Longitudinal Percentage Change in DTI-ALPS and MMSE as Joint Markers of MCI-to-Alzheimer's Disease Conversion

Abstract Early Alzheimer's disease prediction remains an open challenge. DTI-ALPS, an imaging-derived marker of glymphatic clearance, has been proposed as an earlier signal, but absolute ALPS values overlap substantially across diagnostic groups, limiting their value as single-timepoint measurements. This study examined whether longitudinal percentage change in DTI-ALPS from baseline, rather than a single absolute value, better distinguishes subjects who remain stable from those who convert from MCI to Alzheimer's disease, and whether combining it with longitudinal MMSE change adds further value. Unlike prior work linking baseline ALPS to later outcomes, we used ADNI data to compare Stable MCI and MCI-to-AD subjects using longitudinal percentage changes in ALPS and MMSE from each subject's own baseline. Overcoming the overlap inherent to absolute baseline ALPS, both measures differed significantly, showing medium and large effect sizes for ALPS and MMSE changes, and were only weakly and non-significantly correlated with each other. A logistic regression combining them did not significantly outperform MMSE alone on classification accuracy, but ALPS change remained an independent, significant predictor after MMSE was accounted for, indicating that glymphatic decline carries conversion-relevant information of its own. Together, these results support tracking change from an individual's own baseline when assessing early Alzheimer's risk with DTI-ALPS.