BACKGROUND/AIMS: This study aimed to evaluate the diagnostic accuracy of dual-focus narrow-band imaging (dNBI) and Lugol'schromoendoscopy (LCE) combined with probe-based confocal laser endomicroscopy (pCLE) to screen for esophageal squamous cell neoplasms (ESCNs) in patients with a history of head and neck cancer. METHODS: From March to August 2016, dNBI was performed. Next, LCE was performed, followed by pCLE and biopsy. Histology has historically been the gold standard to diagnose ESCN. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of dNBI and LCE adjunct with pCLE were determined. RESULTS: Twenty-four patients were included. Ten ESCNs were found in 8 patients (33%). Forty percent of high-graded intraepithelial neoplasias and all low-grade intraepithelial neoplasias were overlooked by dNBI. The sensitivity, specificity, PPV, NPV, and accuracy of dNBI vs. LCE combined with pCLE were 50% vs. 80%, 62% vs. 67%, 36% vs. 44%, 75% vs. 91%, and 83% vs. 70%, respectively. CONCLUSIONS: The use of dNBI to detect ESCN was suboptimal. LCE with pCLE following dNBI had additional value for detecting esophageal dysplasia not detected by dNBI. The use of pCLE to detect dNBI-missed lesions yielded a high NPV, while pCLE-guided biopsy could reduce the number of unnecessary biopsies.
Biopsy ; Epithelial Cells ; Head and Neck Neoplasms ; Humans ; Neoplasms, Squamous Cell ; Sensitivity and Specificity

Background/Aims: Previous artificial intelligence (AI) models attempting to segment gastric intestinal metaplasia (GIM) areas have failed to be deployed in real-time endoscopy due to their slow inference speeds. Here, we propose a new GIM segmentation AI model with inference speeds faster than 25 frames per second that maintains a high level of accuracy. Methods: Investigators from Chulalongkorn University obtained 802 histological-proven GIM images for AI model training. Four strategies were proposed to improve the model accuracy. First, transfer learning was employed to the public colon datasets. Second, an image preprocessing technique contrast-limited adaptive histogram equalization was employed to produce clearer GIM areas. Third, data augmentation was applied for a more robust model. Lastly, the bilateral segmentation network model was applied to segment GIM areas in real time. The results were analyzed using different validity values. Results: From the internal test, our AI model achieved an inference speed of 31.53 frames per second. GIM detection showed sensitivity, specificity, positive predictive, negative predictive, accuracy, and mean intersection over union in GIM segmentation values of 93%, 80%, 82%, 92%, 87%, and 57%, respectively. Conclusions The bilateral segmentation network combined with transfer learning, contrast-limited adaptive histogram equalization, and data augmentation can provide high sensitivity and good accuracy for GIM detection and segmentation.