Aβ-PET scans may be part and parcel of clinical AD research studies, but beyond the walls of the ivory tower their use is limited. According to a study published June 11 in JAMA Neurology, the time may have come to expand access to the scans, at least for some patients. Researchers led by Arno de Wilde of VU University in Amsterdam reported that for a quarter of patients at a typical memory clinic, Aβ-PET scans prompted a change both in diagnosis and in treatment plan. The study included patients who are typically excluded from receiving the scans—such as people who only have subjective memory complaints, or people whose AD dementia diagnosis seem clear-cut.

  • Five hundred and seven patients, whose clinical diagnoses ranged from subjective cognitive decline to dementia, underwent amyloid-PET scans.
  • For 125 of them, scan results prompted a change in diagnosis and 123 had changes to treatment plans.
  • For people who chose to learn the results, scans did not change anxiety, but did relieve uncertainty.

Aβ PET scans offer invaluable information in the context of clinical trials, especially those testing Aβ-targeted drugs. However, use of the scans outside of the research setting, either to inform diagnosis or to provide disease risk information to concerned patients, has remained controversial. The appropriate-use criteria for Aβ imaging in the clinic limit use of the scans to patients deemed most likely to benefit, including people with objectively measured cognitive symptoms, and people whose AD diagnosis is uncertain (Jan 2013 news). Two massive studies—IDEAS and AMYPAD—in the United States and Europe, respectively, are currently testing the clinical outcomes of the scans. Initial findings from IDEAS, which enrolls people with cognitive symptoms and an uncertain diagnosis, suggest the scans result in a change of diagnosis more often than not (Aug 2017 news). 

De Wilde and colleagues wanted to test the clinical relevance of the scans in a typical, unfiltered population of memory clinic patients. As part of the Alzheimer Biomarkers in Daily Practice (ABIDE) project, they enrolled 507 patients from two memory clinics in the Netherlands. Before giving the participants PET scans, the researchers asked neurologists to submit for each patient a clinical diagnosis, as well as their best guess about the underlying pathology at the root of the symptoms. Participants, who averaged 65 years old, fell into one of three clinical categories (subjective cognitive decline, mild cognitive impairment, or dementia) and either of two etiological ones (suspected AD or non-AD pathology).

Nearly half tested positive for Aβ accumulation. Among people with suspected AD pathology prior to the scans, 78 percent of people with dementia, 45 percent of people with MCI, and 38 percent of people with SCD had a positive scan. Aβ-positive scans were rarer, but still substantial, among participants with suspected non-AD pathology prior to the scans, ranging from 33 to 24 to 21 percent of people with dementia, MCI, and SCD, respectively. Nearly 70 percent of the people with positive scans carried at least one ApoE4 allele, while only about 30 percent of amyloid-negative participants did.

In all, the scan results prompted neurologists to change their etiological diagnoses—i.e., the pathological cause of symptoms—for one in four participants. Aβ-negative scans led to more diagnostic changes than Aβ-positive scans, and people older than 65 received more diagnostic reversals than younger patients did. Neurologists were also asked to gauge their confidence in these etiological diagnoses, and reported that on average, the scans improved diagnostic confidence from 80 percent prior to the scans to 89 percent afterwards. The biggest diagnostic confidence-booster accrued to patients with MCI whose scan results agreed with their suspected pathology prior to scans.

The amyloid-PET scan results motivated neurologists to alter treatment plans in a quarter of participants. For those who tested positive for Aβ accumulation, physicians prescribed cholinesterase inhibitors, and/or referred them to clinical trials. For those who tested negative, the results triggered further investigations, such as genetic screening, FDG-PET scans, and referrals to psychiatrists.

Overall, to de Wilde the findings mean that PET scans are indeed useful in the clinic, particularly for people with objectively measured cognitive symptoms. Particularly for people with MCI, a more certain diagnosis could alter treatment, speed trial recruitment, and improve the patient’s prognosis, he added. He thinks the jury is still out on whether people with SCD should receive PET scans, because it’s still unclear what a positive result means. However, both de Wilde and co-author Wiesje van der Flier told Alzforum that patients in that category are increasingly demanding more information about the cause of their symptoms. Van der Flier said that the researchers are tracking cognitive symptoms in the cohort of SCD patients in the study, hoping to better understand the link between SCD, Aβ accumulation, and cognitive decline. 

