An efficient approach for differentiating Alzheimer's disease from normal elderly based on multicenter MRI using gray-level invariant features

Abstract

Machine learning techniques, along with imaging markers extracted from structural magnetic resonance images, have been shown to increase the accuracy to differentiate patients with Alzheimer's disease (AD) from normal elderly controls. Several forms of anatomical features, such as cortical volume, shape, and thickness, have demonstrated discriminative capability. These approaches rely on accurate non-linear image transformation, which could invite several nuisance factors, such as dependency on transformation parameters and the degree of anatomical abnormality, and an unpredictable influence of residual registration errors. In this study, we tested a simple method to extract disease-related anatomical features, which is suitable for initial stratification of the heterogeneous patient populations often encountered in clinical data. The method employed gray-level invariant features, which were extracted from linearly transformed images, to characterize AD-specific anatomical features. The intensity information from a disease-specific spatial masking, which was linearly registered to each patient, was used to capture the anatomical features. We implemented a two-step feature selection for anatomic recognition. First, a statistic-based feature selection was implemented to extract AD-related anatomical features while excluding non-significant features. Then, seven knowledge-based ROIs were used to capture the local discriminative powers of selected voxels within areas that were sensitive to AD or mild cognitive impairment (MCI). The discriminative capability of the proposed feature was measured by its performance in differentiating AD or MCI from normal elderly controls (NC) using a support vector machine. The statistic-based feature selection, together with the knowledge-based masks, provided a promising solution for capturing anatomical features of the brain efficiently. For the analysis of clinical populations, which are inherently heterogeneous, this approach could stratify the large amount of data rapidly and could be combined with more detailed subsequent analyses based on non-linear transformation.

Figure 1. A 2D LBP test on simulated MRIs.

The first row displays the MRI images and the second row displays their corresponding LBP maps. Images scanned with different flip angles are shown in columns.

Figure 2. A brief illustration of the calculation of LBP-TOP value in pixel p in the axial, coronal, and sagittal orientations.

Pixel p is denoted by the red color, with its 3×3 neighborhood circled by a yellow square in the 2D plane. Bin2Dec is a function for transferring binary code to decimal values.

Figure 3. Seven knowledge-based masks customized for capturing AD-specific morphometry.

These masks covered areas that have been consistently demonstrated as sensitive to AD, including the amygdala (AMG), the entorhinal area (ENT), the hippocampus (HIP), the parahippocampal gyrus (PHG), the temporal lobe (TL), the Lateral ventricle (LV), and a mask that combined the aforementioned six masks (OVALL).

Figure 4. Average feature weights yielded by linear SVM in differentiating a) AD from NC, and b) MCI from NC.

In either a) or b), average weights separately obtained from three orthogonal planes are displayed in rows, and different slices along the axial orientation are displayed in columns. The ascending weights are shown from darkness to brightness.

Figure 5. Classification performances with respect to features selected by seven masks in differentiating a) AD from NC, and b) MCI from NC.

The performances were measured in terms of specificity, sensitivity, and AUC. The ROC curves are also displayed in the blue color with a smooth fitting line shown in red.