Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2411050628
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ramani, A.
Right arrow Articles by Helpern, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ramani, A.
Right arrow Articles by Helpern, J. A.
(Radiology 2006;241:26-44.)
© RSNA, 2006

Special Reviews

Quantitative MR Imaging in Alzheimer Disease1

Anita Ramani, PhD, Jens H. Jensen, PhD and Joseph A. Helpern, PhD

1 From the Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, KIP-600-F, 660 First Ave, New York, NY 10016-3240 (A.R., J.H.J., J.A.H.); and Departments of Psychiatry (J.A.H.) and Physiology and Neuroscience (J.H.J., J.A.H.), New York University School of Medicine, New York, NY. Received April 15, 2005; revision requested June 14; revision received July 25; accepted September 8; final review and update by A.R. April 17, 2006. J.A.H. supported by grants from the Werner Dannheisser Trust and the Institute for the Study of Aging. Address correspondence to A.R. (e-mail: anita.ramani{at}med.nyu.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 
Alzheimer disease (AD) is the most common type of dementia. It currently affects approximately 4 million people in the United States. AD is a progressive neurodegenerative disorder characterized by the gradual deposition of neuritic plaques and neurofibrillary tangles in the brain, which is thought to occur decades before the onset of clinical symptoms. Identification of people at risk before the clinical appearance of dementia has become a priority due to the potential benefits of therapeutic intervention. Although atrophy of medial temporal lobe structures has been shown to correlate with progression of AD, a growing number of recent reports have indicated that such atrophy may not be specific to AD. To improve diagnostic specificity, new quantitative magnetic resonance (MR) imaging methods are being developed that exploit known pathogenic mechanisms exclusive to AD. This article reviews the MR techniques that are currently available for the diagnostic assessment of AD.

© RSNA, 2006


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 
Alzheimer disease (AD) is the most common type of dementia; it is characterized by cognitive impairment, including memory dysfunction, severe enough to interfere with activities of daily living (1). However, cognitive symptoms and brain abnormalities may be present many years before a clinical diagnosis of AD can be made. This preclinical phase of AD is the subject of intense investigation, since prompt diagnosis could allow drug therapy (aimed at treating the underlying pathologic condition) to be started earlier, thereby improving the chances for a positive clinical response. A definitive diagnosis of AD can only be made histologically and is based on the presence of large numbers of neuritic plaques and neurofibrillary tangles in brain tissue. The search for a reliable, noninvasive, and objective method that allows accurate diagnosis within the lifetime of an individual has attracted considerable interest in recent years, especially in view of the prospect of developing agents targeted at modifying disease progression (2). Identification of how the disease process begins and progresses is also important for the development of therapeutic strategies aimed at prevention.

In addition to extensive research focused on the identification and validation of a wide range of biochemical markers (3), the application of in vivo imaging techniques such as computed tomography, magnetic resonance (MR) imaging, and positron emission tomography is growing rapidly. Promising research is also being performed in the field of MR spectroscopy. Each of these techniques could warrant a dedicated review, and several such articles have been published (4–14). In this review, we focus on current MR applications that utilize quantitative parametric approaches, including volumetry, diffusion, magnetization transfer, and nuclear MR relaxation times and their role in the differential diagnosis of AD.


   AD PATHOLOGIC CHARACTERISTICS
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 
Recent anatomy and pathology studies have begun to contribute to our understanding of the location, severity, and specificity of brain changes in the early stages of AD. The two main pathologic structures found within the AD brain are extracellular neuritic plaques, consisting largely of Aß peptide, and neurofibrillary tangles (15), composed primarily of the cytoskeletal protein tau. Under abnormal conditions, these peptides undergo a structural transition from a soluble to an insoluble state. Authors of several studies (16–21) have reported that the {varepsilon}4 allele of apolipoprotein E is a risk factor for AD, and although the mechanism underlying this increased risk is not completely clear, strong evidence (22–26) supports the idea that apolipoprotein E acts as a chaperone protein by interacting with the Aß peptide and influences its conformation and aggregation (Fig 1).


Figure 1
View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a: Diagrams show that (a) in normal aging, ß-amyloid peptides are produced by brain cells and are soluble, but (b) in abnormal conditions, these peptides become insoluble and can be clustered together by chaperone proteins, resulting in the formation of plaques.

 

Figure 1
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b: Diagrams show that (a) in normal aging, ß-amyloid peptides are produced by brain cells and are soluble, but (b) in abnormal conditions, these peptides become insoluble and can be clustered together by chaperone proteins, resulting in the formation of plaques.

 
The formation of plaques is known to be a central event in AD pathogenesis and begins several years before the onset of clinical symptoms, by which stage plaque infiltration is extensive and involves brain regions essential for normal cognition (27–29). The results of several studies (30–33) have shown a correlation between plaque load and severity of dementia. Most vulnerable to the deposition of plaque are the hippocampal formation and the associated entorhinal and perirhinal (transentorhinal) cortex (34–37). The neurons of the hippocampal formation are particularly vulnerable to damage early in the course of AD (38–40).

Recently, iron accumulation in the brain, particularly in cells that are associated with neuritic plaques, has been reported in the AD brain (41–43). Increased iron concentrations in neurons, astrocytes, and microglia, which normally have low iron content up to middle age, are typically present in regions such as the cortex, hippocampus, and substantia nigra, which are particularly susceptible to the neuropathologic changes that characterize AD. There is evidence to indicate that aging and AD are associated with a decline in myelin content, and it has been suggested that this might be related to the amount of iron present in myelin (44,45). Authors of postmortem studies (46) have reported a disruption of iron metabolism in the basal ganglia of such subjects. Processes associated with plaque formation, such as ß-amyloid deposition and oxidative stress, appear to be promoted by iron, and investigators (42,47) have reported the presence of iron in the core of plaques. It seems plausible that processes such as changes in iron content and deposition of ß-amyloid should alter the local biophysical environment of brain tissue water in several ways, and it is reasonable to believe that quantitative MR techniques could be developed that are capable of demonstrating such subtle alterations to allow monitoring of disease progression.


