男性乳腺癌转移模式的研究及远处转移预测模型的构建和验证

Study on metastasis patterns of male breast cancer and construction and validation of a predictive model for distant metastasis

目的探究男性乳腺癌(MBC)的远处转移模式,分析MBC发生远处转移的危险因素,进一步构建并验证MBC发生远处转移的预测模型。方法回顾性分析美国监测、流行病学和结果(SEER)数据库中2010—2020年MBC患者的临床病理学数据,按照7∶3随机分为训练集和测试集,选取2010—2020年就诊于哈尔滨医科大学附属肿瘤医院的MBC患者125例作为外部验证集。训练集采用单因素和多因素分析筛选出MBC远处转移的独立影响因素,进一步构建Logistic回归预测模型,通过绘制受试者工作特征曲线(ROC)、校准曲线以及临床决策曲线(DCA)对模型进行评价。开发基于Web的网页计算器使该模型在线使用。结果在SEER数据库中的3592例MBC患者,有354例患者在初始诊断时即发生了远处转移,占病例数的9.86%,其中骨是最常见的转移部位(3.81%)。多因素Logistic回归分析结果显示:婚姻状态、诊断到治疗的时间、病理学类型、T分期、N分期和分子分型是MBC发生远处转移的独立危险因素。基于这些独立影响因素构建Logistic回归预测模型并进行验证:训练集的曲线下面积(AUC)为0.838,测试集AUC值为0.836,外部验证集AUC值为0.717,模型具有很好的预测能力;校准曲线显示预测模型对MBC患者发生远处转移的预测结果与实际结果之间具有较高的一致性;DCA曲线显示该模型可以使患者有更高的临床获益。结论男性乳腺癌远处转移模式的预测模型对患者发生远处转移具有较好的预测性能,有助于临床医师对于高危患者的早期识别并为患者制定个体化的治疗方案。

Objective To investigate the patterns of distant metastasis in male breast cancer(MBC),analyze the risk factors for distant metastasis in MBC,and further construct and validate a predictive model for distant metastasis in MBC.Methods A retrospective analysis of clinical and pathological data of MBC patients from the Surveillance,Epidemiology,and End Results(SEER)database from 2010 to 2020 was conducted.The patients were randomly divided into a training set and a testing set in a 7∶3 ratio.Additionally,one hundred and twenty⁃five MBC patients treated at Cancer Hospital Affiliated to Harbin Medical University from 2010 to 2020 were selected as an external validation set.Univariate and multivariate analyses were conducted on the training set to identify independent factors influencing distant metastasis in MBC.A Logistic regression prediction model was constructed,and the model was evaluated using receiver operating characteristic curve(ROC),calibration curve,and decision curve analysis(DCA).A web⁃based calculator was developed to enable online use of the model.Results Among 3,592 MBC patients in the SEER database,354 patients had distant metastasis at the initial diagnosis,accounting for 9.86%of the total.Bone was the most common site of metastasis(3.81%).The results of the multivariate Logistic analysis showed that marital status,time from diagnosis to treatment,histological type,T stage,N stage,and molecular subtype were independent risk factors for distant metastasis in MBC.A Logistic regression prediction model was constructed based on these independent factors and validated:the area under the curve(AUC)was 0.838 for the training set,0.836 for the testing set,and 0.717 for the external validation set,indicating good predictive performance of the model.The calibration curve showed high consistency between the predicted and actual results of distant metastasis in MBC patients,and the DCA curve showed that the model could provide higher clinical benefits to patients.Conclusions The predictive model constructed in this study has good predictive performance for distant metastasis in MBC patients,which can help clinicians identify high⁃risk patients early and develop personalized treatment plans for patients.