Baseline and longitudinal patterns of brain atrophy in MCI patients, and their use in prediction of short-term conversion to AD: results from ADNI

Abstract

High-dimensional pattern classification was applied to baseline and multiple follow-up MRI scans of the Alzheimer's Disease Neuroimaging Initiative (ADNI) participants with mild cognitive impairment (MCI), in order to investigate the potential of predicting short-term conversion to Alzheimer's Disease (AD) on an individual basis. MCI participants that converted to AD (average follow-up 15 months) displayed significantly lower volumes in a number of grey matter (GM) regions, as well as in the white matter (WM). They also displayed more pronounced periventricular small-vessel pathology, as well as an increased rate of increase of such pathology. Individual person analysis was performed using a pattern classifier previously constructed from AD patients and cognitively normal (CN) individuals to yield an abnormality score that is positive for AD-like brains and negative otherwise. The abnormality scores measured from MCI non-converters (MCI-NC) followed a bimodal distribution, reflecting the heterogeneity of this group, whereas they were positive in almost all MCI converters (MCI-C), indicating extensive patterns of AD-like brain atrophy in almost all MCI-C. Both MCI subgroups had similar MMSE scores at baseline. A more specialized classifier constructed to differentiate converters from non-converters based on their baseline scans provided good classification accuracy reaching 81.5%, evaluated via cross-validation. These pattern classification schemes, which distill spatial patterns of atrophy to a single abnormality score, offer promise as biomarkers of AD and as predictors of subsequent clinical progression, on an individual patient basis.

Figures

Fig. 1

Representative sections with regions of relatively reduced GM in MCI-C compared to MCI-NC, at baseline. The bottom-left t-maps were obtained by smoothing the RAVENS maps with a Gaussian kernel of 5mm, to better display atrophy in the hippocampus. An 8mm kernel was used in the top row of images. Images are in radiology convention. The scale indicates values of the t-statistic.

Fig. 2

Representative sections showing relatively reduced WM in MCI-C compared to MCI-NC, at baseline. Peri-ventricular WM loss is primarily due to periventricular leukoareosis, albeit it is also observed in peri-hippocampal inferio-medial locations. Images are in radiology convention. The scale indicates values of the t-statistic.

Fig. 3

Representative sections showing baseline ventricular enlargement in MCI-C compared to MCI-NC, primarily in the temporal horns. Images are in radiology convention. The scale indicates values of the t-statistic.

Fig. 4

Regions in which MCI-C had more baseline periventricular abnormal WM tissue, typical of leukoareosis. Images are in radiology convention. The scale indicates values of the t-statistic.

Fig. 5

Regions in which converters had most significant rates of tissue change compared to non-converters. Increases of GM (left) indicate progression of peri-ventricular leukoareosis. CSF increases (right) indicate rapid expansion of the temporal horns. Images are in radiology convention. The scale indicates values of the t-statistic.

Fig. 6

Histograms of the structural abnormality scores for MCI-NC (left) and MCI-C (right). Positive implies AD-like structure whereas negative scores imply normal-like structure.

Fig. 7

Classification rate, evaluated via leave-one-out cross validation, between MCI-C and MCI-NC, as a function of the number of features (brain regions/clusters) used to build the classifier. Baseline images were used for classification.

Fig. 8

ROC curve of the MCI-C vs. MCI-NC classification results of Fig. 7. Area under the curve = 0.77.