Quantitative Gradient Echo MRI Identifies Dark Matter as a New Imaging Biomarker of Neurodegeneration that Precedes Tisssue Atrophy in Early Alzheimer's Disease

Abstract

Background: Currently, brain tissue atrophy serves as an in vivo MRI biomarker of neurodegeneration in Alzheimer's disease (AD). However, postmortem histopathological studies show that neuronal loss in AD exceeds volumetric loss of tissue and that loss of memory in AD begins when neurons and synapses are lost. Therefore, in vivo detection of neuronal loss prior to detectable atrophy in MRI is essential for early AD diagnosis.

Objective: To apply a recently developed quantitative Gradient Recalled Echo (qGRE) MRI technique for in vivo evaluation of neuronal loss in human hippocampus.

Methods: Seventy participants were recruited from the Knight Alzheimer Disease Research Center, representing three groups: Healthy controls [Clinical Dementia Rating® (CDR®) = 0, amyloid β (Aβ)-negative, n = 34]; Preclinical AD (CDR = 0, Aβ-positive, n = 19); and mild AD (CDR = 0.5 or 1, Aβ-positive, n = 17).

Results: In hippocampal tissue, qGRE identified two types of regions: one, practically devoid of neurons, we designate as "Dark Matter", and the other, with relatively preserved neurons, "Viable Tissue". Data showed a greater loss of neurons than defined by atrophy in the mild AD group compared with the healthy control group; neuronal loss ranged between 31% and 43%, while volume loss ranged only between 10% and 19%. The concept of Dark Matter was confirmed with histopathological study of one participant who underwent in vivo qGRE 14 months prior to expiration.

Conclusion: In vivo qGRE method identifies neuronal loss that is associated with impaired AD-related cognition but is not recognized by MRI measurements of tissue atrophy, therefore providing new biomarkers for early AD detection.

Keywords: Alzheimer’s disease; brain atrophy; cognitive impairment; hippocampal subfields; hippocampus; magnetic resonance imaging; neurodegeneration; quantitative Gradient Recalled Echo.

Fig. 1

A) Schematic representation of qGRE biomarkers. Total tissue volume identified on MRI images (e.g., T1-weighted MPRAGE) is separated into two volumes based on qGRE R2t^* measurement: Dark Matter –tissue devoid of neurons (R2t^* < 5.8 s^–1) and Viable Tissue –tissue with relatively preserved neurons (R2t^* > 5.8 s^–1). B) 3D surface views (created by using Slicer 4.5.0 software) of the hippocampus structure of three representative participants from HC, PC, and mild AD groups. Viable Tissue is marked with grey color and Dark Matter is marked with black color. While this figure shows left/right differences in the dark matter distribution in these particular examples, the group analysis shows no statistical left/right differences in any group (HC, PC, and AD).

Fig. 2

Total volume, Viable Tissue volume, Dark Matter fraction, and Neuronal Density Index of the viable tissue in hippocampal subfields. Bars represent mean group values and whiskers show standard deviations. Green bars represent the HC group (n = 34), blue bars represent PC group (n = 19), and red bars represent mild AD (n = 17) group. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. Tukey’s honest significant difference criterion is used to correct for multiple comparisons.

Fig. 3

Percent of volume and neuronal losses in mild AD group in hippocampal subfields evaluated with respect to the group mean values of HC group.

Fig. 4

Correlation of the Episodic Memory test (represented by the z-score of the Free and Cued Selective Reminding Test) with the fraction of Dark Matter, volume of Viable Tissue, and Total Volume of the hippocampus. Each point represents an individual participant, solid lines represent linear regression, and shaded areas are 95% confidence intervals. Additionally, individual cognitive data are presented in Supplementary Figure 2.

Fig. 5

Group differences (A-D) and ROC classifications (E, F) based on qGRE metrics and volumetric measurements in the hippocampus. (A) Dark Matter volume (mm³), (B) Viable Tissue volume (mm³), (C) Relative Neuronal Index (total number of neurons normalized by the mean value of the total number of neurons in the HC group), and (D) Total Volume (mm³). The middle lines of the box plots represent median, ends of the boxes are the 25th and 75th quantiles (quartiles), and interquartile range is the difference between the quartiles. The lines (whiskers) extend from the boxes to the outermost points that fall within the distance computed as 1.5 (interquartile range). All generalized linear models for A, B, C, and D indicated differences among the HC, PC, and AD groups (p≤0.0004) and the p values that resulted from assessing differences between groups are indicated above the horizontal connectors between groups. (E) Result of a classification-tree that was produced using global hippocampal Dark Matter volume and Viable Tissue volume variables as predictors. (F) Result of a classification-tree that was produced using Total Hippocampal Volume as predictor. Receiver operating characteristic (ROC) curves, areas under the curves (AUCs), and a confusion matrix are presented. The confusion matrix presents the numbers of correct and incorrect classifications. ROC analysis is performed by constructing a graph of true and false positive rates (sensitivity and 1 minus specificity, respectively) for a series of cutoff points for a test (in our case, MRI metrics). The ROC curve yields a measure of diagnostic accuracy, independent of the decision criterion. It characterizes the inherent accuracy of the technique. The AUC value represents the probability that a randomly chosen abnormal case is (correctly) rated or ranked with greater suspicion than a randomly chosen normal case.

Fig. 6

Associations between PET Tau pathology and qGRE biomarkers in asymptomatic (CDR = 0) and symptomatic (CDR > 0) study participants. Results for Total Volume measurements are also shown. Data represent mean and standard deviation of Dark Matter fraction (%) (top left panel), Viable Tissue volume (top right panel), relative neuronal density index (bottom left panel), and Total volume (bottom right panel) measurements in Tau–, CDR–(green box plot, n = 21); Tau+, CDR–(blue box plot, n = 11); Tau+, CDR+ (red box plot, n = 10) groups. Each dot in the box plot represents single participant. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. Tukey’s honest significant difference criterion is used to correct for multiple comparisons.

Fig. 7

Data obtained from an 81-year-old male study participant with a clinical diagnosis of dementia (CDR 1) who underwent qGRE MRI 14 months prior to expiration (upper panel). qGRE R2t^* in the hippocampus (outlined in yellow with hippocampal subfields shown in colors, segmented based on FreeSurfer) shows Dark Matter (hypointense lesions with lower R2t^* values) in parts of subiculum, parasubiculum and CA1, indicating the loss of neurons. This is confirmed by direct neuropathological examination shown in the lower panel obtained from the postmortem studies from this participant. Severe neuronal loss in CA1 (hematoxylin and eosin stain) is reflected by the presence of only one remaining definitively identifiable neuron (indicated by the arrow) within this representative image; relative neuronal preservation is shown in a representative photomicrograph from CA2/CA3. Unlike qGRE R2t^*, T1-weighted MPRAGE imaging finds the hippocampal region to be practically homogeneous without any obvious intensity contrast. Hence, the data demonstrate a higher sensitivity of qGRE R2t^* measurements to tissue neuronal loss as compared with standard volumetric measurements. Scale bars are 50 micrometers.