   VOLUMETRIC IMAGING
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 
MR imaging is now recognized as an important tool in the preclinical detection and monitoring of AD. To date, the majority of research in the application of MR imaging to the assessment of AD has been in the field of volumetry—the measurement of brain regional volumes for the detection of atrophy. Although this work has been recently reviewed (48–65), we present a brief overview from a perspective of evaluating the quantitative nature of this approach. As mentioned earlier, pathologic changes in the cortical gray matter (GM) of the AD brain are characterized by the accumulation of ß-amyloid plaques and neurofibrillary tangles, along with neuronal and synaptic loss that produce cerebral atrophy. In established AD, these changes are widely distributed, and the medial temporal lobes and neocortical association areas are severely affected. As with plaque deposition, atrophy has been found to occur first in the hippocampal formation and associated entorhinal cortex before progressing elsewhere (66,67).

This has motivated several volumetric studies of AD brain by using MR imaging. A T1-weighted three-dimensional technique (eg, magnetization-prepared rapid gradient echo) is typically used to obtain high-spatial-resolution anatomic images. This technique is usually referred to as MP-RAGE (Siemens Medical Systems, Erlangen, Germany), three-dimensional Fast SPGR (GE Healthcare, Milwaukee, Wis) or three-dimensional TFE (Philips Medical Systems, Best, the Netherlands). Brain structures are then manually traced on all contiguous sections where the structure of interest is evident (Fig 2); volumes are usually calculated in cubic millimeters by computing the number of voxels within the traced images.


Figure 2
View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2: A–D, Definition of auxiliary guideline traces on selected sagittal T1-weighted three-dimensional spoiled gradient-echo MR sections (repetition time msec/echo time msec, 24/5; section thickness, 1.5 mm; flip angle, 40°; two signals acquired; field of view, 26 x 26 x 18.6 cm; matrix, 256 x 192 x 124). E–N, Definition of hippocampal traces on selected coronal MR sections from most rostral (E) to most caudal (N) aspects of the structure (right hemisphere). O, Three-dimensional reconstruction of hippocampal shape based on coronal traces. (Reprinted, with permission, from reference 218.)

 
Several studies have reported volume losses in mesiotemporal structures, which are indicative of atrophy in patients with AD compared with elderly control subjects (without dementia). Most region-of-interest–based MR studies of these structures have been cross-sectional in design. Such studies (Table 1) are the most feasible to implement but have the disadvantage of dealing with differences in brain size and have generally been limited to short time frames and small study groups (68–78).


View this table:
[in this window]
[in a new window]

  		Table 1. Data from Cross-sectional Studies by Brain Region and Volume Decrease

 
On the other hand, longitudinal studies (79–95) avoid the problems associated with interindividual variation in brain size (Table 2). Serial sections are positionally matched so that differences in the two images (acquired at different times) can be visualized by subtracting one from the other. An automated subtraction algorithm may be used to measure the difference in brain volume. Volume loss is usually expressed as a percentage of initial brain volume and is converted into a rate of atrophy in cubic centimeters per year (Fig 3). Serial sections in the same subject have the advantage that the wide interindividual variability of brain morphology is not an issue, and comparison of before-and-after images in the same subject carries much less error than comparison of cases with controls. However, since the rates of change in brain structure volumes are slow, these studies can be very sensitive to subtle differences in MR image acquisition conditions. Changes in magnetic field homogeneity can produce distortions in signal intensity that impede visual inspection and segmentation (96), and MR images obtained from the same patient by using the same equipment at different times may appear different from each other owing to a variety of imager-dependent variations (97).


View this table:
[in this window]
[in a new window]

  		Table 2. Atrophy Rates from Longitudinal Studies by Brain Region and Annual Volume Decrease

 

Figure 3
View larger version (144K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3: Coronal unenhanced baseline (left) and coregistered year 2 follow-up (middle) T1-weighted MR images (35/9; flip angle, 60°; matrix, 256 x 192; thickness, 1.2 mm) and normalized difference displayed with narrower 20% signal intensity window to enable visualization of brain tissue loss within the boundary region (right). Rectangles are location of the medial temporal lobe (MTL) region. Coronal sections through foot of the hippocampus are shown for three representative study participants. Top: Images in a 72-year-old man who remained healthy at year 6.4 after baseline; annual MTL atrophy rate was 0.2%. Middle: Images in a 70-year-old woman who remained healthy at year 2 and declined to mild cognitive impairment (MCI) by year 6; annual MTL atrophy rate was 0.8%. Bottom: Images in a 77-year-old man with normal findings at baseline who declined to AD by year 2; annual MTL atrophy rate was 1.3%. (Reprinted, with permission, from reference 95.)

 
Recently, techniques such as voxel-based morphometry have been developed to reduce bias in the selection of anatomic regions of interest and observer or operator dependency. This technique involves a voxel size comparison of the local concentration of any tissue structure (GM, white matter [WM], etc) between two groups of subjects. While most current region-of-interest methods have focused on the measurement of preselected regions, atrophy measurement with voxel-based morphometry is accomplished by performing several statistical tests between the two groups. This procedure involves spatially normalizing high-resolution images from all subjects in the study into the same stereotactic space, thus removing spatial and volume differences. The desired tissue of interest (usually GM) is then segmented from these images, and a smoothing algorithm is applied to the segmented tissue. Finally, numerous voxel size parametric statistical tests are performed to compare the smoothed images from the two groups. A detailed explanation of this technique been provided in two excellent articles by Ashburner and Friston (98,99).

A key feature of voxel-based morphometry is that it is sensitive to the local composition of different tissue types while being less sensitive to other large-scale volumetric differences in gross anatomy. However, it has been shown that imperfect registration of MR images to a common template can lead to false estimates of atrophy (100). This is especially true of brain atrophy in AD, where the ventricles are enlarged and registration might thus be imperfect; also, tissue classification errors during automated segmentation of brain tissue classes are likely to produce an artificial thin rim of periventricular GM (101). One should be cautious when using automated image-processing packages for disease description and identification. A few studies have been published in which this technique was used in patients with AD (102–104), but inconsistent results and different study populations and analysis protocols have led to discrepancies in results.

Since the earliest pathologic changes of AD have been found to involve the transentorhinal and/or entorhinal areas and not the hippocampus (37), several new protocols have been developed to enable calculation of the volume of the entorhinal cortex from MR images (74,105–108). Although no major differences have been reported between hippocampal and entorhinal measurements (106) in discriminating between AD patients and control subjects, a recent study (109) showed that the atrophy rate in the entorhinal cortex was higher than that in the hippocampus, which is consistent with the view that AD pathologic changes begin in the entorhinal cortex. In clinical practice, volumetric measurements of the hippocampus are more feasible than are those of the entorhinal cortex, since the former are easier to perform and produce less variability.

