Unfortunately, the 20-min radioactive decay half-life of carbon-11 (¹¹C) limits the use of ¹¹C-PiB to centres with an on-site cyclotron and ¹¹C radiochemistry expertise, making the cost of studies prohibitive for routine clinical use. To overcome these limitations, several Aβ tracers labelled with fluorine-18 (¹⁸F; half-life of 110 min) that permits centralised production and regional distribution have been developed and tested. Recent studies with the newly introduced ⁸F-labelled Aβ-specific radiotracers, ¹⁸F-florbetapir (Wong et al. 2010; Clark et al. 2011), ¹⁸F-florbetaben (Rowe et al. 2008; Barthel et al. 2011; Villemagne et al. 2011a), ¹⁸F-flutemetamol (Nelissen et al. 2009; Vandenberghe et al. 2010b) and ¹⁸F-AZD4694 (Cselenyi et al. 2012), have been successfully replicating the results obtained with ¹¹C-PiB. The latest entry into the list of fluorinated Aβ radiotracers is ¹⁸F-MK-3328, developed by Merck, that has the disadvantage that >17 % of its cortical signal was attributed to binding to monoamine oxidase-B (MAO-B) (Sur et al. 2010).

On visual inspection, cortical ¹¹C-PiB retention is higher in AD, with a regional brain distribution that is highest in frontal, cingulate, precuneus, striatum, parietal and lateral temporal cortices, while occipital, sensorimotor and mesial temporal cortices are much less affected (Fig. 10.1). Both quantitative and visual assessments of PET images present a pattern of radiotracer retention that seems to replicate the sequence of Αβ deposition found at autopsy (Braak and Braak 1997), with initial deposition in the orbitofrontal cortex, inferior temporal, cingulate gyrus and precuneus, followed by the remaining prefrontal cortex and lateral temporal and parietal cortices (Fig.10.2). Multimodality studies in early AD have shown that the regional pattern of Αβ deposition is similar to the anatomy of the ‘default network’ (Buckner et al. 2005; Sperling et al. 2009), a specific, anatomically defined brain system responsible for internal modes of cognition, such as self-reflection processes, conscious resting state or episodic memory retrieval (Buckner et al. 2008; Wermke et al. 2008), an association that is present even in non-demented subjects (Drzezga et al. 2011).

The regional distribution of ¹¹C-PiB retention is not always the same and varies with the specific Αβ distribution characteristic of the underlying pathology. For example, carriers of mutations associated with familial AD (Klunk et al. 2007; Koivunen et al. 2008b; Villemagne et al. 2009a), subjects with posterior cortical atrophy (Ng et al. 2007b; Tenovuo et al. 2008; Kambe et al. 2010; Formaglio et al. 2011) or CAA (Johnson et al. 2007; Dierksen et al. 2010) each presenting a different regional ¹¹C-PiB distribution of retention to the one usually observed in sporadic AD (Klunk et al. 2004; Rowe et al. 2007).

Quantitative PET studies using diverse Αβ radiotracers have shown a robust difference in retention between AD patients and age-matched controls (Klunk et al. 2004; Price et al. 2005; Mintun et al. 2006; Kudo et al. 2007; Pike et al. 2007; Rabinovici et al. 2007; Rowe et al. 2007, 2010; Jack et al. 2008a; Morris et al. 2009; Furst et al. 2012; Barthel et al. 2011; Fleisher et al. 2011; Villemagne et al. 2011a; Camus et al. 2012; Cselenyi et al. 2012).

Longitudinal studies with ¹¹C-PiB are showing that significant, although small, increases in Αβ deposition can be measured but also that these increases in Αβ deposition are present in the whole spectrum of cognitive stages, from cognitively unimpaired individuals to AD dementia patients (Engler et al. 2006; Jack et al. 2009; Okello et al. 2009b; Resnick et al. 2010; Rinne et al. 2010; Sojkova et al. 2011b; Villemagne et al. 2011b; Villain et al. 2012). Αβ accumulation is observed in individuals considered to have low Αβ loads with about 7 % of them progressing above the threshold in about 2.5 years (Vlassenko et al. 2011). Αβ accumulation is a slow process, and the evidence suggests that it remains constant in the preclinical and prodromal stages of the disease in those individuals deemed to have high Αβ burdens while being slightly higher at the early stages of AD (Villemagne et al. 2011b). In individuals with the highest Aβ burdens independent of clinical classification and also as Alzheimer’s dementia progresses, the rates of Aβ deposition start to slow down, although not completely reaching a plateau (Villain et al. 2012). This slower rate of Αβ deposition associated with disease progression might be attributed to a saturation of the process or, more likely, extensive neuronal death that precludes Αβ production.

