Predictive value of axillary lymph node metastasis in breast cancer based on DCE-MRI radiomics

SONG Yabo; LI Jiaojiao; HUANG Ziying; JIAO Guangli; SHI Zixin; LI Xiaodong

doi:10.20258/j.cnki.1006-9011.2025.03.014

Abstract：Objective To investigate the value of radiomics features within breast cancer tumors and containing a 5-mm area around the tumor to predict axillary lymph node(ALN)metastasis in patients.Methods A total of 198 breast cancer patients who underwent preoperative magnetic resonance examination in the First Affiliated Hospital of Hebei North College were retro-spectively collected and randomly divided into a training set(n=138)and a validation set(n=60)according to the ratio of 7:3.The three-dimensional volume of interest(VOI)was sketched on the second-phase images of dynamic contrast-enhanced mag-netic resonance imaging(DCE-MRI),and the peri-tumor area was expanded by 5 mm to form VOI+5 mm.The VOI was ex-panded by 5 mm to form VOI+,and the features were screened by the least absolute shrinkage and selection operator(LASSO)algorithm,after which the intratumor model and the combined intratumor and peritumor model were established by the logistic re-gression(LR)method,respectively.The predictive efficacy of the models was evaluated by plotting the receiver operating char-acteristic(ROC)and calculating the area under the curve(AUC).Decision Curve Analysis(DCA)was used to assess the clinical application value of the model.Results In both the training and validation sets,the AUC values of the intra-tumor combined peri-tumor model(0.861,0.802)were higher than those of the intra-tumor model(0.745,0.741).DCA showed that the intra-tumor model and intra-tumor combined peri-tumor model yielded net when the threshold probabilities were in the ranges of 31.0%to 72.0%and 25.0%to 78.0%,respectively.Conclusion The MRI radiomics model based on the LR classifier and established by DCE-MRI image radiomics features has a certain predictive value for assessing axillary lymph node metastasis in breast cancer,and the intra-tumor combined peri-tumor radiomics features have a greater potential to enhance the predictive efficacy.