There is a particularly great need for biologic indicators of AD that have high sensitivity and specificity in patients with MCI, where the clinical picture is less distinct and the clinical diagnosis is more difficult. The most prevalent type of MCI is the amnestic type, in which subjects have an impairment in delayed verbal or nonverbal recall and a decline in cognitive function from a previous level of functioning (110). Lopez et al (111), however, pointed out that MCI can exist with multiple cognitive deficits (not including memory). These cognitive deficits may or may not affect instrumental activities of daily living and represent a decline from a previous level of functioning, as detected with the aid of annual neuropsychologic testing.

Although MCI is considered a transitional stage in the pathogenesis of AD, not all MCI patients progress to clinically defined AD or decline at an identical rate. Many subjects with MCI remain stable or even revert to a normal cognitive state (112,113). When MCI subjects are observed longitudinally, results have shown that they are likely to progress to clinically probable AD at a considerably accelerated rate, compared with age-matched healthy subjects (110,114).

Authors of cross-sectional and longitudinal studies (74,77,83,86,110,114–117) have used MR imaging in the evaluation of MCI. There is general agreement among the studies that hippocampal atrophy is present in subjects with MCI, as compared with cognitively intact controls; reduced volumes have been reported (118) in the amygdala, hippocampus, and parahippocampal gyrus, and the hippocampal atrophy rate has been used to predict the rate of conversion from MCI to AD (83). Such atrophy in patients with AD has been correlated with autopsy evidence of atrophy and neuronal loss (119). In one study (83), patients were followed up longitudinally with annual cognitive assessments. The primary end point was the crossover of MCI to the clinical diagnosis of AD in individual patients during longitudinal clinical follow-up. During the longitudinal period (32.6 months), 27 of 80 MCI patients developed dementia. In a subsequent study (84), the same group found that the rates of hippocampal atrophy in individuals who undergo conversion from control subject to MCI patient or from MCI to AD patient were comparable, suggesting that the group with MCI may be in transition to early AD. A study of MCI patients who developed AD within 3 years showed significantly decreased perihippocampal volume, compared with the volume in those who did not develop AD (116). A group of subjects with a diagnosis of mild dementia, corresponding to a clinical dementia rating of 0.5, was studied for 3 years; these subjects progressed to AD at a rate of 6% per year (120). This was confirmed in another study (114), where the majority of MCI subjects converted to AD within 6 years. What is clear from these results is that subjects with MCI must be identified and followed up because of their increased risk for developing AD.

Considerable heterogeneity exists in research methods used to study the epidemiology, thresholds for cognitive and functional impairment, and rate of progression. Ambiguity also exists in the definitions of subtypes of MCI. While some investigators believe that a clinical dementia rating of 0.5 is equivalent to MCI, others contend that such a rating actually describes a broader population that includes both subjects with MCI and those with mild AD (121). With the Global Deterioration Scale (122,123), subjects with MCI could have scores that correspond to either 2 or 3 on the scale.

It is important to note that there is a range of normal volumes in MCI patients (57) and that these ranges overlap, which reduces the predictive value of volumetry. The predictive value of hippocampal volume in the early diagnosis of MCI remains controversial. This may be due to the fact that the pathologic processes in AD may be more widespread than focal medial temporal lobe damage and could include, for example, neuritic plaques throughout the cortex or vascular disease.

Other issues surrounding the utility of volumetry in the diagnosis of MCI relate to methodological factors. Since regional volumes are computed by means of manual tracings on MR images, several factors such as section thickness, head positioning, choice of imaging protocol, differences in volume correction procedures to account for individual variations in head size, variations in sample size, and reliability of interrater and test-retest measurements can all contribute to a wide variability in volumetric measurements. Other variables that influence results include the age of the sample population, criteria for selection, size of the sample, and definition of neuroanatomic boundaries (124). These methodological issues have not been sufficiently addressed despite the enthusiasm to identify objective indexes of MCI.

Results from individual studies need to be replicable and generalizable if they are to be of clinical utility in the longer term. The authors of a recent study (125) emphasized the need for uniformity in the use of appropriate structural measures for diagnosis and for reliable methods to determine progression or improvement of cognitive impairment. It must be borne in mind that while volumetric measurements of specific medial temporal lobe structures may provide a more detailed analysis of atrophy, these measurements are laborious and time consuming. Moreover, medial temporal lobe atrophy is not specific to AD and MCI but can be found in other dementias, including dementia in Down syndrome (126), Lewy body disease (127–131), and argyrophilic grain disease (132,133).


   QUANTITATIVE PARAMETRIC MAPPING
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 
Diffusion-weighted Imaging
In diffusion-weighted MR imaging, specific pulse sequences are designed that are sensitive to the microscopic random motion of water molecules in biologic tissue. This is accomplished in spin-echo imaging by applying a pair of strong diffusion-sensitizing magnetic field gradients on either side of the 180° refocusing pulse. Thus, diffusion-weighted imaging provides a form of contrast that enables the diffusional motion of water molecules to be measured and that, as a consequence of the interactions between tissue water and cellular structures, yields information about the microenvironment of a tissue (eg, orientation, and geometry) (134). It is, therefore, possible to quantify alterations in water diffusion that result from microscopic structural changes, which may have little effect on T1- and T2-weighted anatomic MR imaging.

An apparent diffusion coefficient (ADC) can be computed from images with two or more diffusion weightings. If the ADC in any three orthogonal directions is averaged, one obtains the so-called mean diffusivity. In cerebral WM, the ADC depends on the relationship between the orientation of the WM tracts and the direction of the diffusion gradients: The ADC is higher when WM tracts are parallel to the diffusion gradients and lower when the tracts are perpendicular to the diffusion gradients (135,136). This effect, diffusional anisotropy, is thought to be due to the restriction of the random motion of water molecules by the myelin sheaths. Diffusional anisotropy increases with myelination during neonatal brain maturation (137,138) and decreases in demyelinating disorders such as Krabbe disease, Alexander disease–related disorder (139), multiple sclerosis (140), and adrenoleukodystrophy (141). The extent of diffusional anisotropy is usually represented by quantitative indexes referred to as fractional anisotropy and relative anisotropy.

