Comparison of clinical features and survival between thymic carcinoma and thymic carcinoid patients

Abstract

OBJECTIVES: Thymic carcinoma (TC) and thymic carcinoid (TCD) are aggressive thymic epithelial neoplasms with a poor prognosis. Due to rarity, little is known about their comparative clinical characteristics, treatment outcomes and patterns of relapse. METHODS: A retrospective cohort study was performed on 287 patients with TC and 56 patients with TCD who were treated at the Shanghai Chest Hospital between February 2003 and April 2014. Patient demographics, tumour stage, treatment, pathologic findings and postoperative outcomes were compared between the two tumour types using both multivariable Cox regression analysis and propensitymatched analysis. RESULTS: Compared to patients with TC, significantly more patients with TCD were male, had larger tumours, and displayed a greater proportion of lymph node metastases. However, overall survival was similar (60.7% 5-year survival for TC, 80.7% for TCD, P = 0.159), as was disease-free survival (41.1% 5-year survival for TC, 37.6% for TCD, P = 0.696) and patterns of relapse. Multiple Cox regression analysis identified younger patients [hazard ratio (HR) 1.018; 95% confidence interval (CI) 1.000-1.035; P = 0.047], more completeness of resection (HR 1.424; 95% CI 1.105-1.836; P = 0.006), adjuvant radiotherapy (HR 0.455; 95% CI 0.276-0.751; P = 0.002), and no adjuvant chemotherapy (HR 1.799; 95% CI 1.017-3.183; P = 0.044) as independent factors predicting better overall survival. Completeness of resection (HR 1.258; 95% CI 1.022-1.548; P = 0.031) and TNM stage (HR 1.479; 95% CI 1.107-1.977; P = 0.008) were independent predictors of disease-free survival. Propensity matching produced 46 patients in each group and no significant difference on overall survival or diseasefree survival was found. CONCLUSIONS: Patients with TCD have discrete features but share a similar clinical course to those with TC. The importance of complete resection in both of these thymic malignancies is emphasized. Further investigation at multiple centers with the longer follow-up data is required to substantiate our conclusion.