Alzheimer's currently affects approximately 0.5 million Canadians, 5 million in North America, 25 million worldwide and will quadruple in prevalence by 2050 due to aging of the population. The social and financial costs are enormous. At present, it is estimated that the disease will affect 1 in 14 people that live to 80 years old and increase to 1 in 8 with aging baby-boomers.
As in most diseases, early treatment of patients, before they have too much irreversible degeneration of brain tissue, is likely to be more effective. However, it is difficult to test drugs in patients with early Alzheimer's disease, since only 10-15% of patients with symptoms of mild cognitive impairment (MCI) suggestive of possible Alzheimer's disease will actually go on to develop Alzheimer's per year. This percentage is too low to allow for reasonably sized clinical trials.
Our group has developed sophisticated image analysis tools to detect patterns of brain atrophy characteristic of Alzheimer's disease from standard magnetic resonance imaging (MRI) data, and have used this information to generate statistical models that combine MRI with clinical data in order 1) to help diagnose to Alzheimer's disease, and 2) predict which patients with MCI will rapidly go on to develop Alzheimer's disease.
In this project, we propose to extensively test and validate these novel tools on a MRI database from over 2000 subjects to realistically estimate the sensitivity, specificity and accuracy of our diagnostic and prognostic tools. Once validated, these tools will substantially facilitate the development of therapies for early Alzheimer's disease. They will make it feasible to perform drug trials in patients with MCI due to early Alzheimer's disease, and, looking forward, they will make it possible to select patients in the clinic for early treatment