四肢纤维肉瘤病人预后因素分析与列线图的构建

PROGNOSTIC FACTORS FOR PATIENTS WITH FIBROSARCOMA OF EXTREMITIES AND ESTABLISHMENT OF A NOMOGRAM

目的探讨四肢纤维肉瘤病人预后相关因素并构建预测病人总生存率的列线图和危险分层系统。方法回顾性收集美国国家癌症研究所监测、流行病学和最终结果(SEER)数据库(SEER*Stat 8.3.6软件)2004—2015年诊断为四肢纤维肉瘤病人的资料,应用单因素和多因素Cox回归筛选独立预后相关因素并构建列线图。采用C指数、ROC曲线、校准曲线和临床决策曲线评价列线图。应用X-tile软件将病人分为低风险、中风险和高风险3组,并利用Kaplan-Meier(K-M)生存曲线研究3组病人间的生存差异。结果多因素Cox回归分析表明,病人年龄(67~79岁:HR=2.941,95%CI=1.944~4.450;>79岁:HR=8.576,95%CI=5.571~13.200)、肿瘤直径>11.2 cm(HR=3.388,95%CI=2.036~5.637)、肿瘤病理分级Ⅲ级和Ⅳ级(Ⅲ级:HR=2.339,95%CI=1.066~5.133;Ⅳ级:HR=2.891,95%CI=1.345~6.218)、N1分期(HR=3.135,95%CI=1.170~8.403)、M1分期(HR=4.222,95%CI=2.336~7.631)和手术治疗(HR=0.288,95%CI=0.129~0.646)是四肢纤维肉瘤病人的独立预后相关因素(P<0.05)。训练集和验证集的C指数分别为0.792(95%CI=0.757~0.827)和0.780(95%CI=0.717~0.843)。校准曲线和临床决策曲线结果均显示该列线图在训练集和验证集中均表现良好。ROC曲线分析显示,无论在训练集还是验证集中,列线图在1、3和5年的曲线下面积(AUC)始终大于同一时间点所有独立预后相关因素的AUC。K-M生存曲线分析显示,列线图可显著区分低风险、中风险和高风险3组不同预后的病人。结论本研究所建立的列线图模型能有效预测四肢纤维肉瘤病人预后,基于该列线图预测模型的危险分层系统对区分高危病人具有一定临床价值。

Objective To investigate the prognostic factors for patients with fibrosarcoma in extremities,and to establish a nomogram and a risk stratification system for predicting overall survival(OS)rate.Methods The data of patients who were diagnosed with fibrosarcoma of extremities in 2004—2015 were collected from the Surveillance,Epidemiology,and End results(SEER)Database of National Cancer Institute(SEER*Stat 8.3.6),and a retrospective analysis was performed for these data.Univariate and multivariate Cox regression analyses were used to screen out independent prognostic factors and establish a nomogram.The nomogram was evaluated by C-index,receiver operating characteristic(ROC)curve,calibration curve,and decision curve analysis(DCA).X-tile software was used to divide the patients into low-,moderate-,and high-risk groups,and the Kaplan-Meier survival curve was used to investigate the difference in survival between the three groups.Results The multivariate Cox regression analysis showed that age(67-79 years:HR=2.941,95%CI=1.944-4.450,P<0.05;>79 years:HR=8.576,95%CI=5.571-13.200,P<0.05),tumor diameter>11.2 cm(HR=3.388,95%CI=2.036-5.637,P<0.05),pathological gradeⅢandⅣ(gradeⅢ:HR=2.339,95%CI=1.066-5.133,P<0.05;gradeⅣ:HR=2.891,95%CI=1.345-6.218,P<0.05),N1 stage(HR=3.135,95%CI=1.170-8.403,P<0.05),M1 stage(HR=4.222,95%CI=2.336-7.631,P<0.05),and surgical treatment(HR=0.288,95%CI=0.129-0.646,P<0.05)were independent prognostic factors for patients with fibrosarcoma in extremities.The C-index was 0.792(95%CI=0.757-0.827)in the training set and 0.780(95%CI=0.717-0.843)in the validation set.Calibration curve and DCA curve showed that this nomogram had good performance in both training and validation sets.The ROC curve analysis showed that the area under the ROC curve(AUC)of the nomogram was higher than the AUC of all independent prognostic factors at 1,3,and 5 years in both training and validation sets.The Kaplan-Meier survival curve analysis showed that the nomogram significantly distinguished the patients with dif-ferent prognoses in the low-,moderate-,and high-risk groups.Conclusion The nomogram established in this study can effectively predict the prognosis of patients with fibrosarcoma of extremities,and the risk stratification system based on this predictive nomogram model has a certain clinical value in distinguishing high-risk patients.