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.
Small breast cancerMagnetic resonance imagingDynamic contrast enhanceHistogram analysisAxillary lymph node metastasis
国家科技期刊平台
NSTL
万方数据
