SAN DIEGO—Results of several small research studies indicate that the use of positron emission tomography (PET) with an imaging agent known as Pittsburgh Compound-B (PiB) can identify in vivo neuroanatomic evidence of Alzheimer’s disease at a much earlier stage than had been possible previously. Using this finding as a baseline, clinicians would then follow patients by applying the imaging technique to detect any changes in the natural history of amyloid deposition induced by anti-amyloid therapies.

Two of the studies yielded valuable data on patients with mild cognitive impairment (MCI). Because 50% to 60% of patients with MCI will progress to Alzheimer’s disease within five years of diagnosis, according to current estimates, earlier identification of patients who have amyloid plaques in their brain will enable clinicians to initiate treatment earlier, when it is more likely to be effective.

One study assessed patients with MCI in relation to normal controls and patients with Alzheimer’s disease by using the PET amyloid-imaging method to determine the comparative presence, quantity, and distribution of binding with and retention of PiB—thereby determining the status of amyloid plaque deposition among these three patient groups. The investigators found that with respect to retention of PiB, the patients with MCI spanned the entire range of the control and Alzheimer’s disease values, suggesting that those who had significant amounts of binding in the cortex or posterior cingulate were likely to already have pathologic changes characteristic of Alzheimer’s disease present in the brain.

A second study showed the amyloid imaging technique to be useful in identifying pathologic evidence of Alzheimer’s disease in a significant proportion of subjects in a subset of MCI patients—high-IQ older individuals with subtle memory impairment for their level of ability, or "IQ-MI" subjects. The investigators concluded that use of IQ-adjusted memory test scores enabled them to identify individuals at increased risk for Alzheimer’s disease who escaped detection because of high cognitive reserve.

In a third study, researchers at the University of California at San Francisco and Berkeley found that PET imaging with PiB was effective in distinguishing Alzheimer’s disease from frontotemporal lobe dementia (FTLD)—although in some cases PiB results also contradicted the clinical diagnosis. The researchers concluded that their preliminary findings need to be confirmed in larger studies and in autopsy-proven cases.

The results of the studies, which were presented at the 58th Annual Meeting of the American Academy of Neurology, should lead to a major change in the way clinicians think about presymptomatic diagnosis and treatment of Alzheimer’s disease, said William E. Klunk, MD, PhD, Director of Psychiatry at the Alzheimer Disease Research Center (ADRC) of the University of Pittsburgh and coinventor of the PiB imaging technology, with Chet Mathis, PhD.

In his presentation, Dr. Klunk highlighted the crucial role of early detection in the treatment of Alzheimer’s disease while emphasizing how difficult it is, from a research perspective, to find cognitively normal subjects with a high probability of definitive amyloid deposition. He pointed to individuals who have a rare genetic mutation in the presenilin-1 or amyloid-precursor protein (APP) gene as providing a cue to understanding how to properly treat Alzheimer’s disease—with the aid of the new imaging technology he helped to develop.

"PiB binding appears to reflect in vivo levels of amyloid-b deposition," Dr. Klunk said. "About 65% of MCI cases showed PiB retention, and this is interesting because in our center that’s about the percentage that goes on to develop Alzheimer’s disease in a few years. PiB retention may be a good reflection of the efficacy of anti-amyloid therapies; but in order for it to be a really significant advance in the treatment of Alzheimer’s disease, I firmly believe we have to change our thinking to be more in terms of preclinical diagnosis and preclinical treatment.

"I think this group of patients who have these unfortunate mutations in the presenilin-1 or APP gene, much like they’ve led us to much of our understanding of the molecular basis of Alzheimer’s disease, can also lead us to understand how to treat this disorder, because it would be a real shame to take an effective treatment and apply it at a stage of the disease when it just can’t be effective."

DELINEATING PATHOLOGIC CHANGES IN A CONTINUUM OF DISEASE

MCI is now widely considered to be an initial stage of Alzheimer’s disease in an overall trajectory of clinical decline that may begin several years before the onset of overt dementia. What is becoming clear to researchers is that the neuropathologic changes of Alzheimer’s disease, including the accumulation of amyloid-b, are already present by the time that MCI occurs—and may be present at an even earlier stage along the trajectory.

In the study in which patients with MCI were found to overlap with Alzheimer’s disease patients as well as controls with respect to PiB retention, investigators at the University of Pittsburgh and elsewhere began with a question: How early could one detect, using neuroimaging, a pathology implying a continuation of Alzheimer’s disease? In recent years, the pathologic status of MCI has been under intense study, according to principal investigator Steven T. DeKosky, MD, who is Director of the ADRC. One result is that as a cohort of patients with milder levels of cognitive decline has begun to die while taking part in large-scale studies, researchers have been able to see what or how much pathology exists in the brains of people who had undergone neurologic assessment early in the course of their disease.

"We often talk about these entities of change in cognition over a lifetime or over the geriatric span; but, in fact, what we are doing is making a series of categories out of what is fundamentally a multidimensional continuum," observed Dr. DeKosky. "We have always assumed that as we follow our patients from normal to full-blown disease, that their clinical changes are based upon progressive pathologic changes. We have recently been able to examine the changes in the brains of people who happened to die at points at which their disorder is relatively mild. These are extraordinarily valuable pieces [of evidence] to have."