Diffusion-based MR techniques have been applied to the study of patients with AD to achieve in vivo estimates of AD-related structural changes (142–149). An increase in ADC and a decrease in fractional anisotropy values have been reported in temporal lobe WM (142), hippocampus (144), posterior WM (147), and corpus callosum (149) in patients with very mild and moderate AD, reflecting biophysical alterations early in the progression of AD. In one study (148), some variability existed in ADCs and overlapped between subject groups; this prevented the reliable use of ADC measurements to help diagnose MCI or AD or predict the likelihood of progression from MCI to AD. Substantial variability in fractional anisotropy values between groups obscured any possible effects of decreased fiber density that may have been present in patients with MCI or AD. The authors of the majority of these studies (Tables 3, 4) also reported strong correlations between Mini-Mental State Examination score (150) and average overall ADC and fractional anisotropy values in WM (Fig 4).


View this table:
[in this window]
[in a new window]

  		Table 3. Summary Data from Studies of ADC by Brain Region

 

View this table:
[in this window]
[in a new window]

  		Table 4. Summary Data from Studies of Diffusion-Tensor Indices by Brain Region

 

Figure 4
View larger version (5K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4: Scatterplots show strong correlations between Mini-Mental State Examination (MMSE) scores and mean diffusivity (left: ADC, expressed as micrometers squared per millisecond) and fractional anisotropy values (right) in WM of patients with AD. (Reprinted, with permission, from reference 219.)

 
Another tensor index, referred to as the lattice index, has been used to detect disruptions in WM microstructure. The lattice index, like fractional anisotropy, typically decreases in such conditions. Relative to measurements in control subjects, a lower lattice index has been reported in the splenium of the corpus callosum, the superior longitudinal fasciculus, and the left cingulum of AD subjects (145), along with a strong correlation between the lattice index in the splenium and the Mini-Mental State Examination score.

Diffusivity is influenced by structural components (eg, intracellular organelles and macromolecules), physical barriers (eg, cell membranes), and chemical properties (eg, viscosity and temperature) of a tissue. Therefore, diffusivity may well be affected by pathologic alterations of AD such as amyloid deposition, neuritic degeneration, and cytoskeletal destabilization. The reported WM changes are indicative of a net loss of barriers restricting the motion of water molecules and tissue anisotropy of WM, which are consistent with histopathologic data that show partial loss of myelin, axons, and oligodendrial cells with reactive gliosis in WM (151). Other possible explanations for diminished fractional anisotropy in normal-appearing WM include oxidative membrane damage, edema, alterations of ion or fluid homeostasis, and reduced axoplasmic flow related to cytoskeletal dysfunction (152,153).

It is likely that the distribution of WM abnormalities associated with AD is not homogeneous but involves selective regions connected with association cortices (corpus callosum and temporal, frontal, and parietal lobe WM), with a relative sparing of other WM areas that subserve motor (internal capsule) or visual (optic radiations) functions. These results are consistent with those from conventional MR studies that have shown a notable decrease of the area of the corpus callosum in AD patients, compared with the area in age-matched control subjects (154–156).

The pathologic rationale of abnormalities in WM diffusion has not been studied in detail; however, tissue rarefaction of a vascular origin (157) and wallerian degeneration due to cortical disease have been suggested (158,159). Diffusion abnormalities are also seen in animal models of excitatory amino acid–induced neurotoxicity, which may contribute to the neurodegeneration of AD (152,160,161). Higher field strengths and improvements in gradient coils and postprocessing techniques will almost certainly continue to increase the accuracy of diffusion-weighted MR techniques for the evaluation of patients with AD.

In the derivation of ADC maps, a monoexponential relationship is often assumed between the degree of diffusion weighting, as expressed by the b value, and signal intensity. However, studies have shown that in certain biologic tissues (eg, brain), the signal decay due to diffusion does not necessarily follow a simple monoexponential model (162–165). When b values exceed approximately 2000 sec/mm2, a substantial deviation from monoexponential decay is observed (162,166,167). Several theories have been proposed as an explanation. According to one theory, within each imaging voxel in brain parenchyma there are two or more components with different ADCs: a fast component with higher ADC and a slow component with lower ADC, although the origins of these components are not yet fully understood. In a multicomponent diffusion model, the diffusion-weighted signal intensity is weighted toward the fast ADC component at low b values and toward the slow ADC component at high b values. If monoexponential fitting is used, then the measured ADC in both GM and WM will decrease as b increases. The use of a higher b value has been found to substantially improve both the lesion-to–normal tissue contrast and the contrast-to-noise ratio in a study of AD subjects (168), suggesting that diffusion-weighted imaging with a high b value might be more sensitive to AD-related WM degeneration than conventional diffusion-weighted imaging. A major drawback of diffusion-weighted imaging with a high b value is the low signal-to-noise ratio. This is due to the limited gradient strengths obtainable and the long echo time required for strong diffusion weighting. Use of MR imaging units with more powerful gradients should markedly improve the signal-to-noise ratio and may allow higher spatial resolution at high b values.

Magnetization Transfer Imaging
Magnetization transfer imaging is based on the exchange of magnetization between immobile protons (bound to macromolecular proteins) and free protons of tissue. Magnetization transfer imaging can provide pathophysiologic information about the microscopic structure of the brain, reflecting underlying histopathologic changes. The amount of magnetization transfer depends on the concentration, surface chemistry, and biophysical dynamics of macromolecules and may be quantified by calculating the magnetization transfer ratio (MTR). The MTR has been used extensively in the assessment of subjects with multiple sclerosis, since it is believed to provide greater pathologic specificity than conventional MR imaging (169–174). The interpretation of a decrease in the MTR of WM as an indicator of demyelination and tissue damage is supported by histologic studies in which MTR was found to correlate with loss of myelin and/or destruction of axons (175,176). Reduced MTR has been reported in demyelinating plaques of multiple sclerosis (177,178), focal epilepsy (179), brain tumors (180), progressive multifocal leukoencephalopathy (181), wallerian degeneration (182), traumatic brain injury (183), neuromyelitis optica (184), and liver cirrhosis (185).

MTR values are likely to differ on the basis of the macromolecular concentration characteristic of the different underlying histologic structures, thus providing additional information for the detection of structural damage. However, its use to date in assessing AD patients has been limited to only a few studies (143,186–188).

