Background: The prognostic significance of variables related to nodal involvement (node status, number of disease-positive nodes [posNn], node ratio [Nr], and site of nodal metastases) in patients with resected pancreatic head cancer remains poorly defined.
Methods: Clinical, operative, and pathologic data, including indexes of the burden and extent of nodal involvement, were analyzed in a consecutive series of 77 patients who underwent resection with extended lymphadenectomy for adenocarcinoma of the pancreatic head.
Results: Fifty-nine patients (77%) were found to have lymph node (LN) metastases. Median LN count, posNn, and Nr were 28 (10-54), 4 (1-29), and 14% (2%-55%), respectively. Twenty-six patients (44% of N1) had metastases limited to node level (NL) 1 (i.e., peripancreatic nodes); metastases up to NL2 (nodes along main arteries and hepatic hilum) and NL3 (preaortic nodes) were found in 21 (36%) and 12 (20%) patients, respectively. Interestingly, survival of patients with positive LN limited to NL1 was similar to that of node-negative patients (P = 0.407). posNn, Nr, and NL were all significant predictors of survival (P < 0.015). posNn and Nr proved to be an accurate proxy of NL involvement. The best cutoff of posNn was 2 and of and Nr was 10%.
Conclusions: The level of nodal metastatic spread is a statistically significant prognostic factor in cancer of the pancreatic head. Both posNn and Nr are accurate proxy of NL and may improve patients' risk stratification.