Alzheimer’s disease is a progressive and debilitating neurodegenerative disease; one in ten people aged 65 and older have it. As there is no cure for Alzheimer’s disease, early diagnosis is crucial so that mitigating treatments can be initiated as soon as possible.
Reduced cerebral blood flow and blood-brain barrier dysfunction have been reported in the early stages of Alzheimer’s disease. Screening for chronic cerebral hypoperfusion in individuals has been proposed for improving its early diagnosis. Such changes are modulated by the circadian rhythm, which causes blood pressure and flow to undergo 24-hour variability. Changes in these 24-hour profiles have previously been linked with the progression of Alzheimer’s disease. However, 24-hour measurements of cerebral blood flow are difficult to perform outside of a clinical setting.
In this study, we combined mathematical modelling of the systemic circulation with portable blood pressure monitoring and carotid ultrasound imaging to predict 24-hour cerebral blood flow profiles in individuals. One hundred and three participants (53 with mild cognitive impairment, 50 healthy controls) underwent model-assisted prediction of 24-hour CBF. Machine learning classifiers were then trained to assist in the diagnosis of Alzheimer’s disease. These classifiers attempt to find the best combination of available features in the data to separate the cases into two pre-determined classes – in this case, those with mild cognitive impairment and those who are cognitively healthy.
A model-enhanced machine learning classifier that used 24-hour cerebral blood flow values as classification features achieved a positive predictive value of 28% compared to the 14% achieved using standard neuropsychological tests that are used to diagnose cognitive impairment in memory clinics today. Our study provides some evidence that using non-invasive, affordable measurements of blood flow and computational modelling to predict 24-hour cerebral blood flow profiles is feasible and provides additional value in the early diagnosis of Alzheimer’s disease.