Systematic review and meta-analysis: the efficacy and safety of radiofrequency ablation for early superficial esophageal squamous cell neoplasia

Abstract

Abstract Background and Aim Esophageal squamous cell neoplasia (ESCN) is predominant in Asia. Endoscopic mucosal resection and endoscopic submucosal dissection (ESD) have both been recommended worldwide, however the application of endoscopic radiofrequency ablation (RFA) for treatment of early superficial ESCN remains inconclusive. We conducted a meta-analysis to study the effectiveness of RFA for early superficial ESCN. Methods Three major bibliographic databases were reviewed for the enrollment of case series and cohort trials prior to August 23, 2023. We included adults diagnosed with early superficial ESCN who had been receiving endoscopic RFA or ESD if the treatments were available. Our focus was on the 12-month histological complete response rate (CR) and 3-month histological CR, as well as the acute and late postoperative adverse events (AEs) rate during the at follow-up periods. Results Nine studies were enrolled for qualitative synthesis of narrative review, with eight trials involving a total of 447 participants for analysis. The pooled 12-month and 3-month histological CR were 0.83 (95% CI, 0.59–0.94, I2 = 80%) and 0.74 (95% CI, 0.67–0.80, I2 = 0%), respectively. As for safety, the acute and late postoperative AEs were 0.11 (95% CI, 0.05–0.26, I2 = 68%) and 0.19 (95% CI, 0.14–0.26, I2 = 0%), respectively. In subgroup analysis, the incidence of bleeding, laceration and perforation after endoscopic RFA showed 0.06, 0.06 and 0.02, respectively. When compared with ESD, RFA showed lower acute AEs and late AEs without any obvious significance. Conclusions For early superficial ESCN, endoscopic RFA achieved both higher 12-month complete remission and late complication postoperatively when compared to 3-month histological CR and acute AEs separately, while the stricture was encountered most commonly. The choice between endoscopic RFA and ESD remains inconclusive.