|Home | About | Journals | Submit | Contact Us | Français|
Positron Emission Tomography (PET) of brain amyloid-beta is a technology that is becoming more available, but its clinical utility in medical practice requires careful definition. In order to provide guidance to dementia care practitioners, patients and caregivers, the Alzheimer Association and the Society of Nuclear Medicine and Molecular Imaging convened the Amyloid Imaging Taskforce (AIT). The AIT considered a broad range of specific clinical scenarios in which amyloid PET could potentially be appropriately used. Peer-reviewed, published literature was searched to ascertain available evidence relevant to these scenarios, and the AIT developed a consensus of expert opinion. While empirical evidence of impact on clinical outcomes is not yet available, a set of specific Appropriate Use Criteria (AUC) were agreed upon that define the types of patients and clinical circumstances in which amyloid PET could be used. Both appropriate and inappropriate uses were considered and formulated, and are reported and discussed here. Because both dementia care and amyloid PET technology are in active development, these AUC will require periodic reassessment. Future research directions are also outlined, including diagnostic utility and patient-centered outcomes.
Research progress in Alzheimer’s disease (AD) and molecular imaging over the past decade has made it possible to detect human brain amyloid-beta (Aβ) deposition during life using positron emission tomography (PET). Parallel progress has improved our understanding of Aβ as an important and therapeutically targetable component of AD pathology. While Aβ plaques are one of the defining pathologic features of AD, many otherwise normal elderly people have elevated levels of Aβ, as do patients with clinical syndromes other than AD dementia. The potential clinical utility of Aβ PET therefore requires careful consideration so that its role may be identified and placed in the proper clinical context. The Society of Nuclear Medicine and Molecular Imaging (SNMMI) and the Alzheimer’s Association (AA) have jointly developed this document to assist in the appropriate use of this class of PET radiopharmaceuticals. The primary goal of this document is to provide health care practitioners with the information necessary to provide their patients with optimal care, while also considering the cost-effective utilization of limited health care resources.
With the advent of carbon-11 (C-11) labeled Pittsburgh Compound B (PiB), Aβ or “amyloid” PET emerged as a major element in a transformation of AD research that emphasized the development of biomarkers, which could potentially facilitate drug development . Intense efforts were directed at assessing the amyloid status of individuals with AD dementia as well as those with prodromal and preclinical stages of disease, and the technology was rapidly adopted worldwide, albeit largely in specialized research centers. More recently, amyloid PET has been increasingly used in clinical trials for AD therapeutics. Because the short, 20 min half-life of C-11 limits routine clinical use of PiB due to the need for an on-site cyclotron, amyloid-binding radiopharmaceuticals labeled with longer-lived fluorine-18, with a 110 min half-life, were developed and commercialized for wide availability. One such compound, [F-18]Florbetapir, achieved approval by the United States Food and Drug Administration in April, 2012. The European Medicines Agency’s Committee for Medicinal Products for Human Use recommended marketing authorization for [F-18]Florbetapir in October 2012.
To develop this document, an Amyloid Imaging Taskforce (AIT), consisting of experts in the fields of imaging, neurology, and dementing diseases was assembled by the AA and SNMMI to review the available literature and develop consensus recommendations for the clinical use of these promising new radiopharmaceuticals. At the time of this review, experience with clinical amyloid PET imaging is limited. Most published studies to date have been designed to validate this technology and understand disease mechanisms rather than to evaluate applications in clinical practice. As a result, published data are available primarily from highly selected populations with prototypical findings rather than from patients with co-morbidities, complex histories, and atypical features often seen in clinical practice. Despite these limited clinical use data, the members of the task force concluded that the proven sensitivity and specificity of the new radiopharmaceuticals for brain amyloid, and the known association between brain Aβ deposition and AD suggest these new radiopharmaceuticals could potentially be helpful in the workup and diagnosis of patients with cognitive impairment.
Translation of research findings to clinical populations poses substantial challenges. Unlike research subjects, clinical patients can exhibit a wide range of medical and psychiatric problems. Indeed, the prevalence of mixed-cause dementia increases with advancing age and is frequently seen in clinical practice . As the population ages, individuals are increasingly likely to inquire about the usefulness of amyloid PET imaging in a variety of circumstances that are unlikely to be addressed in the scientific literature. In addition, as amyloid imaging agents become more well-known and longitudinal data accumulate, patients and referring physicians may request amyloid PET imaging for individuals who are asymptomatic. Colleagues may also ask for advice about using amyloid PET imaging for purposes such as screening someone with a family history of AD dementia, or use in patients already carrying a diagnosis of a non-AD dementia.
When the peer-reviewed literature is incomplete, as is often the case, expert opinion can be valuable, particularly when considering numerous practical clinical issues and ethical concerns that largely remain overlooked in the design and discussion of published clinical trials. There is often diagnostic uncertainty resulting from the complexities of patient history as well as the inconsistencies in examination results. Incorporating amyloid imaging into clinical decision-making may help to narrow a differential diagnosis and simplify some of the complexities inherent in evaluating patients with mild cognitive impairment and dementia.
Although identifying potential benefits, the AIT concluded that an amyloid PET report will not constitute and is not equivalent to a clinical diagnosis of AD dementia. Imaging is only one tool among many that clinicians should use judiciously to manage patients. Amyloid PET imaging does not substitute for a careful history and examination. Indeed, the history and examination is required to understand the clinical context necessary to incorporate imaging results into clinical decision-making. The diagnosis of dementing diseases has implications that resonate beyond the patient to include family members, particularly those who are genetically related. We hope that these recommendations will be relevant to many patients, even when published evidence may be lacking.
As with most guidelines, the healthcare provider has to make the ultimate judgment regarding the care of each individual patient. The AIT sought to assist this process and identified the following general sequence of events with which amyloid PET could be used according to the AUC set forth below: 1) evaluation by a dementia expert to assess the need for diagnostic testing, possibly to include amyloid PET, if the AUC are met; 2) referral to a qualified provider of amyloid PET services; 3) performance, interpretation and reporting of the amyloid PET according to established standards; 4) incorporation of the PET result into the clinical assessment process by the dementia expert; 5) disclosure of the PET result by the dementia expert to the patient and caregivers, along with discussion of the result and its management consequences. The healthcare provider must bear in mind that amyloid imaging does not make a diagnosis of Alzheimer’s disease, and by itself does not determine that a patient’s cognitive impairment is due to AD pathology.
