Nomograms and risk classification systems predicting overall and cancer-specific survival in primary malignant cardiac tumor.

Abstract

Primary malignant cardiac tumors (PMCTs) are fatal, but up to now, there is still a lack of survival prediction model for prognosis evaluation. We developed nomograms to predict overall survival (OS) and cancer-specific survival (CSS) for PMCTs by the Surveillance, Epidemiology, and End Result (SEER) database. A total of 506 PMCTs participants were identified in the SEER database from 1973 to 2014 and were randomly assigned into the training cohort (N = 354) and the validation cohort (N = 152). The prognostic factors for PMCTs were identified by Kaplan-Meier and multivariate Cox analysis and further incorporated to build OS and CSS nomograms. The nomograms were internally and externally validated via concordance indexes (C-index) and calibration curves. The independent prognostic factors for OS and CSS in PMCTs were associated with age at diagnosis, histopathology, tumor stage, cancer-directed surgery, and chemotherapy (all P < .05). In the internal validation, the C-index values were 0.71 (95% confidence interval [CI]: 0.68-0.75) for OS nomogram, and 0.70 (95% CI: 0.67-0.74) for CSS nomogram. In the external validation, the C-index values were 0.71 (95% CI: 0.66-0.77) for OS nomogram, and 0.71 (95% CI: 0.65-0.77) for CSS nomogram. The calibration curves of internal and external validation showed consistency between the nomograms and the actual observation. The risk stratification of PMCTs was significant distinction (P < .05). We developed and validated credible nomograms to predict OS and CSS in PMCTs. These nomograms can be offered to clinicians to more precisely estimate the survival and identify risk stratification of PMCTs.