A new artificial intelligence (AI)-based tool shows promise for improving surveillance in patients treated with endoscopic eradication therapies for dysplasia associated with Barrett’s esophagus (BE) and early esophageal adenocarcinoma. BE is the only known condition that precedes esophageal adenocarcinoma – an aggressive cancer with high mortality rates.
Developed and validated by US researchers, the AI model was more than 90% accurate in predicting which patients would experience a recurrence of BE after endoscopic eradication therapy (EET) and detecting when it is likely to occur.
The findings were published today in Clinical Gastroenterology and Hepatology.
Early detection of the dysplasia associated with Barrett’s esophagus and associated adenocarcinoma of the esophagus can save lives. Detecting recurrence of BE, BE-related dysplasia, and esophageal adenocarcinoma earlier, especially in high-risk patients who have undergone endoscopic eradication therapy, creates opportunities for early treatment before the cancer develops or progresses.”
Sachin Wani, MD, senior study author, executive director of the University of Colorado Anschutz Cancer Center’s Rady Esophageal and Stomach Center of Excellence
EET is an effective treatment for BE-associated dysplasia and early adenocarcinoma of the esophagus that eliminates 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 therapy, and current follow-up strategies do not distinguish between high- and low-risk patients. Everyone is followed using the same schedule regardless of risk,” said Wani.
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 who had been treated with EET and followed over time to determine if and when BE-related dysplasia or cancer returned. This analysis revealed that nearly 3 in 10 patients relapsed after successful treatment, with the condition returning about two years after treatment on average.
The AI tool was then trained to consider multiple patient factors simultaneously, such as age, body weight, disease severity and treatment details. He learned patterns that people can’t easily see, including how combinations of factors affect risk. They found that relapse was more likely in patients who had:
- A larger area of Barrett’s tissue
- Greater body weight
- Older age
- More treatment sessions were needed to completely remove the abnormal tissue
- More advanced cellular changes at the time of diagnosis
The model was tested in two ways: by checking how well it worked on patients similar to those it was trained on, and by checking performance on different groups of patients from other sources. The tool was accurate for both sets of patients.
This tool could help doctors individualize post-treatment follow-up, rather than using the same schedule for every patient. People with a higher risk of the condition coming back could be monitored more closely, while people with a lower risk may need fewer follow-up procedures. This approach could reduce unnecessary tests, reduce stress for patients and make better use of healthcare resources.
“This work represents many years of effort and collaboration between many institutions. It would not have been possible without the collaboration of our colleagues who shared their data and expertise,” said Wani.
Collaborators include experts at 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 of California Los Angeles, University of Kansas and Hirlanden Clinic Zurich.
The next step is to further validate the model using international datasets through collaborations in the Netherlands, UK, Belgium and Switzerland. The goal is to validate the tool so that it can be widely implemented and used as a reliable, universal aid in clinical care.
Source:
Journal Reference:
