ESR Projects

ESR 5: Endoscopic Multimodal Biophotonic Imaging for Esophageal tracking (ESOTRAC)

Zak Ali, Doctoral Candidate at the Technical University of Munich

Project Description

Presentations of esophageal cancer occur late in the disease, attributing to 400,000 annual deaths worldwide. Today white light endoscopy is a prominent technique in the detection of cancer entailing quadrat random biopsies. This protocol is laborious, plagued with missed cancers, invasive and the cost of analyzing biopsies is spiraling. Recently research has led to the rapid emerging area of early disease detection within the field of biophotonics. We present a panoramic photonic based hybrid endoscope employing multi-spectral optoacoustic tomography (MSOT) and optical coherence tomography (OCT) for esophageal tracking (ESOTRAC). OCT delivers superficial micron-scale morphological imaging and MSOT offers pathophysiological imaging and ability to penetrate deeper than OCT. ESOTRAC performs in-vivo early diagnosis and staging by replacing 2D qualitative and user dependent observations with objective, quantitative, 3D measurements of the entire esophagus wall, delivering morphological and pathophysiological cancer and pre-cancer features not available to white light endoscopy. For further information please visit: https://www.esotrac2020.eu/

ESR 11: Multimodal image analysis for pathology detection in endoscopy

Roger Fonollà, Doctoral Candidate at the Eindhoven University of Technology (TU/e)

Project Description

Colon cancer is one of the top leading cause of cancer death both in women and men. Early screening of the colon can prevent the development of abnormal tissue or polyps into cancer. Visual differentiation of benign and pre-malignant colonic polyps is an on-going challenge in clinical endoscopy routine. White Light Endoscopy (WLE) is the most common technique to visually assess lesions in the intestinal tract but is arguably unreliable due to hampering in polyp classification. Chromoendoscopy techniques are an alternative to enhance visual identification of gastrointestinal lesions by injecting a stain, improving differentiation between the mucosa, vessels and surface patterns in the intestinal tissue, but this requires additional chemical colorization to be injected in the body. In recent years, LED-based enhanced techniques like Blue Laser Imaging (BLI) and Linked Color Imaging (LCI) are potentially promising alternatives to avoid the use of chemical stains and to obtain similar results. In the current work, novel deep learning techniques are being applied to multi-modal gastrointestinal images to detect, classify or segment  polyps to assist clinicians in the diagnosis of colon cancer.