论文部分内容阅读
目的:探讨ALKn +间变性大细胞淋巴瘤（ALCL）预后相关因素，建立一种治疗前预测其总生存（OS）的临床预后模型。n 方法:选取监测、流行病学和最终结果（SEER）数据库（涵盖18个登记处）中2004年1月至2015年12月诊断为ALKn + ALCL的1 602例患者，按简单随机法分成训练组（1 122例，用于模型的构建和内部验证）和验证组（480例，用于外部验证）。应用Cox比例风险模型对OS进行单因素、多因素分析，选取独立预后因素建立ALKn + ALCL 3年、5年的OS列线图。应用受试者工作特征（ROC）曲线、一致性指数（C指数）和校准曲线评估模型的区分度和校准度。与Ann Arbor分期进行对比，使用净重新分类指数（NRI）、综合判别改善指数（IDI）和决策曲线分析（DCA）对建立的模型准确性和效益进行评估。n 结果:多因素Cox比例风险模型分析结果显示，年龄、性别、全身症状、Ann Arbor分期、原发部位是ALKn + ALCL患者OS的独立影响因素，并据此建立预测OS的列线图。与Ann Arbor分期相比，C指数（训练组：0.726比0.597，验证组：0.777比0.639）、ROC曲线下面积（AUC）（训练组3年OS率AUC：0.758比0.620，5年OS率AUC：0.761比0.614；验证组3年OS率AUC：0.819比0.672，5年OS率AUC：0.832比0.656）、校准曲线均提示所建模型具有较好的预测能力；在验证组中，3年、5年OS率的NRI分别为0.866（95%n CI 0.683~1.079）、0.922（95% n CI 0.760~1.099），IDI分别为0.219和0.233（均n P<0.01）。构建的OS预测模型3年、5年DCA比Ann Arbor分期有更大的临床净收益。n 结论:成功建立了ALKn + ALCL的OS预测模型，可个体化预测生存结果。n “,”Objective:To investigate the prognostic factors of ALKn + anaplastic large cell lymphoma (ALCL), and to establish a clinical prognostic model for predicting the overall survival (OS) before treatment.n Methods:A total of 1 602 patients diagnosed as ALKn + ALCL from January 2004 to December 2015 in Surveillance, Epidemiology, and End Results (SEER) database (covering 18 registries) were included. The patients were randomly divided into a training cohort (1 122 cases, used for model construction and internal verification) and a validation cohort (480 cases, used for external verification) by simple randomization. The Cox proportional hazard models were used for univariate and multivariate analyses of OS, and independent prognostic factors were selected to establish 3-year and 5-year OS nomogram of ALKn + ALCL. The discrimination and calibration of the model were evaluated by using the receiver operating characteristic (ROC) curve, concordance index (C index) and calibration curve. Compared with Ann Arbor staging, the accuracy and benefits of the established model were assessed by using the net reclassification improvement (NRI), the integrated discrimination improvement (IDI), and the decision curve analysis (DCA).n Results:The analysis results of multivariate Cox proportional hazard model showed that age, gender, systematic symptoms, Ann Arbor staging, and primary site were independent influencing factors of OS in ALKn + ALCL, and a nomogram for predicting OS was established base on these factors. Compared with Ann Arbor staging, C index (training cohort: 0.726 vs. 0.597, validation cohort: 0.777 vs. 0.639), area under the ROC curve (AUC) (training cohort: AUC of 3-year OS rate 0.758 vs. 0.620, AUC of 5-year OS rate 0.761 vs. 0.614; validation cohort: AUC of 3-year OS rate 0.819 vs. 0.672, AUC of 5-year OS rate 0.832 vs. 0.656) and the calibration curve all indicated that the nomogram had better predicted ability. In the validation cohort, the NRI of 3-year and 5-year OS rates were 0.866 (95% n CI 0.683-1.079) and 0.922 (95% n CI 0.760-1.099), and IDI were 0.219 and 0.233 (both n P < 0.01), respectively. The constructed OS prediction model showed greater clinical net benefits than Ann Arbor staging by 3-year and 5-year DCA.n Conclusion:The OS prediction model of ALKn + ALCL has been successfully established, which can predict the survival outcomes individually.n