Automatic Virtual Colonoscopy Based Diagnostic Decision Support System

Automatic Virtual Colonoscopy Based Diagnostic Decision Support System for Colorectal Cancer Screening Project IKTA 5 -058/2002 Infocommunications Technologies and Applications Research & Development Division, Ministry of Education, Hungary http: //www. om. hu/ikta Project Duration: March 2003 – Sept 2004

Participants Co-ordinator: l Semmelweis University, Department of Diagnostic Radiology and Onkotherapy, Budapest Home page: http: //www. sote. hu Address: H-1082 Budapest, Üllői út 78/a Tel. : (36 -1)-2100307 Project leader: Tarján Zsolt Dr. <tarjan@radi. sote. hu> Consortium members: Budapest University of Technology and Economics, Department of Control Engineering and Information Technology, Budapest Home page: http: //www. iit. bme. hu Team leader: Szirmay-Kalos László Prof. <szirmay@iit. bme. hu> l KDBR Consulting Kft. , Budapest Home page: http: //www. kdbr. hu Team leader: Kuhár János <kuhar@brconsulting. hu> l

Scientific Background Keywords: l Endoscopy, Colonoscopy, Cancer Screening, Radiology, Computer Tomography, Virtual Endoscopy, Virtual Colonoscopy, Computer Graphics, Volume Rendering, Image Processing, Decision Support Related publications by the participants at: l European Congress of Radiology, Wien, 2001 l 3 rd National Conference on Image Processing and Shape Recognition, Domaszék, Hungary, 2002 l First Hungarian Computer Graphics and Geometry Conference, Budapest, 2002 l Winter School on Computer Graphics, 2003 l 3 rd Conference on 3 D Imaging and Virtual Endoscopy, Rome, 2002

Motivation 6000 people die yearly because of colorectal cancer in Hungary. l Early detection could reduce fatalities to one third. l Because of associated discomfort only few asymptomatic patients would undergo traditional screening by colonoscopy. l

Traditional Colonoscopy Video colonoscopy Endoscopic treatment

Virtual Colonoscopy CT scan Axial cross-sectional images set Virtual colonoscopy view Visualization algorithms Volumetric dataset (voxel array)

Goals The aim of the project is to develop a complex alternative screening method for colorectal cancer and preceding conditions. Research focuses on: l less burdening patient preparation, l suitable CT protocol, l automated virtual colonoscopy software, l fecal tagging algorithm, l handling noisy, low dosage CT data, l computer aided semi-automated decision support.

Software Engineering Problems l l l l Processing large volumetric datasets at interactive speed Direct volume visualization Efficient user friendly interface Semi-automated navigation Fecal tagging Automated polyp detection Diagnostic decision support

Virtual Colonoscopy Software

Challenges in Radiology Fast patient preparation l Efficient and comfortable cleansing, contrasting and inflation l CT protocol to minimize X-ray dosage l l Ensuring high diagnostic sensitivity