Atrophy of the medial occipitotemporal, inferior, and middle temporal gyri in non-demented elderly predict decline to Alzheimer's disease.

Our goal was to ascertain, among normal elderly and individuals with mild cognitive impairment, which temporal lobe neocortical regions predicted decline to dementia of the Alzheimer's type (DAT). Individuals received an MRI at baseline and a clinical and cognitive evaluation at baseline and follow-up. By using the baseline MRI we assessed the anatomical subdivisions of the temporal lobe: anteromedial temporal lobe (hippocampus and parahippocampal gyrus), medial occipitotemporal (fusiform) gyrus, middle and inferior temporal gyri, and superior temporal gyrus. We studied two groups of carefully screened age- and education-matched elderly individuals: 26 normal elderly (NL) and 20 individuals with mild cognitive impairment (MCI). Fourteen individuals (12 from the MCI group and two from the NL group) declined to DAT within the 3.2-year follow-up interval. We used logistic regression analyses to ascertain whether the baseline brain volumes were useful predictors of decline to DAT at follow-up after accounting for age, gender, individual differences in brain size, and other variables known to predict DAT. After accounting for age, gender, and head size, adding the volume of the anteromedial temporal lobe (the aggregate of hippocampus and parahippocampal gyrus) and an index of global atrophy raised the accuracy of overall classification to 80.4%. However, the ability to detect those individuals who declined (sensitivity) was low at 57%. When baseline medial occipitotemporal and the combined middle and inferior temporal gyri were added to the logistic model, the overall classification accuracy reached 95.6% and, most importantly, the sensitivity rose to 92.8%. These data indicate that the medial occipitotemporal and the combined middle and inferior temporal gyri may be the first temporal lobe neocortical sites affected in AD; atrophy in these areas may herald the presence of future AD among nondemented individuals. No other clinical baseline variables examined predicted decline with sensitivities above 71%. The apolipoprotein APOE epsilon4 genotype was not associated with decline.