Non-sentinel Lymph Node Metastases In Breast Cancer Patients From Minority Area Of Guangxi In China:Validation Of A Nomogram And Development Of A New Predictive Model

Purpose The aim of our study was to evaluate the predictive values of theMemorial Sloan-Kettering Cancer Center (MSKCC) nomogram for predictingnon-sentinel lymph node (NSLN) metastasis in breast cancer patients fromminority area of GuangXi in China and develop a new predictive model withpossible advantages in loacal people.Methods Eighty-three patients with positive sentinel lymph nodebiopsy(SLNB) followed by ALND were enrolled into our retrospectivestudy.The predicted probability of non-SLN metastasis was calculated for eachpatient by using a computerized model from the MSKCC Web site. The receiveroperating characteristic (ROC) curves were drawn and the areas under the curve(AUCs) were calculated to assess the discriminative power of the nomogram.Univariate and multivariate analysis was used to identify variables predictingnonsentinel node involvement and a multivariable predictive model wasdeveloped. Inorder to assess the accuracy,new model was applied to the original series of83patients and the area under the receiver operating characteristiccurve (AUC) was calculated.Distribution of continuous variables was analyzedusing the Mann–Whitney U test,χ2test or Fisher’s exact test was used forcategorical variables.Results The AUC was calculated for the entire study population,Han ethnicsubgroup,Zhuang ethnic subgroup,macrometastasis subgroup, micrometastasissubgroup in MSKCC nomogram,and the AUC value was0.749,0.768,0.757,0.796,0.585respectively.Size of the primary tumor,histological grade,lymphovascular invasion and size of SLN metastasis were revealed to beindependent predictors of NSLN involvement in multivariate logistic regressionand included in the final predictive model(P＜0.1).A new predictive model wasdeveloped and its AUC value for the entire study population was0.832.Conclusions The MSKCC nomogram provides a fairly accurate predictedprobability for the likelihood of NSLN metastasis in patients with SLNmacrometastasis,but did not provide a reliable predictive model for identifyingpatients with micrometastasis in SLN. A new predictive tool wasdeveloped,more predictive as a model in breast cancer patients from minorityarea of GuangXi in China,but still need further validation.