Laino, CharleneBack to top
The molecular imaging agents florbetaben and flutemetamol are able to detect amyloid beta deposits in the brains of living patients with a high degree of sensitivity and specificity, according to separate emerging science presentations at the AAN annual meeting in New Orleans.
CASE POSITIVE for florbetaben gray matter uptake and amyloid-beta histopathology: (A) Brain slice with region of interest (RO1) 5 (posterior cingulate cortex); (B) co-registered florbetaben PET/T1-weighted MRI scans (coronal view) with ROI depicted as a polygon, where there is massive uptake in gray matter; (C) histological specimen (Bielschowsky silver stain), showing multiple amyloid-beta plaques seen as gray/black deposits; (D) histological specimen (amyloid-beta immunochemistry), showing amyloid-beta pathology denoted as multiple pink deposits.
The experimental agents — both fluorine-18 radiotracers used in conjunction with PET imaging — are being developed as biomarkers to aid in the early diagnosis of Alzheimer disease. A third fluorine-18 agent, florbetapir, was approved by the FDA on April 10 for imaging of amyloid beta (Abeta) plaques in patients with cognitive impairment who are being evaluated for Alzheimer disease. (See “FDA Approves Florbetapir for Imaging Amyloid.”)
In one new study, there was a high degree of agreement between data obtained from florbetaben PET scan images of the brains of end-of-life patients and postmortem Abeta findings, reported principal investigator Marwan N. Sabbagh, MD, director of Banner Sun Health Research Institute in Sun City, AZ, and research professor of neurology at the University of Arizona College of Medicine in Phoenix.
In two other studies, imaging data obtained during flutemetamol PET scans correlated with both Abeta pathology in brain autopsies from end-of-life subjects and with in vivo cortical brain biopsies from patients with suspected normal pressure hydrocephalus (NPH). Those findings were presented by David A. Wolk, MD, assistant professor in the department of neurology at the University of Pennsylvania and assistant director of the Penn Memory Center in Philadelphia.
“All these agents push forward the idea of amyloid imaging, of in vivo diagnosis of Alzheimer disease. Used with symptomology, amyloid imaging improves our confidence of making a diagnosis when a patient is still alive,” Dr. Sabbagh said.
In the patient with progressive dementia whose scan is positive, “there is a high probability of Alzheimer or that Alzheimer is contributing to the patient's dementia,” he said. A negative scan can be used to rule out Alzheimer disease, he said.
The global phase 3 florbetaben study included 204 end-of-life patients recruited from 17 centers across four continents who underwent MRI and florbetaben PET scanning. Ten healthy individuals were also included to support specificity.
At the meeting, Dr. Sabbagh presented findings on 31 patients who died. The florbetaben scans were performed in the six months prior to death, he said. The autopsy was photodocumented in detail to allow for alignment with the MRI/PET images. Abeta in the tissue was detected by silver-staining and immunohistochemistry. The presence or absence of Abeta was determined by a consensus panel of three neuropathologists blinded to all clinical and imaging data.
The researchers analyzed the data in two ways. The first was a visual assessment procedure proposed for routine clinical practice in which whole brain scans were rated positive or negative for amyloid and compared with autopsy findings that were also rated positive or negative for amyloid.
That method demonstrated 100 percent sensitivity and 92 percent specificity with excellent inter-reader agreement (kappa = 0.88), said co-investigator Cornelia Reininger, MD, PhD, associate professor of surgery and senior director of clinical research at Bayer Pharma AG Global Clinical Development in Berlin.
Then the researchers performed a regional analysis, directly comparing six different prespecified regions of the brain.
The goal was to determine “whether florbetaben gray matter uptake in the scan actually correlated with the presence or absence of Abeta in the respective histopathology,” she said, and it did. Based on the regional comparison, florbetaben detected Abeta with a sensitivity of 77 percent and specificity of 94 percent. The inter-reader agreement was 0.66.
This trial was supported by Bayer HealthCare Berlin (Germany), which plans to submit the results for FDA approval, Dr. Reininger said. Piramal Healthcare Ltd. recently signed an agreement to acquire worldwide rights to the molecular imaging research and development portfolio of Bayer Pharma AG.
In one of the phase 3 flutemetamol studies, flutemetamol F18 Injection was administered to 180 end-of-life subjects, 68 of whom were followed to brain autopsy. In the other phase 3 study, flutemetamol F18 Injection was administered to 49 patients with suspected NPH who underwent in vivo cortical brain biopsies.
Abeta tissue was evaluated by silver-staining and immunohistochemistry. [18F]flutemetamol images were read as normal or abnormal for amyloid load by trained readers blinded to subject information: five readers in the autopsy study and three readers in the biopsy study.
Based on “majority rules criteria in which amyloid was considered normal or abnormal if two of three readers agreed,” comparison of the scan results to biopsy findings demonstrated 100 percent sensitivity and 96 percent specificity, Dr. Wolk said.
