Amnestic MCI future clinical status prediction using baseline MRI features

Quantitative MRI Brain Studies in Mild Cognitive Impairment and Alzheimer's Disease: A Methodological Review

“…In their prediction model, only hippocampus was used which, achieved a predictive performance comparable or superior to other studies [45], [99], [122] that employed a multi-region or whole brain approach. This was similar to the accuracy achieved using other predictive models based on non-automated techniques.…”

“…When the anterior cingulate gyrus and orbitofrontal cortex were included in the measurements, the best predictive accuracy obtained was 75%. Duchesne et al [122] used only MTL in their VBM analysis and their results were better compared to other studies (see Table VII) that used the whole brain.…”

Automated MR morphometry to predict Alzheimer’s disease in mild cognitive impairment

“…In the area of predicting conversion from MCI to AD, the methodology is related to a number of other neuroimaging techniques for estimation of atrophic changes [6,[9][10][11][12][13][14] and classification [18][19][20][21][22]. Some of the approaches require manual interaction [9,10,14].…”

“…Many studies do not perform individual patient classification, but group analysis or correlation with clinical scores [6,[9][10][11][12][13][14]. The voxel-wise measure that is derived in this study is mainly based on the gray value intensities found in the image, while other studies are based on features like volumetric measures [6,[9][10][11]13,14], measures of cortical thickness [12], shape-or deformation-based features [18,20,21], or tissue classification [19,22]. Duchesne et al [21] work directly on voxel intensities and combine them with deformation-based features from the registration process.…”

“…Analysis of spatial patterns of brain atrophy has shown that patients with MCI often present patterns of atrophy overlapping with those of patients with AD [19]. Classification rates for the differentiation between converting and non-converting patients with MCI have also recently been reported, reaching classification rates around 85% using different high-dimensional pattern classification techniques [18][19][20][21][22].…”