Risk Factors And Predictive Models Of Breast Cancer In Women

OBJECTIVE:The current reports of risk factors associated with breast cancer are very common.Western countries have established a variety of risk prediction models for screening high-risk groups.However,due to differences in geography,ethnicity,customs,etc.,the relevant models are not applicable in China.This study analyzes and evaluates the risk factors associated with breast cancer in women in this region,and explores and establishes a logistic regression model of major risk factors.It is used to guide the clinical practice by evaluating the sensitivity and specificity of the established model in predicting the incidence of breast cancer.METHODS:This study retrospectively analyzed the clinical data of 394 patients with primary breast cancer diagnosed by pathology in Qingdao Municipal hospital from December 2015 to December 2017,including（1）demographic characteristics:such as current age,body mass index（Body Mass Index,BMI）;（2）Physiological and reproductive factors:age of menarche,menopause,age of first birth,number of births,breastfeeding time,history of abortion and number of abortions,whether to take contraceptives;（3）Breast disease and family history:mainly history of benign breast benign biopsy,history of atypical hyperplasia and history of first-and second-degree relatives breast cancer;（4）Mental depression score:depression caused by various life events in recent 3-6 months 0 points means no suppression,9 points means most suppression.A total of 394 women with no blood relationship and healthy physical examination were selected as the control group.Collect relevant data by filling out questionnaires face to face.Using SPSS22.0 statistical software,after analyzing the patient’s data,factors including single factor logistic regression analysis were screened and analyzed,and factors with statistical significance（P<0.05）were included in multivariate logistic regression.Analysis,if calculated P<0.05,that will be remained in the final model.The Receiver Operating Characteristic curve（ROC）was used to evaluate the performance of the established predictive model.RESULTS:The univariate unconditional logistic regression analysis showed that there were statistically significant differences,age of first live birth,age of menarche,age of first birth,history of abortion,history of menopause,history of oral contraceptives,history of first-degree relative breast cancer,and self-score of mental stress（P<0.05）.Breast feeding time,body mass index（BMI）,biopsy history of benign breast diseases,atypical hyperplasia,and breast cancer history in second-degree relatives were not statistically significant（P>0.05）.Will be included in the single factor analysis of statistically significant factors multi-factor unconditioned logistic regression analysis:for the first time live births（OR=1.14）,and childbearing age menarche age（OR=0.89）,the production number 1（OR=0.18）,the production number 2（OR=0.17）,the number of abortions（OR=2.39）,the menopause（OR=4.6）and oral contraceptives（OR=3.35）,the level of family history of breast cancer（OR=2.41）and mental score（4 to 9）was statistically significant（OR=2.45）.The established regression model was Logit P=-0.91+0.13age of first live birth-0.11 age of first menstruation-1.72 birth times₁-1.77birth times ₂+0.87abortion+1.53menopause+1.21oral contraceptive+0.88breast cancer history of first-degree relatives+0.89mental distress ₂.According to the ROC curve analysis of the prediction model established in this study,the area under the curve（AUC）is 0.752,and its 95%confidence interval is 0.712-0.792,with a sensitivity of0.678,a specificity of 0.706,a consistency rate of 75.5%,and a prediction inconsistency rate of 24.3%.Conclusion:Older age of first live birth,abortion,late menopause,oral contraceptives,first-degree relative breast cancer history and high score of mental depression are risk factors for breast cancer.The age of menarche and birth were the protective factors of breast cancer.