The accuracy of probe-based confocal endomicroscopy versus conventional endoscopic biopsies for the diagnosis of superficial gastric neoplasia (with videos)

Background: Probe-based confocal laser endomicroscopy (pCLE) allows real-time in vivo histologic evaluation of GI mucosal lesions. Although pCLE has been used for various GI disorders, the significance of pCLE for gastric lesions is largely unknown. Objective: We compared the accuracy of conventional endoscopic forceps biopsy and pCLE for the diagnosis of superficial gastric neoplasia before endoscopic resection. Design: Prospective comparative study. Setting: Single tertiary referral center. Patients: This study involved 54 superficial gastric neoplasias in 46 patients. Main Outcome Measurement: Accuracy of in vivo and offline pCLE diagnosis and interobserver agreement. Methods: pCLE was performed before endoscopic resection of superficial gastric neoplasias previously diagnosed by endoscopic biopsy. The overall accuracy of endoscopic, in vivo pCLE, and offline pCLE diagnosis was compared with postendoscopic resection histopathology. Results: Endoscopic resection was performed on 54 lesions. On final histopathology, there were 3 non-neoplastic lesions, 19 gastric dysplasias, 22 differentiated adenocarcinomas, and 10 undifferentiated adenocarcinomas. The overall agreement with the final histopathology was substantial for conventional biopsies (kappa = 0.617) and excellent for in vivo pCLE (kappa = 0.824) (P<.001). The overall accuracy for the diagnosis of adenocarcinoma was 91.7% for pCLE and 85.2% for conventional biopsies (P=.065). The combined accuracy of conventional endoscopic biopsies and pCLE was 98.1%. The interobserver agreement for offline pCLE diagnosis was excellent (kappa = 0.931). Limitations: Single-center study, small sample size. Conclusion: Our study showed that pCLE can provide an accurate diagnosis for superficial gastric neoplasia. pCLE has the potential to compensate for the inherent limitations of a conventional endoscopic biopsy.