Para-aortic lymph node sampling in pancreatic head adenocarcinoma.

Background: The significance of positive para-aortic nodes in patients with resectable pancreatic carcinoma is unclear. This study sought to evaluate the accuracy of intraoperative detection and prognostic significance of these lymph nodes in patients with resected adenocarcinoma of the pancreatic head.

Methods: From 2000 to 2010, para-aortic node sampling was performed prospectively in all patients before pancreatoduodenectomy. Frozen sections were created and nodes categorized as positive or negative for metastases. Surgeons were blinded to the frozen-section results. This was followed by standard histopathological assessment of corresponding paraffin-embedded, haematoxylin and eosin-stained material. Nodes considered uninvolved by this analysis were examined immunohistochemically for micrometastases.

Results: A total of 111 consecutive patients were included, with a median follow-up of 20·8 (range 1·5-126) months. The 1-, 2- and 5-year overall survival (OS) and disease-free survival (DFS) rates were 73·6, 54·0 and 24·7 per cent, and 51·8, 28·1 and 18·8 per cent respectively. Para-aortic node involvement was always associated with peripancreatic lymph node metastasis, and was detected by frozen-section analysis in 12 patients and by haematoxylin and eosin staining in 17. Sensitivity and specificity of frozen-section examination for detecting para-aortic lymph node metastases were 71 and 100 per cent respectively. Median OS for patients with and without para-aortic node involvement on frozen-section analysis was 9·7 versus 28·5 months respectively (P = 0·012), and 15·7 versus 27·2 months (P = 0·050) when assessed by haematoxylin and eosin staining. Median DFS for patients with and without para-aortic node involvement on frozen-section examination was 5·6 versus 12·9 months respectively (P = 0·041), and 8·4 versus 12·9 months (P = 0·038) for haematoxylin and eosin analysis. The presence of micrometastases in para-aortic nodes was not significantly associated with altered OS or DFS.

Conclusion: Para-aortic node sampling with frozen-section examination detects distant lymphatic involvement reliably. It should be performed systematically. When metastases are found, they should be considered a contraindication to pancreatic resection.