The American Joint Committee on Cancer has revised the Tumor-Node-Metastasis (TNM) staging system for papillary thyroid cancer (PTC) patients. We examined the impact of this new classification (TNM-8) on patient stratification and estimated the prognostic value of clinicopathological features for the disease-free interval (DFI) in a cohort of 1148 PTC patients. Kaplan-Meier analyses showed that all clinicopathological parameters analyzed, except age and multifocality, were associated significantly with DFI. Cox regression identified tall cell PTC variant and stage as independent risk factors for DFI. When the stage was replaced with age, tumor size, and lymph node (LN) metastases in the set of covariates, the lateral LN metastases stood out as the strongest independent predictor of DFI, followed by tall cell variant and age. A noteworthy result emerging from these analyzes is that regression models had lower Akaike and Bayesian information criterions if variables were categorized based on the TNM-7. In addition, we examined data from a different PTC patient cohort, acquired from The Cancer Genome Atlas database, to verify whether the DFI prediction could be enhanced by further clinicopathological and molecular parameters. However, none of these was found to be a significant predictor of DFI in the Cox model.
Keywords: TNM; autoimmune thyroid diseases; histology; lymph node metastasis; multifocality; papillary thyroid cancer; prognosis; vascular invasion.