Incidence of macroscopically occult neoplasias in Barrett's esophagus: are random biopsies dispensable in the era of advanced endoscopic imaging?

Abstract

The gold standard for endoscopic surveillance of Barrett's esophagus (BE) includes targeted biopsies (TBs) from abnormalities as well as stepwise four-quadrant biopsies (4QBs) for detection of invisible high-grade intraepithelial neoplasias (HGINs) or early carcinomas (ECs). In a large mixed BE population, we investigated the rate of HGINs/ECs that are macroscopically occult to enhanced visualization with high-resolution endoscopy plus acetic acid chromoendoscopy. From January 2007 to December 2009, 701 consecutive BE patients were enrolled in a prospective study at a tertiary referral center. Of these, 406 patients had a history of HGIN/EC (high-risk group) and 295 patients did not (low-risk group). In 701 patients, 459 TBs and 5,485 4QBs were obtained. Altogether, 92 patients with 132 lesions containing HGINs/ECs were detected. For the diagnosis of HGINs/ECs, patient-related sensitivity and specificity rates of endoscopic imaging with TBs were 96.7 and 66.5%, with a positive and negative predictive value of 30.4 and 99.3%, respectively. In the high-risk group, 4QBs identified three additional patients (3.3%) with macroscopically occult HGINs/ECs. In the low-risk group, no HGINs/ECs were identified with either biopsy approach. Advanced endoscopic imaging identifies the vast majority of BE patients with early neoplasias, and the additive effect of 4QB is minimal. Therefore, in low- and high-risk patients, limiting endoscopic surveillance to guided biopsies is justified in specialized high-volume centers with permanent quality control. However, we do not advocate abandoning 4QB outside this setting.