Abstract
Early diagnosis and discrimination of dementia are very important to determine the drug type and the administration schedule. Dynamic scintigraphy using a radioactive medicine is used for the diagnosis of dementia disease because the examination can estimate mean cerebral blood flow (mCBF). Minimally invasive analysis of mCBF has been proposed as the PATLAK plot method. The PATLAK plot method requires many manual procedures such as determination of two ROIs of aorta and cerebrum of right and left hemisphere on the dynamic scintigram before the estimation of the mCBF. Subjective determination of those ROIs and parameters often cause of inter- or intra-variability of the measurements results. Our developed software will help operators to determine those parameters including segmentation of the ROIs, and will reduce the variability of the measuring results. The software provides semi-automated localization and determination of the shape to fit the aorta and cerebrum, and fully automated alignment of time axis and curve fitting on the plotted graph. The initial variability of our software was investigated by five radiological technologists in nuclear medicine with 45 cases.
