A Nomogram and a Risk Classification System Predicting the Cancer-Specific Survival of Patients With Initially-Diagnosed Osseous Spinal and Pelvic Tumors.

Abstract

Retrospective analysis. Our goal was to provide a predictive model and a risk classification system that predicts cancer-specific survival (CSS) from spinal and pelvic tumors. Primary bone tumors of the spinal and pelvic are rare, thus limiting the understanding of the manifestations and survival from these tumors. Nomograms are the graphical representation of mathematical relationships or laws that accurately predict individual survival. A total of 1033 patients with spinal and pelvic bone tumors between 2004 and 2016 were selected from the Surveillance, Epidemiology, and End Results (SEER) database. Multivariate Cox analysis was used on the training set to select significant predictors to build a nomogram that predicted 3- and 5-year CSS. We validate the precision of the nomogram by discrimination and calibration, and the clinical value of nomogram was assessed by making use of a decision curve analyses (DCA). Data from 1033 patients with initially-diagnosed spinal and pelvic tumors were extracted from the SEER database. Multivariate analysis of the training cohort, predictors included in the nomogram were age, pathological type, tumor stage, and surgery. The value of C-index was 0.711 and 0.743 for the internal and external validation sets, respectively, indicating good agreement with actual CSS. The internal and external calibration curves revealed good correlation of CSS between the actual observation and the nomogram. Then, the DCA showed greater net benefits than that of treat-all or treat-none at all time points. A novel risk grouping system was established for CSS that can readily divide all patients into three distinct risk groups. The proposed nomogram obtained more precision prognostic prediction for patients with initially-diagnosed primary spinal and pelvic tumors. 3.