While clinicopathological studies have shown that the accuracy of clinical assessments for the diagnosis of AD ranges between 70 and 90 %, depending on the specialisation of the centre, the regional retention of ¹¹C-PiB and the novel ¹⁸F radiotracers is highly correlated with regional Αβ plaques as reported at autopsy or biopsy (Bacskai et al. 2007; Ikonomovic et al. 2008; Leinonen et al. 2008b; Villemagne et al. 2009b; Burack et al. 2010; Clark et al. 2011; Kadir et al. 2011; Sabbagh et al. 2011; Sojkova et al. 2011a; Wolk et al. 2011; Wong et al. 2013), with a higher Aβ concentration in the frontal cortex than in hippocampus, consistent with previous reports (Arnold et al. 2000; Naslund et al. 2000). As mentioned before and in contrast with CSF studies, Αβ imaging studies can assess both the quantity and regional brain distribution of Αβ, as observed in the different patterns of regional distribution of Αβ observed in sporadic AD (Klunk et al. 2004; Nordberg 2007), familial AD (Klunk et al. 2007; Koivunen et al. 2008b; Villemagne et al. 2009a), familial British dementia (Villemagne et al. 2010) and CAA subjects (Johnson et al. 2007; Dierksen et al. 2010). The results from Αβ imaging might be considered for the brain regions to be sampled and stained for the neuropathological diagnosis of AD (Montine et al. 2012).

There have been a handful of cases reported with a discrepancy between a positive neuropathological finding or a strong biomarker profile of AD and a low PET signal. To date, only seven discrepant cases have been reported, three of them in mutation carriers (Tomiyama et al. 2008; Schöll et al. 2011) and four in sporadic cases (Leinonen et al. 2008a; Cairns et al. 2009; Sojkova et al. 2011a; Ikonomovic et al. 2012). No ‘false-positive’ cases have been reported. Tomiyama and colleagues reported a case of a novel APP mutation (E693Delta) where the mutant Αβ does not fibrillise in the same way as wild-type Aβ (Inayathullah and Teplow 2011) in which the ¹¹C-PiB study showed low cortical retention (Tomiyama et al. 2008). Schöll and colleagues report on two subject carriers of the Arctic APP mutation (Schöll et al. 2011). These two mutation carriers, while presenting with low ¹¹C-PiB retention, had a clear ‘AD pattern’ of brain hypometabolism as measured by FDG, low Aβ_1–42 and elevated t-tau and p-tau in CSF (Schöll et al. 2011). While neuropathology was not available in these two subjects, a family member carrier of the same mutation presented with atypical plaque morphology—with ringlike plaques lacking a congophilic core—at autopsy (Basun et al. 2008). In regard to sporadic cases, Leinonen et al., while correlating ¹¹C-PiB retention and frontal lobe Αβ burden from biopsies on subjects undergoing intraventricular monitoring for normal pressure hydrocephalus, described a non-demented case with plaques that had a negative ¹¹C-PiB scan (Leinonen et al. 2008a). Cairns et al. reported a longitudinally characterised subject with a low ¹¹C-PiB scan that presented with evidence of cognitive decline and a CSF profile indicative of AD a year later (Cairns et al. 2009). At the time of autopsy, two and a half years after the ¹¹C-PiB PET, diffuse Αβ plaques were observed with only minimal neuritic plaques and NFT pathology. Sojkova et al. reported imaging-neuropathological correlations in five participants from the Baltimore Longitudinal Study of Aging (Sojkova et al. 2011a). While there was agreement in four subjects, even in one that had only sparse plaques at autopsy, a fifth presented with low global and regional ¹¹C-PiB retention despite presenting moderate plaques at neuropathological examination and meeting CERAD criteria for ‘possible AD’ (Mirra et al. 1991). Ikonomovic et al. (2012) reported autopsy findings in a subject with a clinical diagnosis of DLB and possible AD that presented with low ¹¹C-PiB retention 17 months before death. At autopsy, focally frequent neocortical neuritic plaques fulfilled CERAD criteria of ‘definite AD’ (Mirra et al. 1991). Cortical Αβ_1–40 concentration levels were on the same range to those found in a typical AD case with high ¹¹C-PiB retention, but Αβ_1–42 concentrations were lower. Furthermore, the majority of plaques labelled weakly with ¹¹C-PiB and ³H-PiB binding were low, all findings suggesting low fibrillar Αβ.

