Abstract
BACKGROUNDS: Tumor-positive sentinel node(SLN) biopsy results in a risk of nonsentinel node metastases in case of micro and macro metastases ranging from 20 to 50 %, respectively. Therefore, most patients underwent unnecessary axillary lymph node dissections. Thus, the development of a mathematical model for predicting patient-specific risk of non sentinel node(NSLN) metastases is strongly warranted. METHODS: The following parameters were recorded: CLINICAL: hospital, age, medical record number Bio-pathological: tumor (T) size, grading (G), multifocality, histological type, LVI, ER-PR status, HER-2, ki67, molecular classification (luminal A, luminal B, HER2 like, triple negative) Sentinel and nonsentinel lymph node related: number of removed SLNs, number of positive and negative SLNs, copy number of positive sentinel nodes, ratio: number of positive SLNs to number of removed SLNs, number of removed and number of positive nodes after ALND. A total of 2460 patients have been included in the database. All the patients have been provided by the authors of this paper. RESULTS: Multivariate logistic regression analysis demonstrated that only the number of a CK19 mRNA copies (p < 0.0001), T size (p < 0.0001) and LVI (p < 0.0001) were associated with NSN metastases. The discrimination of the model, quantified with the area under the receiver operating characteristics curve, was 0.71 (95 %, C.I. 0.69-0.73), thus confirming a good level of reliability. CONCLUSIONS: The nomogram may be employed by the surgeon as a decision making tool on whether to perform an intraoperative axillary lymph node dissection on breast cancer patients with SLN positive. The large population employed and the standardized method of measuring the value of CK19 mRNA copies are appropiate prerequisites for a reliable nomogram.