Effects of study design and different criterion standards on diagnostic accuracy of ultrathin endoscopy.

摘要

We read the article by Toyoizumi et al,1 who prospectively compared the diagnostic accuracy of ultrathin endoscopy (UTE) and high-resolution endoscopy (HRE) for superficial gastric neoplasia. In contrast, our group is currently comparing the diagnostic accuracies between UTE and standard-resolution endoscopy (SRE) with diagnostic criterion standards obtained from all patients by HRE at least 1 week in advance by using the revised Vienna classification.2 However, in their study, the diagnosis for superficial gastric neoplasia was from the affiliated hospitals or clinics without a reference standard. The "back-to-back" (or the crossover) study design, made tissue biopsy only possible by the second endoscopist, which obviously created procedure bias, increased false negative results (type II error), and thus decreased the sensitivities of both UTE and HRE (Fig. 1A).