Tu1088 Prediction of Tumor Recurrence in Patients With Gastrointestinal Stromal Tumors Following Resection by the Modified NIH Criteria

Abstract

Background: Endoscopic ultrasound (EUS) is often performed to evaluate patients with rectal carcinoid tumor to evaluate tumor size, depth of invasion and lymph node status. A recently proposed surveillance protocol calls for an EUS examination at 3 months following local therapy and then every 6 months for 3 years. The need and utility for surveillance programs in this setting have not been validated. Aims: In patients who have undergone resection of a rectal carcinoid to determine the potential utility of endoscopic surveillance for identifying disease recurrence. Method: A histopathology database was reviewed to identify patients with an endoscopic diagnosis of a well-differentiated rectal carcinoid tumor from 1/1/00-1/9/12. Clinical, radiologic, endoscopic, EUS, and follow-up outcome data were analyzed. Results: Among 87 patients who met inclusion criteria, local excision was performed in 75 (86%) to include: hot snare polypectomy n=56, transanal excision (TAE) n=13, endoscopic mucosal resection (EMR) n=5 and endoscopic submucosal dissection (ESD) n=1. Of 70 patients with local disease management, 55 (79%) underwent post intervention surveillance, 20 (36%) of whom included EUS [1 (IQR 1-2.5)]. Colonoscopy alone was performed in n=23, colonoscopy and non-invasive imaging n=7, non-invasive imaging alone n= 5, colonoscopy and EUS n=4, colonoscopy and EUS and non-invasive imaging n=1. Fifteen (21%) patients had no specific rectal carcinoid surveillance on site but had ongoing on site separate clinical evaluations. Disease progression was identified in 6 (8.6%) patients at 1.4 (IQR 0.38-2.7) years from initial diagnosis. All 6 cases reflected distant metastatic locations (n=4 liver, n=1 pancreas, n=1 lung and liver) rather than that of local recurrence or a combination of local and distant disease. All were identified by non-invasive imaging at 35 (IQR 4.7-33) months following initial therapy. At no point during their course was local recurrence identified. Disease specific mortality (n=3) occurred at 13 (IQR 9-69) months following initial therapy. Conclusions: Our data suggest the lack of local recurrence of rectal carcinoid following local disease management. These data call into question the utility of any endoscopic surveillance program. Additional study is needed to more firmly establish the natural course of disease in such patients to help devise a more accurate and tailored approach to disease surveillance.