Scans Rule Out Alzheimer's

by John Gever for MedPage Today:Autopsy findings in 74 elderly individuals confirmed the accuracy of brain scans for beta-amyloid plaques conducted before they died -- and also showed that the patients' clinical diagnoses were frequently wrong, a researcher said here.Among 57 patients in the cohort who had received diagnoses of Alzheimer's disease on the basis of clinical symptoms, 13 were found to have beta-amyloid plaque burdens too low to be consistent with the disease, either in post-mortem histopathology findings or in PET scans using the recently approved beta-amyloid plaque tracer florbetaben (Neuraceq) performed while the patients were still alive, according to Marwan Sabbagh, MD, director of Banner Sun Health Research Institute in Sun City, Ariz.In the cohort overall, which also included three patients with Lewy body dementia, six with other dementias, and eight nondemented individuals, autopsy findings showed that the PET scans had 97.9% sensitivity for picking up amyloid plaques (95% CI 93.8%-100%) and 88.9% specificity (95% CI 77.0%-100%), Sabbagh told attendees at a late-breaking abstract session at the American Academy of Neurology's annual meeting.He told MedPage Today that the 20% error rate in clinical Alzheimer's disease diagnosis was consistent with earlier studies.Furthermore, these new findings demonstrate that amyloid PET scans, by ruling out Alzheimer's disease diagnosis in patients who would otherwise have received it, can alter patients' management and improve outcomes, Sabbagh said.The question of whether amyloid PET scans with tracers such florbetaben -- other approved products include florbetapir (Amvid) and flutemetamol (Vizamyl)  -- inform clinical decision making was a key sticking point for the Centers for Medicare and Medicaid Services when it decided last year not to allow Medicare coverage except in research studies.CMS argued that although the scans appeared to be accurate in ruling out Alzheimer's disease, there was no specific evidence that clinical management would be changed as a result.The current study does not quite provide such evidence, Sabbagh acknowledged, but it does make it harder to imagine how the scans would not have that effect.Still lacking, he said, was autopsy-confirmed diagnoses in the 13 patients misdiagnosed clinically with Alzheimer's disease. The data also didn't indicate how the PET scan results were incorporated into the patients' treatment.But almost any other diagnosis would probably have led to a different management plan in those patients, he told MedPage Today. A finding of no amyloid plaque in a putative Alzheimer's disease patient should automatically prompt a search for other causes for the symptoms."If the patient has FTD (frontotemporal dementia), I'm not going to give cholinesterase inhibitors because they don't work," Sabbagh said.Neil Graff-Radford, MD, of the Mayo Clinic in Jacksonville, Fla., who was not involved in the study, said he understood CMS' argument, but he agreed that the PET scans can't help but inform clinical management."Refining and being very accurate with the diagnosis is absolutely essential," he said. "Clinical diagnosis, even in very experienced hands, doesn't necessary correlate with the pathological diagnosis."He pointed out that the scans are now being used to select patients for clinical trials of anti-amyloid Alzheimer's disease therapies precisely because they improve the diagnostic accuracy.Study DetailsSabbagh and colleagues were following 205 elderly patients who had agreed to undergo amyloid PET scans, recruited at 17 sites in the U.S., Europe, Japan, and Australia. Their mean age at enrollment was 77. The 74 participants included in the current study were those who died and had brain autopsy results available.Three readers evaluated the PET scans. A standardized uptake value ratio (SUVR, the usual measure of tracer binding) of 1.478 was set as the threshold for a positive result. The "majority read" -- that is, the positive or negative assessment by two or three of the readers -- was used in the sensitivity and specificity analysis.The median SUVR for those below this cutoff and therefore ruled amyloid-negative was 1.2; for those above the cutoff, the median SUVR was 1.7.Among the 27 patients with clinical diagnoses other than Alzheimer's disease, three were found to have significant amyloid plaque burdens. That did not necessarily refute the clinical diagnoses, because extensive beta-amyloid plaques can be present in cognitively normal patients or in those with vascular insufficiency or other causes of dementia symptoms. The other 24 were ruled amyloid-negative, including seven of eight without dementia.Sabbagh said the study was the largest such autopsy-confirmation study performed to date.Source: http://bit.ly/1uiqcFd