- Email: [email protected]
AN EARLY AND LATE PEAK IN MICROGLIAL ACTIVATION IN ALZHEIMER’S DISEASE TRAJECTORY: A LONGITUDINAL PET STUDY
Poster Presentations: Sunday, July 24, 2016 Results: Using our PSQI criterion, 53% (n¼30) of our MCI group were classified as sleep-disturbed. Whereas the total group of MCI subjects and controls demonstrated no significant differences, sleep-disturbed MCIs demonstrated increased connectivity between temporal and parietal regions, and decreased connectivity between the prefrontal cortex and the temporoparietal junction relative to sleep-disturbed controls. Relative to those MCIs without sleep disturbance, sleep-disturbed MCI participants demonstrated significantly diminished DMN connectivity between temporal and parietal regions, a finding that was particularly pronounced in amnestic MCI. Conclusions: Subjective sleep disturbance in MCI is associated with distinct alterations in DMN functional connectivity in brain regions underpinning salient memory and sleep systems. A follow-up study currently in progress builds on these results via objective measurement of sleep parameters using actigraphy. Those with MCI and higher levels of wake after sleep onset (WASO) demonstrated DMN functional connectivity deficits across a number of brain regions, including several of those identified in MCIs with subjective sleep disturbance.
P1-282
P527
Figure 1. Regions with the highest OR values. Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus, Putamen and the Nucleus Accumbens are indicated.
VOXEL-WISE LOGISTIC REGRESSION IMPROVES PREDICTION ACCURACY FOR DEVELOPING ALZHEIMER’S DISEASE
Sulantha S. Mathotaarachchi1,2, Tharick A. Pascoal2, Monica Shin2, Andrea Lessa Benedet3, Min Su Kang2, Thomas Beaudry4, Vladimir S. Fonov5, Serge Gauthier2, Aurelie Labbe1, Pedro Rosa-Neto6,7, 1 McGill University, Montreal, QC, Canada; 2McGill University Research Centre for Studies in Aging, Verdun, QC, Canada; 3Translational Neuroimaging Laboratory, McGill University Research Centre for Studies in Aging, Montreal, QC, Canada; 4McGill Centre for Studies in Aging, Montreal, QC, Canada; 5Montreal Neurological Institute, McGill University, Montreal, QC, Canada; 6Douglas Hospital Research Centre, Verdun, QC, Canada; 7Translational Neuroimaging Laboratory - McGill University, Verdun, QC, Canada. Contact e-mail: [email protected] Background: Predicting clinical trajectories in populations at risk, permit enrich populations for disease modifying clinical trials. Recent developments in machine learning techniques have enabled us to achieve highly accurate predictions in multiple clinical applications. Here we demonstrate a novel data driven method to improve the accuracy of a Random Forest based classifier to predict the development of dementia. Methods: [18F]Florbetapir images were acquired for 275 MCI individuals from the ADNI cohort and the images were processed using an established PET image processing pipeline. Regional SUVr values were extracted from brain regions such as the Angular Gyrus, Supramarginal Gyrus, Posterior Cingulate Cortex and Precuneus which were identified to have a significant amyloid accumulation based on existing literature. 70%(192) and 30% of the population were labeled as the training set and the testing set, respectively. Data driven method included a Voxel-wise logistic regression analysis using the training population to identify anatomically significant brain regions with highest odds-ratios (ORs) to develop dementia. Two Random Forest based predictors were trained with the regional SUVr values based on literature and the data driven method respectively. Their performances were measured against the testing population. Results: Voxel-wise logistic regression analysis indicated that brain regions including Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus,
Figure 2. ROC curves and features importance for the two predictors. a) predictor with the regions based on existing literature b) predictor based on Voxel-wise logistic regression.
