Principal Investigator: Dr Hyojung Paik
KISTI, South KoreaTags: 34303, Alzheimer's Disease, Dementia, genetics-architecture, Machine Learning
The main objective of this study is the development computational model that allows us the identification of a subgroup of disease patients with distinct outcomes, such as severe complications, comorbidity patterns and mortality.
Among neurodegenerative diseases, dementia and associated comorbid conditions including Alzheimer, hypertensive heart disease, and Parkinson will be primarily investigated.
Dementia is a broad category of brain disease that causes gradual decrease in the ability of cognition and daily functionality. Diverse etiologies have been suggested. Each of etiological features has been partially representing the heterogeneity of dementia patients; thus, resulting in the lack of understanding for the early diagnosis and tailored care toward the fewer complication and mortality. By scrutinizing the health record of participants in UK Biobank including genotypes, electronic medical records (EMRs) and brain images, we will develop a novel stratification methods of dementia. First, we will divide all participants of UK Biobank into the cases (i.e. Dementia cases) and controls (non-dementia) using EMRs, brain images and record of death registry. After then, all possible dementia cases will be sub-grouped based on the clinical phenotypes including late/early onsets, Alzheimer?s? disease in Dementia, and an onset of dementia following diverse comorbid diseases (e.g. stroke, hyperlipidemia, drug abuse etc.). In the third step, we will develop a stratification method which allows the identification of poor outcomes of dementia patients, such as severe mortality w/ dementia and genetic risk of dementia.
In principle, we will cover all participants of UK Biobank (up to 500,000) and relevant information including results of physical measurements, genetic screening, body DXA scan, brain/heart MRI images, and linked hospital records. DXA scan images will be used to estimate the obese state of an individual, such as upper body obese. Brain images will be used to measure the brain volume of a dementia patient. Heart images will be used to detect an association between cardiac function and hypertensive dementia patients. However, ~15000 participants have undergone images (MRI, body scans). So, we will initiate our study with partial set of images and other collected data for the full cohort and gradually migrate later collected images. Currently, the medical history partially involves inpatients level, such as HES. However, we may wish to request GP (General Practice) data when available to capture diagnoses of participants made in the community/outpatient visits.