Risk for local recurrence of early gastric cancer treated with piecemeal endoscopic mucosal resection during a 10-year follow-up period.

Background: Endoscopic mucosal resection (EMR) is a therapy for early gastric cancer (EGC) that can be provided relatively easily and safely in any institution. Identification of the resection margin is a problem in EMR, especially in cases of piecemeal EMR. Despite the long-standing widespread use of piecemeal EMR for EGC, its limitation and long-term outcomes in clinical practice have not been fully evaluated. This study aimed to determine the risk factors of piecemeal EMR, the local recurrence rates, and the mortality rate.

Methods: A cross-sectional, retrospective cohort study was performed to investigate the risks of piecemeal EMR for patients with the diagnosis of differentiated adenocarcinoma localized to the mucosa. Local recurrence of EGC was investigated by annual follow-up esophagogastroduodenoscopy (EGD) for 10 years. EMR was performed with snare electrocautery using a two-channel scope. When a resection margin was clearly positive for cancer, additional surgery was performed soon after the initial EMR.

Results: For the 149 EGC patients (mean age, 68.8 ± 9.8; male, 77%) who underwent EMR between 1995 and 2001, EMR was performed en bloc in 66 cases and piecemeal in 83 cases. The comorbid conditions existing in 34 of the 149 patients included other malignancies (n = 12), heart failure (n = 5), pulmonary disease (n = 7), liver cirrhosis (n = 4), and other illness (n = 6). However, EMR was completed without complication. The mean area (length × width) of the lesions was 404 ± 289 mm(2) in the piecemeal group and 250 ± 138 mm(2) in the en bloc groups. The en bloc and piecemeal EMR groups differed significantly in terms of unclear horizontal margins but not in terms of unclear vertical margins. Multiple logistic regression suggested that the adjusted odds ratio for maximum diameters exceeding 20 mm for piecemeal EMR was 2.71 (95% confidence interval [CI], 1.30-5.64). According to Kaplan-Meier estimates, the local recurrence rate was 30% (95% CI, 20-40%) at both 5 and 10 years. No recurrence was observed in the en bloc group. The adjusted hazard ratio of unclear horizontal margins for local recurrence was 1.63 (95% CI, 1.12-2.36). A total of 24 patients died after EMR because of comorbid conditions, including other malignancies (n = 11), cardiovascular disease (n = 6), pulmonary disease (n = 4), liver cirrhosis (n = 2), and traffic accident (n = 1). However, no patient died of gastric cancer during the 10-year follow-up period.

Conclusions: An evaluation of horizontal margins in terms of local recurrence after piecemeal EMR is important, and en bloc resection is recommended. Close follow-up assessment is warranted, especially within 5 years in cases of unclear margin resection after piecemeal EMR. The use of EMR is safe even for patients with severe comorbid conditions.