Lymph node sampling rates and predictors of nodal metastasis in pancreatic neuroendocrine tumor resections: the UCSF experience with 149 patients.

OBJECTIVES

The decision to perform pancreas-preserving procedures or standard resections for pancreatic neuroendocrine tumors (PNETs) is often based on the perceived risk of malignancy, including potential nodal involvement. We sought to identify clinicopathological factors that predict nodal disease.

METHODS

This is a retrospective review of pathology database for PNET resections from January 1, 1988, to March 15, 2010. Univariate analysis and multivariate logistic regression were used to identify predictors of nodal metastasis.

RESULTS

A total of 149 patients were identified. Enucleations had lower lymph node sampling rates compared to major resections. Excluding enucleations, 23% of patients had no lymph nodes sampled. For patients who did have lymph nodes evaluated, a median of 5 lymph nodes were examined. On multivariate analysis, only distant disease predicted nodal metastasis (odds ratio = 3.80, P = 0.02); tumor size did not (P = 0.48). One third of patients with lymph node metastasis had tumors less than 3 cm.

CONCLUSIONS

Lymph nodes are not evaluated in many major pancreatic resections for PNET, and preoperative prediction of nodal metastasis is difficult, even when tumor size is considered. Consequently, many patients may be understaged and undergo potentially inadequate resection. Inconsistent lymph node sampling may explain conflicting conclusions in the literature regarding the prognostic value of lymph node involvement in PNET patients.