The AIT formulated appropriate use criteria (AUC) for amyloid PET imaging using procedures similar to those employed by groups such as the American College of Cardiology Foundation . The process employed: 1) Identification of potential indications / non-indications; 2) Evidence assessment and rating; 3) Group rating of potential indications / non-indications; 4) Discussion and re-voting; and 5) Writing. Three AIT subcommittees were established: 1) Indication Subcommittee; 2) Literature Subcommittee; and 3) Evidence Review Subcommittee.
The Indication Subcommittee, consisting of practicing dementia specialists and imaging experts, discussed 115 potential indications and non-indications based on multiple clinical and non-clinical scenarios with variables including 1) clinical setting; 2) clinical context; 3) evidence of cognitive deficit; 4) family history; 5) knowledge of AD genetic risk; and 6) age. This process is described in Appendix A. Based on the consensus discussion, the Indication Subcommittee consolidated potential indications and non-indications into 14 scenarios that were subsequently incorporated in a Data Extraction Form used for the evidence assessment (see below).
The Literature Subcommittee used a search strategy as established by the American Academy of Neurology (AAN) and the Institute of Medicine (IOM) of the National Academy of Sciences to identify relevant literature. The process is described in the Appendix. The AIT deliberated on the choice of literature screening criteria, and a decision was made on the basis of the types of evidence ultimately needed to establish clinical utility of amyloid PET. The ultimate goal was to determine whether there is evidence that using amyloid PET leads to clinically meaningful improvement in outcomes or is useful in medical or personal decision-making. Because direct evidence linking amyloid PET to health outcomes is lacking at present, the AIT evaluated existing literature according to a possible “chain of evidence” consisting of 3 key questions, adapted from the scheme of Fryback and Thornbury :
1. Technical efficacy (analytic validity or technical test performance): This class of evidence reflects the stability, adequacy and reproducibility of the test itself, and includes both the image data and the qualitative image interpretation. Proof of technical efficacy includes reproducibility of specific amyloid PET acquisition procedures and protocols under standardized conditions, which must be established separately for each amyloid tracer and be applicable to the range of PET instrumentation in common use. In addition, the implementation of amyloid PET requires a qualitative read of images, so that evidence of standardized interpretation protocols that lead to acceptable levels of inter-rater agreement must be considered. These standards and procedures are ultimately the province of the professional certifying organizations, such as the American Board of Nuclear Medicine (ABNM), but reports have already appeared from the Alzheimer Disease Neuroimaging Initiative  and Phase 2 industry-sponsored trials  that describe standardized acquisition protocols for F-18 labeled PET ligands as well as interpretation standards. Such standards and procedures have been implemented by the commercial developers of the F18 amyloid PET ligands, and evidence of validation is required by the FDA before approval for clinical use is granted. Additional analytic validity data should be acquired relating to test-retest and longitudinal stability and change. Web-based instruction programs for readers of F-18 amyloid PET have been developed and validated and should be successfully completed by all imaging specialists prior to reading of clinical amyloid PET. (See also Image Quality and Reporting, below).
2. Diagnostic accuracy (clinical validity), based on an autopsy truth standard: Considerable progress has been achieved to establish clinical validity using histopathology-to-image comparisons. As with other elements of validation, each tracer and its associated interpretation protocol must be assessed separately. The AIT elected to include longitudinal clinical studies as ancillary evidence of clinical or diagnostic validity when the design included a baseline amyloid PET followed by clinical evaluation and assessment of longitudinal decline or conversion in clinical status, according to accepted clinical diagnostic criteria [7, 8].
3. Clinical utility based on a change in management (including change in diagnostic evaluation) and associated improved clinical outcomes: This is the most challenging component of the analytic framework and the evidence for a change in clinical management based on amyloid PET is not yet available. With only medications to treat symptoms currently available and no disease-modifying treatment yet proven, the clinical utility of a diagnostic test to alter patient management and result in a quantifiable benefit is very difficult to establish. However, an accurate diagnosis of the cause of cognitive impairment is often critically important for a practitioner to select appropriate treatments and avoid inappropriate interventions. Furthermore, different dementing diseases have distinctive courses, complications and comorbidities that alter non-pharmacological management and treatment recommendations. While psychological, social, economic and family outcome variables, including value-of-knowing, can be identified as potentially altering management, the data supporting specific outcomes for amyloid PET are not yet available.
Multiple searches were performed using the National Institutes of Health’s (NIH’s) PubMed. The Literature Subcommittee reviewed the list for inclusion and identified a subset of papers by abstract analysis. Papers not relevant to the clinical use of amyloid PET were eliminated based on the primary focus of the reported study and data presented in the paper. In addition, to ensure appropriate papers were captured in initial search and review, the AIT performed a “backward review” to cross-check the literature included in seminal amyloid imaging reviews with those included in the AIT’s initial assessment. For the backward review, the AIT employed the bibliographies of Klunk et al. , Villemagne et al. , and Laforce et al. .
The Literature Subcommittee developed a Data Extraction Form for the evidence assessment. The Evidence Review Subcommittee conducted evidence assessment in two steps. In the Phase I review, valid papers identified in the initial review process were assigned to a pair of reviewers consisting of a dementia specialist and imaging specialist. Each reviewer scored the papers using the Data Extraction Form. Other papers that met the preliminary inclusion criteria, as indicated by both reviewers, but received low scores by both reviewers, or mixed scores, or strong reviewer comments for further assessment were also identified as papers to be discussed. In the Phase II review, additional papers were identified through the backward review and an updated search, as well as new papers in press. A second round of reviews was conducted identical to Phase I. The final inclusion criteria were that the paper must contain data of one of two types, either PET-histopathology correlation or PET correlation with longitudinal clinical follow-up. After the review, Co-Chairs of the Task Force reviewed the findings of both Phase I and Phase II and presented a final list of 23 papers that satisfied the final inclusion criteria and these were presented to the full AIT [5,12-33]. These papers were used as the literature-based evidence for rating the AUC outlined by the Indication Subcommittee.
