Tau-Based PET Imaging Used to Predict Alzheimer’s Progression

Brain imaging of pathological tau-protein tangles reliably predicts the location of future brain atrophy in Alzheimer's patients a year or more in advance, according to a new study out of the University of California San Francisco. In contrast, the location of amyloid plaques, which have been the focus of Alzheimer's research and drug development for decades, was found to be of little utility in predicting how damage would unfold as the disease progressed.

The results, published yesterday in Science Translational Medicine, support researchers' growing recognition that tau drives brain degeneration in Alzheimer's disease more directly than amyloid protein, and at the same time demonstrates the potential of recently developed tau-based PET (positron emission tomography) brain imaging technology to accelerate Alzheimer's clinical trials and improve individualized patient care.

"The match between the spread of tau and what happened to the brain in the following year was really striking," said senior author Gil Rabinovici, MD. "Tau PET imaging predicted not only how much atrophy we would see, but also where it would happen. These predictions were much more powerful than anything we've been able to do with other imaging tools, and add to evidence that tau is a major driver of the disease."

"No one doubts that amyloid plays a role in Alzheimer's disease, but more and more tau findings are beginning to shift how people think about what is actually driving the disease," explained lead author Renaud La Joie, Ph.D. "Still, just looking at postmortem brain tissue, it has been hard to prove that tau tangles cause brain degeneration and not the other way around. One of our group's key goals has been to develop non-invasive brain imaging tools that would let us see whether the location of tau buildup early in the disease predicts later brain degeneration."

Despite early misgivings that tau might be impossible to measure in the living brain, scientists recently developed an injectable molecule called flortaucipir—currently under review by the FDA—which binds to misfolded tau in the brain and emits a mild radioactive signal that can be picked up by PET scans.

The team recruited 32 participants with early clinical stage Alzheimer's disease all of whom received PET scans using two different tracers to measure levels of amyloid protein and tau protein in their brains. The participants also received MRI scans to measure their brain's structural integrity, both at the start of the study, and again in follow-up visits one to two years later.

The researchers found that overall tau levels in participants' brains at the start of the study predicted how much degeneration would occur by the time of their follow up visit (on average 15 months later). Moreover, local patterns of tau buildup predicted subsequent atrophy in the same locations with more than 40 percent accuracy. In contrast, baseline amyloid-PET scans correctly predicted only 3 percent of future brain degeneration.

"Seeing that tau buildup predicts where degeneration will occur supports our hypothesis that tau is a key driver of neurodegeneration in Alzheimer's disease," La Joie said.