Amyloid PET Can Help Predict Cognitive Decline

By analyzing data from hundreds of NACC participants in the National Alzheimer’s Coordinating Center (NACC) database, researchers have found that imaging scans for amyloid proteins can predict a patient’s risk of cognitive decline. The work, published in the journal Alzheimer’s & Dementia, provides strong evidence that amyloid PET, a brain scanning technology increasingly being used to study Alzheimer’s disease, could improve treatment decisions.

Plaques of amyloid proteins in the brain are a hallmark of Alzheimer’s disease, but scientists remain divided on whether the plaques are a cause or an effect of the condition. Nonetheless, investigators and clinicians now use positron emission tomography (PET) scans to track amyloid protein accumulation in patients’ brains. “It’s a marker of this disease, even if we don’t have the exact mechanisms ironed out for its causal pathways, so we put it in the category of disease-specific biomarkers,” says Elizabeth Mormino, Associate Professor of Neurology and Neurological Sciences at Stanford University in Palo Alto, CA, and senior author on the new study.

To determine how reliable this biomarker is, Mormino and her colleagues needed a large, diverse sample of patients who’d received amyloid PET scans and been tested for memory and cognitive declines over time. The NACC database is populated with data from the Alzheimer’s Disease Research Centers (ADRCs). Many of these Centers had been collecting amyloid PET data for years, but the results weren’t standardized. “Some [sites] were collecting these data, some weren’t, but they weren’t being pulled together and shared in any sort of fashion,” says Mormino.

Fortunately, the NACC’s existing systems for coordinating data exchanges across the ADRCs enabled Mormino’s team to start harmonizing the amyloid PET scan results. “The NACC has already done this for the clinical data … but imaging has kind of lagged behind in terms of centralization and sharing,” says Mormino. Working with a large team of collaborators across the Centers, she and her colleagues spent about two years collecting and reformatting imaging data so they could be analyzed in a consistent way.

Next, the researchers asked whether amyloid PET was a useful prognostic tool. “If you are amyloid positive, regardless of your clinical diagnosis, does that tell us something about what happens to you in the future?” says Mormino. The new results show that it does. Higher amyloid levels correlated with progression from unimpaired to mild cognitive impairment, and from mild cognitive impairment to dementia. In addition, those who had amyloid accumulation but were not considered impaired when they were scanned tended to have poorer memory initially, as well as declines later across cognitive tests, compared to those without amyloid accumulation.

It’s a timely result, as the number of patients undergoing amyloid PET has skyrocketed in recent years, especially with the advent of anti-amyloid drug treatments. “Insurance companies are covering these scans as a way of determining eligibility for anti-amyloid therapeutics, so … all of a sudden these scans are actually quite relevant because there are treatment decisions that are linked to them,” says Mormino. She adds that validating amyloid PET as a diagnostic tool will also help researchers find other biomarkers for Alzheimer’s disease, for example by checking blood-based biomarkers against amyloid PET results.

As NACC’s handling of imaging data has evolved, Mormino foresees many more opportunities for this type of analysis. “The centers are now in real time giving their data to a centralized location, so in addition to this older legacy data, we actually have these nice standardized, harmonized imaging data sets,” says Mormino. She adds that “we published almost 900 participants, but we’re already into a couple thousand [more] that have been characterized because of this effort.” In addition to amyloid PET, the growing flood of data includes Tau PET, which images the tangles of Tau protein that have also been linked to Alzheimer’s disease.

By combining these imaging tests with other types of data, researchers and clinicians could soon have much more robust tools for diagnosing and predicting Alzheimer’s disease progression, as well as making treatment decisions. “I think the next big step is integrating these large data sets, continuing to look at them with respect to clinical trajectories and also other data modalities that are collected on these participants, such as genetics and proteomics and eventually postmortem data as well,” says Mormino.