北京、广东两地妇女乳腺癌危险度评价模型的初步研究

目的:筛选北京、广东两地妇女罹患乳腺癌的危险因素;制订适合中国国情的乳腺癌危险度评价模型,探讨其评价标准,与Gail模型进行初步比较。方法:通过Logistic回归筛选乳腺癌的危险因素,在此基础上得到乳腺癌危险度评价模型,通过判别分析评价模型,通过计算ROC曲线下面积比较该模型和Gail模型的诊断效能,并利用ROC曲线寻找乳腺癌低、中、高危险性的合适分界点。结果:妇女罹患乳腺癌的可能危险因素中影响较突出的是:有绝经史、有流产史、有药物避孕史、有乳腺癌家族史、不良情绪、不良事件、情绪调节能力差、劳动强度大、年龄。在危险因素筛选的基础上建立危险度评价模型,建议以预测概率值P≤0.29判为低危险性人群,预测概率值P≥0.58判为高危险性人群,0.29<预测概率值P<0.58判为中危险性人群。结论:乳腺癌重在预防,应结合其危险因素开展积极有效的防治措施。该研究得到的乳腺癌危险度评价模型可用于乳腺癌高风险性人群的筛选。

Preliminary study on risk evaluation model of breast cancer in Beijing and Guangdong

Objective:To filter the risk factors of breast cancer in Beijing and Guangdong and establish the risk evaluation model,then compare it with the Gail model.Methods:The risk factors of breast cancer were screened by logistic regression analysis,then the new risk evaluation model was established on the basis of these important factors.Diagnostic efficiencies of new model and Gail model were compared by the calculation of area below ROC curve,then to find out the appropriate cut-off points of high,moderate and low risk.Results:The potential risk factors of breast caner included menopause,abortion,medicine-contraception,breast cancer-family history,bad emotions,undesirable life,the absence of emotion-adjusting ability,excess working and age.The cutpoints were set up according to the screening indexes and the demand of clinical practice:the prediction probability≤0.29 indicated the low risk,the prediction probability≥0.58 indicated high risk and the prediction probability between 0.29 and 0.58 indicated moderate risk.Conclusion:The Gail and the new model established are just preliminary self-assessment tools,the key point of breast cancer is prevention.