Putamen and the Nucleus Accumbens have the highest OR values for a unit SD increase of [18F]Florbetapir [Figure 1]. The Random Forest predictor trained using regions based on literature achieved 79% and 78% as validation and testing accuracy (0.89 AUC) while the predictor based on the data driven method achieved 84% for both validation and testing accuracy (0.91 AUC). The Temporal Occipital junction, Mid Temporal Sulci and Mid Frontal Sulci regions indicated the highest contribution in predicting the development of dementia [Figure 2]. Conclusions: The data driven method using Voxel-wise logistic regression analysis have increased the accuracy of the Random Forest predictor and have outperforms the methods developed in previously published literature and can be a utilized as a valuable clinical framework. P1-283
AN EARLY AND LATE PEAK IN MICROGLIAL ACTIVATION IN ALZHEIMER’S DISEASE TRAJECTORY: A LONGITUDINAL PET STUDY
Zhen Fan1, David J. Brooks1, Aren Okello2, Paul Edison1, 1Imperial College London, London, United Kingdom; 2Imperial College London, London, United Kingdom. Contact e-mail: [email protected] Background: Beta amyloid deposition, neuroinflammation and tau
tangle formation all play a significant role in Alzheimer’s disease
P528
Poster Presentations: Sunday, July 24, 2016
(AD). However, the relative role of these processes in driving disease progression is still unclear. Methods: In this longitudinal study we have evaluated the relationship between amyloid fibrillar load and microglial activation at baseline and follow up in amnestic MCI subjects. 30 subjects (8 MCIs, 8 ADs and 14 controls) underwent [11C](R)PK11195, [11C]PIB PET and volumetric MRI. MCI subjects were followed up after 1464 months, and findings compared with those of AD subjects. Region of interest analysis and SPM were used to determine longitudinal alterations. Correlations between levels of microglial activation and amyloid deposition at a voxel level were assessed with Biological Parametric Mapping while hippocampal volume changes were analysed using FreeSurfer. Results: Baseline microglial activation was raised by 41% in the MCI cohort compared with controls. There was a longitudinal reduction of 18% in microglial activation in MCI subjects over 14 months, which was associated with a mild elevation in fibrillar amyloid load. Cortical clusters of microglial activation and amyloid deposition spatially overlapped in the MCI subjects. Baseline microglial activation was raised by 36% in AD subjects compared with controls. This increased further by 10% after 14 months and was correlated with progressive reduction in hippocampal volume. Conclusions: This study demonstrated that while neuroinflammation is initially present in MCI subjects, this diminishes over time, while in AD subjects microglial activation continues to rise giving two peaks of glial activation in the disease trajectory suggesting that microglial activation is a dynamic process, which could be detected in vivo. We speculate that activated microglia in MCI subjects initially adopt a protective M2 phenotype but change to a cidal M1 phenotype as disease progresses resulting in two peaks. This results suggests that antimicroglial agents may have their most beneficial effect in later stages of the disease when they target the proinflammatory phenotypes. P1-284
data available selected from the Amsterdam Dementia Cohort. The main outcome parameter was clinical progression (i.e., CDR change 0.5). Single-subject networks were based on grey matter segmentations. We calculated the degree, connectivity density, path length, clustering, and small world parameters. All measures were Z transformed and inverted. ANCOVAs were used for cross-sectional comparisons of disease outcome and baseline diagnosis. Separate Cox proportional hazard models were fitted for each connectivity predictor for time to dementia onset and corrected for age, gender, whole brain volume and scanner. Results: After 2.2 (IQR 1.3–3.1) years 122 (55%) people showed clinical progression (N¼23 SCD; N¼99 MCI). Normalized clustering and
Figure 1. Linear trend analysis of normalized clustering coefficient (gamma) and the small world parameter. Both these network parameters had highest values in patients with subjective cognitive decline who remained stable overtime, and the lowest in MCI patients who progressed to dementia.
GREY MATTER CONNECTIVITY IS ASSOCIATED WITH CLINICAL PROGRESSION IN NONDEMENTED, AMYLOID POSITIVE PATIENTS
Betty M. Tijms1,2, Mara ten Kate2, Sander C. J. Verfaillie1, Alida A. Gouw1,3, Charlotte E. Teunissen1,3, Frederik Barkhof1,4, Philip Scheltens1,3, Wiesje M. van der Flier1,5, 1Neuroscience Campus Amsterdam, Amsterdam, Netherlands; 2Alzheimer Center and Department of Neurology, Neuroscience Campus Amsterdam, VU University Medical Center, Amsterdam, Netherlands; 3VU University Medical Center, Amsterdam, Netherlands; 4Department of Radiology and Nuclear Medicine, Neuroscience Campus Amsterdam, VU University Medical Center, Amsterdam, Netherlands; 5Alzheimer Center, VU University Medical Center, Amsterdam, Netherlands. Contact e-mail: [email protected] Background: Accumulation of amyloid in the brain is among the
first changes leading to Alzheimer’s disease (AD), yet its prognostic value is limited. Grey matter connectivity is disrupted in AD, and these disruptions are associated with worse cognitive functioning. We studied whether grey matter connectivity has prognostic value, by comparing amyloid positive patients with subjective cognitive decline (SCD) and mild cognitive impairment (MCI) and analyzing its association with clinical progression. Methods: CODA (COnnectivity in DementiA) includes 222 non-demented patients (62 (28%) SCD; 160 (78%) MCI; 109 (49%) female, 68 6 8 years; 28 6 2.4 MMSE) with abnormal amyloid CSF (<640 pg/ml), T1weighted structural MRI and over 1 year annual follow up
Figure 2. Survival curves for the time to dementia onset in subjects with subjective cognitive impairment or mild cognitive impairment due to Alzheimer’s disease with separate lines for clustering coefficient tertiles: blue represents subjects with the most lowest values, green represents intermediate values and red line represents subjects with the highest values.