The group rating of potential indications / non-indications was conducted using a rating sheet by individual voting AIT members without knowledge of other members’ rating results. Fourteen scenarios proposed by the Indication Subcommittee were consolidated to 10 possible indications / non-indications by defining a preamble that applies to all indications / non-indications. The rating sheet included 1) final 10 possible indications / non-indications; 2) amount of qualified evidence determined by the evidence assessment; and 3) individual papers that relate to each indication / non-indication. Based on the presented evidence and individual AIT members’ opinions, the AIT members were asked to rate each indication / non-indication with ’Appropriate’, ‘Uncertain’, or ‘Inappropriate’. A non-voting AIT member summarized the rating results.
The following terms are used in the Appropriate Use Criteria:
Dementia Expert: a physician experienced in the assessment and diagnosis of dementia. The AUC depend heavily upon the training, experience and clinical judgment of the dementia experts ordering the test and their application of the published, standardized clinical criteria for mild cognitive impairment (MCI) and Alzheimer’s disease (AD) (see below). Expertise in applying these criteria is typically acquired through formal training and clinical experience in neurology, psychiatry, and geriatric medicine; however, not all physicians in these disciplines are dementia experts.
Alzheimer’s disease (AD): the pathologic process reflected in specific postmortem histopathologic criteria , which is frequently but not necessarily associated with a characteristic dementia syndrome [7, 8].
The AD pathologic process differs conceptually and is uncoupled from the dementia syndrome with which it is associated, as evidenced by the very long preclinical or asymptomatic period preceding the dementia syndrome. The AUC specifically refer to clinical criteria for AD dementia that have recently emerged from the NIA-AA workgroup  and the International Workgroup (IWG; ). Core clinical criteria for AD dementia identify the specific conditions and circumstances under which a dementia expert may determine whether amyloid PET can be used appropriately. While the two international workgroups used different terms, ‘probable AD’ and ‘typical AD’, the underlying principles are quite similar, and either nomenclature may be applied to support the conclusion that amyloid PET would or would not be appropriate (Criterion #4).
Possible AD dementia: a clinical syndrome meeting the core clinical criteria for AD dementia in terms of the nature of the cognitive deficits for AD dementia, but either 1) has a sudden onset of impairment or demonstrates insufficient historical detail or objective documentation of progression; or 2) has an etiologically mixed presentation because of evidence of vascular or Lewy pathology .
Mild Cognitive Impairment (MCI): a clinical syndrome meeting the published core clinical criteria for MCI [36-38](also see [8, 35, 39]). While considerable debate and evolution of the criteria for MCI continue, general agreement exists about the core features. Briefly, these include 1) concern about a change in cognition, 2) impairment in one or more cognitive domains, 3) preservation of independence in functional activities, but 4) not demented . The application of these criteria and their use in determining whether amyloid PET would be appropriate is in the hands of the dementia expert.
Amyloid Positivity/Negativity: the determination by an imaging expert that the amyloid PET scan indicates the presence or absence of Aβ plaque. The imaging expert is a nuclear medicine specialist or radiologist with specific training in the interpretation of amyloid PET. The amyloid PET data must be technically adequate and be acquired at a fully qualified and certified facility (See Image Quality and Reporting, below). The protocol for the qualitative read that determines positivity or negativity must be standardized (e.g., ) and conform to a specific guideline provided by the manufacture if it is available.
While a number of Aβ PET radiopharmaceuticals have been reported with human data [40-43], at present there are five that are in use at multiple sites to image Alzheimer pathology in vivo. Among these, [C-11]-(2-[4-methyl-amino phenyl]-1,3-benzothiazol-6-ol, or Pittsburgh Compound B (PiB) was the first to be described and is the most extensively studied . PiB, a neutral analog of the histological dye Thioflavin T, has been evaluated with respect to clinical syndromic and postmortem histopathologic correlation over approximately 10 years in several clinical populations and in healthy controls (HC). Histopathologic correlation data demonstrate the association between PiB PET and postmortem assessment of Aβ pathology [12-14, 23, 30, 44-47].
The short 20 minute half-life of C-11 limits routine clinical use due to the need for an on-site cyclotron, whereas the 110-minute half-life of F-18 labeled PET ligands allows incorporation of PET into routine clinical practice as has occurred with [F-18]fluorodeoxyglucose (FDG) in clinical oncology. Several F-18-labeled Aβ PET radiopharmaceuticals have been developed, including [F-18]3′-F-PiB (Flutemetamol) , [F-18]AV-45 (Florbetapir) , [F-18]-AV-1 or [F-18]-BAY94-9172 (Florbetaben) [49, 50], and [F-18]-AZD4694 or NAV4694 [51-53]. Postmortem histopathology to PET correlations have been published for Florbetapir  and biopsy histopathology to PET correlations have been published for Flutemetamol .
F-18 ligand PET data have been quantitatively compared to C-11 PiB data acquired in the same subjects with respect to cortical binding, linear regression slope, and diagnostic classification performance (Table 1). Wolk and colleagues performed PiB and Florbetapir PET imaging in 14 cognitively normal adults and 12 AD patients and showed that both ligands displayed highly significant group discrimination and correlation of regional uptake . Landau et al. compared PiB to Florbetapir, found the data were correlated at r = 0.95, slope = 0.60 and that resulting cut points yielded classification agreement in 97% of cases evaluated . A correlation with PiB of r = 0.905 and a slope of 0.99 was reported for Flutemetamol in 20 AD patients and 20 MCI patients, in which the concordance of visual reads between ligands was 100% . Villemagne et al. compared PiB with Florbetaben in 10 healthy control subjects and 10 patients with AD and reported the correlation to be r=0.97, the slope 0.71, and the concordance between ligands to be 100% . Rowe et al. compared NAV4694 to PiB in 7AD, 3 FTD, 10 MCI, and 25 HC and found the correlation to be r= 0.99, the slope 0.95, and classification concordance to be 100% . These findings are consistent with a high correlation of these [F18]-labeled ligands with PiB and they support the translation of PiB PET findings into the domain of these [F-18]-labeled radiopharmaceuticals. Comparison studies of one F-18 agent with another have not yet been reported.
Ratings for each indication / non-indication were obtained from independent voting by eight AIT voting members, and the results were summarized by a nonvoting member. At the time of voting, each member was able to access 1) qualified peer-reviewed papers that potentially concern each possible indication; 2) ratings by AIT members of the quality of the evidence, based on the results of the literature review as described previously. For each indication, the number of supporting publications and the average quality of evidence were indicated on the voting sheet.
