Endoscopic resection for diagnosis and endosonographic follow-up of gastric GI stromal tumors larger than 2cm

Background: Gastric subepithelial tumors originated from the muscularis propria (MP) layer are difficult on differenzial diagnosis before surgical resection, even with EUS or EUS-FNA. Although the small gastric GIST (<Q2cm) carried very low malignant progression (Miettinen and Lasota 2006), it is uncertain for the potential malignancy and prognosis in GIST between 2cm and 5cm, without confirmed mitotic rate.

Aims: The aim of this study was to evaluate feasibility and short-termed effect of endoscopic resection for diagnosis of gastric GIST and EUS follow-up on intended complete eradication cases.

Methods: From April 2005 to July 2010, Eleven patients with EUS-confirmed gastric subepithelial tumors, which were originated from MP layer and endoluminal growth pattern, underwent endoscopic resection. Nine lesions (<R2cm) were confirmed GISTs by immunohistochemical analysis; The EUS features, histopathologic findings, endoscopic resection methods, complete resection rate, complications and EUS follow-ups are analyzed in the retrospective single-center study. The EUS instruments were conventional radial EUS, GF-UMQ240 (2) and/or mini-probe UM-2R (9)/EU-M30 (Olympus, Tokyo, Japan).

Results: The main reasons of endoscopic resection on these 9 patients (3 male, 6 female, mean age 52.8yr, range 37–73 yr) were: histological diagnosis (6), UGI bleeding (2), progressive size change (1). The tumor locations were cardia (2), fundus (3) proximal body (2), distal body (2). Mean tumor size was 26.1mm (range 20–36.5mm). Complete endoscopic resection of GIST was successful in 66.6% (6/9 tumors). Endoscopic submucosal dissection (ESD, 5 cases) had higher complete resection rate (2 endolooping on fundal lesions) as compare with endoscopic mucosal resection method (EMR, 4 cases) (80% vs.50%, P=0.40); however one perforation (11.1%) occurred on ESD group (31mm, fundus) and was closed by multiple hemoclips without surgical intervention. The mitotic rate of all lesions was <5/50 HPF. We used EUS follow-up on all lesions after endoscopic resection and found three hypoechoic tumors on MP layer due to incomplete resection lesions (33.3%). These lesions (21.1mm, 17.3mm, 20.9mm by EUS) were treated by surgical resection 39 days, 57 days, 198 days after previously endoscopic resection respectively without complications or progression. In complete resection cases, there were indistinct, thin hyperechoic submucosa and underlying MP layer around ulcer scar without residual or recurrent GIST on EUS after mean follow-up of 17.92 months (3–53 months).

Conclusions: Endoscopic resection seem to be safe and feasible on histologic diagnosis of gastric GIST>2cm and possible complete eradication (66.6%) on EUS follow-up. However there is still risk of perforation on fundal lesions by ESD technique.