Abstract
Objectives: Developing sensitive imaging biomarkers is critical for capturing early brain alterations in Alzheimer’s disease. This exploratory study aimed to evaluate how multiparametric magnetic resonance imaging (MRI) measures, including chemical exchange saturation transfer (CEST) imaging, can detect early cognitive impairment and to compare their diagnostic performance across multiple brain regions. Methods: 18 cognitively normal participants and 18 patients with mild cognitive impairment or mild dementia were recruited to undergo CEST and other MRI sequences. Imaging metrics included CEST measures (amide proton transfer-weighted or APTw, APT#, nuclear Overhauser enhancement or NOE#), alongside conventional measures (T1, T2) and advanced measures (magnetization transfer ratio or MTR, quantitative susceptibility mapping or QSM, apparent diffusion coefficient or ADC). Quantitative analyses were performed in the hippocampus, caudate, putamen, posterior cingulate cortex, and amygdala. Associations between MRI measures and cognition were evaluated using Pearson correlation analysis, and the diagnostic performance of single-parameter and multiparameter models were assessed using logistic regression and receiver operating characteristic analysis. Results: Among all evaluated MRI measures, APTw, APT#, NOE#, MTR, T1, and QSM demonstrated statistically significant group differences (cognitively normal vs. mild cognitive impairment or mild dementia), with medium to very large effect sizes in at least one region. Across all evaluated regions, APT# and NOE# signals were consistently and positively associated with clinical dementia severity (based on the Clinical Dementia Rating), whereas MTR showed consistently negative associations. In contrast, APTw, T1, T2, QSM, and ADC exhibited weaker or more region-dependent associations. Correlations with the Mini-Mental State Examination were generally weaker across MRI measures and regions. Among these, APT# showed the strongest and most consistent associations with clinical severity, and was the best single-parameter classifier in most regions. Conclusions: Multiparametric MRI with CEST imaging shows potential for detecting microstructural and molecular alterations associated with cognitive dysfunction in individuals at risk for early cognitive decline.