Development and validation of a clinico-histological factor-based nomogram for survival in sinonasal malignancies

Abstract

Abstract
The survival prognosis nomogram is needed for sinonasal malignancies with different histology types. Thus this research collected cases with sinonasal malignancies from The Surveillance, Epidemiology, and End Results (SEER) program. R was the tools for screening the predictors by Accelerated Failure Time model and “backward” elimination. Nomogram was constructed with predictors. C-index, calibration curves and decision curve analysis were applied for the validation of the nomogram. 5795 cases were collected for constructing nomogram. “Age”, “Sex”, “Marriage”, “Site”, “AJCC”, “Surg”, “Radiation”, and “Hist” were retained as predictors. The C-index was 0.77 (95% CI: 0.76–0.79, validation set). Calibration curves were close to the diagonal line. Decision curve analysis showed nonmogram provided positive net benefits relative to the default “treat-all” and “treat-none” approaches. We also conducted survival analysis for seven histology types with treatment (squamous cell carcinoma, adenocarcinoma, adenoid cystic carcinoma, undifferentiated carcinoma, primary mucosal melanoma, neuroendocrine carcinoma, olfactory neuroblastoma ). In conclusion, we constructed a nomogram to predict the survival prognosis of patients with sinonasal malignancies. It can be utilized by clinicians to predict individual survival outcomes and for patient consultation. Besides, we provided survival status for seven histology types of sinonasal malignancies with common and controversial treatment.