Prognostic grading after complete resection for thymic malignancies

OBJECTIVE: Despite the World Health Organization (WHO) and Masaoka classifications have been widely accepted as the main describers of prognosis determinants in thymic malignancies, so far, these have been considered independently from one another. We have reviewed our single-centre 40-year results after surgical treatment of thymic malignancies evaluating the inter-relationships between the clinical, surgical and pathological variables and investigating their prognostic impact in completely resected patients. PATIENTS AND METHODS: A surgical series of 347 patients was reviewed and, of these, 305 with complete resection enrolled. Long-term and disease-free survival (LTS, DFS) analyses were performed. Kaplan-Meir curves for WHO histotypes and Masaoka-stages were inspected and matched with the log-rank test; the Cox regression analysis was adopted in a multivariable approach. RESULTS: Considered independently, the WHO-histotypes did not differentiate clearly from one to another in terms of LTS and DFS; however, types A-AB-B1-B2 and B3-C clustered in 2, statistically different, malignancy groups (LTS, DFS: Cox-p < 0.001). Masaoka staging was confirmed to be a relevant prognostic determinant, even if no evident difference between stages I vs II and stages III vs IV emerged when the Masaoka-classification was factored in. Thus, when investigating 13 surgical and pathological factors of invasiveness, these showed a clustering in 2 groups according to the presence/absence of pathological proven infiltration in the peri-thymic structures (LTS, DFS: Cox-p < 0.001). By matching the WHO-malignancy clusters and infiltration clusters, 4 classes may be identified, which proved to have a distinct prognostic significance: (LTS-Cox: stage-I vs stage-II, p = 0.003; III: p < 0.001, IV: p < 0.001; DFS-Cox: stage-I vs stage-II, p < 0.001; III: p < 0.001; IV: p < 0.001). CONCLUSIONS: When analyzing the long-term outcome of patients underwent complete resection for thymic malignancies, the combination between pathological and surgical variables showed accurate prognosis predictability