A controversial PET tracer may provide a window into the neural devastation wrought by Alzheimer disease, investigators at the University of California-Los Angeles (UCLA) reported in a new study (N Engl J Med 2006;355:2652–2653).
Using the tracer, they could track the progression of amyloid plaques and neurofibrillary tangles in living subjects for the first time, and correlate it with cognitive deterioration, lead author Gary Small, MD, told Neurology Today.
Dr. Small, director of the UCLA Center on Aging and a professor at the Semel Institute of Neuroscience and Human Behavior, and his colleagues synthesized a small molecule, [F-18] FDDNP, with an affinity for the cerebral amyloid and tau proteins that accumulate in the brains of people with Alzheimer disease (AD) and mild cognitive impairment (MCI). In earlier studies, FDDNP binding values in the temporal, parietal, and frontal regions of patients with AD were significantly higher than the values for age-matched, cognitively intact subjects, suggesting that the tracer could help in tracking the development of these disorders.
For this study, the subjects, who ranged from 49 to 84 years of age, were recruited through local advertisements, media coverage, and physician referrals. All had self-reported memory problems, and had had neurologic, psychiatric, and laboratory testing done in addition to the imaging studies.
Twenty-five people had AD, 28 MCI, and 30 were healthy controls. The AD and MCI patients had a mean age of 73 and 70 years, respectively, compared to a mean of 64 years for the control group (p=0.01). Other than that, there were no significant differences between the groups in terms of educational level, gender, or family history of AD.
Each subject was given an intravenous injection of a bolus of FDDNP through an indwelling catheter. Images were then obtained for two consecutive hours, with the final results containing 47 or 63 contiguous slices, depending on the plane separation and the scanner used. Global FDDNP-PET binding values were calculated by combining the average relative distribution volume of the temporal, parietal, posterior cingulate, and frontal regions. Individual binding values were expressed as the average DVR of the left and right region.
In addition to PET, 72 volunteers had MRI; 11 could not have MRI because of claustrophobia or metal objects in the body and were examined with CT.
People with AD had significantly higher global and individual binding values than their counterparts with MCI, and the people with MCI had higher binding values than the control subjects (p<0.001). The reverse was true of FDG glucose metabolism: global and regional metabolism was highest among controls, significantly lower in the MCI group, and lowest of all among those with AD (p<0.001).
Higher values for global FDDNP binding correlated with lower values for FDG-PET in the posterior cingulate and parietal regions (p<0.001 for both regions). Higher global FDDNP binding values also correlated with lower medial temporal volumes on MRI (p=0.02), and larger ventricular volumes (p=0.002). Lower values for FDDNP binding correlated with higher scores on cognitive tests, including the Mini-Mental State Examination and the digit symbol test portion of the Wechsler Adult Intelligence Test (p<0.001 in both cases).
FDDNP was 98 percent accurate at distinguishing between AD and MCI, compared to FDG-PET, which was 87 percent accurate, and MRI, with an accuracy of 62 percent.
The investigators performed follow-up PET on 12 subjects two years later. At the time of the first test, eight had normal cognitive function, and four had MCI. Over the years, one control subject had progressed to MCI, while cognitive function declined in a second but did not quite attain the diagnostic threshold for MCI. Two patients with MCI at baseline had progressed to AD. FDDNP binding reflected these patients' deteriorating cognitive status, increasing from 5 percent in the first scan to 11 percent in the later scan.
This was the first time plaque and tangle development could be followed in living humans and associated with a corresponding decline in cognitive ability, said Dr. Small.
Taken together, these findings indicate that FDDNP-PET can differentiate MCI from normal aging and AD, Dr. Small and his colleagues wrote.
Additional support for this conclusion came from the autopsy findings on a 78-year-old subject with AD who died 14 months after FDDNP-PET scanning and a baseline clinical evaluation. The concentration of plaques and tangles were highest in the regions of the brain with the most FDDNP-PET binding: the medial temporal, lateral temporal, posterior cingulate, and frontal areas.
OTHER TRACERS OF INTEREST
FDDNP is one of a handful of novel imaging ligands for neurofibrillary plaques and tangles currently under investigation. Interest in these compounds is high and the debates about them are heated, because they can track progression of AD and MCI.
Perhaps the best known tracer is the Pittsburgh Compound B (PIB), a fluorescent benzothiazole derivative developed by a team at the University of Pittsburgh in 2003, with the first human findings published in 2004 (Ann Neurol 2004;55:306–319).
Proponents of PIB note that its affinity for the amyloid protein of AD is greater than that of FDDNP. In human PET studies, PIB binding has been increased by as much as 80- to 100-percent in people with AD compared to controls, substantially higher than the 9 percent increase reported for FDDNP.
PIB can produce reliable images after 20- to 30-minutes of scan time, making it easier to use in clinical practice, and “PIB images should be easier for clinicians to read,” said Christopher Rowe, MD, director of nuclear medicine and PET at Austin Hospital in Melbourne, Australia. He and his colleagues have used the PIB tracer in 170 patients, including 50 with MCI. Thirty-five percent of those with MCI showed normal levels of tracer uptake, with the rest displaying levels similar to the lower end of the spectrum observed in AD patients. These findings “may allow us to predict those who will progress to AD from those who will not,” he said.
So far, PIB has been more widely studied than FDDNP, Dr. Rowe added. “There are now over 25 sites around the world using PIB, and all are producing consistent results, with publications from six of the sites and more in press,” he told Neurology Today. “As yet, no one has presented findings with FDDNP other than this UCLA group, even though the ligand has been around as long as PIB.”
On the other hand, FDDNP has at least two important advantages over PIB, Dr. Small told Neurology Today. First, it binds to neurofibrillary tangles as well as plaques, while PIB binds only to plaques. Second, FDDNP has a much longer half-life than PIB, so it can be made at a central facility, and then stored and shipped to multiple PET centers the same day. PIB must be labeled with C11 in the laboratory just prior to use, which makes it harder to work with.
Like many other researchers in this area, Dr. Small contends the most exciting promise of FDDNP and other tracers will be as tools for monitoring patients' response to the secretase inhibitors and other anti-amyloid therapies now in the pipeline. There is even tantalizing preliminary evidence that naproxen and other nonsteroidal anti-inflammatory agents may dissolve plaques, at least in vitro. “That's an interesting story that is not over yet,” he said.
ARTICLE IN BRIEF
A novel PET tracer tracked the progression of amyloid plaques and neurofibrillary tangles in living subjects for the first time, and correlated it with cognitive deterioration.