Pineoblastoma (PB) is a rare neoplasm of the central nervous system. This analysis aimed to identify factors and establish a predictive model for the prognosis of adult patients with PB. Data for 213 adult patients with PB (Surveillance, Epidemiology, and End Results database) were randomly divided into primary and validation cohorts. A predictive model was established and optimized based on the Akaike Information Criterion and visualized by a nomogram. Its predictive performance (concordance index and receiver operating characteristic curve) and clinical utility (decision curve analyses) were evaluated. We internally and externally validated the model using calibration curves. Multivariate Cox regression analysis identified age, year of diagnosis, therapy, tumor size, and tumor extension as independent predictors of PB. The model exhibited great discriminative ability (concordance index of the nomogram: 0.802; 95% confidence interval: 0.78–0.83; area under the receiver operating characteristic curve: ranging from 0.7 to 0.8). Calibration plots (probability of survival) showed good consistency between the actual observation and the nomogram prediction in both cohorts, and the decision curve analyses demonstrated great clinical utility of the nomogram. The nomogram is a useful and practical tool for evaluating prognosis and determining appropriate therapy strategies.
|Original language||English (US)|
|Journal||Frontiers in Oncology|
|State||Published - Jul 24 2020|
- risk factor
ASJC Scopus subject areas
- Cancer Research