Exploring Breast Cancer Data to Find the Risk Factors for Distant Metastasis: Register Data, The National Cancer Institute of Thailand

ABSTRACT

A retrospective cohort of registered breast cancer patients at the National Cancer Institute of Thailand in 2004 was investigated to explore the association of prognostic factors with the probability of distant metastasis. The analysis was done using binary and ordinal logistic regression models. There were 893 breast cancer patients included in a 3 year follow up study. During the study period 129 (14.45%) were reported as metastasis cases, which the highest number of metastasis patients was found in bone (47.3%). The prognostic variables related to the distant metastasis breast cancer using binary logistic regression model were found among vascular invasion (OR=8.8), histologic grade III (OR=2.9), lymphatic invasion (OR=2.4), estrogen receptor (OR=2.2), progesterone receptor (OR=2.0), p53 (OR=1.8), tumor size (OR=1.2), radiation therapy (OR=0.5) and hormonal therapy (OR=0.4). Patients in which metastasis occurred within the 1st, 2nd and 3rd years after initial diagnosis were 75, 28 and 26 cases respectively. Using an ordinal logistic regression model, the prognostic factors included vascular invasion (OR=5.6), estrogen receptor (OR=2.3), lymphatic invasion (OR=2.3), histologic grade III (OR=1.9), p53 (OR=2.0), tumor size (OR=1.2), radiation therapy (OR=0.6) and hormonal therapy (OR=0.5). The prognostic factor of vascular invasion showed the highest OR in this study. Another group of variables with moderate OR were estrogen receptor, lymphatic invasion, histologic grade III, p53 and tumor size. The adjuvant therapies were protective factors (OR \< 1). A follow up system for those at risk for metastasis is suggested.