In their study, Dr. DeKosky and his colleagues found that PiB PET was a better biomarker of mild Alzheimer’s disease than was ¹⁸F-fluorodeoxyglucose PET, separating all of their cases from their controls. They found no difference in retention of the imaging agent between patients diagnosed as "MCI-amnestic" and as "MCI–multiple cognitive domain," and both diagnoses were made in PiB-positive as well as PiB-negative cases.

"Longitudinal follow-up of these MCI cases is under way to see, in fact, if those PiB-negative cases do go on to develop worse cognition and clearcut Alzheimer’s disease," Dr. DeKosky said. "It appears reasonable at this point to say that someone [diagnosed as] MCI-amnestic or MCI-multiple cognitive domain who has significant amounts of PiB binding in the cortex or posterior cingulate cortex is likely to have Alzheimer’s disease, pathologically, setting the stage for earlier definitive diagnosis."

DRAWING ON COGNITIVE RESERVE TO ESCAPE DETECTION

In forming the rationale for their study involving high-IQ older patients with subtle memory impairment, Dorene M. Rentz, PsyD, Keith A. Johnson, MD, and colleagues at Brigham and Women’s Hospital and Massachusetts General Hospital in Boston and the University of Pittsburgh found that it is particularly difficult to differentiate the earlier stages of the pathophysiologic process of Alzheimer’s disease from the trajectory of normal aging.

"Subjects with high cognitive reserve often score within the normal range on memory tests using standard test norms," Dr. Rentz said. "But these standard norms may classify individuals as normal and mask early memory changes."

She and her colleagues devised a method for adjusting memory test scores based on an estimate of premorbid IQ in order to account for cognitive reserve. This method allowed them to identify the individuals classified as IQ-MI. They found that several of the IQ-MI and MCI patients, as well as one normal control, placed above the threshold for PiB positivity or had such binding at a level that was intermediate between normal-control and Alzheimer’s disease values. "Even with a single normal-control subject who showed elevated PiB binding, we still found a significantly higher mean level of PiB binding in the IQ-MI group, and the analysis examining the ratio of PiB positivity in the IQ-MI compared to the normal-control group yielded similar results," Dr. Rentz reported. "These findings suggest that amyloid deposition was identified using the IQ-adjusted method in a [subset of] subjects who would have otherwise gone undetected."

She added, "The amyloid deposition was in the anatomic patterns characteristic of Alzheimer’s disease, and the amount of amyloid present relates to memory performance. Our results suggest that the use of IQ-adjusted norms may provide a means of identifying subjects at increased risk for developing Alzheimer’s disease who escaped detection because of cognitive reserve."

As with the study by Dr. DeKosky et al, Dr. Rentz’s group expects longitudinal follow-up to be particularly useful in evaluating the correspondence of PiB binding to clinical outcome.

DISTINGUISHING DEMENTIAS

Accurately differentiating between Alzheimer’s disease and FTLD has important implications for prognosis and future disease-specific therapies, said Gil D. Rabinovici, MD, lead investigator of a study that used PiB PET in an attempt to distinguish one disease from the other.

FTLD is associated with a more rapid decline and shorter survival than Alzheimer’s disease and is more strongly familial than Alzheimer’s disease. "In certain cases, however, differentiating between these two diseases on clinical grounds can be challenging because of overlapping symptoms, overlapping neuropsychological profiles, and also because Alzheimer’s pathology can mimic any of the FTLD clinical syndromes," he said.

Dr. Rabinovici stated that retrospective autopsy series have consistently demonstrated that a thorough clinical evaluation, including neuropsychological testing and structural imaging, doesn’t always correctly predict the underlying pathology, highlighting the need for tests such as molecular imaging, which may improve the specificity of clinical predictions.

"PiB is a great candidate for this particular diagnostic dilemma since amyloid-b plaques are one of the pathologic hallmarks of Alzheimer’s disease but are not present in the FTLD pathologic spectrum," Dr. Rabinovici said.

In their study, he and his colleagues found that while all six of the Alzheimer’s disease patients had positive PiB scans and most of the FTLD patients had negative ones, as was expected, a surprisingly high number (four of 10) of patients with a clinical diagnosis of FTLD had positive scans. Dr. Rabinovici offered several possible explanations.

"The first is that these are false-positives—that PiB is binding to something other than amyloid in the brains of these patients," he said. "We are still learning about PiB; but at least the preliminary data suggest that the binding to amyloid-b fibrils is very specific, and there is very little affinity for PiB to bind to the brains of patients with non-Alzheimer’s dementias. A second possibility is that there is comorbid Alzheimer’s and FTLD pathology, and that the FTLD pathology is driving the clinical syndrome. This is certainly possible, though I think the relatively young ages of our patients make that somewhat less likely.

"Another possibility, of course, is that PiB is right and the clinical diagnosis is wrong," Dr. Rabinovici concluded, "in other words, that this is Alzheimer’s disease pathology mimicking an FTLD clinical syndrome."

NR

—Fred Balzac

Suggested Reading
Klunk WE, Engler H, Nordberg A, et al. Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann Neurol. 2004;55:306-319.

Return to table of contents