A nomogram for the prediction of cerebrovascular disease among patients with brain necrosis after radiotherapy for nasopharyngeal carcinoma

Background and purpose: This study sought to develop and validate a nomogram to predict cerebrovascular disease (CVD) among patients with brain necrosis after radiotherapy for nasopharyngeal carcinoma (NPC). Materials and methods: A total of 346 eligible patients with brain necrosis after radiotherapy for NPC were divided into a training set (n = 231) and a validation set (n = 115). A multivariate Cox proportional hazards regression model was used to select the significant variables for CVD prediction in the training set. Then, a nomogram was developed based on the regression model. The performance of the nomogram was assessed with respect to discrimination and calibration. All patients were classified into high-or low-risk groups based on the risk scores derived from the nomogram. Moreover, a decision curve analysis was performed with the combined training and validation sets to evaluate the clinical usefulness of the nomogram. Results: Four significant predictors were identified: hypertension, statin treatment, serum level of high-density lipoprotein, and interval between radiotherapy and brain necrosis. The nomogram incorporating these four predictors showed favorable calibration and discrimination regarding the training set, with a C-index of 0.763 (95% CI, 0.694 to 0.832), which was confirmed using the validation set (C-index 0.768; 95% CI, 0.675 to 0.861). Furthermore, the nomogram successfully stratified patients into high-and low-risk groups. The decision curve indicated that our nomogram was clinically useful. Conclusion: The nomogram showed favorable predictive accuracy for CVD among patients with brain necrosis after radiotherapy for NPC and might aid in clinical decision making. (C) 2018 Elsevier B.V. All rights reserved.