On the one hand, these cases illustrate the possibility that different conformations of Aβ deposits (Levine and Walker 2010b) may affect the binding of tracers and that Aβ imaging may not recognise all types of Aβ pathologies with equal sensitivity (Walker et al. 2008; Rosen et al. 2011). On the other, it raises issues not only in regard to the sensitivity of Aβ imaging (e.g. What is the minimal amount of Aβ detected by PET?) but also in regard to the evaluation of brain tissue (e.g. type of antibodies used for immunohistochemistry and ELISA, fluorescent dyes, diagnostic criteria).

Αbout 25–35 % of elderly subjects performing within normal limits on cognitive tests present with high cortical ¹¹C-PiB retention, predominantly in the prefrontal and posterior cingulate/precuneus regions (Mintun et al. 2006; Rowe et al. 2007, 2010; Aizenstein et al. 2008; Villemagne et al. 2008b; Mormino et al. 2009; Reiman et al. 2009b). These findings are in perfect agreement with postmortem reports that ~25 % of non-demented older individuals over the age of 75 have Aβ plaques (Davies et al. 1988; Morris and Price 2001; Forman et al. 2007). Aβ deposition in these non-demented individuals (Mintun et al. 2006; Rowe et al. 2007; Aizenstein et al. 2008; Villemagne et al. 2008b) might reflect preclinical AD (Price and Morris 1999). Furthermore, the prevalence of subjects with high ¹¹C-PiB retention increases each decade at the same rate as the reported prevalence of plaques in non-demented subjects in autopsy studies (Rowe et al. 2010). While cross-sectional examination of these controls essentially shows no significant differences in cognition (Mintun et al. 2006; Aizenstein et al. 2008; Jack et al. 2008a; Villemagne et al. 2008b; Rowe et al. 2010), significant cognitive differences in females were reported (Pike et al. 2011). More recently, Sperling and colleagues reported that high Aβ deposition in the brain was correlated with lower immediate and delayed memory scores in otherwise healthy controls (Sperling et al. 2013). The detection of Aβ pathology at the presymptomatic stage is of crucial importance because if this group truly represents preclinical AD (Price and Morris 1999; Thal et al. 2004; Backman et al. 2005; Small et al. 2007), it is precisely the group that may benefit the most from therapies aimed at reducing or eliminating Αβ from the brain before irreversible neuronal or synaptic loss occurs (Sperling et al. 2011b).

Individuals fulfilling criteria for MCI are a heterogeneous group with a wide spectrum of underlying pathologies (Petersen 2000; Winblad et al. 2004). About 40–60 % of carefully characterised subjects with MCI will subsequently progress to meet criteria for AD over a 3–4-year period (Petersen et al. 1995, 1999; Petersen 2000). Αβ imaging has proven useful in identifying MCI individuals with and without AD pathology. Approximately 50–70 % of the subjects classified as MCI present with high cortical ¹¹C-PiB retention (Price et al. 2005; Kemppainen et al. 2007; Pike et al. 2007; Forsberg et al. 2008; Lowe et al. 2009; Mormino et al. 2009). While most studies found that subjects classified as non-amnestic MCI show low Αβ levels in the brain (Pike et al. 2007; Lowe et al. 2009), other reports did not (Wolk et al. 2009). More specifically, subjects classified as non-amnestic single-domain MCI usually do not show Αβ deposition, consistent with a non-AD underlying pathological process (Pike et al. 2007; Villemagne et al. 2011b).

To date, the relationship between Αβ deposition and the development of clinical symptoms is not fully understood. Although Aβ burden as assessed by PET does not correlate with measures of memory impairment in established clinical AD, it does correlate with memory impairment in MCI and healthy older subjects (Pike et al. 2007; Villemagne et al. 2008b). In subjects with MCI, this Aβ-related impairment is thought to be mediated through hippocampal atrophy (Mormino et al. 2009) while modulated by cognitive reserve (Roe et al. 2008; Rentz et al. 2010) and glucose metabolism (Cohen et al. 2009), although at the AD stage there is no longer correlation between glucose metabolism, cognition and Αβ burden (Furst et al. 2012).

