Introduction

Eosinophilic esophagitis (EoE) is a chronic inflammatory disease of the esophagus. Inflammation in EoE has been associated with potential risk of esophageal cancer (EC), with controversial results by the most recent studies [1] [2] [3].

Aims & Methods

The aim of this study was to estimate the true risk burden of EC in EoE, further investigating the confounding role of overlapping gastroesophageal reflux disease (GERD) or Barrett Esophagus (BE) on EC occurrence. We conducted a retrospective multicenter cohort study on aggregated data of EoE patients accessing the global federated health research network “TriNetX”, with access to electronic medical records from approximately a hundred million patients across large healthcare organizations (HCOs), based on the ICD-10 code (K20.0), from January 1^st 2000 until March 31^st 2025. We inferred the risk of EC, by comparing two different EoE cohorts: cohort A including EoE diagnosis (excluding eosinophilic gastrointestinal disorders [EGIDS]), not applying any other restrictive criteria; cohort B with only pure EoE diagnosis excluding any instance of GERD (K21.0) and Barrett esophagus (BE) (K22.7) or other EGIDS. The control cohort comprised patients encountered in outpatient clinic for undefined issues (Z00), without diagnosis of EoE/BE or pre-cancerous lesions. We used propensity score nearest neighbor greedy matching, adjusting for PPI use, family history of GI cancer, age, sex, other esophageal diseases (K22.0) and GERD status. Kaplan-Meyer (KM) curves with censoring were built for time-to-event analysis. Log Rank test was used to compare KM curves. Hazard Ratios (HRs) and Risk Ratio (RR) estimation for EC risk were calculated and compared. Sub-analyses for squamous-cell cancer (SCC) vs adenocarcinoma, and for EC location in the esophagus (distal, medium and proximal) were implemented.

Results

After 1:1 propensity score matching, 92,447 patients with EoE (with or without GERD) (cohort A) (mean age 38.9±20.7 years, M 59.7%, previous history of malignant GI neoplasm 2.7%, GERD 40.4%, other esophageal diseases 14.9%, PPI use 40.4%; all p>0.05 with control group) and 32,270 patients pure EoE (without GERD/BE) (cohort B) (mean age 33.9±19.3 years, M 64.7%, previous history of malignant GI neoplasm 1.2%, GERD 0%, other esophageal diseases 8.3%, PPI use 23.2%; all p>0.05 with control group), were compared with controls subjects. Mean follow-up of cohort A was 1298.6 days compared to 1046.5 days in cohort B. The total number of EC occurrences in cohort A were 44 in EA group compared to 31 in the control group, resulting in a HR of 1.676 (1.058-2.657, 95%CI), while a non-significant RR of 1.419 (0.896, 2.247, 95%CI). The KM survival analysis resulted significant (log-rank test p=0.026). Interestingly, when comparing cohort B (pure EoE without GERD or BE) with the control group, only 10 instances of EC in both groups where encountered during the follow-up, with a non-significant HR of 0.893 (0.290-2.754 95%CI) with no difference in the KM survival analysis (p= 0.844). No differences in sub-analysis for squamous cell or adenocarcinoma cancer, neither for EC occurrence in different esophageal localizations were detected.

Conclusion

The results from this global retrospective study on more than 90,000 EoE patients, confirmed the irrelevant risk of EC in pure EoE. The mild association reported in other studies may be likely due to EoE overlap with GERD/BE. No specific surveillance for EC seems needed in EoE when GERD and BE are not present.

