RESUMO
BACKGROUND: We aimed to explore whether the anatomic extent of lymph node metastases (AE-LNM) could independently predict prognosis of node-positive major salivary gland carcinoma (MaSGC). METHODS: A total of 376 pathologically node-positive MaSGC patients were identified from the Surveillance, Epidemiology and End Results database and constituted the training cohort. Using the X-Tile program, these patients were divided into three groups based on AE-LNM degrees. Discrimination of overall survival (OS) and disease-specific survival (DSS) was evaluated and compared with the 8th American Joint Committee on Cancer (AJCC) pN classification. The results were externally validated by 220 patients in a Chinese multicenter cohort (Validation cohort). RESULTS: Using the training cohort, AE-LNM was divided into Extent 1 (spread to parotid LNs or level I), Extent 2 (spread to level II-IV) and Extent 3 (spread to level V or bilateral LNs or rare LNs). Regarding both OS and DSS, the AE-LNM model revealed clear separation of survival curves, while the pN classification failed to discriminate the prognosis of pN1 and pN2 patients. When we incorporated both the AE-LNM model and AJCC pN classification into the same multivariate Cox analyses, AE-LNM was still an independent prognostic factor, while the AJCC pN classification lost its significance. These results were externally validated by the validation cohort. CONCLUSION: AE-LNM is an independent nodal prognosticator for node-positive MaSGC and may have improved discriminative ability over the current AJCC pN classification. Integration of anatomic extent of LNM into the current AJCC N classification could be considered.