Endoscopic prediction of recurrence in patients with early gastric cancer after margin-negative endoscopic resection

Although follow-up endoscopy is routinely performed after endoscopic resection (ER) for early gastric cancer (EGC), it remains unclear whether resection scar biopsies should also be taken. This study sought to predict local recurrence at the ER scar on the basis of endoscopic criteria after margin-negative EGC resection and to determine the necessity of taking scar biopsies.

All consecutive patients with EGC who underwent margin-negative ER in June 1995 to December 2011 and developed recurrence at the scar were identified. Each case was matched by four controls with EGC and margin-negative ER but without local recurrence for age, sex, and en-bloc resectability (en-bloc vs. piecemeal resection). Endoscopic data were reviewed by consensus of two endoscopists. Key endoscopic criteria were gross morphology (evenly elevated, unevenly elevated, and flat), hyperemic change, mucosal defect, and spontaneous bleeding.

Of 3037 cases, which underwent margin-negative ER, 22 developed local recurrence (mean age, 63.8 years; 72.9% male). En-bloc resection was achieved in 20 of the 22 (90.9%). Flat endoscopic morphology without hyperemic changes predicted non-recurrence at the scar with a sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 95.5%, 68.2%, 2.6%, 99.94%, and 73.6%, respectively. For patients who underwent en-bloc resection for differentiated EGC, these values were 100%, 71.4%, 3.0%, 100%, and 75.8%, respectively.

Routine follow-up biopsies may be unnecessary when follow-up endoscopy reveals flat mucosa without hyperemic changes at the scar, especially for en-bloc resected and differentiated EGCs.