Computer Algorithm Matches Skilled Specialists in Detecting Early Esophageal Cancer
Using a computer algorithm to scan images of the esophagus for signs of early esophageal cancer is nearly as accurate as top specialists. Early esophageal cancer can be easily missed, and only a handful of top specialists are skilled at detecting it.1
Barrett's esophagus, abnormal tissues in the esophagus due to gastric irritation from prolonged reflux, is a major risk factor for developing esophageal cancer. Reflux is more common in overweight and obese people. Because of the risk of esophageal cancer, people with Barrett's esophagus undergo regular endoscopy. However, only a handful of medical specialists are capable of detecting the earliest stages of esophageal cancer. Thus, the cancer can be missed when it begins and is easiest to treat. After the cancer reaches an advanced stage, its 5-year survival rate is lower than 50%.
This research came from a collaboration between a top specialist in Barrett's esophagus, Erik Schoon, MD, Catharina Hospital, Eindhoven, the Netherlands, and an image analysis expert, Peter de With, PhD, Eindhoven University of Technology, The Netherlands. They combined their expertise in Barrett's esophagus and in image analysis methods.
"Spectacular," Schoon said of the results. "To recognize early forms of cancer in a Barrett's esophagus is one of the most difficult things to do in our field."
The researchers hope that this computer analysis method will become available in every hospital so it can help gastroenterologists to recognize the earliest stages of cancer. However, the software needs to be improved and made suitable for analyzing real-time video frames, then more extensive hospital testing is needed. The researchers predict that broad implementation is still 5 to 10 years away.
1. Van der Sommen F, Zinger S, Curvers WL, et al. Computer-aided detection of early neoplastic lesions in Barrett's esophagus. Endoscopy. 2016 Apr 21. 2016;48(7):617-624.