LYMPH NODE ASSESSMENT INFORMS PROGNOSIS BUT EXTENT OF LYMPHADENECTOMY IS NOT ASSOCIATED WITH SURVIVAL IN PATIENTS UNDERGOING RESECTION FOR PANCREATIC NEUROENDOCRINE TUMORS
Emanuel Eguia*1, Patrick Sweigert1, Majid Afshar1, Gerard V. Aranha1, Gerard Abood1, Constantine Godellas1, PAUL KUO2, Marshall S. Baker1
1Surgery, Loyola University Medical Center, Maywood, IL; 2Surgery, University of Southern Florida, Tampa, FL
BACKGROUND
Most resected pancreatic neuroendocrine tumors (PNETs) are moderate or well-differentiated tumors with relatively indolent biology. The optimal management of regional lymph nodes in the resection of PNETs has not been definitively determined.
METHODS
We queried the National Cancer Database (NCDB) to identify patients undergoing resection for PNET between 2010 and 2014. Patients with poorly differentiated tumors, metastatic disease and those having no nodal assessment were excluded. Parsimonious multivariable (MVR) regression was used to identify factors associated with lymph node involvement (LN+). Cox proportional hazard modeling was used to identify factors associated with overall survival (OS). Candidate variables were chosen a priori and included: age, gender, race, Charlson Comorbidity Index (CCI), insurance status, facility type, procedure, tumor size, histologic grade, margin status, LN+ and number of nodes examined broken by quartiles (1:1-5; 2:6-10; 3:11-16; 4:≥17).
RESULTS
3709 patients met inclusion criteria. 2094 (57%) had tumors <3 cm in size. 3154 (85%) were well differentiated. 2665 (71%) had no LN+. Factors independently associated with LN+ included: male gender, undergoing pancreaticoduodenectomy (PD), moderately differentiated histology, positive margins, African American race (p<0.01) and the number of nodes examined (6-10 nodes: OR 1.72; 11-16: OR 2.21; >16: OR 2.96; p<0.01). In multivariable cox modeling of overall survival, age > 70 (HR 2.82; p<0.001), male gender (HR 1.32; p<0.01), CCI (HR 1.33; p<0.001), having a total pancreatectomy (HR 1.76; p<0.01), positive margins (HR 1.98; p<0.001), undergoing treatment in a community cancer program (HR 1.58; p<0.001) and LN+ (HR 1.32; p = 0.023) were all associated with an risk of death whereas the number of lymph nodes examined was not (5-10 LN: HR 1.15; 11-16 LN: HR 1.39; >16 LN: HR 1.03; p>.05).
CONCLUSION
In patients undergoing reaction for PNET, lymph node involvement is associated with poor prognosis. The number of nodes examined is itself an independent determinate of identifying nodal metastases, but the extent of regional lymphadenectomy does not impact overall survival.
Back to 2019 Posters
