Automated hippocampal subfield measures as predictors of ...

2 downloads 50444 Views 51KB Size Report
Contact e-mail: konstantinos. [email protected]. Background: Several radiotracers, which bind to brain fibrillar amyloid, have been developed and applied to different ...
Poster Presentations: P1 cholesterol, systolic blood pressure, diabetes and smoking) to cortical thickness. Cognition was evaluated using standardized composite scores of episodic memory and executive functions. Results: Among the FCRP factors, age (p ¼.023) and HDL cholesterol (p ¼.020) were correlated with cortical thickness. Both increased age and low HDL cholesterol were associated with thinner frontal and temporal cortex; low HDL was also associated with thinner parietal cortex. When subjects were divided by amyloid status, both high vascular risk (FCRP total score) and low HDL cholesterol were associated with thinner frontal and temporal cortex in both PIB- and PIB+ subjects. In PIB+ subjects, high vascular risk and low HDL cholesterol were also associated with thinner parietal cortex (Figure). Finally, higher beta-amyloid and higher vascular risk were associated with lower memory performance and cortical thickness mediated these relationships. Conclusions: While vascular risk and HDL cholesterol affect cortical thickness in frontotemporal cortex regardless of amyloid status, individuals with beta-amyloid deposition show further involvement of parietal cortex. Furthermore, the data suggest that cortical thickness mediates the impact of vascular risk and beta-amyloid on memory function.

P1-320

COMPARISON OF PiB-PET DATA TO FLORBETAPIR-PET DATA ACQUIRED FROM COHORTS AT DIFFERENT RESEARCH SITES

Konstantinos Chiotis1, Stephen Carter2, Agneta Nordberg3, 1Karolinska Institute, Huddinge, Sweden; 2Karolinska Institute, Stockholm, Sweden; 3 Karolinska Institute, Huddinge, Sweden. Contact e-mail: konstantinos. [email protected] Background: Several radiotracers, which bind to brain fibrillar amyloid, have been developed and applied to different patient cohorts. Recent studies have demonstrated that in the same patients [11C]Pittsburgh compound-B (PIB) and [18F]florbetapir positron emission (PET) scans acquired in close temporal proximity are highly correlated. However, it remains to be determined if PIB and florbetapir scans acquired from different patients and at different research sites are as highly correlated. Methods: PIB-PET data as described before (Nordberg et al, 2012) from 51 healthy controls (HC), 72 mild cognitive impairment (MCI) and 90 Alzheimer’s disease (AD) patients acquired at 5 different European centres was combined and analysed using a PIB-PET population template and atlas. Correspondingly, a population of 51 HC, 72 MCI and 84 AD raw florbetapir data matched for gender and age was downloaded from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database; and a florbetapir-PET population template and atlas was created using the same methodology adopted for the PIB template. A HC derived cut-off of 1.41 was used for the PIB data and a HC derived cutoff for the florbetapir data was established. The prevalence of amyloid ’positive’ scans from each population was determined and compared. The linear equation derived from plotting the HC PIB data against florbetapir data (and vice versa) was used to convert the PIB cut-off into florbetapir units (and vice versa). Results: From the PIB data the AD (mean age 70.1 6 8.1years), MCI patients (mean age 67.5 6 8years) and HCs (mean age 67.4 6 6years) were compared to the florbetapir data of AD (mean age 71.7 6 6.8years), MCI (mean age 67.5 6 7.2years) and HCs (mean age 68.2 6 5.1years). There was no significant difference in age and gender distribution. Conclusions: The data will demonstrate the comparability of amyloid imaging in two unrelated, but matched, populations using two different fibrillar amyloid PET tracers. In addition, it will provide further insight into the possible variations in regional amyloid load between different cohorts regarding age, sex, ApoE genotype, patients’ education, clinical symptoms and progression of disease. P1-321

AUTOMATED HIPPOCAMPAL SUBFIELD MEASURES AS PREDICTORS OF CONVERSION FROM MILD COGNITIVE IMPAIRMENT TO ALZHEIMER’S DISEASE IN TWO INDEPENDENT COHORTS

Wasim Khan1, Eric Westman2, Nigel Jones3, Lars-Olaf Wahlund4, Patrizia Mecocci5, Bruno Vellas6, Magda Tsolaki7, Iwona Kloszewska8,

