Luminal A breast cancer can present with early, unexpected lymph node metastasis, and sentinel lymph node biopsy has been reported false negative in some cases. We aimed to construct a biomarker-based model that predicts lymph node metastasis in luminal A breast cancer, using expression of silent mating type information regulation 2 homolog 1 (SIRT1) and apoptosis-related factors, which are known to be closely related. We selected tissue samples of 278 cases of luminal A invasive ductal carcinoma, constructed tissue microarrays, and performed immunohistochemical staining for SIRT1 and four apoptosis-related proteins. In constructing the best predictive model for lymph node metastasis, six clinicopathological parameters and five molecular markers were considered. Independent factors predictive of lymph node metastasis were pT stage (OR 1.829, p = 0.027), lymphovascular invasion (OR 4.128, p < 0.001), and decreased expression of caspase-3 (OR 0.535, p = 0.034) and of SIRT1 (OR 0.526, p = 0.053). A combination nuclear grade, lymphovascular invasion, increased B-cell lymphoma 2 (Bcl-2) expression, and reduced expression of caspase-3 and of SIRT1 yielded the strongest predictive performance for lymph node metastasis with an area under the curve (AUC) of 0.696. This combination was also predictive of shortened disease-free survival (73.1 vs. 67.7 months, p = 0.003). Our data support a role of SIRT1 protein as tumor suppressor in luminal A breast cancer, in association with apoptosis-related proteins. Our model based upon a combination of these biomarkers is expected to increase accuracy of prediction of lymph node metastasis in luminal A breast cancer. This might serve as a valuable tool in determining the optimal surgical strategy in breast cancer patients.
- Breast cancer
- Predictive model