Decreased MTR of the hippocampus has been found in AD patients, compared with that in non-AD patients with medial temporal lobe atrophy (189), which suggests that MTR measurements may be more specific than visual analysis for the detection of structural damage in the hippocampus of AD patients. Reduced MTR of the hippocampus has also been reported in patients with very mild AD (clinical dementia rating, 0.5), compared with elderly control subjects (190). A decrease in MTR was found in the GM of AD and MCI subjects (188), while a decrease of MTR was found in the WM of only AD subjects. This is likely to be due to the fact that the GM is affected in the early stages of the disease (191), and changes begin to occur in the WM as the disease progresses from MCI to AD. What is particularly interesting is that MTR changes were found in the GM and WM in the absence of notable changes in the volume of these tissues in MCI subjects. MTR histograms have been generated in MCI and AD subjects (Fig 5); lower peak heights have been obtained in such subjects, compared with control subjects, for the whole brain and for the temporal and frontal lobes (187). In MCI patients, structural changes occurred outside the temporal lobe without evidence of atrophy, which suggests that magnetization transfer imaging can provide information about the structure of the brain and reflect underlying histopathologic changes. Damage to the hippocampus and to the cortical GM of AD patients has also been demonstrated with magnetization transfer imaging, suggesting that MTR is sensitive to GM abnormalities (146,190). MTR parameters correlated with the Mini-Mental State Examination score, indicating that volumetric magnetization transfer imaging analysis reflects brain damage that is associated with cognitive decline in MCI and AD. These results have been summarized in Table 5.


Figure 5
View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5: MTR histograms of whole brain after correction for parenchymal volume. Y-axis scale is arbitrary and reflects number of voxels with specific MTR divided by total number of parenchymal voxels. Voxels with MTR < 20 are not shown, because they represent CSF. Lower peak heights in MCI and AD suggest less homogeneity in terms of magnetization transfer characteristics and reflect amount of parenchyma affected by the disease. (Reproduced, with permission, from reference 187.)

 

View this table:
[in this window]
[in a new window]

  		Table 5. Summary Data from Studies of MTR by Brain Region

 
The precise mechanism for reduction of the MTR in the hippocampus of AD patients is not yet clear. In AD, histopathologic findings in the hippocampus show loss of pyramidal cells accompanied by an increase in the number of astrocytes, microglia, and oligodendrocytes, as well as by an accumulation of plaques and neurofibrillary tangles (192). Although it is unknown whether plaques and tangles decrease MTR, pathologic changes, including loss of neurons and gliosis, are likely to result in a decrease in MTR due to a decrease in the bound-proton fraction. In addition, demyelination and axonal loss may be other causative factors for reduced MTR (193,194), since degeneration of intrahippocampal nerve fibers often occurs in AD (195). Because cellular changes, synaptic loss, and neuronal degeneration (for which MTR is proposed as a surrogate measure) are likely to precede gross regional atrophy (35), this finding supports the hypothesis that the underlying biologic changes in brain tissue may be detected in the absence of obvious volumetric changes in subjects with MCI (196). Microscopically, the stratum lacunosum radiatum of the hippocampus has shown demyelination and gliosis in patients with AD (197). Further studies in a larger population are necessary to assess the diagnostic value of MTR measurements for application in the clinical diagnosis of early AD.

T1-weighted MR Imaging
T1-weighted MR images are useful for the assessment of the topographic distribution of cortical and subcortical atrophy. With three-dimensional gradient-echo sequences, high-spatial-resolution images can be acquired within 5 minutes (or less). The three-dimensional data acquisition enables calculation of volumes and coregistration of images during follow-up examinations. Several computer programs are now available that facilitate the calculation of total intracranial volume by using three-dimensional T1-weighted semiautomated volumetric imaging techniques (198) and calculation of volumes of distinct mesiotemporal substructures (eg, hippocampus, amygdala, entorhinal cortex) (84).

There have been very few studies that have used T1 relaxation times with the aim of differentiating AD subjects from control subjects. In one study (199), in vitro measurements of T1 and T2 relaxation times of samples of WM from AD brains were significantly greater in the parietal and temporal WM than in the control group, suggesting an increase in tissue water content in the AD group.

T2-weighted MR Imaging
T2 relaxometry allows the quantitative determination of signal intensity changes on T2-weighted images. However, the abnormal T2 values often encountered in various diseases (including AD) may be associated with different pathologic substrates ranging from edema and inflammation to demyelination and axonal loss, thus making interpretation of such values difficult. Studies in which T2-weighted imaging was used to help differentiate AD subjects from control subjects have been very limited (Table 6), and the results have been mixed.


View this table:
[in this window]
[in a new window]

  		Table 6. Summary Data from Studies of T2 Relaxation Time by Brain Region

 
Prolonged hippocampal T2 values have been reported in AD (200), with a correlation between T2 and clinical severity of the disease (Fig 6). T2 was measured in the hippocampal formation, thalamus, and cortical WM in patients with probable AD and in healthy elderly individuals. In patients with AD, elevated T2 values were found in the hippocampus, and these values were highly correlated with the severity of functional and cognitive impairment, suggesting that hippocampal T2 prolongation may provide a specific marker with which AD pathologic changes can be detected and characterized in vivo. In another study (201), T2 values and periventricular high-signal-intensity foci were significantly correlated with dementia severity, as indicated by the Blessed-Roth Dementia Scale score (202). Subsequently, hippocampal volumes and T2 values were analyzed in control and AD subjects (203). In the AD group, the hippocampal volume was 35% smaller than that in control subjects. T2 relaxation time was slightly prolonged (5–6 msec) in the hippocampus, but there was no correlation between T2 values and hippocampal volumes. T2 was measured in the hippocampus, temporal and parietal WM, amygdala, and thalamus (204), and the relaxation times were correlated with clinical severity. Elevated T2 in the WM of the temporal and parietal lobes and in the thalamus were related to increasing age rather than to diagnostic category. Despite the prolongation of T2 in the AD group, the diagnostic value was compromised by a substantial overlap between the study groups, leading the authors to conclude that T2 relaxometry was not a reliable method for diagnosing early AD. Another study (205) found no difference in hippocampal T2 values between patients with probable AD and elderly control subjects and no correlation between T2 values and severity of disease.


Figure 6
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6: Scatterplot shows T2 relaxation time as function of clinical status and suggests that T2 is sensitive to molecular properties of AD-related pathologic changes. Impairment at different stages of diagnosable dementia are classified as follows: Global Deterioration Scale score of 3, early confusional (eg, decreased performance in demanding settings); score of 4, late confusional (eg, inability to perform complex tasks); score of 5, early dementia (eg, patient cannot survive without assistance); and score of 6A–6E, middle dementia (eg, difficulty dressing at 6A to fecal incontinence at 6E). (Reprinted, with permission, from reference 200.)

