| Comparison of validation in predicting sentinel and non-sentinel lymph node metastasis by prediction models for axillary lymph node metastasis in breast cancer patients |
| Sang Hyup Han, Hae Sung Kim, Jun Ho Kim, Youn Ok Lee, Young Ah Lim, Hee Joon Kang, Lee Su Kim |
| Division of Breast and Endocrine Surgery, Hallym University, Collage of Medicine |
| Corresponding Author:
Lee Su Kim ,Tel: +82-31-380-5930, Fax: +82-31-384-0208, Email: lskim0503@hallym.ac.kr |
|
Received: September 24, 2012; Accepted: September 16, 2012. |
|
|
Share : 
|
| ABSTRACT |
Background: It is current practice to perform a completion of axillary lymph node dissection for breast cancer patients with sentinel lymph nodes (SLN) metastasis. But fewer than half will have non-sentinel lymph node (NSLN) metastasis. To predict NSLN metastasis in patients with a positive SLN biopsy, many different methods have been created. We sought to identify criteria for the intraoperative assessment of SLN and NSLN involvement in breast cancer.
Methods: We performed 279 SLN biopsies in women with breast cancer between 2005 and 2009. We determined factors associated with SLN metastasis and NSLN using univariate and multivariate logistic regression. This study reviews the MSKCC nomogram for predicting SLN and five NSLN predictive models (three nomograms and two scoring systems) reported in the literature. We compare their predictive probability of SLN metastasis and NSLN involvement with SLN-positive patients. In this study, we calculated cut-off value of sensitivity and specificity.
Results: Tumor size (p=0.004) and lymphovascular invasion (p<0.001) were associated with SNL metastasis in multivariate analysis. Also lymphovascular invasion (p=0.007) was related to NSLN metastasis. We found the MSKCC predictive SLN nomogram underestimated our patients’data. Application of five NSLN predictive models to our data set produced the area under the receiver operator characteristic curves similar to each model
Conclusion: Nomograms and scoring systems developed at other institutions should be carefully used when counseling patients about the risk of nodal disease. To improve accuracy of the predictive models, we should investigate identification of molecular parameters not yet in clinical use. |
| Keywords:
breast cancer | sentinel lymph node | predictor |
|
|
|
PDF Links
Full text via DOI
Download Citation 
