**2.2 Imaging Alzheimer's disease with [18F]FDG**

[18F]Fluoro-2-deoxy-D-glucose ([18F]FDG) is the most commonly used radiopharmaceutical for clinical PET imaging to date. Patients receive an injection of [18F]FDG, and then images are typically obtained 30 – 60 min later. As a radiolabeled analog of glucose, [18F]FDG is typically employed as a marker of cell proliferation as it preferentially accumulates in cells with increased glucose consumption (e.g. tumors). Therefore, [18F]FDG finds widespread application in oncology including diagnosis and staging of cancers, and monitoring tumor response to chemotherapy. However, glucose is also the main energy supply for the brain and, reflecting this, levels are closely coupled to neuronal function so that measurement of cerebral glucose metabolism can provide diagnostically relevant information about the neurodegenerative disorders described above. According to Herholz and colleagues (Herholz, et al., 2007), typical resting state cerebral metabolic rate for glucose is 40-60 µmol glucose/100 g/min for grey matter, and 15 µmol glucose/100 g/min for white matter, although this does drop off somewhat with age (Kuhl, et al., 1982). Observed regional differences include higher values in the striatum and parietal cortex. Other phylogenetically older brain structures (e.g. medial temporal cortex and cerebellum) have glucose metabolism rates between grey matter and white matter.

For at least two decades, significant efforts have been made to image patients at various stages of Alzheimer's disease (including high risk, asymptomatic patients; patients with mild cognitive impairment (MCI); and patients with fully developed Alzheimer's disease) with [18F]FDG (e.g. Figure 1). In patients considered high risk for developing AD (for example because of family history and possession of the ApoE ε4 allele (Reiman, et al., 1996; Small, et al., 1995)), impairment of regional cerebral glucose metabolism has been observed decades before likely onset of dementia and certainly while the patients are still asymptomatic (Reiman, et al., 2004).

In 1997, Kuhl and colleagues reported the first example of using posterior cingulate glucose metabolism, determined from [18F]FDG PET scans, to predict progression of disease in patients with MCI (Minoshima, et al., 1997). The results have been echoed by a number of subsequent longitudinal studies, which have confirmed the high predictive power of

Diagnosis of Dementia Using Nuclear Medicine Imaging Modalities 203

sensory processing decline. Such clinical and imaging differences can be used as a means of

More subtle differences between FDG PET scans of normal patients and AD patients are also apparent if more advanced image analysis is employed. Voxel-based analysis can detect differences in FDG PET that are not obvious upon visual interpretation of the scan. For example, comparison of FDG PET scans of AD patients versus normal controls reveals impaired glucose metabolism in the posterior cingulate gyrus as well as the precuneus (Minoshima, et al., 1997). This is not immediately obvious because metabolism in these areas is normally above average and the decrease in AD patients can obscure it into the background. The large amounts of FDG PET data obtained in AD patients now allow advanced analytical techniques to go one step further, and automatically detect the typical pattern of metabolic abnormalities associated with AD. This approach has been successfully used to distinguish AD patients for normal controls with >80% (and frequently >90%) accuracy, and such data have enabled the use of FDG PET in AD therapeutic trials (Heiss, et

As Aβ deposition is considered a hallmark neuropathological sign of AD and is thought to be one of the primary events in the pathogenesis of AD (the "amyloid cascade hypothesis"), Aβ PET imaging agents are at the forefront of this expanding field (Haass and Selkoe, 2007). In AD patients, Aβ deposits are composed of Aβ1-40 and Aβ1-42, peptides that are 40 and 42 amino acids in length, respectively, generated from the sequential proteolytic cleavage of amyloid precursor protein (APP) by β- and γ-secretases (Beyreuther and Masters, 1991; Martins, et al., 1991). Whilst Aβ1-42 deposits have a higher propensity to oligomerize, both isoforms are found in fibrillar amyloid plaques (Tamaoka, et al., 1994), characteristic β-sheet-

