Review Article: Advances in the Diagnosis of Dysplasia in Barrett's Oesophagus

BACKGROUND: Barrett's oesophagus (BE), the precursor to oesophageal adenocarcinoma, progresses through a stepwise dysplastic sequence. Accurate dysplasia detection is critical, as endoscopic eradication therapy is highly effective. The current standard of care is limited by poor real-world adherence to the Seattle protocol, sampling error, and significant interobserver variability in pathologic interpretation. Post-endoscopy oesophageal neoplasia rates approaching 26% underscore the need for improved detection of dysplasia. AIMS: This review evaluates emerging and established technologies for dysplasia detection, procedural efficiency, and risk stratification in Barrett's oesophagus surveillance. METHODS: A narrative review of the published literature was performed, encompassing endoscopic imaging modalities, non-endoscopic sampling devices, artificial intelligence applications, and tissue and molecular biomarkers. RESULTS: Virtual chromoendoscopy increases dysplasia detection over white-light endoscopy and is recommended in current guidelines. Non-endoscopic capsule and balloon-based devices coupled with biomarker testing are currently used for BE screening, though evidence demonstrates strong performance for risk stratification of patients with known BE, which may expand their use into surveillance. Data suggest that wide-area transepithelial sampling with computer-assisted analysis increases adjunctive dysplasia detection. Artificial intelligence systems show high sensitivity and specificity for endoscopic and histopathologic dysplasia detection. Multiple biomarkers have been developed for risk stratification of patients with known BE and appear to aid in estimating individualized progression risk. CONCLUSIONS: A broad array of technologies shows promise in increasing dysplasia yield in patients with BE under surveillance. Most candidate solutions have not yet achieved sufficient validation to replace current practice. A multimodal, biomarker-driven, risk-stratified approach may substantially reduce missed neoplasia and improve survival.