Radiofrequency Ablation of Barrett's Esophagus Containing High-Grade Dysplasia

Abstract

Listen

Translate article

Background: Circumferential balloon-based ablation of non-dysplastic Barrett's esophagus (BE) has been proven by others as safe and effective. This study assessed the efficacy and safety of ablation for BE with high-grade dysplasia (HGD) in patients with and without prior endoscopic resection (ER). Methods: Eligible patients demonstrated BE with HGD on at least 2 prior EGDs. Visible abnormalities were resected with ER prior to ablation. Persistence of dysplasia was confirmed with biopsy after ER. Patients received esomeprazole 40 mg BID during study. A balloon-based electrode (HALO360 System) was used for primary circumferential ablation (CA) and an endoscope-mounted electrode (HALO90 System) for secondary focal ablation (FA) of residual BE. Both systems (BÂRRX Medical, Sunnyvale, CA) use an electrode array that delivers a short burst of high power RF energy (40 W/cm2) at a preset energy density (12 J/cm2). After primary CA, EGD was performed at 2 mo intervals with secondary ablation of residual BE using CA or FA, depending on extent of BE. Two mos after the last ablation, EGD with Lugol's and large cup biopsy (4Q/q 1 cm) was performed. Histopathology was reviewed by a single pathologist. Primary endpoint: complete response dysplasia (CR-D), absence of dysplasia in all biopsies. Secondary endpoints: adverse events (AE); visible BE regression; complete response intestinal metaplasia (CR-IM), absence of IM in all biopsies. Results: Twenty-three pts (17 men, median age 66 yrs, IQR 55-78) were treated (median BE length 7 cm, IQR 4-10). ER was performed in 13 patients: mucosal carcinoma (n = 4), HGD (n = 6) and LGD (n = 3). Worst pathological grade of BE after ER and prior to RFA was LGD (n = 3) and HGD (n = 20). Patients underwent a mean of 1.5 CA and 2.6 FA sessions. CR-D was achieved in 22/23 patients (96%) and CR-IM in 21/23 patients (92%). Patients with residual BE (n = 2) have only small islands remaining (median BE regression: 99%). There were 3 AE's. Fever/chest pain (n = 2) after CA, resolved with narcotics. An ER-related stenosis (n = 1), resolved after 1 dilation. There were no thermally-mediated strictures. After a median additional follow-up of 6 mos and 2.1 endoscopies, no patient with CR-D has shown recurrence of dysplasia and no patient with CR-IM has shown recurrence of IM. None of the 521 biopsies of neosquamous mucosa contained subsquamous BE (“buried Barrett's”). Conclusions: Radiofrequency ablation of BE containing HGD is a safe and effective treatment, with a CR-D and CR-IM of 96% and 92%, respectively. Circumferential and focal ablation using the HALO devices can be safety performed after prior ER for endoscopically visible abnormalities.