Introduction

Modern molecular imaging has provided new exciting tools to investigate the brain and understand functional disturbances as well the time course of different pathological changes. Several neurodegenerative brain disorders are characterized by proteinopathies. For Alzheimer’s disease (AD) which is the most common neurodegenerative disorder the current predominant hypothesis for the cause of AD is related to dysfunction in brain of processing, deposition and clearance of amyloid-β proteins. The introduction of amyloid positron emission tomography (PET) imaging with the radiotracer 11C-Pittsburgh compound B (11C-PIB) 10 years [1] ago have provided new and valuable insight into the dynamic processes and time course of deposition of fibrillar amyloid in brain from preclinical to clinical stages of AD [2–4]. Several 18F amyloid PET tracers suitable for clinical application have now been tested and three 18F-labelled amyloid PET tracers 18F-florbetapir [5], 18F-florbetaben [6], 18F-flutemetamol [7] have during 2012–2014 been approved by US Food and Drug Administration (FDA) and European medical agency (EMA) for use in clinical assessment of patients with cognitive problems to exclude AD. The 18F amyloid PET tracers with longer half life (110 min) in comparison to 11C-PIB (20 min) are more suitable for use at clinical centres to which the 18F-compound have to be transported due to lack of own cyclotron.

PET studies have shown that the deposition of amyloid in brain precedes the decline in regional cerebral glucose metabolism (18F-FDG PET) which in time course is followed by impairment of neurotransmitter function, atrophy and cognitive decline [2, 8]. Patients with mild cognitive (MCI) with a positive amyloid PET scan have a great risk to convert to AD in comparison to MCI patients with a negative amyloid PET scan [9, 10]. A negative amyloid PET scan in demented patients can exclude AD.

The diagnostic imaging biomarkers field is present under rapid development and its robustness and predictive value in clinical praxis are under investigation. While PET imaging has provided deep insight into the understanding of amyloid pathology in and relation to other pathological disease processes there is still lacking methods for in vivo measures of smaller amyloid peptides, oligomers. In addition on-going research studies on development of new PET tracers for imaging of tau deposition in different stages of AD [11] as well as PET tracers for study of inflammatory processes (astrocytosis, microglia activation) [11–13] in brain will further increase the understanding the time course AD disease pathology and clinical progression (Fig. 1).

Fig. 1
figure 1

Tentative time course of pathological processes at different stages of Alzheimer’s disease [33]

Full size image

This manuscript will provide a critical review of the usefulness of PET amyloid imaging in a clinical setting.

New diagnostic criteria for Alzheimer’s disease

Two major sets of new clinical diagnostic criteria for AD have been recently proposed by the International Working Group (IWG) [14–16] and the National Institutet of Aging-Alzheimer Association (NIA-AA) [17–19] with the aim to better define clinical phenotypes and to intergrade biomarkers into the diagnostic procedure and thereby be able to diagnose a pre-dementia stage of AD [14, 15] or MCI due to AD [17, 18]. In the recent published IWG-2 criteria [16], amyloid PET and CSF biomarkers (Aβ42, tau, p-tau) are suggested to be considered as diagnostic AD pathological markers. Decline in cerebral glucose metabolism measured by 18F-FDG and atrophy measured by MRI are suggested by the IWG-2 to reflect time course of disease progression and neurodegeneration (Table 1) [16].

Table 1 IWG-2 criteria for typical AD (A plus B at any stage)
Full size table

Criteria for appropriate use of amyloid PET imaging in clinical setting

The Alzheimer Association and the Society of Nuclear Medicine and Molecular Imaging convened an Amyloid Imaging Taskforce (AIT) to evaluate the potential clinical use of amyloid PET imaging in clinical assessment of patients with cognitive impairment [20, 21]. According to their recommendations (Table 2) it is motivated to perform a clinical amyloid PET scan in patients with persistent or progressive unexplained persistent MCI and in AD patients with an atypical course or an etiologically mixed presentation and in patients with progressive dementia and atypically early age onset (usually defined as 65 years less in age). It is, however, not appropriate to perform amyloid PET scans in patients who fulfil the diagnostic criteria for AD (>65 years) or in people who are asymptomatic or show subjective memory complains (Table 3). Furthermore, the recommendations consider it is inappropriate to perform an amyloid PET scan to determine dementia severity, based upon family history of risk factors as Apolipoprotein E(APOE) genotype or as substitute for genetic testing of autosomal dominant AD [20]. It is also inappropriate to perform an amyloid PET scan for non-medical reasons including assurance or legal, employment decisions [20].

