April 30, 2017
Recommended Topic Related To:


"TORONTO ? For patients with Alzheimer's disease (AD), early treatment with currently recommended medications may prolong survival and save healthcare dollars, compared with nontreatment, according to a longitudinal retrospective study.





Mechanism Of Action

Florbetapir F 18 binds to β-amyloid plaques and the F 18 isotope produces a positron signal that is detected by a PET scanner. In in vitro binding studies using postmortem human brain homogenates containing β-amyloid plaques, the dissociation constant (K d) for florbetapir was 3.7 ± 0.3 nM. The binding of florbetapir F 18 to β-amyloid aggregates was demonstrated in postmortem human brain sections using autoradiographic methods, thioflavin S and traditional silver staining correlation studies as well as monoclonal antibody β-amyloid-specific correlation studies. Florbetapir binding to tau protein and a battery of neuroreceptors was not detected in in vitro studies.


Following intravenous injection, florbetapir F 18 diffuses across the human blood-brain barrier and produces a radioactivity signal detectable throughout the brain. Subsequently, cerebral perfusion decreases the brain florbetapir F 18 content, with differential retention of the drug in areas that contain β-amyloid aggregates compared to areas that lack the aggregates. The time-activity curves for florbetapir F 18 in the brain of subjects with positive scans show continual signal increases from time zero through 30 minutes post-administration, with stable values thereafter up to at least 90 minutes post-injection. Differences in the signal intensity between portions of the brain that specifically retain florbetapir F 18 and the portions of the brain with nonspecific retention of the drug forms the image interpretation methods [see DOSAGE AND ADMINISTRATION].

Clinical studies evaluated the test-retest distribution of florbetapir F 18 within the brains of 21 subjects (11 with probable AD and 10 healthy volunteers) who underwent two injections (with PET scans), separated by a time period of 2 to 30 days. Images were shown to maintain signal distribution reproducibility when evaluated qualitatively (by a reader masked to image time points) as well as quantitatively using an automated assessment of SUV in pre-specified brain regions. A comparison of a 10-minute image acquisition time versus a 20-minute acquisition time showed no difference in the mean cortical to cerebellar SUV ratio results obtained.


Following the intravenous administration of 370 MBq (10 mCi) of florbetapir F 18 to healthy volunteers, the drug was distributed throughout the body with less than 5% of the injected F 18 radioactivity present in the blood by 20 minutes following administration, and less than 2% present by 45 minutes after administration. The residual F 18 in circulation during the 30-90 minute imaging window was principally in the form of polar F 18 metabolites. Whole body scanning following the intravenous injection showed accumulation of radioactivity in the liver within four minutes post-injection, followed by elimination of the radioactivity predominantly through the biliary/gastrointestinal tract with much lower radioactivity detected in the bladder. Essentially all radioactivity collected in the urine was present as polar metabolites of florbetapir F 18.

Clinical Studies

Amyvid was evaluated in three clinical studies that examined images from healthy adult subjects as well as subjects with a range of cognitive disorders, including some terminally ill patients who had agreed to participate in a postmortem brain donation program. All the studies were single arm studies in which subjects underwent an Amyvid injection and scan and then had images interpreted by multiple independent readers who were masked to all clinical information. Image interpretations used co-registration with CT scans when PET scans were performed on dual PET-CT scanners.

In Study One, a semi-quantitative Amyvid image interpretation method, which is not intended for clinical use, was used by three readers to interpret images from 152 terminally ill patients, of whom 35 underwent autopsy (29 included in primary analysis). The median patient age was 85 years (range 55 to 103 years) and 14 of the patients were female. Eighteen of the patients had dementia, 9 had no cognitive impairment and 2 had mild cognitive impairment (MCI). The main study outcome was a comparison of premortem Amyvid images to the findings from a postmortem brain examination (truth standard). The semi-quantitative measures consisted of a five-point whole brain Amyvid uptake image scoring outcome that was compared to a global score of the percentage of the whole brain that contained amyloid, as determined by immunohistochemical microscopy. The percentage of postmortem cortical amyloid burden ranged from 0 to 9% and correlated with the median Amyvid scores (Spearman's rho=0.78; p < 0.0001, 95% CI, 0.58 to 0.89).

Studies Two and Three used a clinically-applicable binary image interpretation method (positive/negative) to evaluate images from a range of patients who had participated in earlier studies. The studies assessed performance characteristics (sensitivity and specificity) among subjects with a postmortem amyloid neuritic plaque density truth standard. Additionally, inter-reader and intra-reader image interpretation reproducibility was assessed among all the subjects, including subjects who lacked a postmortem truth standard. Before image interpretation, all readers underwent special training: Study Two used an in-person tutoring type of training and Study Three used an electronic media-based training method. Five trained readers interpreted images independently within each study. The brain neuritic plaque density in both studies was determined using an algorithm in which microscopic measures of highest plaque density within a brain region were averaged to produce a global brain estimate of neuritic plaque density. The global neuritic plaque density was categorized in the same manner as that for a region (Table 5), where plaques were counted on slides with modified Bielschowsky silver stained tissue sections. For purposes of determining the agreement between the in-vivo Amyvid image results and the post-mortem whole brain amyloid neuritic plaque density, Amyvid results (negative/positive) were pre-specified to correspond with specific plaque density scores, based upon a modification of the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) criteria which use neuritic plaque counts as a necessary pathological feature of AD.

