Preoperative peritumoral edema MRI features combined with serum CA15-3 and CEA levels for constructing a nomo-gram to predict axillary lymph node metastasis burden in invasive breast cancer

ZENG Yurong; LING Meiping; YIN Dongxu; DAI Haiyang

doi:10.20258/j.cnki.1006-9011.2025.09.016

Abstract：Objective To explore the predictive value of a nomogram model constructed based on preoperative peritumoral edema(PE)magnetic resonance imaging(MRI)features combined with serum carbohydrate antigen 15-3(CA15-3)and carcino-embryonic antigen(CEA)levels for axillary lymph node metastasis(ALNM)burden of invasive breast cancer.Methods A total of 397 patients with invasive breast cancer were selected and divided into high-load group with 85 cases of high-load ALNM and low-load group with 312 cases of low-load ALNM according to the number of postoperative ALNM.On the second day after admission,a breast MRI examination was performed to obtain the MRI features of invasive breast cancer and to evaluate the pre-operative PE characteristics(PE type and PE degree).Peripheral venous blood was collected,serum was retained by centrifuga-tion,and serum CA15-3 and CEA were detected by chemiluminescence method.A multivariate unconditional Logistic regression model was used to analyze the risk factors of high ALNM burden in invasive breast cancer,and a nomogram model was con-structed accordingly.The goodness-of-fit of the model was tested by H-L.The ROC curve was used to evaluate the predictive abil-ity of the nomogram.The C-index was employed to evaluate the discriminatory ability of the nomogram.The calibration curve was used to evaluate the consistency of the nomogram.The decision curve was utilized to evaluate the clinical applicability of the no-mogram.Results Univariate analysis revealed that there were statistically significant differences between the two groups in terms of WHO grade,the proportion of HER-2 positive,T stage,N stage,lesion diameter,the proportion of severe PE degree,the proportion of severe PE type,and the levels of serum CA15-3 and CEA(all P<0.05).Multivariate unconditional Logistic re-gression analysis showed that large lesion diameter,severe PE degree,severe PE type,and high levels of serum CA15-3 and CEA were independent risk factors for high ALNM burden in invasive breast cancer(all P<0.05).Based on this,a nomogram model for predicting high ALNM burden in invasive breast cancer based on preoperative PE MRI features combined with serum CA15-3 and CEA levels was constructed.By H-L test,there was no statistically significant difference between the predicted prob-ability and the actual probability of this nomogram model(χ2=6.870,P>0.05).ROC curve analysis showed that the area under the curve of this nomogram model for predicting high ALNM load in invasive breast cancer was 0.901(95%CI:0.867-0.928),and the sensitivity and specificity were 85.9%and 90.1%,respectively.The C-index of this nomogram model was 0.901(95%CI:0.896-0.905).The calibration curve showed that the predicted probability of this nomogram model was in line with the ideal curve.The decision curve showed that when the threshold probability was greater than 0.10,the clinical net benefit of this nomo-gram model was greater than 0.Conclusion The nomogram model constructed based on preoperative PE MRI features com-bined with serum CA15-3 and CEA levels can predict the high burden of ALNM in invasive breast cancer more accurately.This has high predictive ability,discrimination ability and consistency,which has potential application value in clinical decision-making.