In the initial voting, 4 possible indications and 7 possible non-indications were submitted for voting. Substantial disparities in voting results (30-50%) were found for 4 potential indications. Each potential indication / non-indication was reviewed in subsequent AIT discussions, and uncertain language in the proposed AUC were clarified for re-voting. In this process, 2 possible indications were combined into one indication, resulting in total 3 possible indications and 7 possible non-indications.
In the re-voting results, the ratings of ‘appropriate’ or ‘inappropriate’ were unanimous for possible indications #1 and #2 and possible non-indications #4 to #10. For indication #3, 2 voting members voted ‘uncertain’, while the other 6 voting members voted ‘appropriate’.
Amyloid imaging is appropriate in the situations listed below for individuals with all of the following characteristics:
a) a cognitive complaint with objectively confirmed impairment; b) Alzheimer’s disease as a possible diagnosis, but when the diagnosis is uncertain after a comprehensive evaluation by a dementia expert; and c) when knowledge of the presence or absence of amyloid-beta pathology is expected to increase diagnostic certainty and alter management.
The AIT considered whether to specify the patient characteristics for each indication separately, but recognized that there were several elements common to all appropriate indications and set these elements apart in a separate Preamble. The Preamble was intended to characterize all patients for whom the appropriate indications #1 to #3 apply.
The Preamble substantially restricts the set of patients for whom amyloid imaging would be appropriate in several ways. First, the dementia expert, as defined above (see Definitions), must evaluate the patient and determine that there is objective evidence of impairment. The objective evidence may be acquired and interpreted directly by the dementia expert in a detailed mental status examination or obtained from a separate neuropsychological assessment. Second, the expert should evaluate all available clinical evidence, including the history, physical and neurological examinations, and all available laboratory and neuroimaging data in order to consider the possible causes of the illness as well as potentially confounding circumstances such as depression, medication effects, cerebrovascular, endocrine or other medical disorders. This is because the presence of amyloid pathology in the brain, when considered in isolation, is insufficient to determine the cause of the impairment; rather, the presence of amyloid pathology is one factor among many that must be considered. The dementia expert must conclude on the basis of all available evidence that 1) the cause of the impairment is uncertain and 2) that it could be explained on the basis of Aβ pathology, i.e., AD dementia or its prodromal stage must be in the differential diagnosis.
Finally, the expert must conclude that a determination of either amyloid positivity or amyloid negativity would both increase the level of diagnostic certainty and alter the plan for patient management. Empirical evidence for the value of added certainty due to amyloid PET has not yet been reported, however, several patient centered outcome studies are underway and the following should serve to further guide efforts of this type. The AIT considered several situations in which the added certainty of amyloid PET could be useful to patients and caregivers and result in altered management. First, many patients with uncertain diagnoses undergo extensive and repeated testing that would be reduced if the diagnostic certainty were increased by amyloid PET. For others, however, it is also likely that amyloid negativity would require additional diagnostic testing as the dementia expert seeks to identify the underlying Aβ-negative cause of impairment. The relative utility of diagnostic tests should be further evaluated. Second, increased certainty of the diagnosis could provide a basis for earlier and more consistent drug treatment, avoidance of treatments unlikely to afford benefit, and improved monitoring for likely complications and adverse drug effects that are relevant to specific dementing diseases. In addition, improved diagnostic certainty could provide more powerful motivation to make required lifestyle changes and difficult living transitions for which they are otherwise reluctant. Third, a more certain diagnosis can have profound social benefits to patients and families, who may need to identify the required resources and plan for future management. Minimizing diagnostic uncertainty can assist in bringing family members to a uniform understanding of the patient’s condition and needs, facilitating the development of a unified plan of progressive support that best manages financial resources and maximizes quality of life.
Although learning the cause of dementia and the limited efficacy of available treatments may for some cause stress and anxiety, we believe that the value of knowing outweighs the disadvantages. Electing to manage dementing diseases without investigating the cause or with high levels of diagnostic uncertainty often contributes to inconsistent and poor quality of care. In any circumstance patients and their families decide upon their own whether to seek answers by electing or failing to seek care.
This indication refers to a patient who satisfies all the criteria set forth in the preamble and is being evaluated for persistent or progressive cognitive impairment that is still mild, e.g., a patient with MCI as defined above. This means in practice that, while impaired according to objective measures, the patient does not have “significant interference in the ability to function at work or in usual daily activities”  (also see ). In this circumstance, an amyloid PET finding of positivity would, on the basis of it’s known correspondence to brain Aβ, raise the level of certainty that the patient’s mild impairment is on the basis of AD pathology and represents early AD dementia (see Definitions). However, it is important to emphasize again that not all patients with MCI would be appropriate for amyloid PET. Rather, amyloid PET would be appropriate only in those individuals who the dementia expert has concluded would benefit from greater certainty of the underlying pathology and whose clinical management would change as a result of this greater certainty.
The dementia expert should recognize that asymptomatic amyloid deposition is common in older (e.g., >75 years) individuals and may not be related to a patient’s presenting symptoms. Furthermore, the dementia expert will need to consider in older individuals the possibility that amyloid positivity could be present but not be the sole factor in causing the impairment and that comorbid conditions or pathologies such as vasculopathy could be present and account for or significantly contribute to the observed impairment.
The prognostic value of amyloid PET for predicting future outcomes in MCI patients is under active investigation and preliminary studies are suggestive but not complete. Initial reports suggest that the majority of patients with amnestic MCI, variously defined by neuropsychological evaluation, and a positive amyloid PET will progress to AD dementia while the risk of progression to AD dementia is significantly lower in those who are amyloid negative. The available evidence to date has not definitively linked amyloid positivity in individual patients with a future time point when cognitive or functional deterioration can be predicted. Therefore at present the use of amyloid PET to predict the trajectory of a patient’s cognitive decline or the time to any specific outcome is not appropriate because published evidence is limited. (See below, Further Research Questions)
A related, alternative scenario for this Indication is a patient, also satisfying all the criteria set forth in the Preamble, who is amyloid negative, and therefore much less likely to be impaired on the basis of AD. The ‘amyloid negative’ scenario may in practice be the most frequently useful scenario in MCI, given the potential confound of age-associated Aβ, discussed above, among amyloid positive individuals. Thus, in patients with MCI whose clinical picture may be complicated with potential vascular, traumatic, or medical causes of cognitive impairment, amyloid PET may find utility and could be appropriately used to effectively exclude AD pathology as a basis for the clinical syndrome.
