基于决策树法构建病理性乳头溢液患者并发乳腺癌风险的预测模型  

Risk prediction mode of breast cancer in patients with pathological nipple discharge based on decision tree method

作　　者：邵光东 史明明[1] 宋以宁 徐春红[1] 马晓东[1] 郝晓亮 SHAO Guang-dong;SHI Ming-ming;SONG Yi-ning;XU Chun-hong;MA Xiao-dong;HAO Xiao-liang(Department of Thyroid and Breast Diagnosis and Treatment Center,Weifang Hospital of Traditional Chinese Medicine,Shandong Second Medical University,Weifang 261000,China)

机构地区：[1]山东第二医科大学附属中医院(潍坊市中医院)甲状腺乳腺外科,山东潍坊261000

出　　处：《中国现代普通外科进展》2025年第3期175-179,共5页Chinese Journal of Current Advances in General Surgery

基　　金：山东省潍坊市科学技术发展计划项目(2024YX130)。

摘　　要：目的:构建预测病理性乳头溢液患者并发乳腺癌风险的决策树模型。方法:选取2019年1月—2024年4月于潍坊市中医院就诊且符合纳入标准的病理性乳头溢液患者共157例,依据Logistic回归分析构建病理性乳头溢液患者并发乳腺癌的风险预测模型,绘制决策树并根据受试者工作特征(ROC)曲线下面积(AUC)对模型的预测效能进行评价。结果:病理性乳头溢液患者并发乳腺癌的发病率为24.2%,二元Logistic回归分析显示,乳头溢液中CEA、CA153水平和血性溢液为病理性乳头溢液患者并发乳腺癌的独立危险因素(P<0.05);基于上述因素建立了预测病理性乳头溢液患者并发乳腺癌风险的决策树模型,模型验证结果显示,Logistic回归模型的AUC值为0.800,决策树模型的AUC值为0.889。结论:基于影响因素构建的决策树模型,对病理性乳头溢液患者并发乳腺癌的发生风险具有良好的预测能力,有助于提高临床医师对病理性乳头溢液患者的术前精准诊断。Objective:To construct a decision tree model to predict the risk of breast cancer in patients with pathological nipple discharge.Methods:A total of 157 patients with pathological nipple discharge,who were diagnosed and treated at Weifang Municipal Hospital of Traditional Chinese Medicine from January 2019 to April 2024 and met the inclusion criteria,were selected.A risk prediction model for concurrent breast cancer in patients with pathological nipple discharge was developed using Logistic regression analysis.A decision tree was then constructed,and the predictive performance of the model was assessed based on the area under the receiver operating characteristic curve(AUC).Re-sults:The incidence of concurrent breast cancer among patients with pathological nipple discharge was 24.2%.Accord-ing to the results of binary Logistic regression analysis,elevated CEA and CA 153 levels in nipple discharge,as well as bloody discharge,emerged as independent risk factors for the development of breast cancer in such patients(P<0.05).Based on these findings,a decision tree model was constructed to predict the risk of concurrent breast cancer in patients with pathological nipple discharge.The validation results showed that the Logistic regression model had an AUC value of 0.800,while the decision tree model achieved an AUC value of 0.889.Conclusions:The decision tree model,built upon the identified influencing factors,exhibits strong predictive power for the risk of developing concurrent breast can-cer in patients with pathological nipple discharge,thus facilitating more precise preoperative diagnoses by clinicians for these patients.

关 键 词：病理性乳头溢液 LOGISTIC回归分析 决策树模型 乳腺肿瘤 

分 类 号：R737.9[医药卫生—肿瘤]