A new artificial intelligence tool has been developed to help predict the recurrence of Barrett's esophagus after endoscopic eradication therapy, according to an April 7 announcement. The model, created by U.S. researchers and published in Clinical Gastroenterology and Hepatology, was reported to be over 90% accurate in identifying which patients are likely to experience a return of the condition and when it might occur.
Barrett's esophagus is the only known precursor to esophageal adenocarcinoma, a cancer with high mortality rates. While endoscopic eradication therapy is effective at eliminating abnormal tissue and reducing cancer risk, recurrence can still happen for some patients. Currently, all patients receive the same follow-up schedule regardless of their individual risk.
"The challenge is that recurrence of Barrett's esophagus can still occur even after endoscopic eradication therapy and current surveillance strategies don't distinguish between patients at high versus low risk. Everyone is followed using the same schedule regardless of their risk," said Wani.
Wani and colleagues used data from more than 2,500 patients who had undergone treatment for Barrett's esophagus-related dysplasia or early-stage cancer. By analyzing this information with machine learning techniques, they found that about three out of ten patients experienced a recurrence within two years on average after successful treatment. The AI tool considered multiple factors such as age, body weight, disease severity, and treatment details to identify patterns not easily detected by humans.
The model was tested both on patient groups similar to those used for training as well as on different populations from other sources. It performed accurately in both scenarios. According to Wani: "This work represents several years of effort and partnership across multiple institutions. It wouldn't have been possible without the collaboration of our colleagues who shared their data and expertise." Collaborators included 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 Hirlanden Clinic Zurich.
Researchers plan further validation using international datasets through collaborations in countries including the Netherlands, United Kingdom, Belgium and Switzerland so that the tool can be widely adopted as a universal aid in clinical care.