This Indication refers to a patient with an established dementia syndrome who is not typical as to presentation and clinical course or to a patient who is considered to have a mixture of causal pathologic processes. It is intended to explicitly exclude from the category of appropriate use the patient about whom there is little doubt as to the underlying pathology because the onset, course, and examination findings are typical of AD dementia. It is however intended to include those patients for whom substantial uncertainty exists and for whom greater confidence would result from determining whether Aβ pathology is present or not present, as described below.
The AIT chose here to rely on the established concept of possible AD, specifically as it has been recently revised , and again to focus on the dementia specialist as the physician who would apply the criteria based on this diagnostic category. The restriction in this indication to patients with possible AD dementia is based on the well-established existence of patients about whom there is substantial doubt as to whether the dementia is based on AD pathology. The sources of doubt are 1) the presence of an unusual course, e.g., sudden onset or episodic, or because the course cannot be established from the history or from retrospective cognitive test data; or 2) the presence of a comorbid condition that confounds the interpretation of the clinical data, such as cerebrovascular disease, other neurological disease, other medical condition or medication use that is affecting cognition and function. Amyloid PET is not useful in identifying the possible confound of co-existing Lewy pathology (see below).
Amyloid PET is appropriate in the scenario where a relatively young patient, e.g., 50-65 years old, but possibly even younger, presents with a progressive impairment that has features of AD dementia as well as of a non-AD dementia. In the scenario covered by this Indication, the dementia specialist is often called upon to identify the cause of a devastating illness in such a patient, and to manage a complex and comprehensive evaluation. The purpose of the evaluation is to rationally manage the symptomatic treatment, make appropriate employment, driving and lifestyle decisions, possibly refer to clinical trials of candidate disease modifying therapies, and to provide a basis for prognosis and planning for care. The presence or absence of AD pathology in this circumstance is frequently a critical component of the initial differential diagnosis, and it is well known from postmortem studies that clinical diagnosis based upon history and examination is often wrong as to the presence of AD pathology .
The AIT identified seven circumstances or scenarios in which amyloid imaging would be inappropriate.
The AIT recommended against the use of amyloid PET in cases in which core clinical criteria for probable AD dementia were satisfied , and there were typical history and examination findings since the level of uncertainty would be low and the potential benefit from added information and the potential for altered management would be correspondingly low.
Data are lacking to support the use of amyloid imaging to determine the severity of any cognitive disorder. Thus far, the predominance of the evidence is that the level of Aβ burden measured with amyloid PET does not correlate well with severity of deficits in patients with dementia .
There is no data available at present indicating that solely based on family history or APOE genotype that prognosis, course or greater certainty in the cause of cognitive deficits is aided with amyloid PET imaging.
The significance of a cognitive complaint in an elderly person without deficits on examination is currently a topic of active investigation; however, there is insufficient evidence to suggest amyloid PET can aid prognostic judgments or relieve the concerns of such individuals. A negative amyloid PET scan today cannot exclude the possibility of AD dementia in the future.
The use of amyloid PET in lieu of genotyping for suspected autosomal dominant mutation carriers is considered inappropriate. The optimal clinical evaluation in these cases is careful collection of a family history, followed if appropriate by genetic counseling prior to and after genetic testing for known mutations. Future use of amyloid PET in autosomal dominant mutation carriers could include determination of whether or not the amyloid-deposition phase of their illness has yet begun. In the setting of a complete clinical evaluation, including serial neuropsychological testing, this information may be useful in identifying one disease-related milestone that, along with the genetic information, aids decision-making.
The prognostic value of amyloid positivity in normal elderly individuals remains investigational (see below Further Research Questions). There is a significant potential for patients and families to make inaccurate assumptions about risk and future outcomes on the basis of amyloid PET results. At present, the potential harms outweigh the current minimal benefits. The availability of proven preventative therapies undoubtedly would alter this judgment.
The AIT did not find any evidence to support the utility of amyloid PET in a context outside of a diagnostic evaluation to determine the cause of cognitive impairment. In particular, no evidence supported a role for amyloid imaging to inform physicians when they are consulted on legal, disability, and employment related matters. These include assessing competency; screening for insurability; or assessing employability or the ability to perform activities of daily living such as driving, piloting an aircraft, or financial decision-making.
A major limitation of amyloid PET to support a diagnosis of AD dementia is the high prevalence of amyloid positivity in normal older individuals. Population-based studies are only beginning to be reported, but estimates of age-specific positivity rates for amyloid PET are <5% in those 50-60 years old, 10% in 60-70, 25% in 70-80, and >50% in persons aged 80-90 years [59, 60]. This high age-associated prevalence means that the causality of Aβ for a patient’s clinical syndrome cannot be established with amyloid PET by itself without considering the prior probability of positivity based solely upon age. The dementia expert should consider the possibility, prior to ordering amyloid PET, that incidental, age-related amyloid detection may not be related to or relevant to the presenting symptoms of a patient.
Another major caveat is that a positive amyloid scan can also be seen in not only AD, but also other medical conditions. For example, amyloid PET is frequently positive in dementia with Lewy bodies [61, 62]. Amyloid imaging detects both fibrillar amyloid found in blood vessels (cerebral amyloid angiopathy) and interstitial fibrillar amyloid in plaques. Imaging cannot distinguish between amyloid angiopathy and parenchymal fibrillar plaques , and both are highly prevalent in the elderly, with or without dementia. Although usually associated with interstitial amyloid plaques, in rare cases amyloid angiopathy can occur alone . Occasionally, amyloid angiopathy unaccompanied by typical pathological features of AD can cause progressive dementia [64, 65]. More commonly amyloid angiopathy can become clinically manifest as a cause of cerebral hemorrhage, and in such cases carries a high risk of recurrence [66, 67]. It is important to emphasize that amyloid positivity does not establish the diagnosis of AD or differentiate it from Aβ disorders such as dementia with Lewy Bodies and cerebral amyloid angiopathy.