P1-282
P527
Figure 1. Regions with the highest OR values. Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus, Putamen and the Nucleus Accumbens are indicated.
VOXEL-WISE LOGISTIC REGRESSION IMPROVES PREDICTION ACCURACY FOR DEVELOPING ALZHEIMER’S DISEASE
Sulantha S. Mathotaarachchi1,2, Tharick A. Pascoal2, Monica Shin2, Andrea Lessa Benedet3, Min Su Kang2, Thomas Beaudry4, Vladimir S. Fonov5, Serge Gauthier2, Aurelie Labbe1, Pedro Rosa-Neto6,7, 1 McGill University, Montreal, QC, Canada; 2McGill University Research Centre for Studies in Aging, Verdun, QC, Canada; 3Translational Neuroimaging Laboratory, McGill University Research Centre for Studies in Aging, Montreal, QC, Canada; 4McGill Centre for Studies in Aging, Montreal, QC, Canada; 5Montreal Neurological Institute, McGill University, Montreal, QC, Canada; 6Douglas Hospital Research Centre, Verdun, QC, Canada; 7Translational Neuroimaging Laboratory - McGill University, Verdun, QC, Canada. Contact e-mail: [email protected] Background: Predicting clinical trajectories in populations at risk, permit enrich populations for disease modifying clinical trials. Recent developments in machine learning techniques have enabled us to achieve highly accurate predictions in multiple clinical applications. Here we demonstrate a novel data driven method to improve the accuracy of a Random Forest based classifier to predict the development of dementia. Methods: [18F]Florbetapir images were acquired for 275 MCI individuals from the ADNI cohort and the images were processed using an established PET image processing pipeline. Regional SUVr values were extracted from brain regions such as the Angular Gyrus, Supramarginal Gyrus, Posterior Cingulate Cortex and Precuneus which were identified to have a significant amyloid accumulation based on existing literature. 70%(192) and 30% of the population were labeled as the training set and the testing set, respectively. Data driven method included a Voxel-wise logistic regression analysis using the training population to identify anatomically significant brain regions with highest odds-ratios (ORs) to develop dementia. Two Random Forest based predictors were trained with the regional SUVr values based on literature and the data driven method respectively. Their performances were measured against the testing population. Results: Voxel-wise logistic regression analysis indicated that brain regions including Orbitofrontal Cortex, Mid Frontal Sulci, Mid Temporal Sulci, Temporal Occipital junction, PCC, Angular Gyrus, Precuneus,
Figure 2. ROC curves and features importance for the two predictors. a) predictor with the regions based on existing literature b) predictor based on Voxel-wise logistic regression.
Putamen and the Nucleus Accumbens have the highest OR values for a unit SD increase of [18F]Florbetapir [Figure 1]. The Random Forest predictor trained using regions based on literature achieved 79% and 78% as validation and testing accuracy (0.89 AUC) while the predictor based on the data driven method achieved 84% for both validation and testing accuracy (0.91 AUC). The Temporal Occipital junction, Mid Temporal Sulci and Mid Frontal Sulci regions indicated the highest contribution in predicting the development of dementia [Figure 2]. Conclusions: The data driven method using Voxel-wise logistic regression analysis have increased the accuracy of the Random Forest predictor and have outperforms the methods developed in previously published literature and can be a utilized as a valuable clinical framework. P1-283
AN EARLY AND LATE PEAK IN MICROGLIAL ACTIVATION IN ALZHEIMER’S DISEASE TRAJECTORY: A LONGITUDINAL PET STUDY
Zhen Fan1, David J. Brooks1, Aren Okello2, Paul Edison1, 1Imperial College London, London, United Kingdom; 2Imperial College London, London, United Kingdom. Contact e-mail: [email protected] Background: Beta amyloid deposition, neuroinflammation and tau
tangle formation all play a significant role in Alzheimer’s disease
P528
Poster Presentations: Sunday, July 24, 2016
