Endoscopic Mucosal Resection for High-Grade Dysplasia and Intramucosal Carcinoma Occurring in Barrett's Esophagus

Abstract

Introduction: Patients with Barrett's esophagus (BE) are at increased risk of developing esophageal adenocarcinoma (AC). High-grade dysplasia (HGD) is considered the final step toward AC, and esophagectomy is advised. Endoscopic mucosal resection (EMR) could become an alternative to surgery. The limited available data are encouraging. Aim: To evaluate the effectiveness and safety of selective EMR in patients with HGD and/or intramucosal cancer (IMC) occurring in BE patients. Methods: Between October 1998 and July 2003, 38 consecutive patients (mean age 62.5 ± 1.5 years) with HGD (34) and/or IMC (4) in BE underwent EMR. All patients had endoscopically detectable mucosal abnormalities. Long segment BE (>30 mm, LSBE) was present in 18 patients. In 3 patients with short segment BE (SSBE: 15%), HGD was detected in a normal appearing BE. The latter patients were included with the aim of completely removing BE. EUS was performed to assess the depth of the lesion and the status of mediastinal lymph nodes. The median diameter of lesions was 13.5 mm (range 3-40). EMR was carried out by using the cap method (EMRC), after submucosal injection of a diluted epinephrine solution (1:60.000). During the EMR, patients were sedated with propofol. Follow-up was scheduled at 3, 6 months, and then every 6 months thereafter. Results: The average size of EMR was 20.0 ± 9.5 × 14.0 ± 6.4 mm. The results of the histopathologic assessment post-EMR were: 5 LGD (13.2%), 26 HGD (68.4%), 2 IMC (5.3%), and 5 AC with submucosal infiltration (13.2%). EMR changed the pre-treatment diagnosis in 10 patients (26.3%). Three patients with invasive AC underwent surgery, but the histopathologic assessment of the surgical specimen did not show residual disease and lymph node involvement. The remaining 2 patients with AC are cancer free at 8 and 10 months. Among the 26 HGD, one recurrence occurred after 19 months, and a new EMR was performed. Three months later, a further control was negative. Post-EMR bleeding occurred in 4 patients (10.5%), and hemoclips were placed in 2 of them. After a median follow-up of 14.2 months (range 3-55.8), all patients remained in remission. Conclusions: EMRC is effective and safe to treat HGD and/or IMC within BE. Moreover, it is a valuable staging method. EMRC can completely remove BE in patients with SSBE. The complication rate is negligible even in large resections. Careful surveillance is recommended for early detection of residual and metachronous lesions.