OBJECTIVESTo study the factors influencing the non-sentinel lymph node(NSLN) status and to assess performance of Memorial Sloan-Kettering Cancer Center (MSKCC) nomogram in predicting sentinel lymph node(SLN) metastases in a SLN positive Chinese breast cancer population.

METHODSData were collected from breast cancer patients who were diagnosed with pathological positive sentinel lymph node and received further axillary lymph node dissection(ALND) in Shanghai Ruijin Hospital from January 2011 to August 2014. MSKCC nomogram was used to calculate each patient's NSLN metastasis risk score. The receiver operator characteristic curve (ROC curve) and the area under the ROC curve (AUC) was used to assess the predictive accuracy of the model.

RESULTSAmong the 1 147 patients who received sentinel biopsy, 150 SLN positive patients who received ALND were enrolled in this study. By univariate analysis, multifocal breast cancer (χ(2)=5.887, P=0.015), SLN+ /SLN ratio (χ(2)=6.683, P=0.010) and abnormal axillary lymph node displayed by ultrasound (χ(2)=7.736, P=0.005) were the influencing factors of NSLN metastases. By multivariate analysis, multifocal breast cancer (OR=7.25, 95% CI: 1.73 to 30.43, P=0.007), SLN+ /SLN ratio ≥ 0.5 (OR=2.564, 95% CI: 1.22 to 5.39, P=0.013) and abnormal axillary lymph node displayed by ultrasound (OR=2.471, 95% CI: 1.18 to 5.19, P=0.017) were the independent influencing factors of NSLN metastases. The AUC of MSKCC nomogram in this population was 0.677.

CONCLUSIONSFor breast cancer patients with positive sentinel lymph node, multifocality, SLN+ /SLN ratio and axillary lymphadenopathy displayed by ultrasound is related to NSLN metastasis. MSKCC has low accuracy in predicting NSLN status of this population.