Tong San Koh1, Choon Hua Thng2, Septian Hartono1,2, David Collins3, Sophie Richards3, Martin O. Leach3, Aslam Sohaib4, David Dearnaley3, Dow-Mu Koh3
1School of Electrical & Electronic Engineering, Nanyang Technological University, Singapore, Singapore; 2Department of Oncologic Imaging, National Cancer Centre Singapore, Singapore, Singapore; 3Institute of Cancer Research, Sutton, Surrey, UK; 4Department of Radiology, Royal Marsden Hospital, Sutton, Surrey, UK
We aim to extract tumour component maps using Independent Component Analysis (ICA) statistical technique from prostate DCE MRI datasets corresponding to a washout enhancement behaviour for comparison with histology. Five prostate DCE MRI cases with histology comparison were processed with ICA. It is found that ICA extracts spatial component maps corresponding to a few consistent enhancement profiles within the prostate DCE MRI datasets. ICA identifies voxels within the prostate that are associated with a washout signal-time profile which correlated with tumour locations on histology.