Predictors of neoplastic progression in gastroesophageal lesions indefinite for dysplasia.

Abstract

Aims: Current understanding of the risk of neoplastic progression in patients with Barrett's esophagus with indefinite for dysplasia (BE-IND) and gastric indefinite for dysplasia (G-IND) remains limited. This study aims to identify prognostic histological and clinicopathological factors for IND progression in the upper gastrointestinal tract.

Methods and results: Patients with confirmed BE-IND and G-IND, no previous evidence of dysplasia/cancer and follow-up of ≥ 6 months were included. The rate of neoplastic progression was calculated and the multivariate Cox regression model adjusted for demographic and histological features was used to identify risk factors for progression. A total of 719 patients diagnosed with IND (158 BE-IND and 561 G-IND) were identified, 395 of whom were excluded. Progression rates were 4.4 per 100 person-year for BE-IND and 1.6 per 100 person-year for G-IND patients. Progression was observed only in IND of hyperproliferative intestinal metaplasia (HIM) type. Operative link for gastritis assessment (OLGA) stage (III-IV versus 0-II) was the only significant predictor of G-IND progression [hazard ratio (HR) = 47.82, confidence interval (CI) = 5%: 6.24-366.37, P < 0.001]. In BE-IND patients, lack of interval endoscopy significantly increased the risk for BE-IND (HR = 13.45, CI = 95%: 1.24-145.75, P = 0.032).

Conclusion: IND is a challenging diagnosis for pathologists and implies an increased risk for neoplasia, especially in BE patients, without providing definitive information for patient management. This study highlights that neoplastic progression of IND lesions is primarily associated with HIM type. For BE-IND, interval endoscopy significantly reduces progression risk, while in G-IND, OLGA staging (III-IV) is a strong predictor of progression. These findings emphasise the importance of identifying HIM and using OLGA staging in G-IND for better risk stratification and follow-up strategies.