Analysis of variables predicting non-sentinel lymph node involvement in breast cancer patients from Guangxi and development of a new predictive model

doi:10.3969/j.issn.1674-3806.2014.06.02

Abstract：Objective To analyse the clinical pathologic factors predicting non-sentinel node(NSLN) in-volvement and develop a new model to quantify the risk of NSLN metastasis in sentinel lymph nodes ( SLN)-positive breast cancer patients from Guangxi , in order to avoid unnecessary axillary lymph node disection ( ALND) in low-risk patients with positive SLN .Methods Eighty-three patients with positive sentinel lymph node biopsy ( SLNB) followed by ALND were enrolled into our retrospective study .Univariate and multivariate analysis was used to identify variables predicting non-sentinel node involvement and a multivariable predictive model was developed .In order to assess the accuracy ,new model was applied to the original series of 83 patients and the area under the receiver operating charac-teristic curve(AUC) was calculated.Distribution of continuous variables was analyzed using the Mann-Whitney U test, and the χ2 test was used for categorical variables .Results Size of the primary tumor, histological grade, lym-phovascular invasion and size of SLN metastasis were revealed to be independent predictors of NSLN involvement in multivariate logistic regression and included in the final predictive model .For the original series of 83 patients, the AUC was 0.832(95%CI=0.744 to 0.919).Conclusion In our study,new predictive tool was developed to assess the risk of additional axillary metastases after positive SLNB in breast cancer and represent considerable discrimination in our own population .but its accuracy needs to be assessed by prospective validation in both internal center and ex -ternal centers.