Table 2 When is a clinical amyloid PET scan appropriate?
Full size table
Table 3 Inappropriate candidates for amyloid PET imaging
Full size table

Diagnostic value of 11C-PIB PET imaging in memory clinic setting

18F-FGD PET is already an established method for use in diagnostic process of uncertain cases in some memory clinics. A few number of studies have compared the practical value of 18F-FDG and 11C-PIB PET in a clinical setting. When 154 patients underwent 11C-PIB and 18F-FDG at the Memory clinic of the VU University Medical Centre in Amsterdam, 66 % of the AD patients were found to show positive 11C-PIB scan compared to 28 % for patients with frontotemporal dementia (FLTD), 80 % with Lewy body dementia (LDB) and 30 % for other dementia. The 18F-FDG uptake pattern was comparable with the diagnosis of AD in 58 % of the cases compared to 33 % of the FTLD cases [22]. The use of PET imaging in diagnostic procedure changed the diagnosis in 23 % of the patients. It was estimated that the diagnostic confidence increased from 71 to 87 % after the PET examinations [22]. When 39 % of the patients were clinically followed up after 2 year then clinical diagnosis remained in 96 % of the patients [22]. This study illustrates the clinical observation of AD patients where 11C-PIB PET does not show presence of amyloid fibrillar plaque pathology. Usually as also pointed out by the authors the % of PIB negative AD patients is less than 39 % as observed in the present study [22]. A higher concordance of 84 % between clinical diagnosis and 11C-PIB PET was observed in a study of 140 patients from the University of California, San Francisco (UCSF) Memory and Aging Centre [23]. The primary diagnosis was only changed in 9 % of the cases after PET [23]. Autopsy studies were performed in twenty-four of the patients which showed 96 and 96 % agreement with earlier performed 11C-PIB and 18F-FDG investigations, respectively [23]. When fifty-seven patients underwent 11C-PIB scans at the Copenhagen Memory Clinic, Copenhagen University Hospital, Rigshospitalet [24] to confirm or to rule out AD, 47 % of the patients showed presence of amyloid plaque deposition in the brain which led to that 23 % of the patients were reclassified following the 11C-PIB PET investigation [24]. It is estimated from the present study that overall confidence of the diagnosis increased in 49 % of the patients after the PET scan [24]. When thirty-four patients at the Marqués de Valdecilla University hospital, Santander, Spain underwent 11C-PIB and 18F-FDG [25] a clear separation between non-amnestic and amnestic MCI patients was observed with 11C-PIB while 18F-FDG PET was found to be useful in FTLD patients [25]. It was concluded that both 11C-PIB and 18F-FDG can provide valuable information in diagnostic procedure of cognitive impairment [25].

Diagnostic value of 18F-amyloid PET imaging in memory clinic setting

18F-florbetapir was the first 18F-amyloid PET imaging tracer that was approved by FDA 2012 and EMA 2013 and has been followed by the approval of 18F-florbetaben by EMA 2013, FDA 2014 and flutemetamol by FDA 2013, EMA 2014, respectively. Since all three 18F-amyloid PET tracers were approved quite recently there is still quite few scientific reports published regarding their clinical value in diagnostic routine of assessment of cognitive impairment. A limiting factor for introducing these PET tracers for clinical use is for some countries the issue of reimbursement of the costs of the PET scans.

In a large study of 229 subjects from 39 centres in US the impact of 18F-florbetapir PET scanning on the diagnosis and management was evaluated in patients with progressive impairment [26]. The patients were enrolled in the study with a history of cognitive decline and uncertain diagnosis and thirty-six percent had dementia and 64 percent cognitive impairment. Forty-nine percent of the patients showed a positive 18F-florbetapir scan. The diagnosis was changed in 54.6 % of the cases with and increase in the diagnosis of Ad and it was estimated that the diagnostic confidence increased by 21.6 % and the prescription of cholinesterase inhibitors and memantine increased by 17.7 % in cases with positive amyloid PET scan [26].

