統計学的脳画解析における最適しきい値の検討

摘要

脳画像検査は認知症の病理，病態を反映し特徴的な所見を示すために，認知症の診断において欠かせない検査となっている一方で，近年重要視されているmild cognitive impairment(MCI)段階やアルツハイマー病早期における脳画像は，萎縮あるいは血流低下が軽度であるために，専門医であっても診断が困難であった.統計学的脳画像解析(統計解析)は，この問題に対処できると考えられている.そこで統計解析を行っ方法として，statistical parametric mapping(SPM), three mmensional-stereotactic surface projections(3D-SSP)などが開発された.特にSPMは形態診断のためのvoxel based specific regional analysis system for Alzheimer's disease(VSRAD),血流診断のためのeasy Z-score imaging system(eZIS)と複数の診断目的に応用され，臨床の場において広く利用されている.%Purpose:Aim of this study was to investigate optimal threshold of Z score when evaluating statistics image of Alzheimer's disease visually. Method:We classified 53 clinical patients in control and target group, and evaluated the distribution of Z score calculated with statistical brain image analysis for magnetic resonance and perfusion single photon emission computed tomography(SPECT). The optimal Z score threshold was determined from statistical significance that compared previously mentioned groups. Results:Target group was able to classify significantly Z score at 1.25 from control group in wide region of parietal lobe with statistical brain image analysis for perfusion SPECT. Discussion:The optimal threshold is equal or less than 2.0, in the case of Z score variance is close to the standard normal distribution. In contrast, the threshold is over 2.0 in the case of Z score variance is more than 1.0, and then by using ordinary threshold 2.0, it cannot point out abnormality.