References

1. Albaneze N, Cotton CC, Anderson C, Katzka DA, Dellon ES. No Association Between Eosinophilic Oesophagitis and Oesophageal Cancer in US Adults: A Case-Control Study. Aliment Pharmacol Ther. 2025 Jan;61(2):363-368. doi: 10.1111/apt.18431. Epub 2024 Dec 4. PMID: 39629856; PMCID: PMC11680457.
2. Zhou MJ, Hsing AW, Clarke JO. Prevalence and predictors of Barrett's esophagus and esophageal cancer in patients with eosinophilic esophagitis. Dis Esophagus. 2025 Jan 7;38(1):doae120. doi: 10.1093/dote/doae120. PMID: 39775829.
3. Uchida AM, Schuman SS, Pyne A, Peterson K, Carlson M, Garber JJ, Roelstraete B, Ludvigsson JF. Risk of Cancer Diagnosis in Patients With Eosinophilic Esophagitis Using a Nationwide Swedish Population Cohort. United European Gastroenterol J. 2024 Dec;12(10):1378-1387. doi: 10.1002/ueg2.12713. Epub 2024 Nov 20. PMID: 39565026; PMCID: PMC11652337.

Disclosure

E.V. received speaker and/or consulting fees from Sanofi/Regeneron, Falk Pharma, Apollo Therapeutics, and Aurora Biofarma. EVS received consulting fees from Dr Falk Pharma and Reckitt Benckiser; research support from Medtronic and Diversatek Healthcare. DAC.: Medtronic(Speaking, Consulting, License); Diversatek (Consulting); Braintree(Consulting); Medpace (Consulting); Phathom Pharmaceuticals(Speaking); Regeneron/Sanofi(Speaking). AJB received research funding from BMS, Sanofi/Regeneron, SST, and Dr. Falk Pharma and received speaker and/or consulting fees from Uniquity, Laborie, Medtronic, BMS, Dr. Falk Pharma, Calypso Biotech, Eupraxia, Aqilion, Alimentiv, Sanofi/Regeneron, Uniquity, Reckitt, AlfaSigma and AstraZeneca. All other authors did not have any disclosures.

Introduction

Corpus atrophic gastritis (CAG) requires endoscopic-histological surveillance due to the risk of developing gastric neoplastic lesions (GNL). This study aimed to identify variables associated with GNL development at long-term follow-up using a feature ranking method based on a novel One-Class Support Vector Machine (SVM) model.

Aims & Methods

A dataset containing clinical, endoscopic, and histological variables from consecutive CAG patients (2001-2023) adhering to a surveillance program was considered. GNL presence at the longest available follow-up was recorded. Gastric biopsies and histological evaluations were conducted according to the updated Sydney system. CAG patients with any missing data or a GNL diagnosis at the baseline gastroscopy were excluded from the final analysis. A Fisher score-based feature ranking method and One-Class SVM were used to select the optimal subset of baseline variables associated with GNL development, aiming to achieve average sensitivity (Se) and specificity (Sp) above 80%.

Results

Overall, 355 CAG patients were initially considered. Of these, 36 were excluded due to the presence of GNL at baseline gastroscopy, and 216 for missing data. Thus, a total of 103 patients were considered and grouped into: CAG patients with [22 patients (F 68.1%), median age 68 (35-83) years] and without GNL at follow-up [81 patients (F 72.8%), median age 59(26-84) years]. After a median follow-up of 60 (12-192) months, 13 epithelial GNL (gastric adenocarcinoma or high/low-grade dysplasia) and nine type-1 gastric-neuroendocrine-tumors (T1gNET) were recorded. Parietal-cell-antibodies and pepsinogen-I < 30 μg/l were associated with both epithelial GNL and T1gNET. Antral inflammation and age > 60 were linked to epithelial GNL, while anti-thyroperoxidase-antibodies, smoking, and dyspeptic symptoms at CAG diagnosis were linked to T1gNET. Low-dose aspirin and H. pylori eradication therapy showed inverse associations with epithelial GNL and T1gNET, respectively.

Conclusion

This is the first study in which an AI model simultaneously considers clinical, endoscopic, and histological characteristics from a dataset of CAG patients, showing good sensitivity and specificity in identifying those variables associated with developing GNL. Nonetheless, we acknowledge the need for improved variable standardization and data balancing between groups to enhance model performance further.