To identify potential prognostic factors of postoperative small cell lung cancer (SCLC) patients and establish an effective nomogram for prediction of survival outcomes. Methods: Patients who had been diagnosed as SCLC from 2004 to 2012 in the Surveillance, Epidemiology, and End Result database ( SEER) were identified and collected. Kaplan-Meier method was used to estimate the overall survival (OS) in the non-surgery group and the surgery group. Cox regression was performed to identify independ-ent prognostic factors. Four models were established using different staging system and compared by concordance index (C-index) and calibration curve. Two methods were utilized to conduct the comparison. Results: A total of 45226 SCLC patients were identified. Af-ter applying the exclusion criteria, 867 postoperative patients were included and analyzed. In multivariate analysis, prognostic factors were age, sex, surgery, radiation sequence, tumor diameter, tumor extension, T stage, N stage, number of lymph node examination, grade of pathological differentiation and metastasis, respectively. Independent covariates were selected using Akaike's information cri-terion. Nomogram was then formulated based on results of multivariate analysis. C-index of the nomogram incorporating tumor size and extension was 0. 706, which was higher than other conventional classifications such as AJCC TNM Classification (0. 700), VALSG (0. 665) and IASLC (0. 667) staging systems. Validation of the nomogram demonstrated that it had an ideal predictive accuracy. Conclusion: Operation is associated with a better survival of SCLC patients. Tumor size and extension are important independent prog-nostic factors. Nomogram incorporating tumor size, extension and other variables are ideal prognostic prediction tools for OS of postop-erative SCLC patients, which has better predictive accuracy than conventional classifications.%目的:证实手术治疗对于小细胞肺癌患者长期生存的作用.鉴定小细胞肺癌术后患者生存影响因素,构建小细胞肺癌术后患者的生存预测模型.与现有的AJCC分期系统、VALSG分期系统和IASLC的分期系统预测性能进行比较.方法:选取2004年至2012年SEER数据库中确诊为小细胞肺癌的患者(small cell lung cancer, SCLC),提取相应的变量数据.采用Kaplan-Meier比较不同分期下手术组与非手术组患者的生存状况,并绘制生存曲线.针对手术治疗的SCLC患者,利用赤池信息准则(AIC)筛选变量,基于Cox回归模型构建Nomogram预测模型.比较新模型与AJCC分期系统、VALSG分期系统和IASLC分期系统的一致性指数(C-index),评价模型的预测效能.结果: 通过数据检索,共有45 226例SCLC患者入选本研究,其中867例为手术治疗患者.多因素分析发现,影响手术患者预后的因素包括年龄、性别、手术方式、放疗顺序、肿瘤大小、肿瘤侵犯范围、T分期、N分期、淋巴结清扫数量、病理分化程度和远端转移情况.经过赤池信息准则( AIC)筛选,年龄、性别、肿瘤大小、肿瘤侵犯范围、淋巴结侵犯情况、远端转移情况、手术方式、放疗情况、淋巴结清扫数量、病理分化程度共10个变量入选模型.比较4个模型的一致性指数,Nomogram为0. 706,AJCC模型为0. 700,IASLC模型为0. 667,VALSG模型为0. 665.Nomogram模型显示最佳的预测准确度.结论:患者是否接受手术影响小细胞肺癌患者的生存时间.肿瘤的大小和肿瘤侵犯的范围是独立的预后因素. Nomogram生存预测模型的预测性能明显优于其它分期系统.
展开▼
