Combining neuropsychological and structural neuroimaging indicators of conversion to Alzheimer's disease in amnestic mild cognitive impairment.

Morphometric and neuropsychological retrospective studies of amnestic mild cognitive impairment (MCI) have demonstrated that regional atrophies and cognitive impairments may differentiate stable from progressing MCI. No measure has proved helpful prospectively. In this study, twenty five amnestic MCI patients and 25 healthy controls underwent structural MRI and comprehensive neuropsychological assessment. The groups' grey matter volumes were compared with voxel based morphometry and were also correlated with scores obtained on paired associates learning and category fluency tasks. MCI patients had significantly reduced grey matter volume in left mediotemporal and other neocortical regions compared with controls. Atrophy in perirhinal and anterior inferior temporal cortex was associated with poor scores on both category fluency and paired associates learning tasks. After 36 months, 44% of the MCI sample converted to dementia. Converter and non-converter MCI subgroups differed in paired associates learning and in category fluency scores, and showed limited differences in grey matter loss in the hippocampal complex. Variable atrophy in the hippocampus was not a relevant element in the converter/non converter distinction, but converters had significant volumetric reductions in the perhirinal cortex and in other anterior temporal and frontal neocortical areas. A high proportion of converters (91%) could be identified from baseline data using a combination of measures of regional atrophy in left temporal association cortex and poor scores on paired associates learning and category fluency tasks. This combined approach may offer a better option than using each measure alone to prospectively identify individuals at more immediate risk of conversion to dementia in the MCI population. The clinical advantage of this combination of structural MRI and neuropsychological measures in predicting conversion to dementia will need additional prospective validation.