Quality of lymphadenectomy for esophageal cancer in the United States: An analysis of the SEER database.

Abstract

4068 Background: The Worldwide Esophageal Cancer Collaboration (WECC), a consortium of institutions to pool data on EC, reported recommendations regarding the optimal number of lymph nodes removed during esophagectomy. This study was performed to assess how often optimal lymphadenectomy (LAN) was performed in a national EC registry. Methods: From the SEER-17 cancer registry (1988-2007), patients undergoing esophagectomy with “T-classification” and “number of resected lymph nodes (LN)” recorded were analyzed. The WECC recommended optimal LAN as ≥10 LN for T1, ≥20 LN for T2, and ≥30 LN for T3 or T4 tumors. Optimal LAN and survival were primary outcomes. Log-rank was performed for univariate comparisons and Cox regression for multivariate analysis. Results: 5792 patients with resected EC were appropriate for analysis. Median number of lymph nodes resected was 9. Optimal LAN was achieved in 951 patients (16%) of the cohort and in only 164 (5%) of those with T3 or T4 EC. Optimal LAN was more likely achieved with lesser T-classifications (OR for T1=13.43, CI:11.12-16.23), in patients <65 yo (OR 1.26, CI:1.08-1.47), and in patients with adenocarcinoma (OR 1.20, CI:0.98-1.47). Despite identifying more N+ disease, 5-yr OS was improved in each T-classification for patients who underwent optimal LAN (Table). Conclusions: The SEER database provides insight into practice patterns at US hospitals. In that context, LAN for EC is grossly underperformed when held to the standard of the WECC recommendations. Patients who undergo an optimal LAN have improved survival, likely due to identifying the need for further treatment or to a therapeutic effect of the LAN. Efforts should therefore be made by surgeons to perform a more extensive LAN during esophagectomy to ensure that patients are appropriately staged and treated. pT Extent of lymphadenectomy No. of patients (%) Positive lymph nodes (%) 5-year OS p value (OS) T1N=1520(26%) Optimal 46 (42%) 134 (21%) 62.1% <0.001 Suboptimal 874 (58%) 124 (14%) 53.8% T2N=1106 (19%) Optimal 141 (13%) 74 (52%) 39.1% 0.009 Suboptimal 965 (87%) 415 (43%) 31.8% T3/T4N=3166 (55%) Optimal 164 (5%) 122 (74%) 24.3% 0.015 Suboptimal 3,002 (95%) 2,000 (67%) 19.8% *11 patients with missing value.