Endoscopic submucosal dissection (ESD) is a well-accepted method for removing superficial mucosal tumors; however, there is limited data on the use of this method for removing subepithelial tumors.

To investigate the efficacy, safety, and outcome of ESD for gastric subepithelial tumors and determine factors related to treatment success.

Retrospective analysis of a prospectively maintained database.

Single tertiary academic center.

From April 2007 to November 2010, 37 patients with gastric subepithelial tumors were treated with ESD.

Macroscopically and microscopically complete en block resection rate (R0), complication rate, and endosonographic features predictive of R0 resection.

The median tumor diameter was 25.0 mm, (range 10-60 mm, IQR 17-37). The overall rate of R0 resections was 81.1% (30/37, 95%CI: 61.8-90.2%), including 100% (15/15, 95%CI: 78.2-100.0%) of tumors from the submucosa and 68.2% (15/22, 95%CI: 45.1-86.1%) of tumors from the muscularis propria. Seventeen patients had a final diagnosis of gastrointestinal stromal tumor. The severe complication (perforation) rate was 5.4% (2/37, 95%CI: 0.0-9.5%). One patient required surgery; the other was treated conservatively. No recurrence was observed in patients with R0 resections at a median follow up of 21.0 months (IQR 11-35). Successful R0 resections were predicted by the observation of no, or only narrow, tumor connections with the underlying muscle layer during EUS (OR=35.0, 95%CI: 3.7-334.4, p=0.001).

Single-center, retrospective analysis, short follow-up.

ESD is an effective and relatively safe method for removing gastric subepithelial tumors. Endoscopic ultrasonography findings can predict complete tumor resections.