Prediction of breast cancer and lymph node metastatic status with tumour markers using logistic regression models
Hsiao-Lin Hwa1, Wen-Hong Kuo, Li-Yun Chang
1Department of Obstetrics and Gynecology, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan. hwahl013@ms10.hinet.net
Aims:
Early detection of breast cancer can improve disease mortality. The aim of this study was to evaluate the effectiveness of serum biomarkers in the detection of primary breast cancer and lymph node metastatic status.
Methods:
Serum samples were obtained from 55 female patients with breast cancer and 39 women without breast cancer. For these subjects, clinicopathological data were collected and serum levels of carcinoembryonic antigen, breast cancer-specific cancer antigen 15.3 (CA15-3), tissue polypeptide-specific antigen (TPS), soluble interleukin-2 receptor (sIL-2R) and insulin-like growth factor binding protein-3 (IGFBP-3) were assayed. Univariate and multivariate logistic regression were performed to evaluate the association between biomarkers and breast cancer, as well as lymph node metastatic status.
Results:
For breast cancer prediction, the serum level of TPS had the best predictive value, with a sensitivity of 80% at an optimal cut-off value of 69.1 U L(-1). The combination of TPS, CA15-3 and IGFBP-3 with logistic regression model increased the sensitivity to 85%. For lymph node metastasis prediction, the serum level of sIL-2R had the best predictive value, with a sensitivity of 66% at an optimal cut-off value of 286 U mL(-1). The combination of sIL-2R and TPS with logistic regression model increased the sensitivity to 69%.
Conclusion:
TPS may be useful in the detection of primary breast cancer, while sIL-2R may be useful in lymph node metastasis prediction. The combination of more than one biomarker with logistic regression model can improve the predictive sensitivity.