MRI and Dementia: A multi-centre study of autopsy proven cases
Research type
Research Database
IRAS ID
214956
Contact name
Suzie Barker
Contact email
Research summary
Predicting Autopsy Dementia Diagnosis Using MRI
REC name
London - South East Research Ethics Committee
REC reference
16/LO/2046
Date of REC Opinion
3 Jan 2017
REC opinion
Favourable Opinion
Data collection arrangements
All data will be stored at UCL on secure servers protected by comprehensive firewall systems. Data are backed up on a daily basis (on site and off-site). Data will be collected from clinical summaries and post-mortem histology reports and will include demographic information, clinical information and post-mortem diagnoses. No identifiable data will be stored in the database - all cases will be allocated a unique identifier. The CI and delegated team members will hold a link to the identifiable data collected at UCL but will not have access to identifiable data from other centres. Other centres will hold the link to the unique identifiers used in the database and their cases. Nominated individuals from these centres will apply for access to the database to enable them to upload scans and anonymised information.
A delegation log, detailing individual access rights, will be held by the CI.
Researchers from within the host site and from external centres may apply to receive fully anonymised data to be used for specified projects. The CI will oversee the approval of these applications.Research programme
The existing database has resulted in the development of the largest post mortem proven dementia-imaging cohort in Europe (n > 2050 cases including Alzheimer’s disease, dementia with Lewy bodies and a range of frontotemporal lobar degeneration pathologies). As well as collating the data from three centres we have also successfully designed, built and deployed a web application to enable the safe and efficient storage of the data, with the potential to share this valuable data with the wider scientific community. Using a wide range of image analysis techniques, from simple visual assessment to more complex machine learning, various aspects of the dataset have been analysed with a view to maximising the diagnostic value of structural imaging routinely acquired in clinical assessment of dementia. Outputs from this include: - A PhD thesis - Two review papers highlighting the utility of visual image assessment and visual rating scales in routine diagnostic imaging - Three research papers looking at the diagnostic utility of visual rating scales, patterns of atrophy, and brain substructure volume loss associated with different dementia pathologies. Additional novel outputs include a visual rating training tool aimed at clinicians, and a public engagement/citizen science project. The visual rating tool will provide training in the application of visual rating scales designed to assess key regions of the brain affected by the neurodegenerative dementias. Clinicians be able to test and audit their technique against well-established visual rating experts, who have generated a gold-standard set of ratings.Using the same gold-standard visual ratings, we have also piloted a citizen science project. The aim of this project is to further highlight the value and simplicity of visual image assessment to the wider clinical community, whilst engaging the general public in dementia research and increasing their understanding about these devastating diseases.
Research database title
Predicting Autopsy Dementia Diagnosis Using MRI
Establishment organisation
Dementia Research Centre
Establishment organisation address
Institute of Neurology
Queen Square
London
WC1N 3BG