Abstract
BACKGROUND: Thymoma is a rare thymic epithelial tumor, and its prognostic factors remain not fully elucidated. This study aimed to identify simple, practical preoperative predictors of prognosis, focusing on tumor dimensions assessed by computed tomography (CT). METHODS: We retrospectively analyzed 181 patients who underwent complete or partial resection for thymoma between 2004 and 2022. Tumor size was assessed by measuring the maximum transverse and craniocaudal dimensions on preoperative CT. Freedom from recurrence (FFR) and overall survival (OS) were estimated using the Kaplan-Meier method. Cox proportional hazards models were constructed: Model A included only preoperative variables, while Model B additionally incorporated postoperative factors (e.g., TNM stage classification and histology). RESULTS: During a median follow-up of 96 months, the 5- and 10-year FFR rates were 78.4% and 71.4%, and the corresponding OS rates were 96.5% and 87.3%, respectively. In multivariate analysis, TNM stage classification and preoperative steroid pulse therapy were significantly associated with FFR, whereas no variables were significantly associated with OS. Model A demonstrated good discriminatory ability (C-index = 0.839), which improved only modestly after including postoperative factors in Model B (C-index = 0.867). In the steroid-excluded cohort (N = 148), the craniocaudal tumor dimension emerged as a significant predictor of FFR (p = 0.027). CONCLUSIONS: The craniocaudal tumor dimension measured on preoperative CT was consistently associated with recurrence and may reflect prognostic information embedded within the pathological TNM classification. This easily measurable parameter could complement TNM-based evaluation in preoperative risk assessment and surgical decision-making for thymoma.