Author/Editor     Cserni, Gábor; Žgajnar, Janez; Perhavec, Andraž; Gazić, Barbara
Title     Multi-institutional comparison of non-sentinel lymph node predictive tools in breast cancer patients with high predicted risk of further axillary metastasis
Type     članek
Vol. and No.     Letnik 19, št. 1
Publication year     2013
Volume     str. 95-101
ISSN     1219-4956 - Pathology oncology research : POR
Language     eng
Abstract     Although axillary lymph node dissection (ALND) has been the standard intervention in breast cancer patients with sentinel lymph node (SLN) metastasis, only a small proportion of patients benefit from this operation, because most do not harbor additional metastases in the axilla. Several predictive tools have been constructed to identify patients with low risk of non-SLN metastasis who could be candidates for the omission of ALND. In the present work, predictive nomograms were used to predict a high (>50 %) risk of non-SLN metastasis in order to identify patients who would most probably benefit from further axillary treatment. Data of 1000 breast cancer patients with SLN metastasis and completion ALND from 5 institutions were tested in 4 nomograms. A subset of 313 patients with micrometastatic SLNs were also tested in 3 different nomograms devised for the micrometastatic population (the high risk cut-off being 20 %). Patients with a high predicted risk of non-SLN metastasis had higher rates of metastasis in the non-SLNs than patients with low predicted risk. The positive predictive values of the nomograms ranged from 44 % to 64 % with relevant inter-institutional variability. The nomograms for micrometastatic SLNs performed much better in identifying patients with low risk of non-SLN involvement than in high-risk-patients; for the latter, the positive predictive values ranged from 13 % to 20 %. The nomograms show inter-institutional differences in their predictive values and behave differently in different settings. They are worse in identifying high risk patients than low-risk ones, creating a need for new predictive models to identify high-risk patients.
Keywords     rak dojke
tveganje
bezgavke
zasevki
breast cancer
high risk
lymph node
sentinel lymph node