Meeting Banner
Abstract #0867

Automatic Assessment of High Grade Brain Tumour Tissue in MR Images: What Is Tumour & What Radiation Injury

Yaniv Gal1, Stephen Rose2, Pierrick Bourgeat3, Nicholas Dowson3, Zeike Taylor4, Michael Fay5, Paul Thomas5, Olivier Salvado3, Stuart Crozier1

1School of Information Technology and Electrical Engineering, University of Queensland, Brisbane, Queensland, Australia; 2Centre of Clinical Research, University of Queensland, Brisbane, Queensland, Australia; 3CSIRO, Brisbane, Queensland, Australia; 4Department of Mechanical Engineering, The University of Sheffield, Sheffield, United Kingdom; 5Royal Brisbane and Womens Hospital, Brisbane, Brisbane, Queensland, Australia

A method for automatic assessment of high grade brain tumour tissue in MR images is presented. The aim of the method is to differentiate between tumour tissue and radiation injured tissue. The method is validated qualitatively on two high grade brain cancer patients and demonstrates high potential.

Keywords

academic accuracy acknowledgment agent applied assessment assist assisted assumption automatic background benign biology brain bright classification classified classifier clearly clinical clinically computer contrast correlated corresponds creating curve dark dataset datasets defined delineation denote detecting differentiating electrical engineering enhance enhanced enhancement enhancing error expected experienced experiment extent extracted extraction feature features fitted gold gone grade grant growth health hospital hypothesis included indeed independently indicate induced injured injury inspected journal kinetic kingdom limitation limitations linear logistic manual maps mechanical medicine model months moreover namely noise nuclear often operating orange patient patients performance phenylalanine physician places planning poor positron post potential prior process project qualitatively radiation radio radiology radiotherapy received receiver recurrence recurrent registered regression resection resolution risk rose sample samples school selected selection session society source specificity stage strategy suffer suggest supervised supported tagged therapy tissue tomography trained training treated types underlying useful validated validation visually volume wash weeks women world years yield