A computer-assisted algorithm for narrow-band imaging-based tissue characterization in Barrett’s esophagus

Abstract

Background and Aims: The endoscopic evaluation of narrow-band imaging (NBI) zoom imagery in Barrett's esophagus (BE) is associated with suboptimal diagnostic accuracy and poor interobserver agreement. Computeraided diagnosis (CAD) systems may assist endoscopists in the characterization of Barrett's mucosa. Our aim was to demonstrate the feasibility of a deep-learning CAD system for tissue characterization of NBI zoom imagery in BE.Methods: The CAD system was first trained using 494,364 endoscopic images of general endoscopic imagery. Next, 690 neoplastic BE and 557 nondysplastic BE (NDBE) white-light endoscopy overview images were used for refinement training. Subsequently, a third dataset of 112 neoplastic and 71 NDBE NBI zoom images with histologic correlation was used for training and internal validation. Finally, the CAD system was further trained and validated with a fourth, histologically confirmed dataset of 59 neoplastic and 98 NDBE NBI zoom videos. Performance was evaluated using fourfold cross-validation. The primary outcome was the diagnostic performance of the CAD system for classification of neoplasia in NBI zoom videos.Results: The CAD system demonstrated accuracy, sensitivity, and specificity for detection of BE neoplasia using NBI zoom images of 84%, 88%, and 78%, respectively. In total, 30,021 individual video frames were analyzed by the CAD system. Accuracy, sensitivity, and specificity of the video-based CAD system were 83% (95% confidence interval [CI], 78%-89%), 85% (95% CI, 76%-94%), and 83% (95% CI, 76%-90%), respectively. The mean assessment speed was 38 frames per second. Conclusion:We have demonstrated promising diagnostic accuracy of predicting the presence/absence of Barrett's neoplasia on histologically confirmed unaltered NBI zoom videos with fast corresponding assessment time. (Gastrointest Endosc 2021;93:89-98.

Deep learning for diagnosis of precancerous lesions in upper gastrointestinal endoscopy: A review

“…The enhanced system obtained higher accuracy than non-expert endoscopists and with comparable delineation performance[ 33 ]. Furthermore, their team also demonstrated the feasibility of a DL-based system for tissue characterization of NBI endoscopy in BE, and the system achieved a promising diagnostic accuracy[ 34 ].…”

Narrowed-spectrum Technologies in Endoscopic Imaging of The Upper Gastrointestinal Tract

“…For this reason, interest in the field is shifting to the development of computer aided diagnosis algorithms, with some studies showing improved performance compared to expert users. 105,106 FIGURE 1 -Narrow band imaging with near-focus of (A) normal squamous epithelium with branching vessels in the submucosa and intrapapillary capillary loops (IPCLs) that rise to the surface. (B) Esophageal squamous cell carcinoma with IPCLs with aberrant configuration and increased vessel caliber (C) Non-dysplastic Barrett's esophagus (BE) with a regular mucosal pattern and vascular pattern.…”

Artificial intelligence in endoscopy: Present and future perspectives

“…The model achieved sensitivities exceeding 90% and specificities exceeding 80% in both datasets. Struyvenberg et al 26 . developed CADx for BN, applicable to unmagnified WLI and magnified NBI and proved the comparably high diagnostic performances in the datasets of still images and movie clips in per‐image and per‐frame analyses.…”

“…The model achieved sensitivities exceeding 90% and specificities exceeding 80% in both datasets. Struyvenberg et al26 developed CADx for BN, applicable to unmagnified WLI and magnified NBI and proved the comparably high diagnostic performances in the datasets of still images and movie clips in per-image and perframe analyses. Hirasawa 10 first reported CADx for the histological prediction of gastric cancers with a high sensitivity of 92.2% for per-image analysis in a large volume of samples, although the PPV was restricted to 30.6%.…”