Alzheimer’s disease (AD) is the most common cause of dementia. Several million people worldwide are afflicted by the disease, and the number of individuals affected is expected to grow with the increasing life expectancy. Generally, it is diagnosed in people over 65 years of age, although the less-prevalent early-onset AD can occur much earlier. It consist 65% of all demented patients. The first symptoms of the disease are often mistaken as related to aging or stress. At this stage, neuropsychological testing can reveal mild cognitive impairment (MCI). These early symptoms can have an effect on the most complex daily living activities. The most noticeable deficits observed at this stage are memory loss (short-term memory loss), problems with executive functions, and apathy. In patients with AD, the increasing impairment of learning and memory will lead to a definitive diagnosis. Subsequently, language problems and apraxia will appear. At this stage, AD patients can still perform tasks independently, but may need assistance or supervision with the most complicated activities. As the disease progresses further, progressive deterioration hinders independence. Reading and writing skills are progressively lost, urinary incontinence can develop, motor and memory problems worsen, and behavioral changes become evident. During the latest stage of AD, the patient is completely dependent upon caregivers, language is reduced to simple phrases or single words, and mobility deteriorates to the point where the patient is bedridden. Death occurs from complications such as pressure ulcers and pneumonia, and not from the disease itself.
The brains of individuals with AD are characterized by the presence of abundant neurofibrillary tangles (pathological protein aggregates found within neurons) and neuritic plaques (deposits of the betaamyloid in the extracellular spaces). These deposits ultimately cause neuron disintegration, collapsing the neuron’s transport system. Although many individuals develop some plaques and tangles as a consequence of aging, the brains of AD patients usually have a greater number of them in specific brain regions such as the limbic system and temporal neocortex, with a tendency to spread to other brain regions as the disease progresses. However, typically this pathology manifests as clinical AD only after a certain quantitative threshold is reached, and by the time the individual is diagnosed with AD, a significant loss of synapses and neurons has already occurred. In agreement with pathologic findings, most MRI studies have reported global or focal signs of brain atrophy in the brains of AD patients. Gray-matter atrophy was consistently found in the frontal, temporal, and parietal lobes and limbic system of patients with AD, presumably reflecting neuron loss in these regions. In this context, accurate volumetric measures of regional brain volumes on MRI images have demonstrated to support diagnostic decision-making and differential diagnosis. This is particularly useful for detection of cerebral changes in MCI, as these are usually too subtle to be detected by visual inspection of MRI scans alone. Compared with elderly controls, patients with MCI seem to have significant hippocampal and entorhinal cortex volume losses. However, as expected, gray matter volume loss is in general less severe and diffuse in MCI than in clinically evident AD patients.

