June 15, 2021 – Researchers led by Ludwig Oxford’s Jens Rittscher and his University of Oxford colleagues Sharib Ali and Barbara Braden have developed an artificial intelligence (AI) system that can be used alongside endoscopy to get more accurate readings of the precancerous condition Barrett’s esophagus and so identify patients most at risk of developing cancer. The condition develops in the lower esophagus in response to chronic acid reflux.
In a research paper published in Gastroenterology, the researchers report that the new AI-driven 3D reconstruction of Barrett’s esophagus achieved 97.2 % accuracy in measuring the extent of this precancerous condition in real time. This should enable clinicians to assess the risk, the best surveillance interval for patients and the response to treatment more quickly and confidently. There is a less than 0.1-0.4% risk per year of developing cancer with normal Barrett’s esophagus—which translates to roughly one in 200 patients. However, that risk increases with the extent of the condition.
The diagnostic technology developed by the researchers employs computational technology to reconstruct the surface of the Barrett’s area in 3D from endoscopy videos and generates a C&M score, a standardized measure of Barrett’s esophagus. This 3D reconstruction allows the clinician to quantify the Barrett’s area, including patches or “islands” not connected to the main Barrett’s area.
The technique was tested on a purpose-built 3D printed esophagus phantom and high-definition videos from 131 patients scored by expert endoscopists. The endoscopic phantom video data achieved a 97.2 % accuracy for the C&M score measuring the length, while the measurements for the whole Barrett’s esophagus area achieved nearly 98.4had% accuracy. On patient data, the automated C&M measurements corresponded with the endoscopy expert scores.
The news article from which this summary is derived can be found here.