The ability to image and measure Aβ load *in vitro* has considerable implications for the future study of AD. As Aβ deposition commences long before the onset of cognitive deficits, Aβ-targeting probes may support earlier diagnoses and interventions in the pre-dementia stage of this disease (Pike, et al., 2007; Price and Morris, 1999; Thal, et al., 2002). Aβ PET imaging additionally has some utility in accurately differentiating AD from Aβ-negative forms of dementia and, thus, in increasing the specificity of a clinical diagnosis. Objectively monitoring treatments and selecting candidates for particular drugs and clinical trials are other potential uses for this imaging technology (Rafii and Aisen, 2009); for example, patients with low cortical Aβ load, as measured by Aβ PET, are unlikely to qualify for anti-

Due to the large number of possible clinical applications, there has been a dramatic rise in the number of probes that target Aβ plaques over the last decade. The four Aβ PET imaging agents under active commercial development are [11C]PIB, [18F]flutemetamol, [18F]florbetaben, and [18F]florbetapir. Each imaging agent is currently undergoing FDAapproved Phase III clinical trials in the US, except [18F]Florbetapir, which is awaiting FDA approval. Each Aβ PET imaging agent will be discussed individually in further detail below.

The most well-characterized and studied radiopharmaceutical for Aβ pathology is [11C]Pittsburgh Compound B ([11C]PiB, (N-methyl-[11C])2-(4'-methylamino-phenyl)-6-OH-

**2.3 Imaging amyloid-ß (Aß) plaque burden in Alzheimer's disease** 

rich structures that represent the target for Aβ PET ligands.

**2.3.1 [11C]Pittsburgh Compound B ([11C]PiB)** 

differentiating AD from other related disorders.

al., 1994).

amyloid therapies.

Fig. 1. 18F-FDG and PiB group mean images for control, naMCI, aMCI, and AD subjects showing better visual separation of groups using PiB. Scaling shown to right using pons and cerebellar normalization, respectively. Regions with activity similar to these regions of normalization color in 1.0 color ranges (green), whereas regions with greater uptake show up in yellow and red. Color scaling is slightly different for 18F-FDG and PiB groups given different range of cortical ratios. *(Reprinted with permission from Loew VJ, Kemp BJ, Jack CR, et al. Comparison of 18F-FDG and PiB PET in Cognitive Impairment. J Nucl Med. 2009;50:878-886)* 

[18F]FDG PET. For example, Berent and colleagues reported 70% progression of disease in 3 years following an abnormal PET scan, but only 30% progression after a normal PET scan (Berent, et al., 1999). Anchisi and co-workers also showed that a normal FDG PET scan indicated low chance of progression of MCI into full AD within 1 year (Anchisi, et al., 2005). Generally speaking [18F]FDG PET has >80% sensitivity and specificity for prediction of rapid progression. Moreover, a recent report by Drzezga and colleagues discovered that [18F]FDG PET (92% sensitivity, 89% specificity) was superior to ApoE ε4 testing (75% sensitivity, 56% specificity) when used to predict disease progression (A. Drzezga, et al., 2005). Other aspects of [18F]FDG PET scans have also proven pertinent to the assessment of MCI. For example, mesial temporal metabolic impairment (Heiss, et al., 1992), as well as hippocampal and antorhinal metabolic impairment (de Leon, et al., 2001), although the latter can be difficult to assess because of small size.

