Background: The optimal management of small (≤2 cm) pancreatic neuroendocrine tumors (PNETs) remains controversial. We evaluated these tumors in the National Cancer Data Base (NCDB) to determine if resection provides a survival advantage over observation.
Methods: The NCDB was queried to identify patients with nonmetastatic PNETs ≤2 cm treated between 1998 and 2006. Kaplan-Meier survival estimates, stratified by grade and treatment type, evaluated the difference in 5-year overall survival (OS) between patients who underwent resection and observation. Multivariable Cox regression was used to determine the importance of resection in OS.
Results: Three hundred eighty patients met inclusion criteria. Eighty-one percent underwent resection; 19% were observed. Five-year OS was 82.2% for patients who underwent surgery and 34.3% for those who were observed (p < 0.0001). When controlling for age, comorbidities, income, facility type, tumor size and location, grade, margin status, nodal status, surgical management, and nonsurgical therapy in the Cox model, observation [hazard ratio (HR) 2.80], poorly differentiated histology (HR 3.79), lymph node positivity (HR 2.01), and nonsurgical therapies (HR 2.23) were independently associated with an increase in risk of mortality (p < 0.01).
Conclusion: Patients with localized PNETs ≤2 cm had an overall survival advantage with resection compared to observation, independent of age, comorbidities, tumor grade, and treatment with nonsurgical therapies.