Subscribe to eTOC

New Data Support Utility of Amyloid-PET for Dementia Diagnosis

Article In Brief

Researchers found that amyloid PET scans of Medicare beneficiaries suspected to have symptoms of mild cognitive impairment or dementia often changed the way clinicians diagnosed and managed their patients—whether the scans were amyloid positive or not.

Biomarkers to measure brain levels of amyloid-beta (Abeta) have proven their merit in helping to detect Alzheimer's disease (AD), but the Centers for Medicare & Medicaid Services does not reimburse for the technology outside of a research study. Now, a team of scientists have responded to the federal agency's challenge to prove that the scans can help in accurately diagnosing and managing patients in clinical settings.

The scientists, led by Gil D. Rabinovici, MD, the Edward Fein and Pearl Landrith distinguished professor of neurology at the University of California, San Francisco, and Maria C. Carrillo, PhD, chief science officer of the Alzheimer's Association, launched the Imaging Dementia—Evidence for Amyloid Scanning (IDEAS) study two years ago and have collected amyloid PET scanning, diagnosis, and management data on 16,000 patients seen by 964 dementia specialists in almost 600 medical centers and 350 imaging sites. The scientists were able to show that the results of the amyloid PET made an important difference in offering patients a more rigorous diagnosis and a plan for managing their illness. Findings from the study were published April 2 in the Journal of the American Association (JAMA).

“It was quite a large effect,” said Dr. Rabinovici. “In the face of diagnostic uncertainty, doctors often take a wait-and-see approach. With greater diagnostic certainty, physicians can recommend a treatment plan and safety measures that can proactively prevent a poor outcome.”

The US Food and Drug Administration (FDA) has approved three different amyloid PET imaging markers in the last decade. But in 2013, CMS published its decision not to reimburse for scans outside of a research protocol because the agency believed there wasn't sufficient evidence that it offered an advantage over a clinical exam and neuropsychological testing. The agency wanted to see evidence to show it would make a difference in the short-term management and long-term medical outcomes. The Alzheimer's Association convened a number of scientists, government officials, pharmaceutical companies, and clinicians to address the issues and design a large national study. It took two years to design the protocol.

The IDEAS study was born out of the CMS charge. The JAMA study represents the first half of the study, to address the question: Does amyloid PET scan imaging change the short-term management of patients? The second question, that will be addressed in a later paper, is whether the amyloid PET finding and the subsequent change in management led to fewer hospitalizations and emergency department visits a year later.

Study Methods, Findings

The IDEAS study investigators were asked to enroll only Medicare patients with symptoms of cognitive impairment that they believed were part of mild cognitive impairment (MCI) or dementia, but they were not sure if it was AD or another condition.

It took less than two years to enroll 16,008 patients. The last patient was scanned in September 2017 and all patients were followed through January 2018. The dementia specialists were asked to document their diagnosis and their management plan before the scan and then again 90-days after the scan.

The primary outcome measure was a change in management between the pre- and post-scan. Was there a change in Alzheimer's drug therapy or other drug therapy? Was there a change in counseling? The large number of patients enrolled in the study meant that it was powered to detect a 30 percent or greater change in the patient groups, both those with MCI and those with dementia.

A secondary outcome measure was a change in diagnosis. How many diagnoses were changed from Alzheimer's to non-AD dementia? And how many went from a non-AD diagnosis to AD?

All patients enrolled in the study had a comprehensive workup, including a Mini-Mental State Examination and the Montreal Cognitive Assessment, laboratory tests that had been collected within the past year, and a CT scan or MRI that had been done in the past 24 months. They all had to meet the requirements for an amyloid PET scan, that is, that their doctor was still uncertain of an accurate diagnosis, a diagnosis of AD was a strong but not certain possibility, and finally that the neurologist thought that the findings from the scan could alter the diagnosis and management.

The dementia specialists had to complete a comprehensive pre-PET assessment that included the patient's demographics, primary suspected diagnosis (they had to guess on a scale of one to 10 how confident they were that it was AD), and how they were going to manage the care (drugs and counseling) if they did not have an amyloid PET scan. Also, they were asked about any referrals they would make or other diagnostic tests they would order.

Then, the patients underwent the amyloid PET scan. There were 350 brain scanning centers involved in the study and they could use any of the three FDA-approved amyloid-beta ligands. The results of the scan—positive or negative–was recorded, and the results were shared with patients by their treating doctor; immediate management changes were implemented and recorded for study analysis.


“In the face of diagnostic uncertainty, doctors often take a wait-and-see approach. With greater diagnostic certainty, physicians can recommend a treatment plan and safety measures that can proactively prevent a poor outcome.”


