Endoscopic submucosal dissection in patients with early esophageal squamous cell carcinoma: results from a prospective Western series

Background Although endoscopic submucosal dissection (ESD) is becoming accepted as an established treatment for superficial esophageal squamous cell neoplasia, the majority of data on this endoscopic modality has been provided by Japanese series. Objective To assess the efficacy and safety of ESD for esophageal squamous cell neoplasia in a consecutive series of patients treated in a Western setting. Design and Setting Single-center, prospective observational study. Patients and Intervention From January 2005 to July 2008, 20 patients with superficial esophageal squamous cell neoplasia were treated by ESD. Main Outcome Measurements Rates of en bloc resection, complete resection, and complications were evaluated as short-term outcomes. Overall survival, local or distant recurrence, and postoperative stricture rates were evaluated as long-term outcomes. Results ESD was performed in 20 patients (mean age 64 years, range 46-81 years; 16 men). The mean size of the lesion was 32 mm (range 15-60 mm); it was 30 mm or larger in 14 patients (70%). The mean time of ESD was 89 minutes (range 58-180 minutes). En bloc resection with resection-free margins was achieved in 18 patients (90%), whereas 2 patients presented with incomplete or indeterminate resection. Two cases (10%) of mediastinal emphysema without overt perforation and 1 case (5%) of post-ESD symptomatic stricture were reported. No local or distant post-ESD recurrence occurred in those with resection-free margins at a median follow-up of 18 months. Limitations Small number of patients and limited follow-up. Conclusion This Western series study confirms that ESD is a potentially curative treatment for superficial esophageal squamous cell neoplasia. Early and late complication rates were comparable to those of Japanese series. ESD should be probably considered as the treatment of choice in all large lesions amenable to endoscopic treatment.