Background: We previously demonstrated that radioligand binding to the 18 kDa translocator protei... more Background: We previously demonstrated that radioligand binding to the 18 kDa translocator protein (TSPO), a marker for inflammation, was greater in patients with Alzheimer’s disease (AD) than controls and correlated with clinical severity. We sought to determine the relationship between TSPO binding and progression of AD in a longitudinal study. Methods: Eleven patients with either AD or mild cognitive impairment and 8 age-matched controls underwent TSPO PET imaging with C-PBR28 at baseline and again after median follow up of 2.7 years. MRI, Pittsburgh Compound B (PIB) PET, and cognitive testing were also performed at baseline and follow up. PET images were corrected for partial volume effects. Regional C-PBR28 standardized uptake value ratios were determined using cerebellum as pseudo-reference region. PIB distribution volume ratios were determined using Logan graphical method with cerebellar reference. Results: Among patients, there was a significant correlation between increase in C-PBR28 binding and a decline in raw scores on Block Design, Trail Making part B, and Boston Naming Test. The strongest of such correlations occurred in prefrontal and parietal cortices (R > 0.68, P < 0.03). Percent change in C-PBR28 binding negatively correlated with %change in voxel count in prefrontal, parietal, and temporal cortices (R< -0.52, P< 0.03). Percent change in C-PBR28 correlated with %change in PIB binding in occipital cortex only (R 1⁄4 0.63, P 1⁄4 0.037). Controls showed no correlation between C-PBR28 binding and brain atrophy. Among patients, ApoE4 carriers (n 1⁄4 7) had greater increase in C-PBR28 binding than noncarriers (n1⁄4 4) in inferior temporal cortex (P1⁄4 0.042, Mann-Whitney U Test). Conclusions:TSPO increases with worsening clinical severity and cortical atrophy in AD. Larger studies are needed to better determine the relationship between TSPO, fibrillary amyloid load, and ApoE genotype. C-PBR28 PET has promise for monitoring AD progression.
Alzheimer’s disease (AD) is a neurodegenerative disease affecting the brain, with devastating imp... more Alzheimer’s disease (AD) is a neurodegenerative disease affecting the brain, with devastating implications for the afflicted individual due to the gradual loss of cognitive functions and independence in living, which finally result in death. AD generates an ever-growing expenditure for society from the need to care for the individuals who cannot live independently and patients who require institutionalized care.
European Journal of Nuclear Medicine and Molecular Imaging, 2016
Supervised Cluster Analysis for Automatic Extraction of Reference Region in Dynamic [F-18]THK5317... more Supervised Cluster Analysis for Automatic Extraction of Reference Region in Dynamic [F-18]THK5317 PET
[18F]FDG PETAS A MARKER OF DISEASE PROGRESSION IN ALZHEIMER’S DISEASE Elena Rodriguez-Vieitez, An... more [18F]FDG PETAS A MARKER OF DISEASE PROGRESSION IN ALZHEIMER’S DISEASE Elena Rodriguez-Vieitez, Antoine Leuzy, Konstantinos Chiotis, Laure Saint-Aubert, Ove Almkvist, Anders Wall, Agneta Nordberg, Karolinska Institutet, Stockholm, Sweden; Stockholm University, Stockholm, Sweden; 3 Uppsala University, Uppsala, Sweden; 4 Karolinska University Hospital Huddinge, Stockholm, Sweden. Contact e-mail: elena. rodriguez-vieitez@ki.se
healthy connectomes from (Raj, 2015), and tested on ADNI imaging data (Table 1). Table 2 shows th... more healthy connectomes from (Raj, 2015), and tested on ADNI imaging data (Table 1). Table 2 shows the sites of initial pathology as predicted by the model, which closely matches observed pathology. Fig 2 shows the spatiotemporal evolution of the model, which closely recapitulates expected progression of tau and amyloid pathology. Model predictions significantly match empirical ADNI data (Fig 3). Conclusions:Presented theoretical network model successfully recapitulated Braak and Thal stages, respectively, having made no assumptions beyond network spread and healthy regional metabolism patterns. It correctly predicts both the early sites of tau and amyloid pathology, as well as subsequent spread into wider circuits. The model could help explain why amyloid and tau appear to accumulate in certain brain sites, why they spread in such stereotyped fashion, and why the two appear to have different spatiotemporal patterns of spread, even though amyloid drives tau accumulation. The clinical utility of this model is in multi-modal modeling of tau, amyloid and atrophy in neuroimaging data.
