出 处:《疑难病杂志》2020年第9期926-930,共5页Chinese Journal of Difficult and Complicated Cases
基 金:湖北省科技支撑计划项目(2015BCA316);湖北省武汉市科技计划项目(2016060101010045)。
摘 要:目的利用TARGET数据库提供的临床信息分析儿童骨肉瘤患者的预后因素,构建生存预后列线图(Nomogram),并对其预测准确性进行评估。方法从TARGET数据库中获得儿童骨肉瘤患者的原始临床数据共214例,首先通过单因素回归分析确定影响骨肉瘤患者生存预后的相关因素,删除有相关因素空白或缺失的病例,获得101例儿童骨肉瘤患者临床信息的矩阵,随机将101例患者划分为建模组(n=73)和验证组(n=28)。并通过多因素Cox比例风险回归模型分析影响儿童骨肉瘤患者生存预后的独立危险因素。根据筛选的相关因素构建Nomogram图。利用一致性指数和校准图对其预测准确性进行评估。结果患者的性别、人种、原发肿瘤部位、手术方式、转移部位及术前化疗后组织坏死率与生存预后有关(P<0.1),其中,性别、术前化疗后组织坏死率是影响生存预后的独立危险因素(P<0.05)。所有相关危险因素均被用于构建Nomogram图。对Nomogram图预测价值的评估结果显示,建模组的一致性指数(C指数)为0.768(95%CI 0.721~0.815),验证组为0.727(95%CI 0.648~0.806),C指数高于美国癌症联合委员会(AJCC)第六版(0.65)、第七版(0.65),且校正曲线的预测值与真实值显示出良好的一致性。结论利用TARGET数据库构建的Nomogram图可为儿童骨肉瘤患者提供快速、准确的预后评估。Objective To analyze the prognostic factors of children with osteosarcoma and construct a nomogram to predict the long term survival rate by using the clinical information provided by TARGET database,and evaluate the accuracy of the prediction model.Methods The children's original clinical data of 214 patients with osteosarcoma was collected from the TARGET database.Firstly,through the single factor regression analysis to determine the related factors influencing the survival outcomes for patients with osteosarcoma,delete the related factors of blank or missing cases and get a matrix of 101 cases of children patients with osteosarcoma of clinical information.The 101 patients were divided into model group(70%)and validation group(30%)and the independent risk factors affecting the survival and prognosis of patients were analyzed by a multivariate cox proportional risk regression model.The Nomogram was constructed according to the selected relevant factors.The accuracy was evaluated using the consistency index(C index)and the calibration curves.Results The patient's gender,race,primary tumor site,surgical methods,metastatic sites and the rate of tissue necrosis after preoperative chemotherapy were related to survival prognosis(P<0.1),among which gender and rate of tissue necrosis after preoperative chemotherapy were independent risk factors affecting survival prognosis(P<0.05).All relevant risk factors were used to construct the Nomogram.The results of the evaluation of the predictive value of the Nomogram showed that the consistency index(C index)of the modeling group was 0.768(95%CI 0.721-0.815),and that of the verification group was 0.727(95%CI 0.648-0.806),which was statistically higher than the C index values of the following systems:AJCC 6th edition(0.65)and AJCC 7th edition(0.65).The calibration curves showed a high consistency between the predicted value and the real value.Conclusion The constructed Nomogram provides a rapid and accurate prognostic assessment for children with osteosarcoma.
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...