个体化预测视网膜母细胞瘤预后风险预警模型的建立与验证  

Establishment and validation of individualized risk early warning model for predicting the prognosis of retinoblastoma

作　　者：孟佳[1] 杨鑫[1] 胡丽丽[2] MENG Jia;YANG Xin;HU Lili(Department of Ophthalmology,the First Affiliated Hospital of Zhengzhou University,Zhengzhou 450052,China;Department of Radiology,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眼科,河南郑州450052 [2]郑州大学第一附属医院放射科,河南郑州450052

出　　处：《肿瘤基础与临床》2023年第1期52-56,共5页journal of basic and clinical oncology

基　　金：河南省高等学校重点科研计划项目(21A320051)。

摘　　要：目的建立个体化预测视网膜母细胞瘤预后的风险预警模型,同时对其预测效果进行验证。方法选取2018年1月至2020年1月郑州大学第一附属医院收治的视网膜母细胞瘤患者125例作为研究对象。根据患者预后情况将分为预后良好组(98例)及预后不良组(27例),利用Logistic多因素回归分析影响视网膜母细胞瘤患者预后的危险因素。绘制预测视网膜母细胞瘤患者发生不良预后的列线图模型。应用受试者工作曲线(ROC)对列线图模型预测效能进行检验,并利用Bootstrap法检验模型的准确性。另选取2020年1月至2021年1月郑州大学第一附属医院收治的24例视网膜母细胞瘤患者对模型预测效能进行验证。结果单因素分析结果显示:预后良好组和预后不良组肿瘤直径(χ2=4.808,P=0.028)、单/双眼发病(χ2=5.482,P=0.019)、肿瘤分期(χ2=6.302,P=0.042)、肿瘤残留(χ2=9.517,P=0.002)、脑转移(χ2=11.413,P=0.001)、糖类抗原153(CA153)水平(χ2=4.834,P<0.001)、糖类抗原199(CA199)水平(χ2=3.826,P<0.001)及神经元特异性烯醇化酶(NSE)水平(χ2=5.530,P<0.001)比较差异均有统计学意义。多因素Logistic回归分析结果显示:肿瘤直径≥20 mm、双眼发病、具有脑转移、CA153水平、CA199水平及NSE水平均是影响视网膜母细胞瘤患者预后的危险因素(P=0.001,P=0.003,P=0.002,P<0.001,P<0.001,P<0.001)。根据Logistic多因素分析结果建立预测视网膜母细胞瘤患者预后的风险预警模型,Bootstrap法内部验证结果显示,C-index指数为0.837(95%CI:0.766~0.903)。曲线下面积、敏感性、特异性分别为0.812、0.845、0.648。选取2020年1月至2021年1月郑州大学第一附属医院收治的24例视网膜母细胞瘤患者对本研究所建立的风险预警模型预测效能进行验证,视网膜母细胞瘤患者预后风险预警模型预测发生预后不良5例,实际预后不良4例。曲线下面积、敏感性、特异性分别为0.847、0.805、0.703。结论Objective To establish an individualized risk early warning model for predicting the prognosis of retinoblastoma,and to verify its predictive effect.Methods The 125 patients with retinoblastoma admitted to the First Affiliated Hospital of Zhengzhou University from January 2018 to January 2020 were selected as the study subjects.According to the prognosis of the patients,they were divided into the good prognosis group(98 patients)and the poor prognosis group(27 patients).Logistic multivariate regression was used to analyze the risk factors affecting the prognosis of patients with retinoblastoma.A nomograph model to predict the adverse prognosis of patients with retinoblastoma was drew.The area under the subject working(ROC)curve was used to test the prediction efficiency of the nomograph model,and the accuracy of the model was tested by Bootstrap method.In addition,24 patients with retinoblastoma admitted to the First Affiliated Hospital of Zhengzhou University from January 2020 to January 2021 were selected to verify the predictive efficiency of the model.Results Univariate analysis showed that the tumor diameter(χ2=4.808,P=0.028),single/double eye incidence(χ2=5.482,P=0.019),tumor stage(χ2=6.302,P=0.042),tumor residue(χ2=9.517,P=0.002),brain metastasis(χ2=11.413,P=0.001),carbohydrate antigen 153(CA153)level(χ2=4.834,P<0.001),carbohydrate antigen 199(CA199)level(χ2=3.826,P<0.001)and neuron-specific enolase(NSE)level(χ2=5.530,P<0.001)in the good prognosis group were significantly different from those in the poor prognosis group.Multivariate logistic regression analysis showed that tumor diameter≥20 mm,bilateral onset,brain metastasis,CA153 level,CA199 level and NSE level were all risk factors affecting the prognosis of patients with retinoblastoma(P=0.001,P=0.003,P=0.002,P<0.001,P<0.001,P<0.001).The results of logistic multifactor analysis were used to establish a risk early warning model to predict the prognosis of retinoblastoma patients.The internal validation results of Bootstrap method showed that t

关 键 词：个体化预测 视网膜母细胞瘤 预后 

分 类 号：R739.7[医药卫生—肿瘤]