Longitudinal studies are helping clarify the relationship between Aβ burden and cognitive decline in healthy individuals and MCI subjects. High ¹¹C-PiB cortical retention in non-demented elderly subjects is associated with a greater risk of cognitive decline (Villemagne et al. 2008b, 2011b; Resnick et al. 2010). For example, the evaluation of either the cognitive trajectories (Villemagne et al. 2008b; Resnick et al. 2010) or changes in cognition prospectively (Morris et al. 2009; Villemagne et al. 2011b; Doraiswamy et al. 2012) shows that individuals with substantial Αβ deposition are more likely to present with cognitive decline. Aggregating the results from recently reported longitudinal studies reveals that 65 % of MCI subjects with marked ¹¹C-PiB retention progressed to AD over 2–3 years, showing that MCI subjects with high Aβ deposition in the brain are >10 times more likely to progress to AD than those with low Aβ deposition, highlighting the clinical relevance and potential impact in patient management of Aβ imaging (Forsberg et al. 2008; Okello et al. 2009b; Wolk et al. 2009; Villemagne et al. 2011b). The observations that ¹¹C-PiB retention in non-demented individuals relates to episodic memory impairment, one of the earliest clinical symptoms of AD, and that extensive Aβ deposition is associated with a significant higher risk of cognitive decline in HC and with progression from MCI to AD emphasise the non-benign nature of Aβ deposition and support the hypothesis that Aβ deposition occurs well before the onset of symptoms, further suggesting that early disease-specific therapeutic intervention at the presymptomatic stage might be the most promising approach to either delay onset or halt disease progression (Sperling et al. 2011b).

The high prevalence of Αβ deposition in non-demented individuals raises questions in regard to Αβ toxicity and deposition. The evidence points to the fact that Αβ toxicity in the form of oligomers and Αβ deposition in the form of aggregates precede by many years the appearance of clinical symptoms (Price and Morris 1999). These neurotoxic processes eventually lead to synaptic and neuronal dysfunction, manifested as cognitive impairment, grey matter atrophy and glucose hypometabolism (Eckert et al. 2003; Suo et al. 2004; Leuner et al. 2007). Therefore, it is highly likely that ¹¹C-PiB retention in non-demented individuals reflects preclinical AD (Mintun et al. 2006; Rowe et al. 2007; Aizenstein et al. 2008; Villemagne et al. 2008b). This may be associated or attributed to a different susceptibility/vulnerability to Αβ, at either a cellular (frontal neurons seems to be more resistant to Αβ than hippocampal neurons) (Roder et al. 2003; Resende et al. 2007), a regional (compensatory upregulation of choline acetyltransferase (ChAT) activity in the frontal cortex and hippocampi of MCI and early AD subjects) (DeKosky et al. 2002) as well as at an individual or personal level, either due to a particular cognitive reserve (Mortimer 1997; Stern 2002; Kemppainen et al. 2008; Roe et al. 2008), differences in Aβ conformation affecting toxicity and/or aggregation (Lockhart et al. 2005; Deshpande et al. 2006; Levine and Walker 2010a; Walker et al. 2008), or because an idiosyncratic threshold must be exceeded for synaptic failure and neuronal death to ensue (Suo et al. 2004). These factors would help explain why some older individuals with a significant Αβ burden are cognitively unimpaired, while others with lower Αβ burden and no genetic predisposing factors have already developed the full clinical AD phenotype. Ongoing longitudinal studies will help identify those subjects that despite substantial Aβ deposition do no develop the AD phenotype, thus allowing to isolate genetic or environmental factors that may convey resistance to Aβ (Sojkova and Resnick 2011).

The lack of a strong association between Aβ deposition and measures of cognition, synaptic activity and neurodegeneration in AD, in addition to the evidence of Αβ deposition in a high percentage of MCI and asymptomatic HC, suggests that Αβ is an early and necessary, though not sufficient, cause for cognitive decline in AD (Villemagne et al. 2008b; Rabinovici et al. 2010; Sojkova and Resnick 2011), indicating the involvement of other downstream mechanisms, likely triggered by Αβ such as NFT formation, synaptic failure and eventually neuronal loss. Also at play are factors as age, years of education, occupational level and brain volume among others that, under the umbrella of ‘cognitive reserve’, appear to have a modulatory role between cognition and Αβ deposition (Roe et al. 2008; Chetelat et al. 2010b; Tucker and Stern 2011). Perhaps specific in vivo tau imaging will help elucidate or bridge the gap between Aβ deposition and the persistent decline in cognitive function observed in AD (Fodero-Tavoletti et al. 2011; Villemagne et al. 2012a).

