Diagnosing preclinical AD promises to delay or prevent the development of dementia
Current imaging and molecular/cognitive biomarker techniques enabling an asymptomatic diagnosis of AD are expensive, time-consuming and difficult to administer or of limited sensitivity. However, two new techniques that can be carried out quickly and easily at home or in the physician’s office and repeated as necessary promise to change the natural history of AD through effective intervention and monitoring of patients with asymptomatic AD before the neurodegenerative pathology becomes irreversible.
Digitization of the classic clock drawing test
Ability to draw a clock face deteriorates with advancing AD until the clock face is almost unrecognizable. This ability is evaluated in the classic clock drawing task using pen and paper. Using a digital pen, however, enables a much more informative analysis of a variety of cognitive features associated with producing the drawing in real-time, including:
- drawing efficiency
- information processing
- motor control
- spatial reasoning
The DCTClock takes only 2 minutes to complete
This is the basis for the DCTclock described at AAIC 2019.1 Machine learning techniques have been used to derive a composite score of ranging from 0 to 100 based on the data obtained from digital clocks drawn by thousands of individuals in different diagnostic groups.
The clinical potential of the DCTclock was investigated in 159 individuals who were cognitively normal and 24 who had mild cognitive impairment (MCI), and was found to be highly correlated with the Preclinical Alzheimer’s Cognitive Composite (PACC5), which is sensitive to amyloid burden in asymptomatic AD.
The diagnostic discriminability provided by the DCTclock between individuals who were cognitively normal and those with MCI was 37% higher compared to traditional clock scoring methods and comparable to the PACC5.1 However, whereas the PACC5 takes 30 minutes to complete, the DCTClock takes 2 minutes.
The Integrated Cognitive Assessment (ICA) — a 5-minute, self-administered, computerized test
The ICA is based on grayscale animal shape recognition and is independent of language, cultural background and education. It targets cognitive domains affected in the initial stages of AD before the onset of memory symptoms and uses artificial intelligence (AI) to analyze high-dimensional clinical and demographic data to continuously improve its predictive power.
Head-to-head studies have been carried out with widely used cognitive assessments in individuals with mild AD, MCI, and cognitive impairment secondary to multiple sclerosis (MS).
The ICA can be accessed by large populations and used as an accurate endpoint
The ICA had convergent validity with other cognitive assessments, and achieved high accuracy in detecting cognitive impairment, discriminating MCI from mild AD with 96% accuracy.2
- engaged cognitive areas among the earliest affected by tau pathology in asymptomatic AD on functional magnetic resonance
- strongly correlated with neurofilament light chain protein and severity of cognitive impairment
- demonstrated excellent test-retest reliability and no evidence of learning bias3