The Limitations of Using Cognitive Cutoff Scores for Enrollment in Alzheimer Trials

Temporal association of neuropsychological test performance using unsupervised learning reveals a distinct signature of Alzheimer's disease status

INTRODUCTION

Subtle cognitive alterations that precede clinical evidence of cognitive impairment may help predict the progression to Alzheimer's disease (AD). Neuropsychological (NP) testing is an attractive modality for screening early evidence of AD.

METHODS

Longitudinal NP and demographic data from the Framingham Heart Study (FHS; N = 1696) and the National Alzheimer's Coordinating Center (NACC; N = 689) were analyzed using an unsupervised machine learning framework. Features, including age, logical memory-immediate and delayed recall, visual reproduction-immediate and delayed recall, the Boston naming tests, and Trails B, were identified using feature selection, and processed further to predict the risk of development of AD.

RESULTS

Our model yielded 83.07 ± 3.52% accuracy in FHS and 87.57 ± 1.19% accuracy in NACC, 80.52 ± 3.93%, 86.74 ± 1.63% sensitivity in FHS and NACC respectively, and 85.63 ± 4.71%, 88.41 ± 1.38% specificity in FHS and NACC, respectively.

DISCUSSION

Our results suggest that a subset of NP tests, when analyzed using unsupervised machine learning, may help distinguish between high- and low-risk individuals in the context of subsequent development of AD within 5 years. This approach could be a viable option for early AD screening in clinical practice and clinical trials.

The Utility of Episodic Memory Cut-Off Scores for Inclusion in Clinical Trials for Early Symptomatic Alzheimer Disease

Correctly diagnosing early symptomatic Alzheimer disease in the context of multinational clinical trials poses a significant challenge. Subjective complaints of memory are fairly ubiquitous in an older population, and establishing the presence of definitive cognitive decline from clinical assessments is difficult. Most such trials have adopted the use of standardized episodic memory measures as an inclusion criterion, typically setting the cutoff at one standard deviation below age normal means. This is useful in terms of establishing the presence of an objective impairment of memory, thereby excluding subjects with purely subjective complaints and increasing the probability that clinical outcome measures will be sensitive to disease progression. Further demographic adjustments are unnecessary as other demographic variables are not strongly associated with memory performance, are difficult to equate across cultures, and will not eventuate in reduced screen fail rates and would be challenging to implement. Not all episodic memory measures are equivalent for this purpose, however, and existing data suggest significant variability in terms of specificity for identifying true (i.e., amyloid positive) early symptomatic AD.