Abstract
Analysis of amyloid amount deposited to the cerebral cortex, which can be estimated from an amyloid PET image, is effective for early diagnosis of Alzheimer's disease (AD). Nevertheless, its quantifiable measure is hampered by the partial volume effect due to the low spatial resolution, because the amount of amyloid deposition is thereby underestimated. In this work, we proposed a method using SPM8 (statistical parametric mapping 8) to correct a partial volume effect of PET image and to improve its quantifiable measure of amyloid deposition amount based on high-resolution morphological information of CT image, which is simultaneously acquired by a PET/CT scanner. The procedure was applied to clinical data of nine AD patients and eleven normal volunteers, and then emphasized the difference between the normal and AD groups in SUVR (standardized uptake value ratio) values in the eight volumes of interest effective for AD diagnosis. In addition, we demonstrated the effectiveness by comparing the AUC (area under ROC curve) between the results with and without the partial volume correction.