As MCI progresses and develops into Alzheimer's disease, numerous studies over the last two decades have shown that both cerebral blood flow and glucose metabolism are reduced in a number of areas of the brain. For example, impairment of glucose metabolism in the temporoparietal association cortices is typical in AD (on the order of 16-19% over 3 years (Smith, et al., 1992)), whilst no significant decline is apparent in the corresponding normal controls. These cortices are also prone to amyloid deposition during AD (Bartzokis, et al., 2007). Reduced glucose metabolism may also occur in the frontal association cortex, although more so as AD progresses. In contrast to the other dementia diseases discussed herein, the rate of glucose metabolism in other areas of the brain including the visual and motor cortex, basal ganglia and cerebellum is unaffected (Herholz, et al., 2007). This is in agreement with the clinical manifestation of AD, as primary motor and sensory function remain intact, whilst memory, associative thinking, planning of action and other higher

Fig. 1. 18F-FDG and PiB group mean images for control, naMCI, aMCI, and AD subjects showing better visual separation of groups using PiB. Scaling shown to right using pons and cerebellar normalization, respectively. Regions with activity similar to these regions of normalization color in 1.0 color ranges (green), whereas regions with greater uptake show up in yellow and red. Color scaling is slightly different for 18F-FDG and PiB groups given different range of cortical ratios. *(Reprinted with permission from Loew VJ, Kemp BJ, Jack CR, et al. Comparison of 18F-FDG and PiB PET in Cognitive Impairment. J Nucl Med. 2009;50:878-886)*  [18F]FDG PET. For example, Berent and colleagues reported 70% progression of disease in 3 years following an abnormal PET scan, but only 30% progression after a normal PET scan (Berent, et al., 1999). Anchisi and co-workers also showed that a normal FDG PET scan indicated low chance of progression of MCI into full AD within 1 year (Anchisi, et al., 2005). Generally speaking [18F]FDG PET has >80% sensitivity and specificity for prediction of rapid progression. Moreover, a recent report by Drzezga and colleagues discovered that [18F]FDG PET (92% sensitivity, 89% specificity) was superior to ApoE ε4 testing (75% sensitivity, 56% specificity) when used to predict disease progression (A. Drzezga, et al., 2005). Other aspects of [18F]FDG PET scans have also proven pertinent to the assessment of MCI. For example, mesial temporal metabolic impairment (Heiss, et al., 1992), as well as hippocampal and antorhinal metabolic impairment (de Leon, et al., 2001), although the latter can be difficult to

As MCI progresses and develops into Alzheimer's disease, numerous studies over the last two decades have shown that both cerebral blood flow and glucose metabolism are reduced in a number of areas of the brain. For example, impairment of glucose metabolism in the temporoparietal association cortices is typical in AD (on the order of 16-19% over 3 years (Smith, et al., 1992)), whilst no significant decline is apparent in the corresponding normal controls. These cortices are also prone to amyloid deposition during AD (Bartzokis, et al., 2007). Reduced glucose metabolism may also occur in the frontal association cortex, although more so as AD progresses. In contrast to the other dementia diseases discussed herein, the rate of glucose metabolism in other areas of the brain including the visual and motor cortex, basal ganglia and cerebellum is unaffected (Herholz, et al., 2007). This is in agreement with the clinical manifestation of AD, as primary motor and sensory function remain intact, whilst memory, associative thinking, planning of action and other higher

assess because of small size.

sensory processing decline. Such clinical and imaging differences can be used as a means of differentiating AD from other related disorders.

More subtle differences between FDG PET scans of normal patients and AD patients are also apparent if more advanced image analysis is employed. Voxel-based analysis can detect differences in FDG PET that are not obvious upon visual interpretation of the scan. For example, comparison of FDG PET scans of AD patients versus normal controls reveals impaired glucose metabolism in the posterior cingulate gyrus as well as the precuneus (Minoshima, et al., 1997). This is not immediately obvious because metabolism in these areas is normally above average and the decrease in AD patients can obscure it into the background. The large amounts of FDG PET data obtained in AD patients now allow advanced analytical techniques to go one step further, and automatically detect the typical pattern of metabolic abnormalities associated with AD. This approach has been successfully used to distinguish AD patients for normal controls with >80% (and frequently >90%) accuracy, and such data have enabled the use of FDG PET in AD therapeutic trials (Heiss, et al., 1994).