(AD). However, the relative role of these processes in driving disease progression is still unclear. Methods: In this longitudinal study we have evaluated the relationship between amyloid fibrillar load and microglial activation at baseline and follow up in amnestic MCI subjects. 30 subjects (8 MCIs, 8 ADs and 14 controls) underwent [11C](R)PK11195, [11C]PIB PET and volumetric MRI. MCI subjects were followed up after 1464 months, and findings compared with those of AD subjects. Region of interest analysis and SPM were used to determine longitudinal alterations. Correlations between levels of microglial activation and amyloid deposition at a voxel level were assessed with Biological Parametric Mapping while hippocampal volume changes were analysed using FreeSurfer. Results: Baseline microglial activation was raised by 41% in the MCI cohort compared with controls. There was a longitudinal reduction of 18% in microglial activation in MCI subjects over 14 months, which was associated with a mild elevation in fibrillar amyloid load. Cortical clusters of microglial activation and amyloid deposition spatially overlapped in the MCI subjects. Baseline microglial activation was raised by 36% in AD subjects compared with controls. This increased further by 10% after 14 months and was correlated with progressive reduction in hippocampal volume. Conclusions: This study demonstrated that while neuroinflammation is initially present in MCI subjects, this diminishes over time, while in AD subjects microglial activation continues to rise giving two peaks of glial activation in the disease trajectory suggesting that microglial activation is a dynamic process, which could be detected in vivo. We speculate that activated microglia in MCI subjects initially adopt a protective M2 phenotype but change to a cidal M1 phenotype as disease progresses resulting in two peaks. This results suggests that antimicroglial agents may have their most beneficial effect in later stages of the disease when they target the proinflammatory phenotypes. P1-284
data available selected from the Amsterdam Dementia Cohort. The main outcome parameter was clinical progression (i.e., CDR change 0.5). Single-subject networks were based on grey matter segmentations. We calculated the degree, connectivity density, path length, clustering, and small world parameters. All measures were Z transformed and inverted. ANCOVAs were used for cross-sectional comparisons of disease outcome and baseline diagnosis. Separate Cox proportional hazard models were fitted for each connectivity predictor for time to dementia onset and corrected for age, gender, whole brain volume and scanner. Results: After 2.2 (IQR 1.3–3.1) years 122 (55%) people showed clinical progression (N¼23 SCD; N¼99 MCI). Normalized clustering and
Figure 1. Linear trend analysis of normalized clustering coefficient (gamma) and the small world parameter. Both these network parameters had highest values in patients with subjective cognitive decline who remained stable overtime, and the lowest in MCI patients who progressed to dementia.
GREY MATTER CONNECTIVITY IS ASSOCIATED WITH CLINICAL PROGRESSION IN NONDEMENTED, AMYLOID POSITIVE PATIENTS
Betty M. Tijms1,2, Mara ten Kate2, Sander C. J. Verfaillie1, Alida A. Gouw1,3, Charlotte E. Teunissen1,3, Frederik Barkhof1,4, Philip Scheltens1,3, Wiesje M. van der Flier1,5, 1Neuroscience Campus Amsterdam, Amsterdam, Netherlands; 2Alzheimer Center and Department of Neurology, Neuroscience Campus Amsterdam, VU University Medical Center, Amsterdam, Netherlands; 3VU University Medical Center, Amsterdam, Netherlands; 4Department of Radiology and Nuclear Medicine, Neuroscience Campus Amsterdam, VU University Medical Center, Amsterdam, Netherlands; 5Alzheimer Center, VU University Medical Center, Amsterdam, Netherlands. Contact e-mail: [email protected] Background: Accumulation of amyloid in the brain is among the
first changes leading to Alzheimer’s disease (AD), yet its prognostic value is limited. Grey matter connectivity is disrupted in AD, and these disruptions are associated with worse cognitive functioning. We studied whether grey matter connectivity has prognostic value, by comparing amyloid positive patients with subjective cognitive decline (SCD) and mild cognitive impairment (MCI) and analyzing its association with clinical progression. Methods: CODA (COnnectivity in DementiA) includes 222 non-demented patients (62 (28%) SCD; 160 (78%) MCI; 109 (49%) female, 68 6 8 years; 28 6 2.4 MMSE) with abnormal amyloid CSF (<640 pg/ml), T1weighted structural MRI and over 1 year annual follow up
Figure 2. Survival curves for the time to dementia onset in subjects with subjective cognitive impairment or mild cognitive impairment due to Alzheimer’s disease with separate lines for clustering coefficient tertiles: blue represents subjects with the most lowest values, green represents intermediate values and red line represents subjects with the highest values.