Abstract
BACKGROUND: Colon neuroendocrine tumors (NETs) are uncommon. Currently, the impact of the number of metastatic lymph nodes (LNs) and lymph node ratio (LNR) on survival has been well investigated in other colon malignancies, but both remain nebulous for patients with colon NETs. METHODS: Surgically resected patients with histologically proven nonmetastatic colon NETs were queried from the Surveillance, Epidemiology, and End Results database between 1988 and 2011. Patients with lymph nodes involved were investigated and categorized into four LNs-based classifications (≤4, >4-10, >10-13, and >13) or three LNR-based subgroups (≤0.51, >0.51-0.71, and >0.71) according to the threshold, determined by Harrell's C statistic. Univariate and multivariate survival analyses were performed by log-rank test and Cox stepwise regression analysis, respectively. RESULTS: Eight hundred fifty-one patients met the inclusion criteria. Among them, higher LNR and LNs classification are associated with a worse prognosis. The 10-year NETs-specific survival rate was 78.3% (74.2-82.6%), 61.3% (52.4-71.7%), 40.8% (20.7-80.7%) for patients in the ≤4, >4-10, and 10-13 LNs groups, respectively. When patients were classified with LNR, the observed 10-year NETs-specific survival rate was 79.9% (74.8-85.5%) for ≤0.51, 57.4% (43.8-75.2%) for >0.51-0.71, and 40.0% (31.0-51.5%) for >0.71. In stratified analysis, higher LNs and LNR groups have worse prognosis only in patients with advanced T stage (T3-T4). Regarding stage migration, the LNR-based system did not show superiority to LNs-based classification. CONCLUSIONS: Current TNM staging classification could be improved by considering the count of metastatic nodes and LNR instead of a simple record of lymph node status (N1 or N0) for colon NETs.