Diagnostic Inaccuracies of Barrett's Oesophagus on Gastroscopy: Are We Performing Unnecessary Surveillance?

Background It is common for patients to enter Barrett's oesophagus (BO) surveillance based on endoscopic appearances before the diagnosis is histologically confirmed. We set out to review this practice by establishing the accuracy of endoscopic diagnoses of BO. Methods All gastroscopy reports in which a diagnosis of BO was recorded were reviewed over one year. These were compared to the histopathological reports to assess diagnostic accuracy. Results BO was diagnosed in 84 procedures. This diagnosis was incorrect according to histology in 42.9% (n=36) of cases. Diagnostic accuracy was higher with gastroenterologists (38.8% incorrect, n=21) compared to surgeons (50% incorrect, n=15). Diagnostic accuracy was higher with consultants (34.9% incorrect, n=22) compared to registrars (66.7% incorrect, n=14). The dose of sedation used had no impact on accuracy. Unnecessary surveillance was booked in 36.1% (n=13) of cases. Conclusion It is insufficient to rely on endoscopic appearances alone to diagnose BO, irrespective of speciality or experience. The diagnosis should only be made after reviewing the histopathology report. This can eliminate unnecessary repeat endoscopy procedures, sparing patients from unjustifiable risk and helping to cut down on long waiting lists in endoscopy departments. The implementation of the Prague classification and Seattle protocol can improve diagnostic accuracy.