Gastrointestinal cancers, such as esophageal- and colorectal cancer, are among the most common and lethal cancers. Therefore, endoscopic screening programs aim to detect such cancers at an early stage, at which the prognosis is considerably better, with five-year-survival rates close to 100 percent. However, the subtle visual signs of early tumors and their low exposure to gastroenterologists at peripheral medical centers render them hard to detect and often overlooked. Artificial Intelligence (AI) can help in detecting those early signs of cancer by real-time inspection during endoscopic surveillance, considerably improving patient care while simultaneously reducing costs.

Dr. Fons van der Sommen is an assistant professor working on computer vision for healthcare applications at the Department of Electrical Engineering of the Eindhoven University of Technology. Fons heads a cluster of 8 Ph.D. researchers within the VCA computer vision research group, addressing technical aspects of modern Convolutional Neural Network (CNN) architectures, such as efficiency, robustness, and interpretability, by leveraging methodological tools from the image- and signal processing. Fons’ cluster works on a variety of healthcare applications of CNNs, ranging from Computer-Aided Detection and Diagnosis (CADe/CADx) in endoscopy to denoising in Cone-Beam Compute Tomography (CBCT).

In this talk, Fons will show his work on computer vision for early esophageal cancer detection, addressing the clinical background, employed technologies over recent years, their clinical validation, and the road to implementation and adoption.

• Event Link: Youtube