Risk Factors Of Non Sentinel Lymph Node Metastasis And Comparative Validation Of Three Nomograms In Breast Cancer Patients

Background Breast cancer is one of the most common cancer in women.Sentinel lymph node biopsy(SLNB)is important for evaluating the state of axillary lymph nodes in early breast cancer patients.However,for some patients who have performed ALND because of positice SLNB results,there are no significant lymph node metastasis after operation,and these patients not only benefit little from ALND,but face the increased risk of complications.Therefore,in order to exclude non-sentinel lymph node metastasis patients from patients with low risk of positive sentinel lymph nodes so that these patients could avoid ALND,global researchers are constantly exploring models to predict the risk of NSLN transfer,and there have several models established,including MSKCC,Tenon,MD Anderson(MDA),Mayo,Cambridge,Stanford,Helsinki and so on,However,these models are based on the research of Western people,whether they are suitable for Chinese population need to be further studied..Objective To explore the ri node(NSLN)metastasis whose sentinel lymph node biopsy(SLNB)are proved to be positive,and to analyze and validate the efficacy and clinical value in predicting NSLN metastasis of three predictive models of MSKCC,Cambridge and Tenon.Methods The clinical data of 89 breast cancer cases confirmed positive by SLN biopsy and underwent axillary lymph node dissection(ALND)in our hospital from January 2014 to June 2018 were enrolled and analyzed retrospectively,including 27 cases with NSLN metastasis.The related risk factors of NSLN metastasis of breast was calculated by MSKCC model,Cambridge model and Tenon model respectively,and predictive efficacy was evaluated by area under ROC curve(AUC);The predicted values and real values of MSKCC,Cambri dge and Tenon model were compared byResults Univariate analysis showed that the size of primary tumors,histological grade,lymphovascular invasion,the size of SLN metastases,the number of positive SLNs,and the ratio of positive SLNs/total and SLNs were correlated with NSLN metastasis in breast cancer patients(P<0.05),while there was no statistical difference between the positive and negative groups of NSLN metastasis in age(P=0.197),location of tumors(P=0.157),pathological type(P=0.247),number of negative SLNs(P=0.104),multifocal tumors(P=0.266),ER(P=0.226),PR(P=0.162),HER2(P=0.129)and tumor,the number of positive SLNs,lymphovascular invasion,the size of SLN metastases and the ratio of positive SLNs/total SLN,were the independent risk factors of NSLN metstasis.The AUC values of NSLN metastasis predicted by MSKCC,Cambridge and Tenon model were 0.712,0.600 and 0.654,respectively.The calibration curves show that the predicted value curves of MSKCC and Tenon nomogram were similar to the true value curves,which is better than Cambridge nomogram.The value was correlated with the actual value(χ2 = 7.728,P=0.006),while the Cambridge and Tenon predicted value was not correlated with the actual value(χ2=0.732,P=0.269,or χ~2=1.059,P=0.326).Conclusion The prediction ability of NSLN metastasis risk of MSKCC was slightly superior to the Cambridge and Tenon models,and it can provide a clinical accessory in the preoperative discussion of the breast cancer patients with positive SLN,which means it has a certain clinical value.