T. J. Weatherall1, J. W. Denbo1, J. P. Sharpe1, M. Martin2, M. Ismail3, T. O’Brien6, K. Groshart4, R. Gupta5, S. W. Behrman1, J. L. Deneve1, G. Munene1, P. V. Dickson1 1The University Of Tennessee Health Science Center,Division Of Surgical Oncology,Memphis, TN, USA 2The University Of Tennessee Health Science Center,Division Of Hematology & Oncology,Memphis, TN, USA 3The University Of Tennessee Health Science Center,Division Of Gastroenterology & Hepatology,Memphis, TN, USA 4Trumbull Laboratories,Pathology Group Of The MidSouth,Memphis, TN, USA 5The University Of Tennessee Health Science Center,Division Of Pathology,Memphis, TN, USA 6Duckworth Pathology Group,Memphis, TN, USA
Introduction:
Duodenal neuroendocrine tumors(dNETs) are rare neoplasms and specific treatment recommendations are less clear than for other gastroenteropancreatic NETs. The purpose of the current study was to examine an institutional experience with well-differentiated, non-functional, non-ampullary dNETs and identify factors to help guide appropriate evaluation and management.
Methods:
We performed a retrospective review of patients with a diagnosis of dNET from 1993-2015, excluding those with hormonally functional or ampullary tumors. Clinical data was obtained from medical records and evaluated to identify predictors of regional lymph node(LN) metastases. Tumor grade was determined by Ki-67 index according to WHO definitions(G1≤2%; G2 3-20%). Regional lymph node dissection(RLND) was defined as an anatomic operation that included hepatoduodenal, peri-pancreatic, and hepatic arterial LNs.
Results:
During the study period, 36 patients were identified with a diagnosis of well-differentiated, non-functional, non-ampullary dNET. Resection was performed via surgery in 28 patients and endoscopy in 8. Operations included antrectomy(15), partial duodenectomy(6), transduodenal submucosal resection(4), and pancreaticoduodenectomy(3). Eight patients underwent formal RLND as part of their operation, yielding a median of 8(5-12) LNs vs 3(1-6) in those in whom LNs were sampled incidentally(p=0.003). Depth of tumor invasion was evaluable in 33 patients and involved the mucosa in 4(12%), submucosa in 22(67%), and muscularis propria in 7(21%). Ki-67 index was evaluable in 30 patients, revealing G1 tumors in 20(67%) and G2 in 10(33%). LNs were included in the specimen in 19/28(68%) with LN metastases identified in 5/19(26%). Of those with LN metastases, all had tumors ≤2cm and there was no significant difference in tumor size between LN+ and LN- patients(1.4cm vs 1.0cm, p=0.13). When compared to patients with no LN involvement, those with positive LNs were more likely to have muscularis propria invasion(80% vs 23%, p=0.04) and have undergone formal RLND(80% vs 31%, p=0.03). Among LN+ patients, 40% had G2 tumors vs 50% in LN- pts(p=0.99). Of LN+ patients, suspicious LNs were identified on preoperative CT in 40%. Preoperative EUS was not performed in any LN+ patients, however, when utilized was found to have 100% accuracy in predicting tumor depth of invasion. No patients were found to have distant metastases. At a mean follow up of 28 months, no patient developed recurrence or experienced disease-specific death.
Conclusion:
Non-functional, non-ampullary dNETs have a propensity to metastasize to regional LNs. Although preoperative CT and EUS may assist in treatment planning, tumor size and grade were not predictive of LN involvement. Therefore, performance of an anatomic RLND at the time of resection likely results in more accurate disease staging. Longer follow up is necessary to further evaluate the prognostic impact of LN+ disease and therapeutic benefit of RLND.