Computer-Aided Evaluation System for Parkinson's Disease Using Image Registration and Labeling

Single photon emission computed tomography image (SPECT) of dopamine transporter with ^99mTc-TRODAT-l has recently been proposed to be a valuable and feasible means of assessing the integrity of dopamine neurons. In order to measure the specific-to-nonspecific binding ratio of the nuclear medicine within the specific-binding tissues, i.e. putamens and caudate nuclei, the corresponding MRI is needed to be registered to SPECT for bounding the regions of interest. Therefore, an automatic labeling algorithm which enables contouring the putamens and caudate nuclei is necessary because segmenting these tissues manually from MRI costs tons of time and energy of physicians. We have built a computer-aided clinical diagnosis system which integrates MRI/SPECT registration and MRI labeling for the evaluation of Parkinson's disease. Clinical MRI and SPECT data including eighteen healthy subjects and thirteen patients were involved to validate the performance of the proposed system.