Introduction: Societal guidelines require ≥1cm of esophageal intestinal metaplasia (IM) to diagnose Barrett’s esophagus (BE), and discourage endoscopists from biopsying Z-lines with <1cm of variability, primarily because IM in irregular Z-lines (IMIZL) is assumed to have negligible risk for cancer development. Nevertheless, endoscopists frequently biopsy irregular Z-lines, and it recently has been proposed that many (perhaps most) esophageal adenocarcinomas (EACs) develop from IMIZL. We aimed to compare the frequency of dysplasia (the precursor of EAC) in IMIZL versus BE found in endoscopies performed in clinical practice.
Methods: We searched the database of a large, community practice-owned pathology laboratory for all cases classified as Barrett’s esophagus (BE) between January 1, 2012 and December 31, 2018. We included only patients who had IM with goblet cells documented on histology, and who had medical records with adequate demographic information and documentation of length of columnar-lined esophagus (CLE). Patients were classified as SSBE (1-3cm CLE), LSBE (≥3cm CLE), IMIZL (<1cm CLE), or IM esophagitis (IM in biopsies of distal esophagus taken for “esophagitis” with location not specified). Medical records were reviewed up to March 2023 for subsequent endoscopies showing development of high-grade dysplasia (HGD) or low-grade dysplasia (LGD). Dysplasia was considered prevalent if found on first endoscopy and incident if found on subsequent endoscopies.
Results: 1,997 individual cases were classified as BE; 862 had documented IM with goblet cells and adequate medical records. Among those 862 study patients, 136 (15.8%) had LSBE, 496 (57.5%) had SSBE, 119 (13.8%) had IMIZL, and 111 (12.9%) had IM esophagitis (Table 1). Demographic data were similar for the 4 groups, but females were more frequent in the IMIZL (48.7%) and IM esophagitis (51.4%) groups than in the LSBE (20.5%) and SSBE (29.2%) groups. The mean length of follow up was 67.7 months. There were 38 total cases of HGD [15 (39.5%) prevalent, 23 (60.5%) incident] and 67 total cases of LGD [45 (67.2%) prevalent, 22 (32.6%) incident]. As expected, rates of prevalent and incident HGD and LGD were considerably higher in LSBE than in the other 3 groups. However, there were no significant differences between the IMIZL and SSBE groups in the rates of prevalent HGD (0.84% IMIZL vs 0.4% SSBE, p=.54), and in the rates of prevalent and incident LGD (prevalent 3.36% IMIZL vs 4.44% SSBE, p=.60; incident 2.52% IMIZL vs 2.62% SSBE, p=.95).
Conclusions: In a large, community-based practice, patients with IM in irregular Z-lines had rates of prevalent HGD, prevalent LGD and incident LGD similar to those of patients with SSBE. These data suggest that the risk of cancer for IMIZL may not be negligible, and support the notion that IMIZL may well be the source of many EACs.

Table 1: Study Patient Demographic Data and Frequency of Dysplasia

Aleeza Ali

Appearances