基于SEER数据库构建原发性纵隔生殖细胞瘤患者生存预测的列线图

Background and objective Primary mediastinal germ cell tumor (PMGCT) is a rare but occasionally highly invasive mediastinal tumor. At present, there are few related disease special survival (DSS) studies on PMGCT, rare large data analysis, and uncommon DSS prognostic models. This study was to investigate the prognostic factors of DSS of the PMGCT patients, and build a simple and effective nomogram to predict the DSS prognosis in patients with PMGCT. Methods Retrospective clinicopathological data of 325 patients with PMGCT from 1975 to 2019 were extracted from Surveillance, Epidemiology, and End Results (SEER) database. The Kaplan-Meier method along with the Log-rank test were utilized to estimate the DSS. Cox proportional hazard regression model was used to screen the independent risk factors affecting prognosis, from which an individualized nomogram was constructed to predict 3-yr, 5-yr and 8-yr DSS of patients with PMGCT. The prediction accuracy of the model is evaluated by receiver operating characteristic (ROC) curve, correction curve and decision curve analysis (DCA) curve. Results The 3-yr, 5-yr and 8-yr survival rates of PMGCT were 84.6%, 83.6% and 83.3%, respectively. Univariate analysis showed that histology, surgery, age, tumor size, metastasis and stage could affect the prognosis of PMGCT. Multivariate analysis showed that histology, surgery, age and tumor size were independent risk factors for the prognosis of PMGCT patients, and the nomogram was constructed using these independent risk factors. The area under the curve (AUC) of ROC curve was 0.824. The results of the correction curves of 3-yr, 5-yr and 8-yr survival time and DCA, indicated that there was a good consistency between the predicted results of the nomogram evaluation and the real results. Conclusion Patients with histological classification of seminoma in PMGCT have a better prognosis than patients with non-seminoma. The prognosis of patients with over the age of 40 yr, tumor size ≥15 cm and without surgical treatment was even worse. The nomogram model can accurately and intuitively predict the DSS of patients with PMGCT.