Endoscopic Indication of Endoscopic Submucosal Dissection for Early Gastric Cancer Is Not Compatible with Pathologic Criteria in Clinical Practice.

Abstract

The inappropriate selection of patients with early gastric cancer (EGC) for endoscopic submucosal dissection (ESD) may lead to additional surgery because of a non-curative resection. This study was performed to assess the accuracy of clinical decisions in ESD for EGC. A total of 607 cases of EGC treated by ESD were prospectively enrolled from January 2011 to June 2014 at a single academic hospital. The 607 EGCs were divided into three groups (overestimated, same-estimated, and underestimated) based on pre-procedure endoscopic findings (indication) and pathological diagnosis after ESD (criteria). We evaluated the discrepancy rates between pre-procedure indication and pathological criteria, and then analyzed the pre-procedure factors that could influence the occurrence of the discrepancies. The absolute, expanded, and beyond the expanded indication has its accuracy on curability criteria in 87%, 77.6%, and 55.6% of cases, respectively. The ratio of overall indication-criteria discrepancies was 250/607 (41.2%). The curability was significantly lower in the underestimated group compared to the overestimated and same-estimated groups (41.6% vs. 94.6%, 94.4%, p < 0.001). In multivariate analysis examining the predictive factors for discrepancies in the 598 EGCs with absolute/expanded indications, the endoscopic size ≥ 20mm [odds ratio (OR) 2.493, confidence interval (CI) 1.546-4.022, p < 0.001], presence of ulcers (OR 1.712, CI 1.070-2.738, p = 0.025), patient age < 60years (OR 1.689, CI 1.044-2.733, p = 0.033), and undifferentiated type EGC on forceps biopsy (OR 5.397, CI 2.027-14.369, p = 0.001) were all associated with discrepancies. Indication judged by pre-procedural endoscopy is not sufficiently accurate to be used as a good measurement for post-procedural criteria.