Location and appearance of dysplastic Barrett’s esophagus recurrence after endoscopic eradication therapy: no additional yield from random biopsy sampling neosquamous mucosa

Abstract

Surveillance after complete remission of intestinal metaplasia (CRIM) is essential. Current recommendations are to sample visible lesions first, followed by random 4-quadrant biopsy sampling of the original Barrett's esophagus (BE) length. To inform post-CRIM surveillance protocols, we aimed to identify the anatomic location, appearance, and histology of BE recurrences. We performed an analysis of 216 patients who achieved CRIM after endoscopic eradication therapy for dysplastic BE at a Barrett's Referral Unit between 2008 and 2021. The anatomic location, recurrence histology, and endoscopic appearance of dysplastic recurrences were evaluated. After a median of 5.5 years (interquartile range, 2.9-7.2) of follow-up after CRIM, 57 patients (26.4%) developed nondysplastic BE (NDBE) recurrence and 18 patients (8.3%) developed dysplastic recurrence. From 8158 routine surveillance biopsy samplings of normal-appearing tubular esophageal neosquamous epithelium, the yield for recurrent NDBE or dysplasia was 0%. One hundred percent of dysplastic tubular esophageal recurrences were visible and in BE islands, whereas 77.8% of gastroesophageal junction dysplastic recurrences were nonvisible. Four distinct endoscopic features suspicious for recurrent advanced dysplasia or neoplasia were identified: buried or subsquamous BE, irregular mucosal pattern, loss of vascular pattern, and nodularity or depression. The yield of routine surveillance biopsy sampling of normal-appearing tubular esophageal neosquamous epithelium was zero. BE islands with indistinct mucosal or loss of vascular pattern, nodularity or depression, and/or signs of buried BE should raise clinician suspicion for advanced dysplasia or neoplasia recurrence. We suggest a new surveillance biopsy sampling protocol with a focus on meticulous inspection, followed by targeted biopsy sampling of visible lesions and random 4-quadrant biopsy sampling of the gastroesophageal junction.