Preliminary Study Of Predicting Axillary Lymph Node Metastasis Of Invasive Breast Cancer With Synthetic MRI Quantitative Parameters

Purpose:To evaluate the potential value of synthetic magnetic resonance imaging（syMRI）quantitative parameters in predicting axillary lymph nodes metastasis（ALNM）of invasive breast cancer.Materials and methods:The clinical and imaging data of 56 patients with invasive breast cancer diagnosed by pathology were analyzed retrospectively.In addition to conventional MRI scans,all patients underwent syMRI sequence scans before and after enhancement,and the axillary region was completely covered.Region of Interest（ROI）was manually delineated on the synthetic T2-weighted images and T1、T2 and PD values before and after the enhancement were obtained using GE AW 4.7workstation.Meanwhile,T1map、T2map and proton density（PD）map were generated.At the maximum slice of tumor lesion,the ROI was delineated and recorded as"tumor".At the maximum slice of the axillary lymph node（ALN）,the ROI was delineated and recorded as"LN".Quantitative parameters before and after enhancement were represented by T and T⁺,PD and PD⁺.The quantitative parameters values of tumor lesion(T1_tumor、T2_tumor、PD_tumor and T1⁺_tumor、T2⁺_tumor、PD⁺_tumor)and these of axillary lymph node(T1_LN、T2_LN、PD_LN and T1⁺_LN、T2⁺_LN、PD⁺_LN)were obtained,and then calculated the changes of parametersΔT、ΔPD and the ratios T1/T2、T1⁺/T2⁺.The independent samples t test or Mann-Whitney U test were used to analyze the quantitative parameters of tumor lesions and lymph nodes between the ALNM group and non-ALNM group.Those that were found to be significant in the univariate analysis were analyzed by logistic regression analysis.The prediction performance of syMRI quantitative parameters on ALNM was evaluated by receiver operating characteristic（ROC）curve analysis and the areas under the curve（AUC）.Results:There were 30 cases in the ALNM group and 26 cases in the non-ALNM group among the 56 patients of invasive breast cancer.The short diameter of lymph nodes and the presence of lymphatic hilum showed significant differences between the two groups（P<0.05）.The AUC of short diameter of lymph nodes was 0.894,and the sensitivity and specificity were 86.7%and 90.2%respectively.In syMRI quantitative parameters analysis of tumor lesions,the differences of T2_tumor value、ΔT2_tumor value andΔPD _tumor value between the two groups were statistically significant（P<0.05）.T2_tumor value had the best efficacy in predicting ALNM（AUC=0.712）,and the optimal cutoff was 90.12ms,the sensitivity and specificity were 65.0%and 83.6%respectively.The AUC ofΔT2_tumor andΔPD_tumorwere 0.669 and 0.687,respectively.In terms of syMRI quantitative parameters of lymph nodes,univariate analysis showed that T1_LN、T2_LN、T1_LN/T2_LN、T2⁺_LN andΔT1_LN values were significantly different between the two groups（P<0.05）,and the AUC were 0.785、0.840、0.886、0.702 and 0.754,respectively.Multivariate logistic regression analysis indicated that only T1_LNvalue was an independent variable for predicting ALNM（AUC=0.785,95%CI:1.130-1.798,P=0.039）.The diagnostic sensitivity and specificity of T1_LNwere 86.7%and 69.4%at the best cutoff point of 1371.00ms.The prediction efficacy of T1_LN、T2_LN、T1_LN/T2_LNandΔT1_LN combined with short diameter of lymph nodes was further improved,with AUC of 0.905、0.957、0.964 and 0.897,respectively.Conclusions:1.T2_tumorvalue derived from syMRI of tumor lesions had the best efficacy in predicting ALNM of breast cancer.2.The syMRI quantitative parameter T1_LN value of axillary lymph nodes was an independent predictor of ALNM in breast cancer.Short diameter of lymph nodes combined with syMRI quantitative parameters could further improve the prediction efficiency.