背景:探讨影响IIIC1期宫颈癌患者总生存期(OS)和癌症特异性生存期(CSS)预后的因素,并建立列线图模型来预测这种预后。
方法:来自Surveil-lance患者的数据,流行病学,符合纳入标准的最终结果(SEER)计划被归类为一个培训组,和验证数据于2010年至2019年在安徽医科大学第一附属医院获得。发病率,卡普兰-迈耶曲线,评估训练组IIIC1期宫颈癌患者的OS和CSS。根据单变量和多变量Cox回归模型的结果建立列线图。Harrell的C-index,校准图,计算受试者工作特征(ROC)曲线和决策曲线分析(DCA)以验证预测模型。
结果:盆腔淋巴结转移的发生率,宫颈癌预后的高危因素,随着时间的推移略有下降。确定了OS的八个独立预后变量,包括年龄,种族,婚姻状况,组织学,扩展范围,肿瘤大小,放疗和手术,但是只有七个被确定为CSS,婚姻状况被排除在外。基于结果建立OS和CSS的列线图。OS和CSS的列线图的C指数分别为0.687和0.692,分别使用SEER数据集和0.701和0.735的随机抽样,使用外部数据集的随机抽样。OS列线图的AUC分别为SEER数据集为0.708和0.705,外部数据集为0.750和0.750。分别。此外,当使用SEER数据集验证时,CSS的列线图的AUC为0.707和0.709,以及使用外部数据集验证时的0.788和0.785。列线图的校准图几乎与实际观察结果相同。DCA还指出了两种模型的价值。
结论:确定了OS的8个独立预后变量。同样的因素预测CSS,除了婚姻状况。验证后OS和CSS列线图均具有良好的预测和临床应用价值。值得注意的是,肿瘤大小对OS和CSS列线图的贡献最大.
BACKGROUND: To explore the factors that affect the prognosis of overall survival (OS) and cancer-specific survival (CSS) of patients with stage IIIC1 cervical cancer and establish nomogram models to predict this prognosis.
METHODS: Data from patients in the Surveil-lance, Epidemiology, and End Results (SEER) programme meeting the inclusion criteria were classified into a training group, and validation data were obtained from the First Affiliated Hospital of Anhui Medical University from 2010 to 2019. The incidence, Kaplan-Meier curves, OS and CSS of patients with stage IIIC1 cervical cancer in the training group were evaluated. Nomograms were established according to the results of univariate and multivariate Cox regression models. Harrell\'s C-index, calibration plots, receiver operating characteristic (ROC) curves and decision-curve analysis (DCA) were calculated to validate the prediction models.
RESULTS: The incidence of pelvic lymph node metastasis, a high-risk factor for the prognosis of cervical cancer, decreased slightly over time. Eight independent prognostic variables were identified for OS, including age, race, marriage status, histology, extension range, tumour size, radiotherapy and surgery, but only seven were identified for CSS, with marriage status excluded. Nomograms of OS and CSS were established based on the results. The C-indexes for the nomograms of OS and CSS were 0.687 and 0.692, respectively, using random sampling of SEER data sets and 0.701 and 0.735, respectively, using random sampling of external data sets. The AUCs for the nomogram of OS were 0.708 and 0.705 for the SEER data sets and 0.750 and 0.750 for the external data sets, respectively. In addition, AUCs of 0.707 and 0.709 were obtained for the nomogram of CSS when validated using SEER data sets, and 0.788 and 0.785 when validated using external data sets. Calibration plots for the nomograms were almost identical to the actual observations. The DCA also indicated the value of the two models.
CONCLUSIONS: Eight independent prognostic variables were identified for OS. The same factors predicted CSS, with the exception of the marriage status. Both OS and CSS nomograms had good predictive and clinical application value after validation. Notably, tumour size had the largest contribution to the OS and CSS nomograms.