Novel Combined Modality Therapy for Barrett's Esophagus Containing High-Grade Dysplasia: Endoscopic Mucosal Resection Followed By Circumferential and Focal Ablation Using the HALO System

Abstract

Background: Endoscopic mucosal resection (EMR) allows for removal of high-grade dysplasia (HGD) and early mucosal cancer (EMC) in Barrett's esophagus (BE) with histological verification of resection margins. Focal EMR, however, is associated with recurrent lesions in non-resected areas, while radical EMR is associated with a significant incidence of stricture formation. Ablative therapy with PDT or APC may allow for treatment of the remaining BE after EMR, but is associated with residual BE and dysplasia, “buried Barrett's”, and stricture formation. A newer ablative therapy (HALO System) may be successful in eliminating residual BE with dysplasia after focal EMR. Aims: Assess the safety and efficacy of focal EMR followed by ablation using the HALO System in patients with BE containing HGD or EMC. Methods: To be eligible, patients had BE <10 cm, visible abnormalities with HGD or EMC, and no signs of submucosal infiltration or lymphnode metastases on endoscopy or EUS. EMR was performed with cap technique or multi-band mucosectomy (MBM) device. Circumferential ablation (CA) was performed with the balloon-based HALO360 System (BÂRRX Medical, Sunnyvale, CA, USA) and secondary focal ablation (FA) with the endoscope-mounted HALO90 System. Six weeks after EMR, CA was performed, followed every 2 months by CA or FA sessions until endoscopically clear of BE. Two months after the last ablation, EGD with lugol's staining and large cup biopsy (4Q/1 cm) was performed. Histopathology was reviewed by a single pathologist. Results: 13 pts (9 men, mean age 59 ± 10 yr, median BE length 7(5-10)cm) underwent EMR (n = 7 cap, n = 6 MBM). There were 7 en-bloc and 6 piecemeal resections (median 2.5/pt). One patient had acute bleeding after EMR, treated with a hemoclip. Resection specimens: EMC(n = 4), HGD(n = 6), LGD(n = 3); all with clear vertical margin. Remaining BE after EMR: HGD (n = 11), LGD (n = 2). Complete histological and endoscopic elimination of dysplasia and IM occurred in 12/13 patients (92%) after 1.5 CA and 2.5 FA sessions (mean/patient). None of the 284 biopsies obtained from neosquamous mucosa contained subsquamous BE. One patient developed a stenosis at the site of the EMR after a procedure in which CA was combined with a 2nd EMR for a new focal lesion. The stenosis resolved with 1 dilation. Conclusions: Patients with BE containing HGD or EMC can be effectively treated with this combined modality therapy of focal EMR followed by ablation of the remaining BE using the HALO System. A complete response rate for IM and dysplasia of 92% was achieved and there was no ablation-related stenosis. These results compare favorably with other regimens, such as radical EMR, PDT, or APC.