Diagnostic efficiency of Aß protein PET brain imaging model in patients with AD and MCI

Objective To evaluate the diagnostic value of Aβ protein PET brain imaging model in distinguishing patients with Alzheimer's disease （AD） from mild cognitive impairment （MCI）. Methods Thirty-two subjects with cognitive impairment were enrolled in this study. All participants were diagnosed with AD or MCI by clinical examination and underwent 18F-AV-45 （2.96~4.44 MBq/kg）PET/CT scan（heart and head）. The fitting curves of specific activity of left ventricle and region of interest （ROI） at different time points were delineated, and then the metabolic rate K was calculated by Logan model. The difference in metabolic rate K value was compared between AD and MCI groups by using Mann-Whitney U test. The diagnostic efficiency of Logan model was evaluated by the receiver operator characteristic （ROC） curve. Results The degree of Aβ protein deposition in the cerebral cortex of AD and MCI patients was different. AD and MCI patients could be distinguished with the K value of Logan model. The K value in the AD group was 3.96（2.66, 4.26）, significantly higher than 2.62（1.41, 2.96） in the MCI group （P < 0.05）. According to ROC curve analysis, the optimal diagnostic threshold of K value was 3.23, the diagnostic sensitivity was 100%, the specificity was 58.33%, and the area under ROC curve （AUC） was 0.781 （Z = -2.350, P < 0.05）, respectively. Conclusion Aβ protein PET imaging with Logan model has high diagnostic efficacy in distinguishing AD from MCI.