Objectives: Complete Barrett's eradication endoscopic mucosal resection (CBE-EMR) is the endoscopic removal of all Barrett's epithelium with the curative intent of eliminating high-grade dysplasia (HGD)/intramucosal carcinoma (IMC) and reducing the risk of metachronous lesion development. We report our single tertiary referral center's long-term clinical experience using this modality in HGD/IMC management.
Methods: In this study, we retrospectively reviewed all patients who had CBE-EMR for Barrett's esophagus (BE) with HGD/IMC who had been entered into our center's prospectively collected database. High-definition white-light and narrow-band imaging examinations were used according to the protocol. Staging endoscopic ultrasound was done before CBE-EMR to exclude invasive disease or suspicious lymphadenopathy. High-dose proton pump inhibition was instituted after initial treatment, and Seattle-type surveillance biopsies were performed on follow-up every 6 months once the CBE-EMR procedure was completed.
Results: A total of 49 patients (mean age 67 years, median 65, s.d. 11; 75% men) with histologically confirmed BE and HGD (33), IMC (16), underwent CBE-EMR from August 2003 to August 2008. The mean BE segment length was 3.2 cm (median 2, s.d. 2.2); 26 patients had short-segment BE, and 30 had visible lesions. A total of 106 EMR procedures were performed. On initial EMR, two patients had superficial submucosal carcinoma invasion (sm1) and two had IMC with lymphatic channel invasion. All four patients were referred for esophagectomy, but one opted for continued endoscopic management, without evidence of residual or recurrent carcinoma. A total of 14 patients await completion of EMR (9) or first follow-up endoscopy (5). CBE-EMR therapy was completed in 32 patients by an average of 2.1 sessions (median 2, s.d. 0.9). Surveillance biopsies showed normal squamous epithelium in 31 of 32 (96.9%) patients (mean remission time 22.9 months, median 17, s.d. 16.7, interquartile range 11-38). In all, 10 of 46 patients who continued in the endoscopic protocol had subsquamous Barrett's epithelium on EMR specimens and/or treatment endoscopy biopsies. Overall, 1 of these 10 patients had Barrett's underneath squamous mucosa on most recent surveillance biopsies. CBE-EMR upstaged pre-EMR pathology results in 7 of 49 (14%) of patients and downstaged pathology in 15 of 49 (31%) patients. In all, 18 of 49 (37%) patients developed symptomatic esophageal stenosis after a mean of 24.4 days (median 13.5, s.d. 27.8); all were successfully managed by endoscopic treatment. No perforations or uncontrollable bleeding occurred.
Conclusions: To our knowledge, this is the largest American single-center experience demonstrating that CBE-EMR with close endoscopic surveillance is an effective treatment modality for BE with HGD/IMC. Although the rate of stenosis development is significant, it is easily treated by endoscopic dilation. Patients considering endoscopic ablation should be counseled appropriately. The role of CBE-EMR in patients with lymphatic invasion or superficial submucosal invasion remains to be defined.