Abstract

Background: The cross-sectional visual rating scale (c-VRS) is an easy and effective method of evaluating medial temporal atrophy (MTA). In evaluating longitudinal morphological change, c-VRS cannot express slight changes as precisely as volumetry.

Methods: We selected 87 people 65 years old from the Alzheimer's Disease Neuroimaging Initiative database. Their high resolution magnetic resonance images, brain structural volume data, and cognitive performances, were downloaded. We evaluated the temporal lobe atrophy in the baseline images, and the longitudinal alterations by judging the widening of the parahippocampal cerebrospinal fluid space one year and two years later. These results from c-VRS and longitudinal VRS (l-VRS) were also compared with the volume data derived from the FreeSurfer.

Results: The cross-sectional and longitudinal visual rating assessments showed significant differences between mild cognitive impairment (MCI) and cognitively normal subjects. In a receiver operating characteristic (ROC), the longitudinal assessment showed higher sensitivity/specificity and area-under curve than did the cross-sectional assessment.

Conclusions: Our findings demonstrated that longitudinal visual rating is a useful tool for detecting longitudinal morphological alterations caused by MCI, and would improve diagnostic accuracy.