Novel artificial intelligence (AI)-based tools have shown potential to improve surveillance of patients undergoing endoscopic eradication therapy for Barrett’s esophagus (BE)-related dysplasia and early esophageal adenocarcinoma. BE is the only known condition that precedes esophageal adenocarcinoma. Esophageal adenocarcinoma is an aggressive cancer with a high mortality rate.
The AI model, developed and validated by researchers in the United States, showed more than 90% accuracy in predicting which patients will experience recurrence of BE after endoscopic eradication therapy (EET) and detecting when recurrence is most likely to occur.
The findings were published today in the journal Clinical Gastroenterology and Hepatology.
Early detection of Barrett’s esophagus-associated dysplasia and associated esophageal adenocarcinoma can save lives. Early identification of recurrence in BE, BE-related dysplasia, and BE-related esophageal adenocarcinoma provides an opportunity for timely treatment before cancer develops or progresses, especially in high-risk patients who have undergone endoscopic eradication therapy. ”
Sachin Wani, MD, senior author of the study and executive director of the Rady Esophageal and Gastric Center of Excellence at the University of Colorado Anschutz Cancer Center
EET is an effective treatment for BE-associated dysplasia and early esophageal adenocarcinoma that removes abnormal Barrett’s tissue and significantly reduces the risk of progression to esophageal cancer.
“The challenge is that recurrence of Barrett’s esophagus can occur even after endoscopic eradication treatment, and current surveillance strategies do not distinguish between high- and low-risk patients. Everyone is followed up on the same schedule, regardless of risk,” Wani said.
Using artificial intelligence and data from more than 2,500 patients, Wani and a team of leading experts from across the country developed the machine learning tool. To create this, they analyzed detailed clinical data from patients treated with EET and followed over time to determine if and when BE and BE-associated dysplasia or cancer had returned. This analysis revealed that nearly 3 in 10 patients experienced a relapse after successful treatment, with symptoms returning on average approximately 2 years after treatment.
The AI tool was then trained to look at many patient factors at once, including age, weight, disease severity, and treatment details. We learned patterns that humans cannot easily recognize, such as how combinations of factors affect risk. They found that recurrence was more likely in patients who:
- Longer areas of bullet tissue
- higher weight
- senior citizen
- More treatment sessions were required to completely remove the abnormal tissue
- More advanced cellular changes at diagnosis
This model was tested in two ways. One is to see how well it performs on patients similar to those it was trained on, and the other is to see how well it performs on different patient groups from other sources. This tool was accurate for both patients.
This tool helps doctors customize post-treatment follow-up care instead of using the same schedule for all patients. Those at high risk of having symptoms return may be able to be monitored more closely, while those at lower risk may require fewer follow-up steps. This approach has the potential to reduce unnecessary testing, reduce stress for patients, and better utilize healthcare resources.
“This study represents several years of effort and partnership across multiple institutions. It would not have been possible without the collaboration of our colleagues who shared their data and expertise,” Wani said.
Collaborators include experts from Johns Hopkins University, Mayo Clinic (UZ Leuven), University of North Carolina at Chapel Hill, Washington University School of Medicine, Cleveland Clinic London, Northwestern Feinberg School of Medicine, University College London, University of California Los Angeles, University of Kansas, and Hilllanden Clinic Zurich.
The next step is to further validate the model using international datasets through a collaboration between the Netherlands, UK, Belgium and Switzerland. The goal is to validate the widespread application of this tool so that it can be used as a reliable and universal adjunct in clinical care.
sauce:
University of Colorado Anschutz
Reference magazines:
Aksintala, V. Others. (2026). Machine-based learning model for recurrence prediction and timing after endoscopic eradication treatment of Barrett’s esophagus. Clinical gastroenterology and hepatology. DOI: 10.1016/j.cgh.2026.03.026. https://www.sciencedirect.com/science/article/abs/pii/S1542356526002363
