Proposal of a TNM classification-based staging system for esthesioneuroblastoma: More precise prediction of prognosis.

Abstract

Esthesioneuroblastoma (ENB) is a rare malignant neoplasm. Currently, no consistent and universal staging system for ENB exists. The aim of this study is to propose a TNM-based classification. Hundred and forty-two patients from our institution, with ENB pathologically confirmed between July 1978 and December 2018, were reviewed. All patients were restaged according to the Kadish stage, Morita stage and American Joint Committee on Cancer (AJCC) T classification from clinical and radiological data. Multivariate Cox proportional hazard regression analyses were performed to determine the impact of various factors. The goodness-of-fit and predictive accuracy of the different staging systems were calculated using R software. The median follow-up time was 57 months (range: 4-229 months). According to the Kadish system, the 5-year overall survival (OS) for patients with stage A, B and C was 100%, 83.6% and 64.2%, respectively (P = .055). With respect to the Morita classification, 5-year OS for stages A, B, C and D was 100%, 83.6%, 70.7% and 50.0%, respectively (P = .004). Analysis based on the proposed staging model demonstrated 5-year OS for stage I, II, III and IV disease was 100%, 88.9%, 75.9% and 49.0%, respectively (P < .001). In separate multivariate Cox regression models, only the novel staging system exhibited independent effects on OS (P = .004); the Akaike information criterion and Harrell's concordance index were also superior to those calculated for the Kadish or Morita systems. The proposed TNM-based staging system offers an improved prognostic assessment for patients with ENB. Further verification and refinement from additional dataset application is required.