1H-MRSpectroscopy

Large decreases in brain NAA have been observed since the earliest proton MRS and MR spectroscopic imaging (MRSI) studies of AD patients. Reduction of NAA concentration, or the ratio of NAA to other metabolites such as Cr, has been consistently found in the mesial temporal lobe, posterior cingulate gyrus, parietotemporal region, frontal lobe, occipital lobe, and hippocampus. These NAA decreases are much less evident in the white matter, probably due to the fact that AD affects the cortical regions primarily. Measurements that account for atrophy in acquired voxels show that the decreases in NAA are independent of CSF content. NAA is a sensitive marker for neuronal density or viability, but its loss is certainly not specific for AD. However, it is interesting to note that decreases in NAA seem more pronounced in anatomic locations that show higher severity of neuropathologic findings (e.g. amyloid plaques and neurofibrillary tangles) on postmortem studies. This suggests that in AD patients, brain NAA changes do correlate with neuronal loss or dysfunction in patients with AD.
Proton MRS studies have reported conflicting results on choline (Cho) metabolite levels in brains of AD patients, with increases, no changes or decreases of this metabolite in different brain locations. In contrast, increases in myo-inositol (mI) have consistently been reported in several anatomic locations of brains of AD patients, with again an effect more pronounced in gray matter (e.g. mesial temporal lobe, anterior and posterior cingulate gyrus, and parietal lobe) than in white matter. As most of the mI in the brain is present in glial cells, it is likely that persistent elevation of mI levels reflects microglial proliferation in AD. In patients with MCI, brain regional levels of mI are often increased without a large decrease in NAA, suggesting that MRS may be sensitive to the biochemical changes in the pathological progression of prodromal AD even before there is a significant loss of neuronal integrity.
Owing to the consistent coexistence of high mI and low NAA brain levels in patients with AD, the ratio of NAA/mI has been proposed by several authors as the most accurate 1H-MRS measurement in this disease. Whereas the clinical specificity of the NAA decline in AD is poor, the addition of mI information increases accuracy. Thus, the NAA/mI ratio is able to distinguish clinically diagnosed patients with AD from normal elderly people, with good sensitivity and specificity. For the more challenging task of discriminating AD cases from other possible dementia diagnoses, the NAA/mI ratio can still be useful.
Whereas NAA levels are almost constantly decreased in the different dementia types, levels of mI are elevated predominantly in dementias that are pathologically characterized by gliosis, such as AD. Thus, higher levels of NAA/mI are expected in patients with, for example, vascular dementia or in dementia with Lewy bodies where gliosis is little or absent.
Furthermore, in patients with AD, neuropsychological measures of cognitive function may correlate with levels of the NAA/mI ratio, with region-specific association between neuropsychological performance and MRS metabolite changes depending on the cognitive domain being studied. The intriguing suggestion that 1H-MRS may have a useful role in prognosis of mental function and tracking of disease progression seems to be confirmed by the close correlation recently reported between antemortem metabolite changes found on 1H-MRS examination and AD-type pathology subsequently seen in the same brains at autopsy (e.g. strong association of NAA/mI and Braak stage, a histopathological estimate of extent of neurofibrillary tangle involvement).
Despite its potential to monitor the temporal evolution of metabolite changes, longitudinal 1H-MRS studies are uncommon. There may be several reasons for this, including technical difficulties.
For example, it is technically demanding to obtain reproducible measurements from the anteromedial temporal lobe (a region which is known to be affected earlier and more severely than any other brain regions in AD patients), as this brain region is in proximity to the tissue–air interface near the petrous bone and, consequently, difficult to achieve a homogenous magnetic field and an optimal water suppression within that VOI. In addition, the ~ 8 cm³ voxel size generally used to obtain the sufficient SNR is much larger than the volume of the cortical gray matter structures, with the consequent lack in the anatomic specificity of the measurements.
Since short TE (30–35 ms) 1H-MRSI acquisition has become easier to use and routinely available on clinical MR systems, this should be the preferred method for quantification of spectroscopic metabolites in AD brains. Furthermore, more sophisticated postprocessing using segmented MRIs to calculate percentages of gray and white matter in each 1H-MRSI voxel can allow estimation of “pure” cortical changes, improving the reliability of the spectroscopic measure. Alternatively, short TE single voxel (usually 8 cm³ size) 1H-MRS acquisitions from the posterior cingulate region and mesial occipitoparietal cortex provide good quality data and seem to show changes specific enough to AD.

Familial Alzheimer’s disease
Familial Alzheimer’s disease (FAD) accounts for only 5% of cases, but has provided invaluable insights into the pathogenesis of sporadic disease. All forms show autosomal dominant inheritance with a high degree of penetrance. Presentation is usually below the age of 65 and the term early-onset Alzheimer’s disease is therefore used. Most cases of familial Alzheimer’s disease are due to mutations in one of the three main genes involved in amyloid processing:
■ APP, on chromosome 21, which encodes amyloid precursor protein.
■ PSEN1, on chromosome 14, which encodes presenilin-1 (PS-1).
■ PSEN2, on chromosome 1, which encodes presenilin-2 (PS-2).
Mutations in the genes for PS-1 or PS-2 account for nearly half of familial Alzheimer’s disease cases. There is usually increased Aβ production, with alteration of the Aβ40:Aβ42 ratio, so that more of the plaque-forming Aβ42 species is produced. In many FAD families with known mutations in the APP gene, the region of the protein flanking the amyloid beta sequence is affected, which is also
associated with increased Aβ42 production. This is in contrast to mutations affecting the central portion of the amyloid sequence, which are more likely to cause cerebral amyloid angiopathy.

Finally, it must be stressed that the diagnosis of AD as well as other forms of dementia is difficult and relies mainly on clinical information. Quantitative MR techniques are not recommended for routine use of dementia evaluation at this time because superiority to clinical criteria has not been demonstrated. However, the early diagnosis of AD using a combination of quantitative MR methodologies, including 1H-MRSI, seems possible in the near future.