P275

Hilkka Soininen9, Christian Spenger10, Simon Lovestone11, J-Sebastian Muehlboeck3, Andy Simmons12, 1King’s College London, London, United Kingdom; 2Karonlinska Institute, Stockholm, Sweden; 3King’s College London, London, United Kingdom; 4Karolinska Institute, Stockholm, Sweden; 5University of Perugia, Perugia, Italy; 6Clinic of Internal Medicine and Gerontology, Toulouse, France; 7Aristotle University of Thessaloniki, Thessaloniki, Greece; 8Medical University of Lodz, Lodz, Poland; 9Kuopio University and University Hospital, Kuopio, Finland; 10Karolinska University Hospital, Stockholm, Sweden; 11King’s College London, London, United Kingdom; 12King’s College London, Institute of Psychiatry, London, United Kingdom. Contact e-mail: [email protected] Background: Hippocampal atrophy has been frequently observed in Alzheimer’s disease (AD) and has been demonstrated in Mild Cognitive Impairment (MCI) subjects bearing an increased risk of future conversion to the disease. However, hippocampal atrophy in AD is non-uniform and some sub-regions of the hippocampus may be more strongly affected than others. As a result, segmenting the hippocampal subfields has become an attractive approach for improving early diagnosis and predicting future disease progression. Methods: 1069 subjects were selected from the European AddNeuroMed study and the US-based Alzheimer’s Disease Neuroimaging Initiative (ADNI) study. The Freesurfer 5.1.0 image analysis pipeline was used for the automated segmentation of the hippocampal subfields. Data from both cohorts was used to (1) investigate patterns of subfield volume loss in MCI and AD subjects, (2) determine the pattern of subfield volume ^ genotype, and neuroloss in relation to age, gender, education, APOE O‘4 psychological test scores, (3) compare combined subfield volumes to hippocampal volume alone in discriminating between AD and healthy controls (HC), and predicting future MCI conversion to AD at 12 months. Orthogonal partial least squares to latent structures (OPLS) was used to train models on AD and HC subjects using one cohort for training and the other for testing. The combined AddNeuroMed and ADNI cohorts were subsequently used to predict future MCI conversion to AD at one year follow up. Results: MANCOVA and linear regression analyses showed multiple subfield volumes including CA1, subiculum and presubiculum were atrophied in AD ^ genotype, and MCI and were related to age, gender, education, APOE O‘4 and neuropsychological test scores. For classifying AD from HC, combined subfield volumes achieved comparable classification accuracy (81.7%) to total hippocampal (80.7%), subiculum (81.2%) and presubiculum (80.6%) volume. Further analysis showed combined subfield volumes and presubiculum volume were more accurate (81.1%) in predicting MCI subjects converting to AD than total hippocampal volume (76.7%). Conclusions: Hippocampal subfield volumetry appears to be superior to total hippocampal volume in predicting MCI conversion to AD. These findings are of interest for both research and clinical applications and suggest that hippocampal subfield measurement using high field ultra-high resolution MRI may hold future promise. P1-322

MYELIN INTEGRITY, COGNITIVE FUNCTION AND HYPERTENSION IN MCI AND ALZHEIMER’S DISEASE

Athene Lee1, Julia Rao2, Steven van Huiden3, Stephen Correia4, Jonathan O’Muircheartaigh4, Sean Deoni4, Stephen Salloway4, Paul Malloy4, 1Alpert Medical School of Brown University, Providence, Rhode Island, United States; 2Alpert Medical School of Brown University, Providence, Rhode Island, United States; 3Vrije Universities, Amsterdam, Netherlands; 4Brown University, Providence, Rhode Island, United States. Contact e-mail: [email protected] Background: Declines in white matter integrity including white matter hyperintensities (WMH), alterations in diffusion-tensor imaging (DTI) metrics , and myelin degradation, have been associated with cognitive decline in aging and Alzheimer’s disease (AD). Recent research shows elevated blood pressure is associated with disrupted white matter integrity (Salat et al., 2012). The current study examines whether hypertension affects the association between myelin integrity in normal-appearing white matter (NAWM) and cognitive function in patients with mild cognitive impairment (MCI) and AD. Methods: Among 30 participants (aged 55-87) with either MCI