去分化脂肪肉瘤预后风险模型的建立与分析

Establishment and Analysis of a Prognostic Model for Dedifferentiated Liposarcoma

目的 基于美国国家癌症研究所监测、流行病学和最终结果(SEER)数据库,构建预后列线图用于预测去分化脂肪肉瘤(DDLPS)3、5和8 a的总生存(OS)率。方法 从SEER数据库中提取DDLPS患者的临床信息,随机分为建模组和验证组。通过LASSO回归和多因素Cox回归筛选变量,将筛选出的变量纳入预测模型并构建列线图。使用一致性指数(C指数)、时间依赖曲线下面积(时间依赖AUC)和校准曲线评估模型的区分度和校准度。通过决策曲线分析(DCA)、净重新分类指数(NRI)、综合判别改善指数(IDI)进一步比较列线图和AJCC分期之间的净收益和预测准确性。根据列线图计算总得分,对患者进行风险分层,通过Kaplan-Meier曲线和log-rank检验比较分层间OS率差异。结果 共纳入1 172例DDLPS患者,筛选出年龄、原发位置、T分期、N分期、M分期、FNCLCC分级、手术和放疗8个预后变量。时间依赖AUC(>0.7)和C指数(建模组为0.741,验证组为0.764)表明列线图有良好的区分度。校准曲线表明预测生存概率和实际生存率具有良好的一致性。IDI和NRI表明,列线图的准确性优于AJCC分期系统(P<0.05)。DCA显示,列线图与AJCC分期系统相比具有更多的净收益。风险分层显示低、中和高风险组3组之间的OS率差异有统计学意义(P<0.05)。结论 建立了DDLPS患者OS率的列线图,对于评估患者预后有一定临床指导作用。

Objective To construct a nomogram to predict 3, 5 and 8-year overall survival(OS) rate for dedifferentiated liposarcoma(DDLPS) based on the surveillance, epidemiology and end results(SEER) database of the National Cancer Institute.Methods The clinical information of DDLPS patients was extracted from SEER database. They were further randomly divided into a modeling group and a validation group. Least absolute shrinkage and selection operator(LASSO) regression and multivariate Cox regression were used to screen variables. The selected variables were included in the prediction model and the nomogram was constructed. Concordance index(C-index), area under the time-dependent receiver operating characteristic curve(time-dependent AUC) and calibration plots were used to evaluate the discrimination and calibration of the nomogram. Decision curve analysis(DCA), net reclassification index(NRI) and integrated discrimination improvement(IDI) were used to further compare the net benefits and prediction accuracy between nomogram and AJCC staging system. Total score was calculated according to the nomogram and patients were risk stratified. The difference in OS rate among stratified groups was compared by Kaplan-Meier curve and log-rank test.Results A total of 1 172 patients with DDLPS were enrolled and 8 variables including age, primary site, T stage, N stage, M stage, FNCLCC grade, surgery and radiotherapy were screened. Time-dependent AUC(>0.7) and C index(0.741 for modeling group and 0.764 for validation group) indicated that the nomogram had good discrimination. The calibration curves showed that the predicted survival rate was in good agreement with the actual survival rate. IDI and NRI indicated that the nomogram was more accurate than the AJCC staging system(P<0.05). DCA showed that the nomogram has more net benefits than the AJCC staging system. Risk stratification showed significant differences in OS rate among low-, middle-and high-risk groups(P<0.05).Conclusion The nomogram for DDLPS patients was established, which has a certain clinical guiding role in evaluating the prognosis of patients.