Reliability of endoscopic ultrasonography and endoscopy in measurement of gastric subepithelial tumor size.

Abstract

Subepithelial tumor (SET) size is important in determining the treatment plan; however, size estimation for gastric SETs has not been well investigated. We aimed to investigate which method predicts SET size most accurately by retrospectively analyzing surgically removed SETs. From January 2015 through June 2020, patients who underwent surgical gastric SET removal at Asan Medical Center, Seoul, Korea, were enrolled. SET sizes measured by pathologists and endoscopists were retrospectively reviewed. The reliability of SET size measurement by endoscopic ultrasonography (EUS) and endoscopy was calculated using intraclass correlation coefficient (ICC), with pathologic size as the gold standard. Overall, EUS was highly reliable (ICC 0.86, P < 0.001), and endoscopy was moderately reliable (ICC 0.75, P < 0.001). When analyzed according to SET location, endoscopy was highly reliable in the lesser curvature's lower third (ICC 0.86, P = 0.014), middle third (ICC 0.88, P < 0.001), and upper third (ICC 0.90, P < 0.001); as well as the anterior wall's middle third (0.84, P < 0.001) and the posterior wall's upper third (ICC 0.80, P < 0.001). EUS (ICC 0.96, P = 0.005) and endoscopy (ICC 0.95, P = 0.008) both were most reliable for lower-third posterior wall lesions, whereas endoscopy was unreliable for middle-third greater curvature lesions (ICC 0.41, P = 0.05). Both EUS and endoscopy were reliable methods for measuring gastric SET size, and overall, EUS was more reliable than endoscopy. In terms of SET location, EUS was consistently reliable, whereas endoscopy showed variable reliability. When measuring SET size by endoscopy, additional size measurements with EUS should be considered in certain locations.