Using The Cancer Genome Atlas data to refine the 8th edition of the American Joint Committee on Cancer staging for papillary thyroid carcinoma

PURPOSE: The 8th edition of the American Joint Committee on Cancer (AJCC) staging led to a significant downstaging of well differentiated thyroid cancer patients. However, some patients who had been downstaged still experienced death. By using data from the thyroid cancer dataset of The Cancer Genome Atlas (TCGA), we aimed to find molecular features that could improve survival prediction. METHODS: TCGA data were downloaded from cBioPortal. Restaging of cases was performed according to the pathological reports. RESULTS: Out of 496 cases, 204 (41.1%) were downstaged, and the proportion of deaths increased in stages III and IV. TERT promoter mutations were no longer enriched in stage IV only, but significantly redistributed also in stages II and III. TERT mutation was the only alteration predictive of poor survival; however, in this series it was not independent from the AJCC staging. Five proteins (4E-BP1_pT70, Chk1_pS345, Snail, STAT5 alpha and PAI-1) were significantly associated with survival, and their use as a panel refined the risk stratification independently from the AJCC staging, with a hazard ratio for a positive result of 21.2 (95%CI 3.7–122.2, P = 0.0006). CONCLUSIONS: In the TCGA series, the proportion of deaths is in line with the expected survival of the latest AJCC staging, with a neat separation of risk among stages. Nevertheless, the use of protein expression can be useful in refining the stratification. Finally, after the restaging, a considerable number of tumors with TERT mutations will be allocated in lower stages; hence, dedicated studies should define the prognostic usefulness of these mutations in low-stage diseases.