Another rapidly growing area is the exploration of the potential association between biomarkers of Aβ deposition or neurodegeneration (Wahlund and Blennow 2003; Morris et al. 2005; Sunderland et al. 2005; de Leon et al. 2006; Thal et al. 2006; Blennow et al. 2007; Shaw et al. 2007; Clark et al. 2008; Jack et al. 2010; Albert et al. 2011; McKhann et al. 2011; Storandt et al. 2012) and Aβ burden as measured by PET (Jack et al. 2008a, 2010).

Age is the strongest risk factor in sporadic AD, with the prevalence of the disease increasing exponentially with age. These risks are increased in the presence of the ApoE ε4 allele (Farrer et al. 1997). Both risk factors have been directly associated with Aβ burden as measured by PET (Reiman et al. 2009b; Morris et al. 2010; Rowe et al. 2010).

To date, ApoE ε4 is the most consistent genetic risk factor associated with sporadic AD, and its presence has been associated with an earlier age of onset and a dose-dependent—either one or two ApoE ε4 alleles—higher risk of developing AD (Farrer et al. 1997). Examination of ApoE ε4 allele status revealed that, independent of clinical classification, ε4 carriers present with significantly higher ¹¹C-PiB retention than non-ε4 carriers, further emphasising the crucial role that ApoE plays in the metabolism of Αβ (Reiman et al. 2009b; Morris et al. 2010; Rowe et al. 2010). When specific clinical groups are examined, the results—from both autopsy and Αβ imaging studies—are not that clear cut. While some studies found significant differences in Αβ burden between AD ε4 carriers and non-ε4 carriers (Gomez-Isla et al. 1996; Drzezga et al. 2008; Grimmer et al. 2010), others did not (Berg et al. 1998; Klunk et al. 2004; Rowe et al. 2007; Rabinovici et al. 2010).

Both symptomatic and asymptomatic individual carriers of mutations within the APP or presenilin 1 or 2 genes associated with familial AD present with very high and early ¹¹C-PiB retention in the caudate nuclei, retention that seems to be independent of mutation type or disease severity (Klunk et al. 2007; Koivunen et al. 2008b; Remes et al. 2008; Villemagne et al. 2009a). Larger multicenter studies are helping elucidate the value of several biomarkers in these autosomal mutation carriers, assessment that might lead to better targeted disease-specific therapeutic strategies (Bateman et al. 2012).

While a similar pattern of Aβ deposition as seen in AD has been reported in older adults with Down syndrome (Landt et al. 2011), a completely different pattern of ¹¹C-PiB retention was observed in a mutation carrier of familial British dementia (Villemagne et al. 2010).

There are several other factors that have been associated with AD. Among these, the most relevant are vascular risk factors (VRF), such as diabetes, hypercholesterolemia, hypertension and hyperhomocysteinemia. The influence of these VRF on AD pathology is an area that, despite the long and intense scientific debate, has not been clearly established (Viswanathan et al. 2009). While some studies support a direct role of VRF in the development of AD (Qiu et al. 2010), others suggest that they may accelerate its clinical presentation (Riekse et al. 2004). Recent studies have used MRI white matter hyperintense lesions as surrogate markers of vascular disease in association with assessment of Aβ deposition and cognition (Kuczynski et al. 2008; Marchant et al. 2012).

As with AD, VRF are strongly age-dependent factors that share similar pathogenic pathways, such as oxidative stress or inflammation, so it is not easy to extricate one from the other in order to ascertain a causal, casual or co-morbid association between them (Rosendorff et al. 2007). On the other hand, in light of the absence of disease-modifying therapies for AD, the relevance of these risk factors rests in the fact that early intervention over them, especially at middle age, might have an impact on either the onset or progression of AD.

Aβ imaging in a wide spectrum of neurodegenerative conditions has yielded diverse results. Although lower than in AD, similar patterns of ¹¹C-PiB retention are usually observed in DLB (Rowe et al. 2007; Gomperts et al. 2008; Maetzler et al. 2009). Cortical ¹¹C-PiB retention is also higher in subjects diagnosed with CAA (Johnson et al. 2007), while there is usually no cortical ¹¹C-PiB retention in patients with FTLD (Rabinovici et al. 2007; Rowe et al. 2007; Drzezga et al. 2008; Engler et al. 2008). Aβ imaging has also facilitated differential diagnosis in cases of patients with atypical presentations of dementia (Ng et al. 2007b; Wolk et al. 2012a).