Confocal endomicroscopy (CLE) allows real-time in-vivo high-resolution and high-magnification imaging of the gastrointestinal epithelium, which is comparable to histopathology. Previous studies have investigated the accuracy of pCLE for diagnosis and differentiated of colorectal polyps, Barrett's esophagus and pancreaticobiliary strictures. However, to date there are limited data exploring the application of pCLE to gastric lesions, and this is the first study comparing the diagnosis of conventional forceps biopsy with that of pCLE using the final specimens obtained from endoscopic resection as a reference standard. The aims of this study were (1) the accuracy of pCLE compared to conventional forceps biopsy using histopathology results following endoscopic resection as a reference, and (2) comparison of "real time" in-vivo pCLE diagnosis with that of blinded "off-line" pCLE diagnosis, and off-line interobserver agreement.

Inclusion Criteria:

• Patients with biospy proven superficial gastric neoplasia, suitable for endoscopic resection.

Exclusion Criteria:

• Patients with uncorrectable coagulopathy,
• liver cirrhosis,
• acute gastrointestinal bleeding,
• pregnancy,
• breast feeding,
• documented allergy to fluorescein,
• patients with lesions that were deemed unsuitable for endoscopic resection, and
• patients without documented conventional biopsy results.

Consecutive patients referred to a tertiary medical center for endoscopic resection of gastric superficial lesions.