22 November 2011. Beacons for possible future dementia often go unnoticed because seeing them requires imaging technology only available in specialized research centers. The tide may soon turn, though. New developments in amyloid imaging with F18-labeled positron emission tomography (PET) ligands and in a new magnetic resonance imaging (MRI) technique that measures brain metabolism could make these tools more widely available.
One need not be a neuroimaging specialist to know of Pittsburgh Compound B (PIB)—the best-studied PET reagent for detecting brain amyloid. While growing evidence confirms its reliability as a readout for brain Aβ, PIB remains out of reach to many researchers because the short half-life (20 minutes) of the carbon-11-labeled PIB limits its use to state-of-the-art facilities that can synthesize the compound on site. Scientists developed 18F-labeled tracers with a longer half-life (110 minutes), in part to facilitate commercial availability of amyloid imaging. Three papers in the November Archives of Neurology describe advances for two such reagents.
In one study, Adam Fleisher of Banner Alzheimer’s Institute in Phoenix, Arizona, and colleagues analyze 18F-florbetapir PET data on 210 elderly people with Alzheimer’s disease (AD), mild cognitive impairment (MCI), or normal brain function, and show that this compound behaves similarly to 11C-PIB. Florbetapir, aka AV-45, was developed by Avid Radiopharmaceuticals, which Eli Lilly and Company acquired in late 2010. 18F-florbetapir also measured brain amyloid reliably in a Down's syndrome patient with AD, as reported by Marwan Sabbagh, Banner Sun Health Research Institute in Sun City, Arizona, and colleagues in a separate publication. And in a third paper, a team led by David Wolk of the University of Pennsylvania, Philadelphia, presents the first-ever correlation of Aβ histological data with PET imaging using flutemetamol, the 18F version of PIB.
As reported in the first paper, Fleisher, senior investigator Eric Reiman, and colleagues analyzed pooled data from four Phase 1 and 2 florbetapir trials that imaged 68 people with probable AD, 60 people with MCI, and 82 normal elderly. The main goal of the study—funded by Avid and initially published online in July—was to test in a large dataset whether an 18F compound could distinguish AD, MCI, and normal subgroups as well as 11C reagents such as PIB.
The researchers defined thresholds for what is considered amyloid “positive” or “negative” based on a recent Phase 3 study that correlated florbetapir imaging data with histopathology done in the near-death patients (ARF related news story on Clark et al., 2011). Fleisher and colleagues set the “positive” threshold as the minimum amyloid load measured in pathologically proven cases from the Phase 3 autopsy cohort. They set the “negative” threshold as the maximum level of amyloid seen in ApoE-negative members of a young cohort used to determine the rate of false positives in the autopsy study. Applying these thresholds to the pooled data, the researchers found that average florbetapir measurements differed in the expected direction for the three patient subgroups.
Though promising, the accumulating data on florbetapir may still fall short of approval by the U.S. Food and Drug Administration for now. “The FDA wants to know that a radiologist can look at a single image and consistently say whether it’s positive or negative,” Fleisher said. “The paper does not address this issue.” That might be because scientists and physicians have different goals for amyloid, suggests William Jagust of the University of California, Berkeley, in an Archives of Neurology editorial on the new florbetapir research. “As physicians, we need to define individuals as having a disease or not, whereas, as scientists, we are tracking a fundamental biological process (Aβ deposition) that occurs on a continuum,” Jagust writes. Another recent editorial also hints at this divergence (Moghbel et al., 2011).
To address the FDA’s concerns about scan readability, Avid and Lilly scientists worked closely with the agency on a training program to help radiologists read florbetapir PET scan images using a binary scale. “The training program was developed based on the scans used in the florbetapir Phase 3 study, and we believe the updated program, which was recently resubmitted for FDA review, can train readers to give a reliable interpretation,” Mark Mintun, Avid’s chief medical officer, noted in an e-mail to ARF.
