AI Tool Offers New Possibilities for Predicting Barrett’s Esophagus and Esophageal Cancer

Mayo Clinic researchers have developed a new artificial intelligence (AI) tool that may help identify individuals at risk of Barrett’s esophagus and esophageal cancer. Barrett’s esophagus, a condition caused by gastroesophageal reflux disease (GERD), is a known risk factor for esophageal cancer. Although screening tools are available, uptake remains low, leaving many cases undiagnosed. This AI-powered tool, which uses electronic health record (EHR) data, aims to improve early detection and close the screening gap.

Predictive Model Analyzes EHR Data to Flag At-Risk Patients

The study, led by Dr. Prasad Iyer, a gastroenterologist and researcher at Mayo Clinic, used de-identified EHRs from over six million patients to develop a predictive AI model. The tool analyzes clinical, endoscopy, pathology, and laboratory notes to estimate an individual’s risk of developing Barrett’s esophagus or esophageal cancer. According to the research, the tool can identify at-risk patients at least one year before a potential diagnosis.

In their analysis, the team identified 8,476 patients with Barrett’s esophagus, 1,539 with esophageal cancer, and a control group of 252,276 individuals. These datasets formed the basis for training and testing the predictive models. Results showed that the tool demonstrated an area under the receiver-operating curve (AUROC) of 0.84 for both conditions, indicating high accuracy. For Barrett’s esophagus, it achieved 76% sensitivity and specificity. For esophageal cancer, sensitivity was 84% and specificity was 70%.

The model also identified known risk factors for Barrett’s esophagus and esophageal cancer, including GERD, as well as novel factors such as coronary artery disease, triglyceride levels, and electrolyte imbalances. These findings could provide new insights for future research and screening protocols.

Why Early Risk Prediction Matters

Barrett’s esophagus and esophageal cancer are often detected at advanced stages, limiting treatment options and worsening outcomes. Screening for Barrett’s esophagus, a key step in early detection, is currently underutilized. The AI tool’s ability to integrate into EHR systems offers a potential solution: it could automatically prompt healthcare professionals to consider screening for at-risk patients, even in primary care settings.

“This tool could be integrated into the electronic health record and combined with a minimally invasive (nonendoscopic) screening tool and used by health care professionals in primary care,” said Dr. Iyer. By identifying patients earlier, the tool could enable timely interventions and improve outcomes.

While the tool shows promise, further research is needed to validate its clinical application and evaluate its impact in real-world settings. If successful, it could represent a significant step forward in the early detection and prevention of esophageal cancer.