Table 5: Global and Regional Neuritic Plaque Densitya Correlates to Amyvid Image Results

Neuritic Plaque Counts CERAD Score Amyvid Image Result
< 1 none Negative
1-5 sparse
6-19 moderate Positive
20+ frequent
aJ of Neuropathology and Experimental Neurology 1997; 56(10):1095.

Study Two examined images only from terminally ill patients who had premortem Amyvid scans and postmortem brain examinations to determine a truth standard. Among the 59 patients, 35 of whom were also in Study One, the median age was 83 years (range 47 to 103 years), half were females and most were Caucasian (93%). Twenty-nine patients had an AD clinical diagnosis, 13 had another type of dementing disorder, 12 had no history of cognitive impairment and 5 had MCI. The time interval between the Amyvid scan and death was less than one year for 46 patients and between one and two years for 13 patients. Among the subset of patients who died within one year of Amyvid scanning (a prespecified outcome), the sensitivity using the majority interpretation of the readers was 96% (95% CI: 80% to 100%) and specificity was 100% (95% CI: 78% to 100%). W ith the entire dataset of 59 patients, the sensitivity using the majority interpretation of the readers was 92% (95% CI: 78% to 98%) and specificity was 100% (95% CI: 80% to 100%). At autopsy, the global brain neuritic plaque density category (CERAD score, as in Table 5) was: frequent n=30; moderate n=9; sparse n=5; and none n=15. Tables 6 and 7 show the Amyvid performance characteristics among all the patients. Among the subset of patients who died within one year of Amyvid scanning (n=46; 28 positive and 18 negative based on histopathology) the median (and range) of correct read results, false negatives, and false positives were 44 (37 to 45), 1 (0 to 7), and 1 (0 to 2), respectively, for In-Person Training (Study Two); and were 43 (38 to 44), 3 (0 to 7), and 1 (0 to 2), respectively, for Electronic Media Training (Study Three).

Table 6: Amyvid Scan Results by Reader Training Method among Autopsied Patients (n = 59)

Test Performance In-Person Training (Study Two) Electronic Media Training (Study Three)
Sensitivity (%) Median 92 82
Range among the 5 readers (69 – 95) (69 – 92)
Specificity (%) Median 95 95
Range among the 5 readers (90 – 100) (90 – 95)

Table 7: Amyvid Correct and Erroneous Scan Results by Reader Training Method among Autopsied Patients

Read Result In-Person Training (Study Two) Electronic Media Training (Study Three)
Reader Reader
1 2 3 4 5 6 7 8 9 10
All Scans with Autopsies (N=59a) Correct 55 56 53 56 45 49 54 46 53 51
False Negative 3 2 5 3 12 8 3 12 5 7
False Positive 1 1 1 0 2 2 2 1 1 1
a39 positive and 20 negative based on histopathology

Study Three included images from subjects who did not have a truth standard (20 healthy volunteers, 52 patients with mild cognitive impairment, 20 patients with AD) as well as all 59 of the patients who underwent an autopsy (same patients as in Study Two) and provided a truth standard. Duplicate images of 33 patients were included within the total pool of images in order to assess intra-reader image reproducibility. Among the 151 subjects, the median age was 76 years (range 47 to 103), half were females and most were Caucasian (93.4%). Performance characteristics for patients with a truth standard are shown above (Tables 6 and 7). The major reproducibility results are shown in Table 8 for various groups of subjects. Inter-reader reproducibility analyses for all images showed an overall Fleiss' kappa statistic of 0.83 (95% CI: 0.78 to 0.88); the lower bound of the 95% CI exceeded the pre-specified success criterion (95% CI lower bound > 0.58). Intra-reader reproducibility analyses showed that, between the two readings for each of the 33 patients with duplicate images, one of the five readers had complete agreement for all 33 patients, two readers had discrepant reads for a single patient, one reader had discrepant reads for two patients and another reader had discrepant reads for three patients.

Table 8: Number of Positive Amyvid Scan Results within Study Three Subject Groups and Reproducibility of Scan Results Among Readers

Subject group by cognitive and truth standard (TS, autopsy) status Positive Scans, na Kappa (95% CI) Percent of Scans with Inter-reader Agreement
3 of 5 readers agree 4 of 5 readers agree 5 of 5 readers agree
All subjects with a TS, n=59 33 0.75 (0.67, 0.83) 14 10 76
All subjects without a TS, n=92 33 0.88 (0.82, 0.94) 2 11 87
AD, n=49 (29 with TS; 20 no TS) 38 0.67 (0.58, 0.76) 10 14 76
MCI, n=57 (5 with TS; 52 no TS) 17 0.91 (0.83, 0.99) 2 7 91
Cognitively normal without TS, n=20 4 0.83 (0.69, 0.97) 5 5 90
Cognitively normal with TS, n=12 1 0.73 (0.55, 0.87) 0 8 92
Other (non-AD) dementia with TS, n=13 7 0.52 (0.35, 0.69) 23 23 54
aShown is the median number of scans interpreted as positive across the 5 readers for each subgroup of patients listed in the first column.

Last reviewed on RxList: 12/28/2016
This monograph has been modified to include the generic and brand name in many instances.

Report Problems to the Food and Drug Administration


You are encouraged to report negative side effects of prescription drugs to the FDA. Visit the FDA MedWatch website or call 1-800-FDA-1088.

WebMD Daily

Get breaking medical news.