Abstracts
2002 Digestive Disease Week

# 100614 Abstract ID: 100614 Duodenal Adenocarcinoma: The Need for Accurate Lymph Node Staging
Abeezar Sarela, Murray F Brennan, Martin S Karpeh, David Klimstra, Kevin C Conlon, New York, NY

Duodenal adenocarcinoma is often included with either peri-ampullary or small bowel adenocarcinoma when determining prognosis or therapy. We hypothesized that duodenal cancer is prognostically more similar to gastric antrum cancer and thus should be pathologically staged in a similar fashion. Methods: Prospectively accrued experience with duodenal adenocarcinoma and distal gastric adenocarcinoma, between 10/1983 and 12/2000, was examined. Results: Some 137 patients with adenocarcinoma of the duodenum (excluding ampullary lesions) and 545 patients with gastric antrum cancer were identified. Of 137 patients, 56 (41%) underwent pancreaticoduodenectomy for tumors in the first or second portion of the duodenum, 16 (12%) underwent segmental resection of tumors in the third or fourth portion and 65 (47%) had unresectable tumors. For the operable patients (all R0), pathological stage distribution was I (T1 or T2, N0; 12, 17%), II (T3 or T4, N0; 29, 40%) or III (any T, N1; 31, 43%). The median number of lymph nodes examined in each case was 13 (IQR 7 - 22), with 6 or more nodes available in 59 (82%) patients and 15 or more nodes in 34 (47%) patients. Of patients with gastric antrum cancer, 331 (61%) underwent R0 resection with a median of 23 (16 -32) lymph nodes examined and 15 or more nodes available for 256 (77%) patients. Disease-specific survival characteristics were similar for both groups (table). The presence of nodal metastasis was a significant prognostic indicator for duodenal cancer (hazard ratio 2.6, P = 0.03; table) and its predictive value was substantially enhanced by examination of 15 or more lymph nodes (5-yr survival: 100% for N0 vs 47% for N1; P = 0.01) but was negated with less than 15 nodes (69% vs 64%; P = 0.5). Prevalence of nodal metastasis was similar in patients with < or ?15 nodes examined (16, 42% vs 15, 44%, respectively; P = 1.0). For patients with N1 disease, the number of involved lymph nodes (1 - 2 nodes, 16 patients vs 3 or more nodes, 13 patients) did not substratify survival characteristics (60% vs 50%, respectively; P = 0.5). Conclusion: Resectable cancers of the duodenum are associated with prognosis similar to gastric antrum cancer and examination of at least 15 nodes is required for accurate prognostic discrimination. This study suggests that duodenal adenocarcinoma should be included with gastric cancer when designing investigative therapies.