Early Validation of a Structural Magnetic Resonance Imaging Metric for Tracking Dementia‐Related Neurodegeneration

Abstract

AbstractBackgroundCurrent research is focused on noninvasive biomarkers of neurodegeneration that can detect dementia early, benchmark disease severity, monitor disease progression, and aid in testing the efficacy of therapeutic interventions utilizing a single number1–3. The objective of this study is to create and begin to validate an intuitively meaningful biomarker called the ADNeuro‐Score.MethodCognitively normal (CN) individuals and patients with a mild cognitive impairment (MCI) or dementia diagnosis were drawn from the Alzheimer's Disease Neuroimaging Initiative (ADNI)4. Eighty‐four cortical and subcortical regional volumes were estimated from T1‐weighted MR images using FreeSurfer5. To determine which regional volumes were associated with cognitive decline, we randomly selected 150 participants (N=50 each CN, MCI, and dementia) and performed an ANOVA with an alpha=0.05, Bonferroni corrected. A vector containing the resulting 41 regions, consistent with the existing literature6,7,8, was extracted for experimental and template cohorts (Table 1). To compute ADNeuro‐Score, a Hausdorff distance metric was used to estimate the differences between individual and average template vectors. ADNeuro‐Score was benchmarked using adjusted hippocampal volume (AHV), which is a NIA‐AA diagnostic biomarker for Alzheimer’s disease9, the most common form of dementia10. Both ADNeuro‐Score and AHV were harmonized for intracranial volume, age, sex, and scanner model11. Validation used an experimental cohort (N=929, mean age=72.67 years) and tested sensitivity to diagnosis using pairwise t‐tests and an alpha=0.001, Bonferroni corrected. Results were converted to a z‐score. We also tested association with disease severity, operationalized here as MMSE and ADASCog scores, using linear regression.ResultBoth ADNeuro‐Score and AHV differed between all three cognitive groups. ADNeuro‐Score (z=10.0) better differentiated MCI from dementia than AHV (z=8.9). AHV (z=7.3) was slightly better at distinguishing MCI from CN than ADNeuro‐Score (z=6.7). ADNeuro‐Score and AHV were similarly associated with MMSE (RADNS=0.41; RHA=0.41) and ADAS‐Cog (RADNS=0.49; RHA=0.47) scores.ConclusionThe newly developed ADNeuro‐Score seems to be a robust and reliable way to distinguish CN, MCI and AD patients and performs equivalently to AHV in predicting MMSE and ADAS‐Cog assessment scores. We hope ADNeuro‐Score will be highly specific to AD because of its focus on AD‐effected regions. Future research will determine if ADNeuro‐Score can improve the differential diagnosis of AD.