Molecular Testing Refines Barrett’s Esophagus Progression Risk Beyond Biopsy Findings

Routine endoscopic biopsy findings guide surveillance and therapy, but histopathology may not fully capture an individual patient’s biological risk of progression. This can complicate management when findings are nondysplastic, inconclusive, or discordant. New findings demonstrate that molecular risk assessment can identify elevated risk beyond pathology findings alone.

Castle Biosciences (Friendswood, TX, USA) presented data from three studies evaluating its TissueCypher Barrett’s Esophagus test at the American College of Gastroenterology (ACG) 2026 Annual Scientific Meeting and Postgraduate Course. TissueCypher analyzes protein biomarkers and tissue architecture in routine endoscopic biopsy samples. The test provides an individualized five-year probability of progression to high-grade dysplasia or esophageal adenocarcinoma.

The test is designed for patients with nondysplastic Barrett’s esophagus, indefinite for dysplasia, or low-grade dysplasia. Castle Biosciences describes TissueCypher as using an AI-derived spatialomics approach to identify molecular signatures that may precede dysplasia. The assay is performed on standard esophageal pinch biopsies and is designed to integrate into routine endoscopic practice without requiring additional procedures.

Two case reports described TissueCypher results that differed from conventional biopsy-based risk expectations. In one presumed low-risk patient with nondysplastic Barrett’s esophagus, TissueCypher returned a high-risk result indicating a 62% probability of progression to high-grade dysplasia or esophageal adenocarcinoma within five years. A repeat endoscopy six months after the initial examination identified previously missed prevalent high-grade dysplasia, after which the patient underwent endoscopic eradication therapy with radiofrequency ablation.

A second case involved successively classified biopsies of nondysplastic Barrett’s esophagus, low-grade dysplasia, and indefinite for dysplasia. TissueCypher testing of the indefinite-for-dysplasia biopsy returned a high-risk result indicating a 47% five-year probability of progression. Interventional endoscopy with radiofrequency ablation two months later confirmed previously undetected high-grade dysplasia, and subsequent treatment led to follow-up biopsies confirming complete eradication of intestinal metaplasia.

Additionally, a multicenter cohort study compared TissueCypher molecular risk stratification with the Progression in Barrett’s clinicopathologic-based risk stratification approach. Agreement between the approaches was slight, with Cohen’s kappa reported as 0.04. TissueCypher classified 83% of patients as low risk and 17% as intermediate or high risk, while the Progression in Barrett’s score classified 31% as low risk and 69% as intermediate or high risk.

Following TissueCypher testing, management recommendations changed in a manner consistent with the test’s risk classification for 96% of patients in the cohort. Physicians recommended a longer five-year surveillance window for 38 patients and endoscopic eradication therapy for eight patients, including five with nondysplastic Barrett’s esophagus whose TissueCypher results indicated intermediate or high risk despite no dysplasia on biopsy. The studies were conducted at Emory University School of Medicine and the University of Kansas School of Medicine.

“Endoscopic biopsy findings do not always tell the whole story when it comes to a patient’s risk of progression,” said Virendra Joshi, M.D., FAGA, FACG, board-certified gastroenterologist and associate professor in the Division of Digestive Diseases in the Department of Medicine at Emory University School of Medicine. 

“The studies being presented at ACG demonstrate how TissueCypher can complement standard pathology by providing objective molecular risk information that helps guide surveillance and treatment decisions for patients with Barrett’s esophagus, particularly when a patient’s molecular risk suggests a different approach than their biopsy diagnosis alone would indicate,” added Dr. Joshi.

Related Links
Castle Biosciences