Abstract
This study was aimed at developing a predictive model for assessing the breast cancer risk of Korean women under the assumption of differences in the risk factors between Westerners and Koreans. The cohort comprised 384 breast cancer patients and 2 control groups: one comprising 166 hospitalized patients and the other comprising 104 nurses and teachers. Two initial models were produced by comparing cases and the 2 control groups, and the final equations were established by selecting highly significant variables of the initial models to test the accuracy of the models in terms of disease probability and predictability. Both the initial models and the final disease-probability models were confirmed to exhibit high degrees of accuracy and predictability. Major risk factors determined by comparing the patients with hospitalized controls were a family history, menstrual regularity, total menstrual duration, age at first full-term pregnancy, and duration of breastfeeding. Major risk factors determined by comparing patients with nurse/teacher controls were age, education level, menstrual regularity, drinking status, and smoking status. The predictive model developed here shows that risk factors for breast cancer differ between Korean and Western subjects in the aspect of breastfeeding behavior. However, identifying the relationship between genetic susceptibility and breast cancer will require further studies with larger samples. In a model with nurse/teacher controls, drinking and higher education were found to be protective variables, whereas smoking was a risk factor. Hence the predictive model in this group could not be generalized to the Korean population; instead, breast cancer incidence needs to be compared among nurses and teachers in a nurse-and-teacher cohort.