Nomogram based on TNM stage to predict the prognosis of thymic epithelial tumors (TETs) patients undergoing extended thymectomy.

Abstract

Thymomas and thymic carcinoma are thymic epithelial tumors (TETs) of the anterior mediastinum. On the basis of The AJCC 8th Edition of TNM classification, no prognostic prediction model has been established for TETs patients undergoing surgical resection. In this study, based on data from Qilu Hospital of Shandong University, we identified prognostic factors and developed a nomogram to predict the prognosis for TETs patients undergoing extended thymectomy. Patients with TETs who underwent thymectomy between 2010 and 2020 were consecutively enrolled. An analysis of multivariate Cox regression and stepwise regression using the Akaike information criterion (AIC) was conducted to identify prognostic factors, and a nomogram for TETs was derived from the results of these analyses. The model was validated internally with the Kaplan-Meier curves, ROC curves and calibration curves. There were 350 patients with TETs enrolled in the study, and they were divided into a training group (245,0.7) and a validation group (105,0.3). Age, histological type, tumor size, myasthenia gravis, and TNM stage were independent prognostic factors for CSS. The Kaplan-Meier curves showed a significant difference between high nomorisk group and low nomorisk group. A nomogram for CSS was formulated based on the independent prognostic factors and exhibited good discriminative ability as a means of predicting cause-specific mortality, as evidenced by the area under the ROC curves (AUCs) of 3-year, 5-year, and 10-year being 0.946, 0.949, and 0.937, respectively. The calibration curves further revealed excellent consistency between the predicted and actual mortality when using this nomogram. There are several prognostic factors for TETs. Based on TNM stage and other prognostic factors, the nomogram accurately predicted the 3-, 5-, and 10-year mortality rates of patients with TETs in this study. The nomogram could be used to stratify risk and optimize therapy for individual patients.