Henrik Zetterberg of the University of Gothenburg in Sweden commented that the study supports the idea that diagnoses based on clinical symptoms alone are insufficient. “If we can make a more correct diagnosis, of course we should use these tools if the patient wants to know,” he said. He added that both CSF biomarkers and amyloid-PET could serve this purpose. Although CSF measurements have the advantage of measuring multiple biomarkers, PET and CSF are equivalent as far as Aβ is concerned, Zetterberg said. Which one is used depends largely on which equipment is available, as well as the expertise and preference of physicians and patients, he said.

De Wilde, van der Flier, and Zetterberg all noted that while the scans will become pivotal once an amyloid-targeted drug is approved, biomarker knowledge serves an important role even in the absence of such a therapy.

In an editorial accompanying the paper, Steven Salloway of Brown University in Providence, Rhode Island, wrote that de Wilde’s study supports the idea that the scans will benefit people with MCI or dementia, regardless of the availability of an amyloid-targeting drug. "The foundation for excellent medical care rests on an accurate diagnosis,” he wrote. "The improvement in diagnostic accuracy provided by amyloid PET scans will allow families to make important decisions about work, finances, and driving safety more knowledgeably and to take early advantage of both approved medications and opportunities to participate in clinical trials."—Jessica Shugart


  1. From my perspective, de Wilde et al. provides important, novel and high-level evidence for the clinical utility of amyloid PET imaging.

    Our current knowledge in this regard was so far based on a number of smaller, sometimes retrospective or observational studies in selected patient cohorts. It is, thus, extremely valuable, not only for discussing potential reimbursement of amyloid PET with payers, to confirm these previous data by a prospective study and in an unselected population as achieved by de Wilde et al.

    It is important to stress, however, that a change of diagnosis, diagnostic confidence, and management is, while encouraging based on our knowledge on amyloid PET imaging accurately reflecting histopathology, per se not attributable to an improved outcome. As a control group without amyloid imaging is, understandably, not available in this ABIDE project, the ultimate questions of whether, and to which degree, amyloid imaging improves patient outcome cannot be addressed here. In this regard, the community will, thus, need to wait for the results of the ongoing IDEAS and AMYPAD studies.

    I agree with the comment made on the somewhat limited generalizability of the study results to the “real clinical world.” This refers, in my eyes, to both the scenarios in which amyloid imaging was carried out, and in which the clinical diagnosis was established. For the amyloid imaging, as the nuclear physician performing the PET scan is bound to collect all patient information beforehand, the information on the pretest clinical diagnosis and on that of the CSF/APOE testing results (where available) are includable in clinical routine (other than in this study) into the amyloid image read. For establishing the pre-PET clinical diagnosis, the same holds true for CSF/APOE testing results.

    However, it would definitely be interesting to, in a comparative design, investigate in future studies how amyloid imaging compares to CSF sampling in regard to clinical utility/outcome effects/cost-efficiency. Such research should definitely also consider the frequency of borderline/inconsistent results, the frequency of side effects, and the interlab reproducibility for both modalities.

    Regardless, the ABIDE study results provide one important missing link for establishing amyloid imaging in a wider clinical routine scenario. This PET approach to accurately detect/exclude amyloid pathology in cognitively impaired patients in a noninvasive manner can, already by today, be considered without doubt a valuable addition to the diagnostic toolbox in dementia.                     

Make a Comment

To make a comment you must login or register.


News Citations

  1. HAI—Amyloid Imaging in the Clinic: New Guidelines and Data
  2. In Clinical Use, Amyloid Scans Change Two-Thirds of Treatment Plans

Further Reading

Primary Papers

  1. . Association of Amyloid Positron Emission Tomography With Changes in Diagnosis and Patient Treatment in an Unselected Memory Clinic Cohort: The ABIDE Project. JAMA Neurol. 2018 Sep 1;75(9):1062-1070. PubMed.
  2. . Improving Evaluation of Patients With Cognitive Impairment With Amyloid Positron Emission Tomography. JAMA Neurol. 2018 Sep 1;75(9):1045-1046. PubMed.