Current understanding of the risk of neoplastic progression in patients with Barrett’s esophagus samples harboring indefinite for dysplasia (BE-IND) stems from small retrospective and pathology-registry studies. In this multicenter cohort study, we aimed to determine the incidence and prevalence of neoplasia in BE-IND.
Patients with confirmed BE-IND from 2 academic centers were included if they had no previous evidence of dysplasia and underwent endoscopic follow-up (FU) of ≥1 year. Rate of progression to neoplasia was calculated and categorized as prevalent (progression <1year of FU) and incident (progression ≥1 year of FU). Multivariable regression adjusted for relevant clinical features was performed to identify risk factors for progression.
Four hundred sixty-five patients diagnosed with BE-IND were identified between 1997 and 2017, of which 223 (48.0%) were excluded. Of the remaining 242 patients, 184 (76.0%) had no evidence of dysplasia during FU. In 23 patients (9.5%) prevalent neoplasia occurred (20 low-grade dysplasia [LGD], 2 high-grade dysplasia [HGD], 1 intramucosal cancer [IMC]), whereas 35 patients (14.5%) developed incident neoplasia (27 LGD, 5 HGD, 3 IMC), after median 1.5 years (IQR 0.6-3.2 years). The incidence rates of any neoplasia and HGD/IMC were 3.2 and 0.6 cases/100 patient-years, respectively. BE-length correlated with an increased risk of prevalent (OR, 1.18 per 1 cm; 95% CI, 1.02-1.38; P=0.033) and incident neoplasia (OR, 1.02; 95% CI, 1.00-1.03; P=0.016).
Patients with BE-IND should be closely monitored, as nearly a quarter harbor or will shortly develop dysplasia. BE length is a clinical predictor of neoplastic progression; however, more accurate molecular biomarkers for risk stratification are warranted.