Abstract
Background: Guidelines for the current National Comprehensive Cancer Network recommend radical cholecystectomy, including hepatic resection and portal lymph node (LN) dissection, for patients with early stage gallbladder (GB) cancer. We sought to determine the survival benefit conferred by adequate LN evaluation. Methods: We used the surveillance, epidemiology and end results (SEER) neoplasm registry to identify patients who had an operation for GB cancer between 1988 and 2004. Patients were classified by stage of disease, operative procedure performed (cholecystectomy alone or radical resection), number of LNs evaluated (0, 1, >1), and receipt of radiation (RT). We included patients with T1B, T2, and T3 neoplasms who were LN positive or negative. Patients with T4 neoplasms and those with metastatic disease were excluded. Multivariate analysis included adjustment for age, race, sex, neoplasm grade, stage, operation performed, receipt of RT, and neoplasm registry. Results: We identified 4,614 patients who underwent operative treatment for stage 1-2B GB (including T1B-T3 and LN positive or negative) cancer between 1988 and 2004. Of 4,614 patients, 9.6% (442) had radical resection, whereas 90.4% (4,172) had cholecystectomy alone. Among patients undergoing radical resection, 56% had LNs evaluated as compared with 28% of patients after cholecystectomy. For patients with T1B and T2 neoplasms who underwent radical resection, pathologic evaluation of at least 1 LN was associated with a significant improvement in median overall survival (OS) compared with those who had no LN evaluated (123 months vs 22 months; P < .0001). Radical resection with no LN evaluation provided similar OS compared with cholecystectomy alone (22 months vs 23 months; P = NS). For patients with T3 neoplasms, radical resection, including pathologic evaluation of at least 1 LN, was also associated with improved OS compared with radical resection with no LN evaluation (12 months vs 7 months; P = .0014). Again, individuals who had radical resection without LN evaluation had similar OS compared with those who had cholecystectomy alone (7 months vs 6 months; P = NS). Individuals who had radical resection with LN evaluation were more likely to receive RT than those who had radical resection without LN evaluation (33.1% vs 19.1%; P = .002). In multivariate analysis (including adjustment for RT), however, LN evaluation was still associated with a decrease in mortality compared with no LN evaluated (HR = 0.611; 95% CI = 0.484, 0.770). The pathologic evaluation of additional LN (>1) did not provide any additional benefit compared with the evaluation of a single node (HR = 0.795; 95% CI = 0.571, 1.107). Radical resection alone (without LN evaluation) did not provide any benefit over cholecystectomy alone (HR = 1.098; 95% CI = 0.971, 1.241). Conclusion: LN evaluation is a critical component of radical resection for GB cancer. In the absence of LN evaluation, radical resection provides no benefit over cholecystectomy alone. © 2009 Mosby, Inc. All rights reserved.