 
Iron-dependent T2 Contrast
Recently, there has been a growing interest in assessing the role of T2 values as a potential biomarker for characterizing the presence of iron. The basal ganglia contain high levels of iron, and it has been suggested that these iron deposits lead to T2 shortening of bulk water protons by means of a mechanism involving diffusion of water through local magnetic field gradients generated by iron (206). In the presence of ferritin (the primary form of tissue iron storage), T2 values are decreased; this sensitivity of the effect of ferritin iron is evident on T2-weighted images, where iron-rich regions such as the globus pallidus, substantia nigra, red nuclei, and others appear darker at higher field strengths. As a consequence, T2-weighted MR imaging has been proposed as a means of imaging regional cerebral iron levels at field strengths of 1.5 T or higher. A recent study with 3-T MR imaging (207) showed a linear relation between T2 relaxation rate measured in healthy volunteers and the iron concentration predicted on the basis of reported results from postmortem studies (208).

It seems reasonably plausible that this contrast mechanism can be exploited in assessing the involvement of iron in the pathophysiologic features of AD. Oxidative damage to the brain caused by free radical reactions (catalyzed by iron) has been implicated in AD. As Haacke et al (209) pointed out, the entorhinal cortex and substantia innominata are thought to play an important role as markers for iron in AD. T2 shortening in the entorhinal cortex (210) and in the globus pallidus, putamen, and caudate nucleus (211) have been observed in AD patients, compared with T2 in age- and sex-matched control subjects. These MR observations are consistent with results from postmortem biochemical studies (41,46) in which excessive brain iron was measured in cortical and basal ganglia regions in AD brain, suggesting that AD may involve disturbances in brain iron metabolism. Although visualization of iron-containing plaque has recently become possible in mice (212), there have been constraints in sensitivity and spatial resolution, thus preventing a perfect one-to-one correspondence between every plaque seen on MR images and histologic slices. With the advent of MR imagers with a very high field strength and multichannel coils, human applications of this approach may be feasible in the near future.


   TECHNICAL LIMITATIONS OF QUANTITATIVE MR IMAGING
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
SUMMARY
 ESSENTIALS
 References
 
Currently established quantitative MR techniques have several important limitations and associated pitfalls. For instance, inaccuracies in nuclear MR relaxation times can occur because of radiofrequency pulse imperfections (213). In addition, both T1 and T2 may be ill defined owing to multiexponential signal decay, which can be caused by either partial volume effects or microscopic tissue heterogeneity. This is particularly important in the brain where WM, GM, and cerebrospinal fluid reside in proximity and have substantially different T2 values. Consequently, apparent T2 changes in GM have often been attributed to contamination from the presence of cerebrospinal fluid, and the large spread of T2 values reported in the literature may, in part, reflect variations and shortcomings in experimental technique.

Echo-planar MR imaging is most commonly used in diffusion-weighted and diffusion-tensor imaging, resulting in artifacts in areas of large discontinuities in bulk magnetic field susceptibility, such as those that occur near interfaces between brain tissue, bone, and air. These may produce local field gradients that notoriously degrade and distort diffusion-weighted images, particularly at higher field strengths (3 T and higher). In addition, factors such as non-Gaussian diffusion and cerebrospinal fluid pulsation can confound accurate diffusion measurements (214). In the case of magnetization transfer imaging, the measured MTR depends on pulse sequence parameters such as the pulse shape, bandwidth, amplitude, offset frequency, and duration.

At present, the degree to which measured MR imaging parameters can be regarded as truly quantitative depends critically on the hardware and techniques employed. Despite these issues, quantitative MR techniques allow changes in disease progression and the response to potential therapies to be studied with a higher degree of reproducibility. In particular, these methods result in a substantial reduction in the problems associated with bias and interpretation.


   SUMMARY
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
ESSENTIALS
 References
 
While atrophy of the hippocampus, amygdala, and entorhinal cortex is well established in AD, research has more recently focused on determining how long these underlying pathologic conditions are present prior to the onset of AD. Progress in our clinical knowledge of AD has led to more reliable diagnostic criteria and diagnostic accuracy, and research efforts are expanding to uncover the earliest manifestations, and even the presymptomatic phases, of the disease. One of the most likely uses of MR imaging in the future will be in the identification of patients at risk for developing AD. In AD, biochemical changes precede macroscopic structural abnormalities, and quantitative parametric MR imaging may, therefore, be more sensitive than conventional MR imaging in the early stages of the pathologic process and may augment the specificity with which MR-visible abnormalities can be defined.

One of the main goals of all of the quantitative MR methods detailed in this review has been to improve our understanding of the distribution of microstructural damage in AD. Multimodal MR imaging holds promise in facilitating the diagnosis of AD at an early clinical stage and in monitoring the progression of the disease.


   ESSENTIALS
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
References
 


   FOOTNOTES
 

Abbreviations: AD = Alzheimer disease • ADC = apparent diffusion coefficient • GM = gray matter • MCI = mild cognitive impairment • MTR = magnetization transfer ratio • WM = white matter


   References
TOP
 ABSTRACT
 INTRODUCTION
 AD PATHOLOGIC CHARACTERISTICS
 VOLUMETRIC IMAGING
 QUANTITATIVE PARAMETRIC MAPPING
 TECHNICAL LIMITATIONS OF...
 SUMMARY
 ESSENTIALS
 References
 