It is important to note several clinical circumstances in which amyloid PET would not be expected to be useful. First, it would not add any useful information in differentiating disorders that are not associated with Aβ, such as the various frontotemporal dementia syndromes. Second, amyloid PET would not be expected to detect the rare forms of AD in which ligand binding is greatly reduced due to unusual forms of Aβ [14, 68]. The appropriate use of amyloid PET requires knowledge of all relevant findings of clinical evaluations, laboratory tests and imaging relating how each component of the accumulated evidence should be weighed. Thus, clinical amyloid PET should be performed in the context of a comprehensive evaluation undertaken by a clinician with expertise in evaluating cognitive neurodegenerative disorders.
The AIT did not consider other proposed diagnostic biomarkers for AD and therefore did not draw any conclusions as to the relative value of amyloid PET compared to CSF, MRI and FDG-PET (See Further Research Questions below).
The AIT considered broader social and psychological implications of amyloid status determination. While empirical data have not yet been evaluated, the AIT concluded that certain steps should probably be taken by the dementia expert to avoid psychological harm to patients and families that could follow upon the initial disclosure of amyloid status. These steps include pretest counseling about the emotional and social implications ofboth a positive and a negative amyloid PET. Implications in the realms of legal and insurance status, including health, life and long term care, and employment ramifications are even less well understood at this time, and policy makers should consider whether existing laws such as the Americans With Disabilities Act provide adequate protection for these patients. Notably, the U.S. Genetic Information Nondiscrimination Act only applies to genetic tests.
Although published data concerning amyloid PET results and impact on patient care outcome is extremely limited, amyloid PET is likely to contribute to better patient care under specific circumstances. These are described in the following 3 domains.
Greater physician confidence in the diagnosis of or exclusion of AD can result in better medication management. An amyloid positive PET result that raises confidence in the diagnosis of AD is likely to result in earlier and appropriate use of specific medications for symptomatic treatment of AD such as acetylcholinesterase inhibitors and memantine. In contrast, a plan to commence or continue medications developed for the treatment of AD such as the acetylcholinesterase inhibitors and possibly memantine in patients with a negative amyloid scan may be inappropriate. However, there are no studies to date that have assessed the value of these medications in amyloid negative persons with a clinical phenotype suggestive of AD. Furthermore, there is some evidence that acetylcholinesterase inhibitors can benefit patients with vascular dementia . Exclusion of AD should result in consideration of alternate diagnoses including depression, and in some cases of patients with atypical cognitive impairment who are amyloid negative, it may be appropriate to undertake a trial of antidepressant medication.
An amyloid PET cannot answer all diagnostic questions that are encountered in clinical dementia evaluation. It can, however, reduce the use of other tests, which are burdensome to patients and their caregivers. For example, a positive amyloid PET may obviate repeat imaging for the purpose of establishing a clinical diagnosis of dementia while a negative amyloid PET may guide clinicians to order tests that can help differentiate amyloid-negative dementing disorders. Amyloid PET may reduce the use of neuropsychological testing for the purpose of clinical diagnosis.
Under the circumstances outlined above, the results of amyloid PET will increase physician confidence in the clinical diagnosis and allow better planning for patients and caregivers. The following data is from a survey conducted by the Harvard School of Public Health on the public perceptions and awareness of Alzheimer’s disease . The poll was commissioned by Alzheimer Europe through a grant provided by Bayer. These data and facts can be found at http://www.hsph.harvard.edu/news/press-releases/2011-releases/alzheimers-international-survey.html and http://www.alz.org/documentscustom/public-health/valueofknowing.pdf
1) Nearly 89 percent of Americans say that if they were exhibiting confusion and memory loss, they would want to know if the cause of the symptoms was Alzheimer’s disease.
2) Of those aged 60 and over, 95 percent say they would want to know if they had Alzheimer’s disease.
3) Over 97 percent say that if they had a family member exhibiting problems with memory loss, they would want them to see a doctor to determine if the cause was Alzheimer’s.
4) The convergence of evidence from numerous sources indicates that as many as half of people with dementia have never received a diagnosis.
5) A formal diagnosis allows individuals and their caregivers to have access to available treatments, build a care team, participate in support services and enroll in clinical trials.
6) Participating in planning early in the disease process allows individuals with Alzheimer’s to create advance directives regarding their care and finances - so that their wishes can be carried out when they are no longer cognitively able to make such decisions.
7) Early diagnosis also allows individuals with the disease and their caregivers to better manage medications, receive counseling, and address driving and safety issues in advance.
10) Undertaking the diagnostic process early potentially allows cognitive impairment to be reversed in some people. For nearly one in every four individuals who reported to a memory clinic with cognitive problems, their cognitive impairment was due to a reversible cause, such as depression or a vitamin B12 deficiency.
The clinical value of amyloid PET imaging is entirely dependent upon the quality of the images and accuracy of interpretation. Amyloid PET imaging is technically challenging and should be performed only when there is strict attention to quality control. Clinical PET scanning is widely available, but the experience of PET facilities with brain imaging is quite variable. Amyloid imaging is an evolving modality; therefore image interpretation criteria, the clinical significance of positive and negative scans, and technical imaging considerations are evolving. The following recommendations are based on current knowledge, and may require modification in the future.
The safe performance and accurate interpretation of amyloid imaging require physician training as described in the nuclear medicine program requirements of the Accreditation Council for Graduate Medical Education or the equivalent. All nuclear medicine examinations should be performed under the supervision of and interpreted by a physician certified in nuclear medicine or nuclear radiology by the ABNM or the ABR in the U.S.A. or equivalent organizations outside the U.S.A.
The individual performing the scan must be familiar with brain anatomy and have adequate specific training in amyloid PET interpretation. Training specific to the interpretation of amyloid imaging such as provided by the manufacture of the radiopharmaceutical (if available) should be completed, and preferably augmented by training programs offered by professional societies such as the SNMMI and the European Association of Nuclear Medicine (EANM).
Imaging procedures should be performed by a qualified nuclear medicine technologist with appropriate training and certification. All nuclear medicine examinations should be performed by a qualified nuclear medicine technologist who is registered/certified in nuclear medicine by the Nuclear Medicine Technology Certification Board (NMTCB), the American Registry of Radiologic Technologists (ARRT), or equivalent organizations outside the U.S.A. The nuclear medicine technologist works under the supervision of a physician with qualifications outlined above. Imaging should be performed in an imaging facility certified by the Intersocietal Commission for the Accreditation of Nuclear Laboratories (ICANL), the American College of Radiology (ACR) or other equivalent accrediting agency. A procedure guideline for amyloid PET is currently being developed by the SNMMI and EANM.