In a smaller series of 30 patients, 18F-florbetapir PET investigations were performed at a cognitive evaluation centre of urban dementia centre at Mount Sinai, New York [27]. The clinical assessment of the memory impairment also included neuropsychological assessment as well as neurological examination [27]. The 18F-florbetapir imaging caused change in diagnosis of 10 of the 30 patients (33 %) and clarified the diagnosis in 9 patients (30 %) [27]. It was concluded that the 18F-florbetapir scan in patients with positive scan added important diagnostic certainty and also possibility to help patients and family in the planning of future care and treatment [27]. Similar observations were also observed in evaluation of a small population of 11 patients at Duke University Medical Centre, Durham, US [28]. The clinician surveys were performed before and after 18F-flutemetamol PET scan in 11 patients with memory decline [28] in which six patients were amyloid PET positive and the diagnosis was changed in four of them and the treatment in three patients. In the five patients with negative amyloid PET scan the diagnosis was changed in four patients and the treatment plan in two patients [28]. It was concluded that the both a positive and negative 18F-florbetapir PET may have an impact on clinical decisions especially in cases where the cognitive decline has an uncertain aetiology [28].

When the use of AD biomarkers were investigated in large number of European Alzheimer Consortium Centres, seventy-nine percent of the responders expressed that they felt “very/extremely” comfortable delivering a diagnosis of MCI due to AD when both the biomarkers of amyloid (CSF Aβ 42, amyloid PET) and neuronal injury (18F-FDG) were abnormal and they also agreed that a combination of these biomarkers was a strongly indicative AD signature [29].

In a recent study in cognitive healthy subjects from 59 to 89 years age Jack et al. [30] studied the presence of brain amyloidosis (11C-PIB) and neurodegeneration (18F-FDG/MRI). Eighty-nine percent of the cognitive normal subjects at 89 years of age showed both positive 11C-PIB PET scan and reduced 18F-FDG PET uptake indicating that normal cognitive is possible despite pathological changes in brain at high age [30]. It has recently been suggested that a primary age-related tauopathy (PART) type of pathology might characterize a subset of SNAP (suspected non-Alzheimer pathophysiology) group [31]. Neurodegeneration as measured by FDG PET and CSF tau biomarkers appears to better reflect cognitive decline than amyloid PET imaging [32].

Conclusions

Molecular PET imaging techniques allow measurement of fibrillar amyloid plaque deposition in brain of patients with progressive cognitive impairment. The 18F amyloid PET tracers are presently introduced into the clinical setting and still more experience has to be obtained before significant conclusions can be made of their role in the management of patients with memory impairments. Other biomarkers including CSF biomarkers measuring Aβ42, tau, p-tau are also presently introduced at many memory clinics and other clinics already using CSF biomarkers as a part of the normal assessment. This is also the case for structural imaging with MRI as well as 18F-FDG. It is, therefore, important that international/national guidelines are developed as well as each memory clinic develops recommendations in which order these different biomarkers should be used. However, as illustrated in the present study, the clinical experience from different memory centres, the addition of amyloid PET scan to routine memory assessment battery provide valuable information which also led to change of diagnosis (in most cases from MCI to AD) in a varying proportion of the patients as well as treatment strategies. In addition, positive amyloid PET scan made the physician more confident in the diagnosis of AD which also increases the possibility for future plans with patient and their family. As pointed out by several authors a negative PET amyloid scan which rule out the fact that the patient express AD pathology is valuable for the further clinical management of the patient, especially patients where the diagnosis has been unsure. Visual reading is recommended as routine evaluation of the 18F amyloid PET tracers but quantitative techniques will probably be of great value in cases where the amyloid PET tracer uptakes are close to cutoff values and, therefore, might be difficult to judge with sole visual inspection whether the PET scan is negative or positive. Up to now most clinical studies have been performed using 18F-florbetapir in addition to e 11C-PIB but more clinical experience is needed from the use of 18F-florbetaben and 18F-flutemetamol. Important issues to discuss further involve the ethical issues concerning the disclosure of amyloid PET scans in assessment at early stages of memory assessment.