Abstract
The aim of this study was to perform a detailed clinical and pathologic analysis of biopsies from 180 BE patients with the goal of understanding the clinical significance and outcome of patients diagnosed with crypt dysplasia (CD). Biopsies from 58 progressors and 122 nonprogressors (372 biopsies) were graded by 3 gastrointestinal pathologists. A consensus biopsy diagnosis (agreement by ≥2 pathologists) was used for outcome analysis. The overall diagnostic agreement among all 3 observers was high (0.75) and was observed in 83.9% cases, with a moderate level of agreement (0.44) noted for CD. BE progressors had a significantly higher proportion of index biopsies with CD (17% vs. 2%) and LGD (9% vs. 0%; overall P <0.0001) compared with nonprogressors. Compared with an index consensus biopsy diagnosis of NDBE, a consensus biopsy diagnosis of IND, CD, and LGD had significantly increased odds of developing HGD/EAC by univariable (OR: 4.05; P <0.0001) as well as multivariable analysis (OR: 10.3; P =0.01). Furthermore, Kaplan-Meier analysis showed that patients diagnosed with CD or LGD on an index biopsy had a higher probability of developing HGD/EAC than those with NDBE ( P <0.0001), and the timeline for progression among patients with CD was found to be intermediate between those diagnosed with NDBE and LGD. BE patients with an index biopsy diagnosis of CD are more likely to develop HGD/EAC compared with those without dysplasia, and these patients may benefit from surveillance and management strategies, similar to LGD.