  1. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 1984;34(7):939–944.[Abstract/Free Full Text]
  2. Esiri MM. Is an effective immune intervention for Alzheimer's disease in prospect? Trends Pharmacol Sci 2001;22(1):2–3.[CrossRef][Medline]
  3. Turner RS. Biomarkers of Alzheimer's disease and mild cognitive impairment: are we there yet? Exp Neurol 2003;183(1):7–10.[CrossRef][Medline]
  4. d'Asseler YM, Koole M, Lemahieu I, et al. Recent and future evolutions in NeuroSPECT with particular emphasis on the synergistic use and fusion of imaging modalities. Acta Neurol Belg 1997;97(3):154–162.[Medline]
  5. Masterman DL, Mendez MF, Fairbanks LA, Cummings JL. Sensitivity, specificity, and positive predictive value of technetium 99-HMPAO SPECT in discriminating Alzheimer's disease from other dementias. J Geriatr Psychiatry Neurol 1997;10(1):15–21.[Medline]
  6. Wahlund LO. Magnetic resonance imaging and computed tomography in Alzheimer's disease. Acta Neurol Scand Suppl 1996;168:50–53.[Medline]
  7. Smith CD. Quantitative computed tomography and magnetic resonance imaging in aging and Alzheimer's disease: a review. J Neuroimaging 1996;6(1):44–53.[Medline]
  8. Burns A. Computed tomography in Alzheimer's disease. Lancet 1993;341(8845):601–602.[CrossRef][Medline]
  9. Mayberg HS. Clinical correlates of PET- and SPECT-identified defects in dementia. J Clin Psychiatry 1994;55(suppl):12–21.[Medline]
  10. Scheltens P, Weinstein HC, Leys D. Neuro-imaging in the diagnosis of Alzheimer's disease. I. Computer tomography and magnetic resonance imaging. Clin Neurol Neurosurg 1992;94(4):277–289.[CrossRef][Medline]
  11. DeCarli C, Kaye JA, Horwitz B, Rapoport SI. Critical analysis of the use of computer-assisted transverse axial tomography to study human brain in aging and dementia of the Alzheimer type. Neurology 1990;40(6):872–883.[Free Full Text]
  12. Engelhardt E, Moreira DM, Laks J, Marinho VM, Rozenthal M, Oliveira AC Jr. Alzheimer's disease and magnetic resonance spectroscopy of the hippocampus [in Portuguese]. Arq Neuropsiquiatr 2001;59(4):865–870.[Medline]
  13. Valenzuela MJ, Sachdev P. Magnetic resonance spectroscopy in AD. Neurology 2001;56(5):592–598.[Abstract/Free Full Text]
  14. Rhinehart DL, Cox LA, Long BW. MR spectroscopy of Alzheimer disease. Radiol Technol 1998;70(1):23–28.[Medline]
  15. Hauw JJ, Duyckaerts C, Delaere P, Lamy C, Henry P. Alzheimer's disease: neuropathological and etiological data. Biomed Pharmacother 1989;43(7):469–482.[CrossRef][Medline]
  16. Beffert U, Cohn JS, Petit-Turcotte C, et al. Apolipoprotein E and beta-amyloid levels in the hippocampus and frontal cortex of Alzheimer's disease subjects are disease-related and apolipoprotein E genotype dependent. Brain Res 1999;843(1-2):87–94.[CrossRef][Medline]
  17. Berg L, McKeel DW Jr, Miller JP, et al. Clinicopathologic studies in cognitively healthy aging and Alzheimer's disease: relation of histologic markers to dementia severity, age, sex, and apolipoprotein E genotype. Arch Neurol 1998;55(3):326–335.[Abstract/Free Full Text]
  18. Chen L, Baum L, Ng HK, Chan LY, Pang CP. Apolipoprotein E genotype and its pathological correlation in Chinese Alzheimer's disease with late onset. Hum Pathol 1999;30(10):1172–1177.[CrossRef][Medline]
  19. Corder EH, Saunders AM, Strittmatter WJ, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 1993;261(5123):921–923.[Abstract/Free Full Text]
  20. Marz W, Scharnagl H, Kirca M, Bohl J, Gross W, Ohm TG. Apolipoprotein E polymorphism is associated with both senile plaque load and Alzheimer-type neurofibrillary tangle formation. Ann N Y Acad Sci 1996;777:276–280.[Abstract]
  21. Ohm TG, Kirca M, Bohl J, Scharnagl H, Gross W, Marz W. Apolipoprotein E polymorphism influences not only cerebral senile plaque load but also Alzheimer-type neurofibrillary tangle formation. Neuroscience 1995;66(3):583–587.[CrossRef][Medline]
  22. Chan W, Fornwald J, Brawner M, Wetzel R. Native complex formation between apolipoprotein E isoforms and the Alzheimer's disease peptide A beta. Biochemistry 1996;35(22):7123–7130.[CrossRef][Medline]
  23. Strittmatter WJ, Weisgraber KH, Huang DY, et al. Binding of human apolipoprotein E to synthetic amyloid beta peptide: isoform-specific effects and implications for late-onset Alzheimer disease. Proc Natl Acad Sci U S A 1993;90(17):8098–8102.[Abstract/Free Full Text]
  24. Wisniewski T, Golabek A, Matsubara E, Ghiso J, Frangione B. Apolipoprotein E: binding to soluble Alzheimer's beta-amyloid. Biochem Biophys Res Commun 1993;192(2):359–365.[CrossRef][Medline]
  25. Sanan DA, Weisgraber KH, Russell SJ, et al. Apolipoprotein E associates with beta amyloid peptide of Alzheimer's disease to form novel monofibrils: isoform apoE4 associates more efficiently than apoE3. J Clin Invest 1994;94(2):860–869.[Medline]
  26. Yang DS, Smith JD, Zhou Z, Gandy SE, Martins RN. Characterization of the binding of amyloid-beta peptide to cell culture-derived native apolipoprotein E2, E3, and E4 isoforms and to isoforms from human plasma. J Neurochem 1997;68(2):721–725.[Medline]
  27. Morris JC, Storandt M, McKeel DW Jr, et al. Cerebral amyloid deposition and diffuse plaques in "normal" aging: evidence for presymptomatic and very mild Alzheimer's disease. Neurology 1996;46(3):707–719.[Free Full Text]
  28. Cataldo AM, Peterhoff CM, Troncoso JC, Gomez-Isla T, Hyman BT, Nixon RA. Endocytic pathway abnormalities precede amyloid beta deposition in sporadic Alzheimer's disease and Down syndrome: differential effects of APOE genotype and presenilin mutations. Am J Pathol 2000;157(1):277–286.[Abstract/Free Full Text]
  29. Haroutunian V, Perl DP, Purohit DP, et al. Regional distribution of neuritic plaques in the nondemented elderly and subjects with very mild Alzheimer disease. Arch Neurol 1998;55(9):1185–1191.[Abstract/Free Full Text]
