安徽省女性乳腺癌危险因素分析与风险预测模型研究

目的尝试建立乳腺癌风险预测模型, 为乳腺癌预防工作提供依据。方法以癌症登记或医保报销系统为基础, 调查安徽省2015年以来确诊的35岁及以上癌症病人, 对照为同时期癌症病例所在社区的未确诊过任何一种癌症的居民。采用logistic回归模型, 筛选乳腺癌的危险因素, 得到相应的OR值并赋分, 分别计算乳腺癌风险因素指数和保护因素指数; 利用"改良哈佛风险指数"和"年龄-因子加权指数", 建立风险预测模型。结果筛选出乳腺癌的危险因素包括乳头溢液、乳房肿块累计年、已绝经、饮食重盐、经常生闷气及乳腺癌家族史(P < 0.01);高血压/高血糖/高血脂、患疟疾次数、失眠多梦、首次月经>13岁、经常吃新鲜水果及体育锻炼为保护因素(P < 0.05~P < 0.01)。最终构建的乳腺癌风险评估预测模型为: R=ln(乳腺癌发病指数)×2+乳腺癌风险因素指数-1.3×乳腺癌保护因素指数, AUC值为0.779。结论成功构建安徽省女性乳腺癌风险评估模型, 且预测效果较好, 有助于评估、发现肿瘤风险较高的居民并进而集中资源开展重点预防干预。

Risk factors analysis and risk prediction model of female breast cancer in Anhui Province

ObjectiveTo establish a risk prediction model of breast cancer and provide evidence for the prevention of female breast cancer.MethodsBased on the cancer registration or medical insurance reimbursement system, the cancer patients aged 35 and above who were diagnosed with cancer in Anhui province since 2015 were investigated, and the residents who had not been diagnosed with any cancer in the community where the cancer cases were located during the same period were selected as the comtrol.Logistic regression model was used to screen the risk factors of breast cancer, and the corresponding OR value was obtained and assigned scores.The risk factor index and protective factor index of breast cancer were calculated, respectively.The risk prediction model of breast cancer was established by using "modified Harvard risk index" and "age-factor weighted index".ResultsThe risk factors of breast cancer were as follows: nipple discharge, accumulated years of breast mass, menopause, heavy diet salt, sulking and family history of breast cancer(P < 0.01).Hypertension / hyperglycemia / hyperlipidemia, incidence of malaria, frequent insomnia and dreams, age of first menstruation >13 years old, frequent eating of fresh fruit and regular physical exercise were protective factors(P < 0.05 to P < 0.01).The final risk prediction model for breast cancer was R=ln(breast cancer incidence index) ×2+breast cancer risk factor index-1.3×breast cancer protective factor index, and the AUC value was 0.779.ConclusionsThe risk prediction model of female breast cancer in Anhui province is successfully constructed.The prediction effect is good, which is helpful to evaluate and find residents with high tumor risk, and then concentrate resources to carry out key preventive intervention.