A genomic biomarker-based model for cancer risk stratification of non-dysplastic Barrett's esophagus patients after extended follow up; Results from Dutch surveillance cohorts.

S. J. M. Hoefnagel, N. Mostafavi, M. R. Timmer, C. T. Lau, S. L. Meijer, K. K. Wang, K. K. Krishnadath

Research output: Contribution to journalArticleAcademicpeer-review

9 Citations (Scopus)

Abstract

Barrett's esophagus is the only known mucosal precursor for the highly malignant esophageal adenocarcinoma. Malignant degeneration of non-dysplastic Barrett's esophagus occurs in < 0.6% per year in Dutch surveillance cohorts. Therefore, it has been proposed to increase the surveillance intervals from 3 to 5 years, potentially increasing development of advanced stage interval cancers. To prevent such cases robust biomarkers for more optimal stratification over longer follow up periods for non-dysplastic Barrett's patients are required. In this multi-center study, aberrations for chromosomes 7, 17, and structural abnormalities for c-MYC, CDKN2A, TP53, Her-2/neu and 20q assessed by DNA fluorescence in situ hybridization on brush cytology specimens, were used to determine marker scores and to perform clonal diversity measurements, as described previously. In this study, these genetic biomarkers were combined with clinical variables and analyzed to obtain the most efficient cancer prediction model after an extended period of follow-up (median time of 7 years) by applying Cox regression modeling, bootstrapping and leave-one-out analyses. A total of 334 patients with Barrett's esophagus without dysplasia from 6 community hospitals (n = 220) and one academic center (n = 114) were included. The annual progression rate to high grade dysplasia and/or esophageal adenocarcinoma was 1.3%, and to adenocarcinoma alone 0.85%. A prediction model including age, Barrett circumferential length, and a clonicity score over the genomic set including chromosomes 7, 17, 20q and c-MYC, resulted in an area under the curve of 0.88. The sensitivity and specificity of this model were 0.91 and 0.38. The positive and negative predictive values were 0.13 (95% CI 0.09 to 0.19) and 0.97 (95% CI 0.93 to 0.99). We propose the implementation of the model to identify non-dysplastic Barrett's patients, who are required to remain in surveillance programs with 3-yearly surveillance intervals from those that can benefit from less frequent or no surveillance.

Original languageEnglish
Article numbere0231419
JournalPLOS ONE
Volume15
Issue number4
DOIs
Publication statusPublished - Apr 2020

Cite this