Objective:To explore the value of quantitative parameters from synthetic MRI in differentiating triple-negative breast cancer(TNBC)from other subtypes.Methods:A total of 186 patients with breast lesions confirmed by pathology were prospectively collected.All patients underwent conventional MRI and synthetic MRI before surgery.ICC and Bland-Altman plots were used to analyze the consistency of synthetic MRI quantitative parameters between different observers.The value of quantitative parameters from synthetic MRI in distinguishing TNBC from other types was analyzed by univariate and multivari-ate logistic regression,and the receiver operating characteristic curve(ROC)was plotted.The area under the curve(AUC),sensi-tivity,and specificity were calculated.Results:ICC3D of T1,T2 and PD based on 3D-ROI(0.994,0.996 and 0.989)were bet-ter than ICC2D based on 2D-ROI(0.848,0.850 and 0.898).The AUC value of the T1-mean and T2-mean combined model for distinguishing TNBC from Luminal A type was 0.79(95％CI:0.65～0.88),with a sensitivity of 56.67％and specificity of 88.00％.The AUC value of the T2-mean ROC curve distinguishing TNBC from Luminal B breast cancer was 0.74(95％CI:0.64～0.82),with a sensitivity of 63.66％and specificity of 80.88％.The AUC value of the PD-mean ROC curve distinguishing TNBC from Luminal B breast cancer was 0.70(95％CI:0.55～0.82),with a sensitivity of 80.00％and specificity of 55.00％.Conclusions:The repeatability of synthetic MRI quantitative parameters based on 3D-ROI was higher than that of 2D-ROI between different observers.Quantitative parameters from synthetic MRI may provide a basis for distinguishing TNBC from oth-er types.
维普
万方数据