Skip to main content

Article Category

News

Article available in the folowing languages:

Diagnostic tool for esophageal cancer

Images taken using an endoscope are not always easy to read. A new image-processing system automatically searches for similar-looking pictures, accompanied by a diagnosis, to serve as a comparative reference. The system can be seen at Medica 2007 in Düsseldorf from November 14 to 17 (Hall 10, Stand F05).

The meal was great, but afterwards a burning sensation at the back of the throat spoils the pleasure – we call it heartburn. This unpleasant reaction occurs because the sphincter is unable to perform its normal function of holding back the gastric acid in the stomach, allowing it to rise up into the esophagus. If this recurs frequently over many years, it is not merely discomforting but can actually have serious consequences. The esophagus often responds to the repeated attacks from gastric acid by replacing some of its cells with less sensitive stomach cells – experts refer to this as Barrett’s ulcer, which can develop into cancer if left untreated. In order to detect these health-damaging changes, the physician performs an endoscopic examination of the esophagus. But the resulting images are not always easy to interpret. Is it a definite case of Barrett’s ulcer, or is the esophagus healthy? Medical practitioners will soon be able to obtain help with their diagnosis. Researchers at the Fraunhofer Institute for Integrated Circuits IIS in Erlangen have developed a system named EndoCAD for this purpose. “The physician marks an area of tissue that appears to be abnormal, and EndoCAD compares it with a set of images stored in a database. The system selects the images showing the greatest similarity with respect to coloration and structural features,” explains IIS project manager Dr. Christian Münzenmayer. “The selected images are displayed together with the corresponding, confirmed diagnosis. This gives the physician some useful points of reference for establishing a diagnosis.” The system can be used similarly to analyze endoscopic images of the larynx as a starting point for a diagnosis. “EndoCAD’s performance is particularly good because it compares the marked sections of the images with respect to both structure and coloration. Our system has performed reliably using the clinical data that has been available up to now, with an overall classification rate of over 85 percent,” says Münzenmayer. The research team will be demonstrating a prototype of the system at Medica 2007 in Düsseldorf from November 14 to 17 (Hall 10, Stand F05).

Countries

Austria, Belgium, Czechia, Germany, Denmark, Estonia, Greece, Spain, Finland, France, Hungary, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, Netherlands, Poland, Portugal, Sweden, Slovenia, Slovakia, United Kingdom