  30. Blessed G, Tomlinson BE, Roth M. The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects. Br J Psychiatry 1968;114(512):797–811.[Abstract/Free Full Text]
  31. Caramelli P, Robitaille Y, Laroche-Cholette A, et al. Structural correlates of cognitive deficits in a selected group of patients with Alzheimer's disease. Neuropsychiatry Neuropsychol Behav Neurol 1998;11(4):184–190.[Medline]
  32. Naslund J, Haroutunian V, Mohs R, et al. Correlation between elevated levels of amyloid beta-peptide in the brain and cognitive decline. JAMA 2000;283(12):1571–1577.[Abstract/Free Full Text]
  33. Bussiere T, Friend PD, Sadeghi N, et al. Stereologic assessment of the total cortical volume occupied by amyloid deposits and its relationship with cognitive status in aging and Alzheimer's disease. Neuroscience 2002;112(1):75–91.[CrossRef][Medline]
  34. Arnold SE, Hyman BT, Flory J, Damasio AR, Van Hoesen GW. The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer's disease. Cereb Cortex 1991;1(1):103–116.[Abstract/Free Full Text]
  35. Hyman BT, Van Horsen GW, Damasio AR, Barnes CL. Alzheimer's disease: cell-specific pathology isolates the hippocampal formation. Science 1984;225(4667):1168–1170.[Abstract/Free Full Text]
  36. Hyman BT, Van Hoesen GW, Damasio AR. Memory-related neural systems in Alzheimer's disease: an anatomic study. Neurology 1990;40(11):1721–1730.[Abstract/Free Full Text]
  37. Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol (Berl) 1991;82(4):239–259.[CrossRef][Medline]
  38. Braak H, Braak E, Bohl J. Staging of Alzheimer-related cortical destruction. Eur Neurol 1993;33(6):403–408.[Medline]
  39. Gomez-Isla T, Price JL, McKeel DW Jr, Morris JC, Growdon JH, Hyman BT. Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer's disease. J Neurosci 1996;16(14):4491–4500.[Abstract/Free Full Text]
  40. West MJ, Coleman PD, Flood DG, Troncoso JC. Differences in the pattern of hippocampal neuronal loss in normal ageing and Alzheimer's disease. Lancet 1994;344(8925):769–772.[CrossRef][Medline]
  41. Connor JR, Menzies SL, St Martin SM, Mufson EJ. A histochemical study of iron, transferrin, and ferritin in Alzheimer's diseased brains. J Neurosci Res 1992;31(1):75–83.[CrossRef][Medline]
  42. Smith MA, Wehr K, Harris PL, Siedlak SL, Connor JR, Perry G. Abnormal localization of iron regulatory protein in Alzheimer's disease. Brain Res 1998;788(1-2):232–236.[CrossRef][Medline]
  43. Smith MA, Harris PL, Sayre LM, Perry G. Iron accumulation in Alzheimer disease is a source of redox-generated free radicals. Proc Natl Acad Sci U S A 1997;94(18):9866–9868.[Abstract/Free Full Text]
  44. Bartzokis G, Sultzer D, Lu PH, Nuechterlein KH, Mintz J, Cummings JL. Heterogeneous age-related breakdown of white matter structural integrity: implications for cortical "disconnection" in aging and Alzheimer's disease. Neurobiol Aging 2004;25(7):843–851.[CrossRef][Medline]
  45. Bartzokis G. Age-related myelin breakdown: a developmental model of cognitive decline and Alzheimer's disease. Neurobiol Aging 2004;25(1):5–18.[CrossRef][Medline]
  46. Connor JR, Snyder BS, Arosio P, Loeffler DA, LeWitt P. A quantitative analysis of isoferritins in select regions of aged, parkinsonian, and Alzheimer's diseased brains. J Neurochem 1995;65(2):717–724.[Medline]
  47. Lovell MA, Robertson JD, Teesdale WJ, Campbell JL, Markesbery WR. Copper, iron and zinc in Alzheimer's disease senile plaques. J Neurol Sci 1998;158(1):47–52.[CrossRef][Medline]
  48. DeCarli C. The role of neuroimaging in dementia. Clin Geriatr Med 2001;17(2):255–279.[CrossRef][Medline]
  49. Burggren AC, Bookheimer SY. Structural and functional neuroimaging in Alzheimer's disease: an update. Curr Top Med Chem 2002;2(4):385–393.[CrossRef][Medline]
  50. Felber SR. Magnetic resonance in the differential diagnosis of dementia. J Neural Transm 2002;109(7-8):1045–1051.[CrossRef][Medline]
  51. Hsu EW, Buckley DL, Bui JD, Blackband SJ, Forder JR. Two-component diffusion tensor MRI of isolated perfused hearts. Magn Reson Med 2001;45(6):1039–1045.[CrossRef][Medline]
  52. Kumari V, Mitterschiffthaler MT, Sharma T. Neuroimaging to predict preclinical Alzheimer's disease. Hosp Med 2002;63(6):341–345.[Medline]
  53. Scheltens P, Fox N, Barkhof F, De Carli C. Structural magnetic resonance imaging in the practical assessment of dementia: beyond exclusion. Lancet Neurol 2002;1(1):13–21.[CrossRef][Medline]
  54. Anstey KJ, Maller JJ. The role of volumetric MRI in understanding mild cognitive impairment and similar classifications. Aging Ment Health 2003;7(4):238–250.[CrossRef][Medline]
  55. Atiya M, Hyman BT, Albert MS, Killiany R. Structural magnetic resonance imaging in established and prodromal Alzheimer disease: a review. Alzheimer Dis Assoc Disord 2003;17(3):177–195.[CrossRef][Medline]
  56. Chetelat G, Baron JC. Early diagnosis of Alzheimer's disease: contribution of structural neuroimaging. Neuroimage 2003;18(2):525–541.[CrossRef][Medline]
  57. DeKosky S. Early intervention is key to successful management of Alzheimer disease. Alzheimer Dis Assoc Disord 2003;17(suppl 4):S99–S104.[CrossRef][Medline]
  58. Good CD. Dementia and ageing. Br Med Bull 2003;65:159–168.[Abstract/Free Full Text]
  59. Kantarci K, Jack CR Jr. Neuroimaging in Alzheimer disease: an evidence-based review. Neuroimaging Clin N Am 2003;13(2):197–209.[CrossRef][Medline]
  60. Lee BC, Mintun M, Buckner RL, Morris JC. Imaging of Alzheimer's disease. J Neuroimaging 2003;13(3):199–214.[CrossRef][Medline]
  61. Norfray JF, Provenzale JM. Alzheimer's disease: neuropathologic findings and recent advances in imaging. AJR Am J Roentgenol 2004;182(1):3–13.[Free Full Text]
  62. Petrella JR, Coleman RE, Doraiswamy PM. Neuroimaging and early diagnosis of Alzheimer disease: a look to the future. Radiology 2003;226(2):315–336.[Abstract/