Introduction Endoscopic therapy is safe and effective for Barrett’s Oesophagus (BE) patients with high grade dysplasia (HGD) or early (T1A) adenocarcinoma (EAC). Endoscopic resection (ER) removes visible neoplasia but metachronous lesions occur in up to 30%.1 RCTs of RFA2 & APC3 show reduced recurrence but the 2 techniques have not been compared. We aim to test the feasibility of recruiting & retaining up to 100 BE patients with HGD or T1A EAC to RFA or APC after ER.
Methods Patients with BE & HGD or T1A EAC confirmed by ER (single tongues ≤2 cm excluded) were recruited over 1 year in 6 centres, stratified into 3 groups as per BE length (<5cm, 5–10 cm, >10 cm) & randomised to 4 interventions with RFA or APC every 2 months. All received high dose twice-daily PPIs. ER was allowed for further visible lesions. Exit endoscopy with biopsies was at 12 months. Persistence of BE, recruitment, retention & adverse events(AEs) were recorded. Quality of life (QL) was assessed with EQ-5D, EORTC QLQ-C30 & OES18 at 0, 6 & 12 months.
Results 171 patients were screened. 41.5% were excluded & 14% declined participation. 76 patients (84.2% males, mean age 69.7) were randomised to receive RFA (36) or APC (40) and stratified according to BE length (<5 cm (27); 5–10 cm(44); >10 cm(4)). 13 withdrew, 4 for more advanced cancer and 9 for other reasons. Persistence of HGD/T1 EAC at 12 months (including 3/4 withdrawn patients) was RFA: 15.6% & APC: 12.9% (OR 1.25; 95%CI 0.3–5.17). Endoscopy duration was longer for RFA especially at 2 months (Median, mean (SD) mins= RFA: 30.0, 38.0 (25.7) and APC 23.0, 27.2 (15.8)). Residual BE at 12 months occurred in 38.7% (RFA) & 31% (APC). Median, mean (SD) residual BE were: RFA 0, 1.3 (2.6) cm & APC 1.0, 1.7 (2.6) cm, slightly favouring RFA for BE >5 cm. AEs for APC: 2 dysphagia/strictures, 5 bleeds & for RFA: 1 perforation due to ER, 2 dysphagia/strictures, 1 syncope, 4 bleeds. There were no difference in QL scores. Exit histology showed EAC (1 RFA; 0 APC), buried glands (1 RFA;3 APC), HGD (3 RFA, 2 APC). Sample size for a non-inferiority trial comparing APC to RFA (with a non-inferiority margin of 6%) would be 461 subjects/arm.
Conclusion Patients with BE are willing to enrol & remain in studies comparing RFA & APC over 12 months. This feasibility study suggests no difference in outcome between APC & RFA. A fully powered RCT requires a large sample size to show non-inferiority of APC compared to RFA. (NIHR RfPB Grant No PB-PG-0711-25066)
References 1 Pech et al. Gut 2008;57:1200–6.
2 Shaheen et al. NEJM 2009;360:2277–88.
3 Manner et al. Endoscopy 2014;46:6–12.
Disclosure of Interest M. F. Peerally: None Declared, H. Barr: None Declared, L. Lovat: None Declared, P. Bhandari: None Declared, K. Ragunath Conflict with: Speaker Honoraria, Educational grants, Consultancy, Research grants and support: Olympus, Pentax, Cook, Boston Scientific, ERBE, Medtronics/Covidien, Astra Zeneca, Ferring, H. Smart: None Declared, R. Harrison: None Declared, C. Stokes: None Declared, J. Decaestecker: None Declared