Because a correlation between tau pathology and the clinical symptoms of Alzheimer disease (AD) h... more Because a correlation between tau pathology and the clinical symptoms of Alzheimer disease (AD) has been hypothesized, there is increasing interest in developing PET tracers that bind specifically to tau protein. The aim of this study was to evaluate tracer kinetic models for quantitative analysis and generation of parametric images for the novel tau ligand (S)-18F-THK5117. Methods: Nine subjects (5 with AD, 4 with mild cognitive impairment) received a 90-min dynamic (S)-18F-THK5117 PET scan. Arterial blood was sampled for measurement of blood radioactivity and metabolite analysis. Volume-of-interest (VOI)–based analysis was performed using plasma-input models; single-tissue and 2-tissue (2TCM) compartment models and plasma-input Logan and reference tissue models; and simplified reference tissue model (SRTM), reference Logan, and SUV ratio (SUVr). Cerebellum gray matter was used as the reference region. Voxel-level analysis was performed using basis function implementations of SRTM, reference Logan, and SUVr. Regionally averaged voxel values were compared with VOI-based values from the optimal reference tissue model, and simulations were made to assess accuracy and precision. In addition to 90 min, initial 40- and 60-min data were analyzed. Results: Plasma-input Logan distribution volume ratio (DVR)-1 values agreed well with 2TCM DVR-1 values (R2 = 0.99, slope = 0.96). SRTM binding potential (BPND) and reference Logan DVR-1 values were highly correlated with plasma-input Logan DVR-1 (R2 = 1.00, slope ≈ 1.00) whereas SUVr70–90-1 values correlated less well and overestimated binding. Agreement between parametric methods and SRTM was best for reference Logan (R2 = 0.99, slope = 1.03). SUVr70–90-1 values were almost 3 times higher than BPND values in white matter and 1.5 times higher in gray matter. Simulations showed poorer accuracy and precision for SUVr70–90-1 values than for the other reference methods. SRTM BPND and reference Logan DVR-1 values were not affected by a shorter scan duration of 60 min. Conclusion: SRTM BPND and reference Logan DVR-1 values were highly correlated with plasma-input Logan DVR-1 values. VOI-based data analyses indicated robust results for scan durations of 60 min. Reference Logan generated quantitative (S)-18F-THK5117 DVR-1 parametric images with the greatest accuracy and precision and with a much lower white-matter signal than seen with SUVr70–90-1 images.
Background: We previously demonstrated that radioligand binding to the 18 kDa translocator protei... more Background: We previously demonstrated that radioligand binding to the 18 kDa translocator protein (TSPO), a marker for inflammation, was greater in patients with Alzheimer’s disease (AD) than controls and correlated with clinical severity. We sought to determine the relationship between TSPO binding and progression of AD in a longitudinal study. Methods: Eleven patients with either AD or mild cognitive impairment and 8 age-matched controls underwent TSPO PET imaging with C-PBR28 at baseline and again after median follow up of 2.7 years. MRI, Pittsburgh Compound B (PIB) PET, and cognitive testing were also performed at baseline and follow up. PET images were corrected for partial volume effects. Regional C-PBR28 standardized uptake value ratios were determined using cerebellum as pseudo-reference region. PIB distribution volume ratios were determined using Logan graphical method with cerebellar reference. Results: Among patients, there was a significant correlation between increase in C-PBR28 binding and a decline in raw scores on Block Design, Trail Making part B, and Boston Naming Test. The strongest of such correlations occurred in prefrontal and parietal cortices (R > 0.68, P < 0.03). Percent change in C-PBR28 binding negatively correlated with %change in voxel count in prefrontal, parietal, and temporal cortices (R< -0.52, P< 0.03). Percent change in C-PBR28 correlated with %change in PIB binding in occipital cortex only (R 1⁄4 0.63, P 1⁄4 0.037). Controls showed no correlation between C-PBR28 binding and brain atrophy. Among patients, ApoE4 carriers (n 1⁄4 7) had greater increase in C-PBR28 binding than noncarriers (n1⁄4 4) in inferior temporal cortex (P1⁄4 0.042, Mann-Whitney U Test). Conclusions:TSPO increases with worsening clinical severity and cortical atrophy in AD. Larger studies are needed to better determine the relationship between TSPO, fibrillary amyloid load, and ApoE genotype. C-PBR28 PET has promise for monitoring AD progression.
Alzheimer’s disease (AD) is a neurodegenerative disease affecting the brain, with devastating imp... more Alzheimer’s disease (AD) is a neurodegenerative disease affecting the brain, with devastating implications for the afflicted individual due to the gradual loss of cognitive functions and independence in living, which finally result in death. AD generates an ever-growing expenditure for society from the need to care for the individuals who cannot live independently and patients who require institutionalized care.