Results of amyloid PET imaging should be communicated to the referring physician by the imaging physician by way of a written report according to a standard diagnostic imaging practice as outlined in the SNMMI General Imaging Guideline. The final reading should indicate whether Aβ was found to be present (amyloid positive) or was found not to be present (amyloid negative)(see Definitions above). Indeterminate results may arise due to technical or physiological factors and should be reported as such.
The report should not confound amyloid positivity with AD dementia, i.e., it should not by itself advance or rule out a diagnosis of AD dementia. The dementia specialist should then communicate with patients and family members after comprehensive review of the clinical assessment and test results.
The AIT recognized that studies suggest amyloid imaging may have a role to stratify patients into their risk of developing cognitive decline and that, someday, as longitudinal research studies accumulate data, amyloid imaging may become useful to predict future clinical conditions, such as the risk of developing cognitive decline or of transitioning into clinical states such as MCI or dementia [10, 16, 21, 24, 27, 33, 71, 72]. However, these studies need further replication and their results analyzed in a pooled meta-analysis . Therefore, at this point, data are simply incomplete to support using amyloid imaging to provide prognostic information to persons with AD risk factors such as age, family history of dementia, APOE4 status, genetic mutation carrier status, cognitive complaint that is unconfirmed on clinical examination, or to asymptomatic persons.
Recent data from longitudinal studies of normal elderly cohorts with positive amyloid scans show a very slow rate of decline in memory function and suggest that the process of amyloid accumulation may extend for 20 years before dementia is apparent [10,17]. These studies also have shown considerable variation in the rate of amyloid accumulation among individuals. The proportion of healthy elderly persons with a positive amyloid scan who will develop dementia in their lifetime is not known at this time. For this reason, scanning for amyloid in an asymptomatic person in the absence of an effective disease modifying therapy is discouraged.
Multiple imaging modalities and fluid biomarkers have been investigated in clinical and research contexts. Brain FDG PET has been used in a clinical setting and can be diagnostically useful in certain circumstances when a characteristic pattern of hypometabolism is detected for specific neurodegenerative disorders [74, 75], in particular differentiation of dementing disorders in which amyloid PET are similarly positive (such as AD versus DLB) or negative (such as subtypes of FTD). Cerebrospinal fluid measures of amyloid peptides and tau have been extensively investigated and applied to research populations for the purpose of establishing the presence of AD pathology [7, 38, 76, 77], and these fluid assays are beginning to be used in clinical settings. Combined use of presynaptic dopaminergic and amyloid imaging is now being studied as a diagnostic stratification approach to aid differential diagnosis of AD, DLB and FTD . However, the effective use of these diagnostic tests in relation to amyloid PET should be investigated further in the context of patient outcome, benefit and resource use.
Computer-aided analysis software for amyloid imaging is under development and several programs are available for use in the clinic. These can provide quantitative information about the amount of radiopharmaceutical in different brain regions, and may be particularly valuable for sites with limited experience in the reading of amyloid scans and to provide more information than a binary visual read. However, more research is required to validate their use in the clinical environment and demonstrate that they improve the accuracy of clinical reports.
We thank William Klunk for his expert advice to the Amyloid Imaging Task Force meetings and teleconferences. We thank the following individuals for rating the papers used in the evidence-based literature review: Michael Devous, PhD, Alex Drezga, MD, Karl Herholz, MD, FRCP, Gil Rabinovici, MD, Victor Villemagne, MD, David, Wolk MD. Additionally, we thank Michelle Bruno, Avalere Health, for her administrative assistance in managing the Conflict of Interests and the literature searches.
Task Force Members (Alphabetical)
Keith Johnson, MD
Satoshi Minoshima, MD, PhD (Co-chair)
Voting Members (alphabetical)
Nicolaas Bohnen, MD, PhD
Kevin Donohoe, MD
Norman Foster, MD
Peter Herscovitch, MD
Keith Johnson, MD
Jason Karlawish, MD
Satoshi Minoshima, MD, PhD
Christopher Rowe, MD
William Klunk, MD, PhD
SNMMI and AA Members (alphabetical)
Maria Carrillo, PhD (AA)
Dean Hartley, PhD (AA)
Saima Hedrick, MPH (SNMMI)
Kristi Mitchell MPH (Avalere Health)
Virginia Pappas, CAE (SNMMI)
William Thies, PhD (AA)
Indication Subcommittee (alphabetical)
Norman L. Foster, MD (chair)
Jason Karlawish, MD
Christopher Rowe, MD
Literature Subcommittee (alphabetical)
Nicolaas Bohnen, MD, PhD (chair)
Peter Herscovitch, MD
Kristi Mitchell MPH
Literature Reviewers (alphabetical)
Nicolaas Bohnen, MD PhD
Michael Devous, PhD
Kevin Donahoe, MD
Alex Drezga, MD
Karl Herholz, MD, FRCP
Peter Herscovitch, MD
Keith Johnson, MD
Jason Karlawish, MD
Satoshi Minoshima, MD, PhD
Gil Rabinovici, MD
Chris Rowe, MD
Victor Villemagne, MD
David, Wolk MD
The Expert Work Group consisted of three experienced clinicians, two geriatric cognitive neurologists and a geriatrician. The developed guidelines for appropriate and inappropriate clinical use are based upon available literature as well as expert opinion using a modified Delphi procedure. The first task of the Work Group was to individually rate appropriateness of 115 different clinical scenarios based upon the 7 variables listed in table B1. Beginning with the premise that there is value in determining the cause of cognitive impairment, each expert weighed the potential clinical value of amyloid PET against the expense and potential for misuse. After reporting the outcome of these individually considered judgments, the experts came to a consensus about each of the scenarios and used these conclusions to draw generalizations which should be applicable in many different scenarios.
A second task of the work group was to consider the utility of amyloid PET in 8 situations where syndrome classification would be clinically important (e.g. delirium vs. dementia) and in 40 clinically relevant differential diagnosis decision points (e.g. Alzheimer’s disease vs. frontotemporal dementia). Finally, the three clinicians jointly reviewed 10 actual anonymous clinical cases to test whether the guidelines they constructed accurately reflected their own clinical judgment in real rather than strictly theoretical situations. All voting members of the AIT reviewed and discussed the expert-recommended guidelines, and comments were elicited from within and beyond the Work Group. The final expert guidelines reflect this entire process and are the joint opinion of the entire AIT.
