Allaying uncertainty in diagnosing buried Barrett's esophagus

Subsquamous intestinal metaplasia (SSIM) in the setting of Barrett's esophagus (BE) is a technically challenging diagnosis. While the risk for progression of BE involving the surface mucosa is well documented, the potential risk for development of advanced neoplasia associated with SSIM has been controversial. This study aimed to determine the effects of specimen adequacy, presence of dysplasia, and interobserver agreement for SSIM interpretation. Adult patients (n = 28) who underwent endoscopic therapy for BE with high-grade dysplasia or intramucosal carcinoma (HGD/IMC) between October 2005 and June 2013 were included. Initial evaluation (n = 140 slides) by an experienced gastrointestinal pathologist was followed by an interobserver study by 8 pathologists. Forty-seven (34%) slides had insufficient subsquamous tissue to assess for SSIM. SSIM was found in 19% of all slides and 29% of slides with sufficient subsquamous tissue. At least one slide had SSIM in 54% to 64% of patients. Subsquamous low grade dysplasia (LGD) was found in 4 (15%) slides with SSIM and subsquamous HGD/IMC was found in 5 (19%) slides with SSIM. At the patient level, 8 (53%) had no dysplasia, 4 (27%) had LGD and 3 (20%) had HGD/IMC. Overall agreement for SSIM by slide was 92% to 94% (κ = 0.73 to κ = 0.82, moderate to strong agreement), and by patient was 82% to 94% (κ = 0.65 to κ = 0.87, moderate to strong agreement). This study confirms the need for assessing specimen adequacy and assessing the prevalence of SSIM and is the first to assess interobserver agreement for SSIM and dysplasia within SSIM.