European Journal of Nuclear Medicine and Molecular Imaging, 2016
Supervised Cluster Analysis for Automatic Extraction of Reference Region in Dynamic [F-18]THK5317... more Supervised Cluster Analysis for Automatic Extraction of Reference Region in Dynamic [F-18]THK5317 PET
[18F]FDG PETAS A MARKER OF DISEASE PROGRESSION IN ALZHEIMER’S DISEASE Elena Rodriguez-Vieitez, An... more [18F]FDG PETAS A MARKER OF DISEASE PROGRESSION IN ALZHEIMER’S DISEASE Elena Rodriguez-Vieitez, Antoine Leuzy, Konstantinos Chiotis, Laure Saint-Aubert, Ove Almkvist, Anders Wall, Agneta Nordberg, Karolinska Institutet, Stockholm, Sweden; Stockholm University, Stockholm, Sweden; 3 Uppsala University, Uppsala, Sweden; 4 Karolinska University Hospital Huddinge, Stockholm, Sweden. Contact e-mail: elena. rodriguez-vieitez@ki.se
healthy connectomes from (Raj, 2015), and tested on ADNI imaging data (Table 1). Table 2 shows th... more healthy connectomes from (Raj, 2015), and tested on ADNI imaging data (Table 1). Table 2 shows the sites of initial pathology as predicted by the model, which closely matches observed pathology. Fig 2 shows the spatiotemporal evolution of the model, which closely recapitulates expected progression of tau and amyloid pathology. Model predictions significantly match empirical ADNI data (Fig 3). Conclusions:Presented theoretical network model successfully recapitulated Braak and Thal stages, respectively, having made no assumptions beyond network spread and healthy regional metabolism patterns. It correctly predicts both the early sites of tau and amyloid pathology, as well as subsequent spread into wider circuits. The model could help explain why amyloid and tau appear to accumulate in certain brain sites, why they spread in such stereotyped fashion, and why the two appear to have different spatiotemporal patterns of spread, even though amyloid drives tau accumulation. The clinical utility of this model is in multi-modal modeling of tau, amyloid and atrophy in neuroimaging data.
Because a correlation between tau pathology and the clinical symptoms of Alzheimer disease (AD) h... more Because a correlation between tau pathology and the clinical symptoms of Alzheimer disease (AD) has been hypothesized, there is increasing interest in developing PET tracers that bind specifically to tau protein. The aim of this study was to evaluate tracer kinetic models for quantitative analysis and generation of parametric images for the novel tau ligand (S)-18F-THK5117. Methods: Nine subjects (5 with AD, 4 with mild cognitive impairment) received a 90-min dynamic (S)-18F-THK5117 PET scan. Arterial blood was sampled for measurement of blood radioactivity and metabolite analysis. Volume-of-interest (VOI)–based analysis was performed using plasma-input models; single-tissue and 2-tissue (2TCM) compartment models and plasma-input Logan and reference tissue models; and simplified reference tissue model (SRTM), reference Logan, and SUV ratio (SUVr). Cerebellum gray matter was used as the reference region. Voxel-level analysis was performed using basis function implementations of SRTM, reference Logan, and SUVr. Regionally averaged voxel values were compared with VOI-based values from the optimal reference tissue model, and simulations were made to assess accuracy and precision. In addition to 90 min, initial 40- and 60-min data were analyzed. Results: Plasma-input Logan distribution volume ratio (DVR)-1 values agreed well with 2TCM DVR-1 values (R2 = 0.99, slope = 0.96). SRTM binding potential (BPND) and reference Logan DVR-1 values were highly correlated with plasma-input Logan DVR-1 (R2 = 1.00, slope ≈ 1.00) whereas SUVr70–90-1 values correlated less well and overestimated binding. Agreement between parametric methods and SRTM was best for reference Logan (R2 = 0.99, slope = 1.03). SUVr70–90-1 values were almost 3 times higher than BPND values in white matter and 1.5 times higher in gray matter. Simulations showed poorer accuracy and precision for SUVr70–90-1 values than for the other reference methods. SRTM BPND and reference Logan DVR-1 values were not affected by a shorter scan duration of 60 min. Conclusion: SRTM BPND and reference Logan DVR-1 values were highly correlated with plasma-input Logan DVR-1 values. VOI-based data analyses indicated robust results for scan durations of 60 min. Reference Logan generated quantitative (S)-18F-THK5117 DVR-1 parametric images with the greatest accuracy and precision and with a much lower white-matter signal than seen with SUVr70–90-1 images.
Uploads