“Individual readers did vary, with 88 percent to 100 percent sensitivity and 93 percent to 96 percent specificity,” he said.
The researchers went a step further to look at quantitative results and tried to “correlate how hot the signal was on PET to plaque density,” Dr. Wolk said. There was a very significant correlation between uptake on the scan and plaque load, with p values ranging from 0.0001 to 0.005 depending on the method used, he said.
In the other study, comparison of [18F]flutemetamol scan results to autopsy findings demonstrated a sensitivity of 86 percent and a specificity of 92 percent.
Blinded visual assessment of all images showed a high level of agreement between all readers (κ=0.86 for biopsy and similar for autopsy). Within-reader agreement for image interpretation was also very good, he said.
The study was funded by GE Healthcare.
Rachelle S. Doody, MD, PhD, director of the Alzheimer's Disease and Memory Disorders Center at Baylor College of Medicine in Houston, said that overall, all the F18 tracers seem to do “an accurate job” at detecting Abeta.
The scans will be most useful for patients “with symptoms that we think probably have Alzheimer, but are not sure. If patients have Alzheimer by clinical criteria, we might be able to skip the test,” she said.
Indeed, a clinician skilled and experienced in working with people with Alzheimer disease dementia can make an accurate diagnosis roughly 90 percent of the time, according to William Thies, PhD, chief medical and scientific officer of the Alzheimer's Association.
Ronald C. Petersen, MD, PhD, professor of neurology, Cora Kanow professor of Alzheimer's disease research, and director of the Mayo Alzheimer's Disease Research Center at Mayo Clinic College of Medicine in Rochester, MN, said: “While these tracers are very good, they are not perfect.”
A concern is that they may not pick up the earliest manifestations of disease when plaque is just forming, he said. That is best illustrated by the 77 percent sensitivity in the regional analysis of florbetaben, he said.
“That suggests it is not picking up one-fourth or so [of amyloid],” Dr. Petersen said. “And it's a question I have about all the F18 tracers, not just florbetaben,” he said, noting that the FDA approval letter on florbetapir states that it detects moderate to severe plaques.
Dr. Doody agreed. “In a person with patchy amyloid, you may not see it in a particular region of the brain, but it doesn't mean they don't have Alzheimer's,” she said.
Dr. Petersen said he is also concerned that clinicians will tell patients with symptoms and positive scans that they have Alzheimer disease.
“That's a risk because we don't know that. What we do know is whether they have amyloid or not,” he said. People can have amyloid and not have Alzheimer disease, he added.
“But the real problem would be use in a person with no symptoms,” said Dr. Petersen, adding that he expects patients with family histories to ask for the test. “You can have a positive scan, have amyloid, and be clinically normal late in life. There can be a subset of patients that tolerate amyloid,” Dr. Petersen said.
The F18 tracers may be most useful as a research tool, he continued. “If you are doing a clinical trial with an agent that targets amyloid, it would be useful to know that the patient has the target,” he explained.
Before-and-after scans to determine if plaque is being removed from the brain may also be helpful, Dr. Petersen said.
Dr. Doody was unsure what results of the biopsy part of the flutemetamol study will mean clinically. “They only biopsy an area the size of a pin in patients with NPH,” she said. “I'm not sure it is an accurate reflection of what is going on in the brain,” he said.
Still, she praised efforts to diagnose AD earlier and said she hoped publicity about the findings “will motivate doctors to refer more patients for early diagnosis.”
While reader agreement was fairly high in the studies, another question is whether clinicians in the field can read scan results accurately and consistently. Eli Lilly has established online and in-person programs to train doctors to interpret the results for florbetapir and the manufacturers of the other agents, if approved, will presumably do the same.
As for whether one agent is better than another, all the experts noted that comparisons cannot really be made across clinical trials as trial methodologies are different and no head-to-head trials have been conducted.
But all the agents have an advantage over the carbon 11–labeled Pittsburgh compound B (PiB), another imaging agent that has been shown to detect beta-amyloid plaques, Dr. Wolk noted. That's because PiB has a half-life of only 20 minutes, limiting its use to centers that can produce it on site.
The half-life of the F18 tracers is 90 minutes. And florbetaben has a shelf life of 10 hours before it loses its radioactivity, Dr. Reininger said.
“At this point all the agents seems to do a fairly good job at detecting pathology in the brain. There might be subtle differences but until we do a head-to-head comparison we won't know. My suspicion is that they are similar,” Dr. Wolk said.
Are the new radiotracing imaging agents ready for prime time — the detection of Alzheimer disease? Watch this video as Neurology Today Editor-in-chief Steven P. Ringel, MD, and Neurology Today associate editor Robert Holloway, MD, respond to the question: http://bit.ly/aNQ4KB