Offering further support for florbetapir’s reliability, researchers led by Sabbagh and senior author Thomas Beach at Banner Sun Health Research Institute, Sun City, Arizona, used the 18F ligand to visualize amyloid in the brain of a 55-year-old man with Down's syndrome and AD. Since people with Down’s carry three copies of chromosome 21, which contains the gene for amyloid precursor protein (APP), they rack up amyloid pathology and develop early-onset AD. In the recent case study, the florbetapir PET imaging fit with histopathological data obtained when the patient died 14 days later. Overall, the location and amount of detectable amyloid also seemed consistent with what has been reported for late-onset AD.
In the third Archives of Neurology paper, first published online in July, Wolk and colleagues showed that PET imaging with another 18F amyloid tracer, flutemetamol, agrees with brain histology done on the same people. They analyzed seven patients who received brain biopsies as part of their treatment for normal pressure hydrocephalus (NPH), a condition caused by accumulation of cerebrospinal fluid in the brain ventricles. Since more than two-thirds of people with NPH have AD-like cortical damage (Hamilton et al., 2010), these patients were a convenient sample to study, Wolk said.
GE Healthcare is developing flutemetamol and funded the current study. The company has biopsy data from more than 40 NPH patients at other centers, as well as an ongoing autopsy study similar to that done for florbetapir, Wolk told ARF. The results of these analyses are “fairly consistent with what we report in the Archives of Neurology paper,” Wolk said, noting that abstracts of the pooled biopsy data have been submitted to the Human Amyloid Imaging 2012 and American Academy of Neurology (AAN) 2012 meetings.
Along with florbetapir and flutemetamol, a third 18F amyloid PET tracer is jockeying for FDA approval. It is Bayer Healthcare’s florbetaben (see ARF related conference story). Though florbetapir appears furthest along in the FDA process at present, “there’s a possibility that all three compounds may end up getting approval within the next 12 months,” Fleisher told ARF. Moreover, he said, “there’s no reason to think at this time that they wouldn’t compete equally in the market. There may be slight technical differences in how they bind and how they show up on the film, but ultimately, they’re doing the same thing—showing us how much amyloid is in the brain.”
Wolk agrees that all three 18F compounds could potentially compete in the clinical arena. However, “it remains to be seen whether they vary in regard to sensitivity and specificity, or have somewhat different properties that may be useful in different situations,” he said. Several research groups have unpublished data comparing PIB to an 18F PET tracer, but Wolk was unaware of head-to-head evaluations between the F18 compounds. FDA-approved amyloid PET tracers would help propel the field toward earlier detection, though some question the push to screen early for a disease as yet lacking a cure (see 8th Annual Symposium on Early Alzheimer’s Disease on early AD; ARF Webinar on AD biomarkers; ARF related news series on the Alzheimer’s Prevention Initiative).
New Spin on MRI—FDG-PET Alternative?
Whereas amyloid PET tracers reveal the brain’s burden of Aβ, fluorodeoxyglucose (FDG)-PET measures how much glucose the brain uses up in a given region. A relatively new MRI technique called arterial spin labeling (ASL) tracks changes in cerebral blood flow, which is tightly linked to the brain’s metabolism. Since the sagging metabolism measured by FDG-PET correlates with cognitive decline in AD, scientists have wondered whether ASL-MRI could supply similar information. Prior studies suggested so, but no one had directly compared the two methods in the same people at the same time.
John Detre, University of Pennsylvania, Philadelphia, and colleagues did just that in a recent analysis of 15 AD patients and 19 healthy controls. He and colleagues used a new imaging protocol to collect ASL-MRI and FDG-PET data in a single session. Participants received an FDG injection at the start of their MRI scan, and by the time that scan was finished, the FDG had coursed through their bodies in preparation for PET imaging right afterward.
In a study reported online October 24 in Alzheimer’s & Dementia (see also ARF conference story), first author Erik Musiek and colleagues collected ASL-MRI and FDG-PET scans and asked two nuclear medicine physicians to score them as “AD” or “normal.” Both physicians reported similar regional abnormalities, as well as comparable sensitivities and specificities. In addition, Norbert Schuff of the University of California, San Francisco, told ARF that he and UCSF colleague Howard Rosen have unpublished data showing good correlation between ASL and FDG-PET imaging data in frontotemporal dementia (FTD) patients.
