Background & aims: Patients with esophageal high-grade dysplasia or mucosal esophageal cancer can be successfully treated by endoscopy. We performed a systematic review of the literature to determine whether endoscopic ultrasound (EUS) correctly predicts the T-stage of early esophageal cancers, compared with pathology specimens obtained by using endoscopic mucosal resection (EMR) or surgery.
Methods: Standard systematic review methods were used to perform reference searches, determine eligibility, abstract data, and analyze data. When possible, individual patient-level data were abstracted, in addition to publication-level aggregate data.
Results: Twelve studies had sufficient information to abstract and review for quality; 8 had individual patient-level data (n = 132). Compared with surgical or EMR pathology staging, EUS had T-stage concordance of 65%, including all studies (n = 12), but only 56% concordance when limited to individual patient-level data. Factors such as initial biopsy pathology (high-grade dysplasia vs early-stage cancer) did not appear to affect the concordance of staging between EUS and EMR/surgical staging.
Conclusions: EUS is not sufficiently accurate in determining the T-stage of high-grade dysplasias or superficial adenocarcinomas; other means of staging, such as EMR, should be used.
Copyright © 2010 AGA Institute. Published by Elsevier Inc. All rights reserved.