In our deliberations, we assumed that the expert clinician would receive an interpretation of the amyloid PET study as either positive or negative (See Definitions above). Briefly, a positive amyloid PET result typically means that the image demonstrates uptake of amyloid radiotracer in the gray matter. In contrast, a negative scan means that the image does not demonstrate any gray matter cortical uptake that is above the level of nonspecific binding. The determination of cortical uptake relies on inspection of the contrast between white matter and cortical gray matter. In white matter, imaging ligands are routinely visible not as a result of their binding to amyloid, but due to lipophilic interactions with myelin. Some ligands have greater white matter uptake than others. Histopathology-PET correlation studies relating amyloid PET results with neuropathologic measures of amyloid plaques, using both immunohistochemistry or silver staining, have shown that if amyloid plaques are “none” or “sparse”, amyloid imaging ligand specific binding in cortical grey matter is very low and reflects nonspecific, i.e.., non-amyloid binding. In contrast, when histopathologic evidence at autopsy confirms the presence of “frequent” amyloid plaque, the amyloid image contrast between white and grey is greatly reduced or visually undetectable.
The AIT did not consider the potential impact on the AUC of the use of quantitative PET data, i.e., data from automated numerical measurements of specific ligand binding. Quantitative measurement entails image processing steps in which specific brain regions are identified in each individual data set for the purpose of comparing regional cortical tracer uptake with an unaffected reference region. At present, there is insufficient published data to recommend a specific implementation of amyloid PET quantitation that could be identified in the AUC.
|Scenario variable||Variations considered|
|Memory or cognitive complaint|
|Clinical context||N/A (Non-medical)|
|Full evaluation, AD not suspected|
|Full evaluation, AD suspected|
|Evidence of deficit on exam||Not applicable (Non-medical)|
|Family history of AD or ApoE4 positive||Negative|
|AD genetic mutation carrier||Negative|
The Literature Subcommittee used a search strategy as established by the American Academy of Neurology and the National Institutes of Medicine to identify relevant literature. Multiple searches were performed using the National Institute of Health’s (NIH’s) PubMed where 408 publications were initially identified. Literature search limits and parameters were as follow: Human, English, and Publication date: 01/01/2002-Present. The search terms determined to be the most useful for identifying the pertinent literature were: 1) (Florbetapir AND AV-45) OR Amyvid; 2) PIB OR Pittsburgh Compound B; 3) Flutemetamol OR AV1; 4) (F-18 FDDNP) OR (F18 FDDNP); and 5) Florbetaben OR 8F-BAY94-9172.
Using the PubMed generated list of 408 papers, the Literature Subcommittee reviewed the list for inclusion and identified a subset of papers by abstract analysis. Papers not relevant for clinical use of amyloid PET were eliminated based on the primary focus of the study and data presented in the paper. These include radiochemistry study; in vitro study; animal toxicity study; biodistribution study; image and kinetic analysis study; dosimetry study; pathophysiological investigation; correlational study; study with a small number of subjects; surrogate marker study in therapeutic trials; and review and editorial commentary. In addition, to ensure appropriate papers were captured in initial search and review, the AIT Task Force performed a “backward review” to cross-check the literature included in seminal amyloid imaging reviews with those included in the AIT’s initial assessment. For the backward review, the AIT employed the bibliographies of Klunk , Villemagne , and Laforce .
The Literature Subcommittee developed a Data Extraction Form for the evidence assessment. The Data Extraction Form included multiple questions and data extraction sections including: 1) if the paper addresses one or more of the potential indications/non-indications proposed by the Indication Subcommittee; 2) individual data points for recalculation of the data; 3) study design; 4) study logistics; 5) patient recruitment setting; 6) inclusion/exclusion criteria; 7) criteria used for diagnosis; 8) inclusion of control subjects; 9) subject characteristics; 10) type of radiopharmaceuticals used; 11) type of PET scanner used; 12) method of PET interpretation and analysis; 13) 2×2 data extraction for histopathologically confirmed study as well as MCI to AD conversion study; 14) 19-point Quality Score of Paper; 15) the AAN Level of Evidence for a Diagnostic Study Article; 16) the AAN Level of Evidence for a Prognostic Study Article; and 17) if the paper addressed changes in physician confidence.
The Alzheimer’s Association and the Society of Nuclear Medicine and Molecular Imaging rigorously attempted to avoid any actual, perceived, or potential conflicts of interest that might have arisen as a result of an outside relationship or personal interest of the writing committee members of the Amyloid Imaging Task Force or of external reviewers used to review specific papers. Both organizations reviewed their own Industry Relationship Policies to ensure that the ensuing process adhered to both standards.
Amyloid Imaging Task Force (AIT) members were required to provide disclosure statements of all relationships that might be perceived as real or potential conflicts of interest (COI). These statements were reviewed and discussed by the Task Force co-chairs, and updated and reviewed by an objective third party at the beginning of every Task Force meeting and/or teleconference. A table of disclosures for Task Force members and external literature reviewers can be found below, Table D1.
To adjudicate the conflicts of interest, the leadership from the Alzheimer’s Association and the Society of Nuclear Medicine and Molecular Imaging first determined the threshold for a real COI. Following consultation with various experts and review of other policies used, the team defined COIs as the following: An individual that had relationships with industry, including consulting, speaking, research, and other non-research activities, that exceed $5,000 in funding over the previous or upcoming twelve-month period.
Additionally, if external expert reviewers of the papers were either a principle investigator or other key study personnel on a study, their participation in the review would likely present a COI. All reviewers completed COI forms. Paper authors and sponsors were identified, and then cross-checked against reviewers’ financial and intellectual COIs. Conflicted individuals were noted as unable to review papers in which there was a real COI present.
Of note, William Klunk, MD, co-inventor of PiB-class and Chrysamine-G-class amyloid imaging agents, was appointed as an advisor to the Task Force, contributing expertise as requested, but recused himself from any and all discussions that resulted in a vote among writing committee members.
|Name||Reported Relationships with Industry or Other|
The Amyloid Imaging Taskforce solicited information from all communities through the SNMMI and AA websites and by direct solicitation to members of these societies. The comments and input helped to shape the development of these appropriate use criteria and the consensus recommendation for the appropriate use of amyloid imaging for clinical indications of the detection of fibrillar amyloid in the brain.