Whereas Musiek’s team tested the reliability of visual reads, Yufen Chen and colleagues did a more quantitative analysis of the MRI and PET images from the same UPenn cohort. Even without physicians reviewing the scans, quantitative processing alone suggests that the two approaches provide largely overlapping information. These data appeared in Neurology online on November 16.
Together, the two papers make the case for ASL-MRI as a potential stand-in for FDG-PET in measurements of brain metabolism. Given that PET imaging costs three to four times more than MRI and is more invasive because it requires injection of radioactive FDG, Wolk thinks “it would be great, as a clinician, to be able to acquire the information obtained from an FDG-PET scan in a routine MRI.” Wolk is a coauthor on both ASL-MRI papers.
The current data are preliminary because the studies are small. However, if replicated in larger cohorts, the findings could pave the way for a cheaper, more widely available brain imaging tool, Schuff noted. Though developed almost two decades ago, ASL has been relatively obscure—in part because prior patent disputes kept ASL software out of the public domain, Schuff said. Now, under a new agreement, scientists who purchase an MRI scanner from Siemens, GE, or Phillips can purchase an ASL software package. Beyond the legal issues, though, ASL “is not a trivial technique,” Schuff said. “You need substantial post-processing of the data, and experienced raters to interpret the signal.”
Given that the current studies only involved AD patients and normal elderly, it will be important to determine whether ASL-MRI can achieve reasonable sensitivity in larger studies that include patients with milder pathology, Wolk said. The second phase of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) may help address this issue, as some ADNI2 participants will get ASL as part of their MRI scan.—Esther Landhuis.
Fleisher AS, Chen K, Liu X, Roontiva A, Thiyyagura P, Ayutyanont N, Joshi AD, Clark CM, Mintun MA, Pontecorvo MJ, Doraiswamy PM, Johnson KA, Skovronsky DM, Reiman EM. Using positron emission tomography and florbetapir f 18 to image cortical amyloid in patients with mild cognitive impairment or dementia due to Alzheimer disease. Arch Neurol. 2011 Nov;68(11):1404-11. Abstract
Wolk DA, Grachev ID, Buckley C, Kazi H, Grady MS, Trojanowski JQ, Hamilton RH, Sherwin P, McLain R, Arnold SE. Association Between In Vivo Fluorine 18-Labeled Flutemetamol Amyloid Positron Emission Tomography Imaging and In Vivo Cerebral Cortical Histopathology. Arch Neurol. Nov 2011;68(11):1398-1403. Abstract
Sabbagh MN, Fleisher A, Chen K, Rogers J, Berk C, Reiman E, Pontecorvo M, Mintun M, Skovronsky D, Jacobson SA, Sue LI, Liebsack C, Charney AS, Cole L, Belden C, Beach TG. Positron Emission Tomography and Neuropathologic Estimates of Fibrillar Amyloid-β in a Patient With Down's syndrome and Alzheimer Disease. Arch Neurol. Nov 2011;68(11):1461-1466. Abstract
Jagust WJ. Amyloid Imaging: Liberal or Conservative? Let the Data Decide. Arch Neurol. Nov 2011;68(11):1377-78.
Musiek ES, Chen Y, Korczykowski M, Saboury B, Martinez PM, Reddin JS, Alavi A, Kimberg DY, Wolk DA, Julin P, Newberg AB, Arnold SE, Detre JA. Direct comparison of fluorodeoxyglucose positron emission tomography and arterial spin labeling magnetic resonance imaging in Alzheimer’s disease. Alzheimer’s & Dementia. Nov 2011. Abstract
Chen Y, Wolk DA, Reddin JS, Korczykowski M, Martinez PM, Musiek ES, Newberg AB, Julin P, Arnold SE, Greenberg JH, Detre JA. Voxel-level comparison of arterial spin-labeled perfusion MRI and FDG-PET in Alzheimer disease. Neurology. Nov 2011. Abstract