乳腺癌骨转移的影响因素分析及预测模型构建  

Influencing Factors for Bone Metastasis of Breast Cancer andConstruction of A Predictive Model

作　　者：李奕颖[1] 苏静[1] 李孟圈[1] LI Yiying;SU Jing;LI Mengquan(Department of Breast Surgery,the First Affiliated Hospital of Zhengzhou University,Zhengzhou 450052,China)

机构地区：[1]郑州大学第一附属医院乳腺外科,河南郑州450052

出　　处：《河南医学研究》2021年第2期211-215,共5页Henan Medical Research

摘　　要：目的分析乳腺癌骨转移患者的临床病理特征,构建乳腺癌骨转移发生的预测模型。方法回顾性分析2019年6月1日至2019年12月31日郑州大学第一附属医院收治的201例乳腺癌患者的临床资料。根据首次转移部位,将其分为骨转移组与非骨转移组。通过t检验、Mann-Whitney U检验、χ^2检验分析各临床病理特征与患者发生骨转移之间的关系。采用logistic回归筛选变量构建预测模型,并制作列线图。采用受试者操作特征曲线(ROC曲线)评估该模型的可靠性。结果单因素分析结果显示,淋巴结转移情况、激素受体状态、Ki-67、pS2、CK5/6、TOP-Ⅱ、P53及E-cadherin表达差异有统计学意义(P<0.05)。多因素分析结果显示,患者有淋巴结转移(OR=3.115,95%CI:0.992~9.785,P=0.043)、Ki-67高表达(OR=1.993,95%CI:1.975~2.012,P=0.032)、pS2表达阴性(OR=0.040,95%CI:0.040~0.430,P=0.002)、P120表达阴性(OR=0.261,95%CI:0.078~0.877,P=0.025)、E-cadherin表达阴性(OR=0.129,95%CI=0.037~0.455,P=0.001)及TOP-Ⅱ高表达(OR=1.491,95%CI:1.620~1.020,P=0.001)是患者发生骨转移的独立危险因素,也是构建列线图的显著变量。该列线图通过了校正和验证步骤,训练组和验证组的曲线下面积(AUC)分别为0.767和0.855。结论该模型是预测乳腺癌患者骨转移的可靠工具,有助于临床医生为患者提供更加准确的决策依据并优化治疗方案。Objective To analyze the clinical characteristics of breast cancer bone metastasis patients,and to construct a prediction model of breast cancer bone metastasis.Methods The clinical data of 201 breast cancer patients from June 2019 to December 2019 in the First Affiliated Hospital of Zhengzhou University were obtained.Based on the first site of metastasis,they were divided into bone metastasis group and non-bone metastasis group.The data were determined retrospectively by using t-test,Mann-Whitney U-test andχ2 test.Logistic analysis was used to establish prediction model based on clinal factors.Meanwhile,these factors were used to construct the nomogram.The model was assessed by using receiver operating characteristic curve(ROC curve).Results The result of univariate analysis showed that lymph node status,hormone receptor status and the expression of Ki-67,pS2,CK5/6,TOP-Ⅱ,P53 and E-cadherin were risk factors which affect the occurrence of breast cancer bone metastasis.Multivariate analysis found that patients with lymph node metastasis(OR=3.115,95%CI:0.992-9.785,P=0.043),high Ki-67 expression(OR=1.993,95%CI:1.975-2.012,P=0.032),negative pS2 expression(OR=0.040,95%CI:0.040-0.430,P=0.002),negative P120 expression(OR=0.261,95%CI:0.078-0.877,P=0.025),negative E-Cadherin expression(OR=0.129,95%CI:0.037-0.455,P=0.001)and high TOP-Ⅱexpression(OR=1.491,95%CI:1.620-1.020,P=0.001)were independent risk factors of bone metastases in breast cancer patients.They were also the variables which construct the nomogram.According to the calibration and validation steps,the area under the curve(AUC)of the modeling group and the verification group were 0.767 and 0.855 respectively.Conclusion The nomogram model provided an insightful and applicable tool to evaluate the occurrence of breast cancer bone metastasis.It also can assist clinical decision effectively.

关 键 词：乳腺癌 骨转移 预测模型 

分 类 号：R737.9[医药卫生—肿瘤]