Optimal ROI setting on the anatomically normalized I-123 FP-CIT images using high-resolution SPECT

OBJECTIVE The aim of this study is to establish the optimal regions of interest (ROIs) in anatomically normalized I-123 FP-CIT SPECT images for the quantification of dopamine transporter binding. METHODS The subjects comprised 16 normal controls and 14 Parkinsonian patients. All of the normal control subjects underwent I-123 FP-CIT SPECT and MRI. The SPECT device used in this study was a Toshiba GCA-9300R with triple head detectors. I-123 FP-CIT (148 MBq) was intravenously administered as a bolus, and the SPECT scan started 4 h after the administration. The data were collected over 20 min for each subject, and reconstructed using a 3D-OSEM algorithm. The data were analyzed using SPM8. I-123 FP-CIT SPECT images were anatomically normalized to the MNI space using an I-123 FP-CIT template, and then divided by the background counts automatically measured using the ROIs set for the cerebral cortices. RESULTS In the normal control subjects, the specific binding ratios of the MRI-based ROIs were lowest in the caudate nucleus, while the ratios of the I-123 FP-CIT-based ROIs were almost the same throughout all three parts. In contrast, in Parkinsonian patients, the specific binding ratios of the I-123 FP-CIT-based ROIs revealed rostrocaudal decline, while those of the MRI-based ROIs were highest in the anterior putamen. CONCLUSION We created an ROI template on the anatomically normalized MRI and I-123 FP-CIT images, and concluded that I-123 FP-CIT-based ROIs are more suitable for obtaining quantitative values than MRI-based ones.