PET scans using radio-labeled flutemetamol are safe and confirm the presence or absence of neuritic β-amyloid levels “with high sensitivity and specificity,” according to research reported in the March issue of JAMA Neurology. “In vivo detection of brain β-amyloid may help increase diagnostic accuracy in cognitively impaired patients, compared with clinical diagnosis alone,” said Craig Curtis, MD, Chief Medical Officer at Compass Research in Orlando, and his associates.
The investigators conducted a phase III trial involving terminally ill older adults. The PET scanning agent used in the study, [18F]flutemetamol, binds with high affinity to β-amyloid but rapidly clears from normal brain tissue. The trial included 203 consecutive patients attending 19 dementia clinics, memory centers, and hospices in the United States and England who were age 55 or older and who had a life expectancy of one year or less. A total of 180 participants who underwent [18F]flutemetamol PET scanning made up the population for the safety analysis. A subset of 68 patients who died during the study underwent autopsy so that brain pathology could be compared with the PET images. These patients were the basis for the sensitivity and specificity analysis. Five specially trained readers who were blinded to patient information independently interpreted the PET scans as positive or negative for neuritic β-amyloid.
On postmortem examination, 25 of the 68 brains (37%) were β-amyloid negative, and 43 (63%) were β-amyloid positive. The sensitivity of [18F]flutemetamol PET scanning ranged from 81% to 93% (median, 88%) across the five scan readers, and the specificity ranged from 44% to 92% (median, 88%). One of the readers frequently overestimated the level of β-amyloid in the scans and was considered an outlier regarding specificity. Between-reader agreement was 90% or higher when that reader’s interpretations were excluded from the analysis.
Six (14%) of 43 brains found to be β-amyloid positive at autopsy were read as false negatives, and two (8%) of 25 brains found to be β-amyloid negative at autopsy were read as false positives.
The IV [18F]flutemetamol injections were well tolerated, and no adverse events were attributed to the agent or the PET procedure. However, four participants were withdrawn from the study because they could not remain motionless in the scanner.
“A scan that accurately detects β-amyloid is not the same thing as a test for Alzheimer’s disease,” said Gil D. Rabinovici, MD, Associate Professor of Neurology at the University of California, San Francisco’s Memory and Aging Center, in an accompanying editorial. “Approximately 25% of cognitively normal older adults and nearly 50% of people older than 80 years also show β-amyloid on PET scans, as do patients with conditions such as dementia with Lewy bodies and cerebral amyloid angiopathy.
“It is concerning that one of the five rigorously trained PET scan readers frequently overestimated the amount of β-amyloid and showed worryingly low specificity (44%) in interpreting these scans,” he continued. “PET amyloid imaging should be considered only for patients in whom there is considerable diagnostic uncertainty after a comprehensive clinical evaluation and in whom knowledge of β-amyloid status is likely to change diagnosis and management. This would include patients with unexplained mild cognitive impairment, with atypical or mixed presentations, and those who present with dementia at an unusually early age.”
—Mary Ann Moon