALN无转移(无转移组)。患者术前或治疗前均行MRI检查。采用FireVoxel软件勾画ROI并行直方图分析,采用t检验及Mann-WhitneyU检验分析比较2组的直方图参数。采用Pearson及Spearman相关检验分析ALN转移与直方图参数之间的相关性,采用二元logistic回归分析对筛选出ALN转移组和无转移组间差异有统计学意义的参数构建联合模型,并利用ROC曲线分析直方图参数对ALN状态的诊断效能。结果:年龄和绝经状态均与ALN转移存在相关性(均P<0.05)。直方图参数中最小值、第5百分位数、变异系数、体积均与ALN转移显著相关(均P<0.05)。在直方图参数中,体积预测ALN状态的AUC最大,为0.729,最佳阈值为1.765。最小值、第5百分位数、变异系数、体积和年龄构建的联合模型的AUC最高,为0.780。结论:基于全病变DCE-MRI直方图分析可用于预测小乳腺癌的ALN转移,且最小值、第5百分位数、变异系数和体积是预测ALN状态有价值的影像学指标。Objective:To predict axillary lymph node(ALN)metastasis of small breast cancer with whole-volume dynamic contrast-enhanced MRI(DCE-MRI)histogram analysis.Methods:The clinicopathological data of 93 patients with small breast cancer were retrospectively analyzed,of which 30 cases had ALN metastasis(the metastasis group)and 63 cases had no metastasis(the non-metastasis group).All patients underwent MRI examination before operation or further treatment.Fire Voxel software was used to outline ROI and histogram analysis was performed,and the histogram parameters of the two groups were compared by t test or Mann-Whitney U test.Pearson and Spearman correlation analysis were used to analyze the correlation between ALN metastasis and histogram parameters.Logistic regression was used to obtain the statistically different parameters between the two groups to develop the combined model.The ROC curve was used to analyze the diagnostic efficiency of histogram parameters for ALN metastasis.Results:Age and menopausal state were associated with ALN metastasis(both P<0.05).The histogram parameters,the minimum value,5th percentile,coefficient of variation and volume had significant correlations with ALN metastasis(all P<0.05).Among the histogram parameters,volume had the highest AUC of 0.729 in predicting ALN metastasis,and the best cut-off value was 1.765.The combination model of the minimum value,5th percentile,coefficient of variation,volume and age had the highest AUC of 0.780.Conclusions:Histogram analysis based on whole lesion DCE-MRI can be used to predict ALN metastasis of small breast cancer,and the combination of the minimum value,5th percentile,coefficient of variation and volume may be useful index for distinguishing ALN status.

关 键 词：小乳腺癌 磁共振成像 动态增强扫描 直方图分析 腋窝淋巴结转移 

分 类 号：R445.2[医药卫生—影像医学与核医学] R737.9[医药卫生—诊断学]