Metastasis pattern and prognosis of large cell neuroendocrine carcinoma: a population-based study

Abstract

Purpose As a rare type of tumor, the metastasis pattern of large cell neuroendocrine carcinoma (LCNEC) is still unclear. Our aim was to investigate metastatic patterns and develop a predictive model of prognosis in patients with advanced LCNEC. Methods Patients of LCNEC diagnosed between 2010–2015 from the Surveillance, Epidemiology and End Results (SEER) database were retrospectively included. Chi-square test was used for baseline characteristics analysis. Survival differences were assessed using Kaplan–Meier curves. Independent prognostic factors identified by multivariate Cox proportional risk model were used for the construction of nomogram. Results 557 eligible patients with metastasis LCNEC (median (IQR), 64 (56 to 72) years; 323 males) were included in this research. Among patients with isolated metastases, brain metastases had the highest incidence (29.4%), and multisite metastases had worse OS (HR: 2.020: 95% CI 1.413–2.888; P < 0.001) and LCSS (HR: 2.144, 95% CI 1.480–3.104; P < 0.001) in all age groups. Independent prognostic indicators including age, race, T stage, N stage, chemotherapy, radiotherapy and metastatic site were used for the construction of nomogram. Concordance index (C-index) and decision-curve analyses (DCAs) showed higher accuracy and net clinical benefit of nomogram compared to the 7th TNM staging system (OS: 0.692 vs 0.555; P < 0.001; LCSS: 0.693 vs 0.555; P < 0.001). Conclusions We firstly established a novel comprehensive nomogram to predict the prognosis of metastasis LCNEC. The prognostic model demonstrated excellent accuracy and predictive performance. Chemotherapy and metastasis pattern were the two strongest predictive variables. Close follow-up of patients with LCNEC is necessary to make individualized treatment decisions according to different metastasis patterns.