At the end of the study, 71 percent of patients completed the study protocol, which meant that the researchers had data on 11,409 patients. The mean age of the group was 75 years old. There was an equal number of men and women. Sixty percent had a pre-scan diagnosis of MCI. The amyloid PET results were positive in 3,817 patients with MCI (55.3 percent) and 3,154 patients with dementia (70.1 percent). The management plan changed in 4,159 of the 6,905 patients with MCI (60.2 percent) and 2,859 of 4,504 patients with dementia, representing a 63 percent change.

There was a 25 percent change from Alzheimer's to a non-AD diagnosis, and a 10.5 percent change from non-AD to AD.

Three months after the PET scans, patients had a follow-up visit with the referring dementia specialist for another assessment. Here again, the management changes were documented as well as any changes in the diagnosis. They were asked to fill out the diagnosis confidence scale now that they had findings from the amyloid PET scan. They were also asked to explain the management changes they ordered.

The researchers also collected information on anyone who died during the study period, even if they never made it to amyloid PET imaging. They wanted to see whether there were any suicides that might have occurred after patients were told of the findings. There were not.

Post-scan changes included starting, stopping or modifying drugs used in the treatment of dementia disorders.

Doctors involved in the study said that about 85 percent of the changes they made were a result of the new information from the scan.

Interestingly, the proportion of AD diagnoses went from 80.3 percent pre-scan to 95.5 percent post-scan in patients with a positive scan and from 71.5 percent to 10.2 percent in those patients whose scans came back negative.

The physician's confidence in their diagnoses also improved once they had the scan results. Pre-scan, 72.4 percent were in the uncertain range (between 4 and 7 on the scale) and this dropped to 16.2 on the post-scan 90 day visit.

Not surprisingly, prescription changes for AD drugs increased in those who tested positive on the scan—from 40.4 percent to 81.5 percent in patients with MCI and from 63.2 percent to 91.2 percent in those with dementia. For those with a negative scan, there were reductions in the use of AD drugs, as well. There is evidence from other studies that certain dementias that are not associated with amyloid-beta, such as frontotemporal dementia, may do worse on some of the AD medications.


“There are other biomarkers that measure amyloid (and tau) that may be less expensive. We need to be sure that this is the best way to structure a diagnostic decision and a management plan.”


Patients were also counseled accordingly. There was a change in 20 percent of the counseling information in patients with dementia and a 25 percent change in counseling for MCI patients.

The next phase of the study will try to answer the question of whether amyloid PET is associated with long-term improvements in clinical outcomes. The researchers will use Medicare claim records to assess hospitalization rates and emergency department visits one year after the amyloid PET scan. They are analyzing additional health outcomes, as well, and will compare the results with information from age- and disease-matched Medicare recipients who never had an amyloid PET scan. Dr. Rabinovici said that the available treatments for AD may not impact these long-term health outcomes, but they are hoping to see some effect.

Dr. Rabinovici said it is important for patients and their families to have an accurate diagnosis. “They want validation that their symptoms are real,” he said. “Also, having a diagnosis at the MCI stage is important for treatment decisions.”

The findings also support the use of amyloid PET as a biomarker for recruiting AD patients into clinical studies. Without information from an amyloid PET scan, this study and others show that the findings from the scan allowed them to rule out a diagnosis of AD in a third of patients who were given a probable AD diagnosis and treated for AD. “Using a scan means the right people can be enrolled into trials,” Dr. Rabinovici said.

There are several caveats to the study findings, said Dr. Rabinovici. Amyloid-beta deposition occurs in other neurodegenerative conditions and also in cognitively fit older people. While strong in numbers and in real-world clinical care, this study is non-randomized and lacking a control group. Also, minorities were scarce in the recruitment for this study and the researchers are now planning a second study to bring in more minorities for scanning.

The trial was funded by the Alzheimer's Association, the American College of Radiology, Avid Radiopharmaceuticals Inc., a wholly owned subsidiary of Eli Lilly and Company, General Electric Healthcare, and Life Molecular Imaging. CMS provided coverage for amyloid PET scans in a program the agency set up to conduct research to develop evidence to support payment.

Expert Commentary

“The study findings are quite impressive,” said Zaven Khachaturian, PhD, senior science advisor to the Alzheimer's Association and editor-in-chief of Alzheimer's & Dementia: Journal of the Alzheimer's Association. “It can serve as a starting point for larger studies to validate biomarkers. The amyloid PET scan is an important tool that can add to the clinical assessment in the earliest stages of Alzheimer's. Clinical assessment is highly variable and often not accurate. Adding biomarkers can enhance a doctor's decision about what is causing their symptoms.” “The diagnosis of Alzheimer's disease requires the presence of amyloid and tau,” said Stephen Salloway, MD, director of neurology and the memory and aging program at Butler Hospital and the Martin M. Zucker professor of psychiatry and human behavior and professor of neurology at the Alpert Medical School of Brown University. “If a clinician thinks someone has MCI or dementia and the amyloid scan is negative the chance that they have Alzheimer's is low. It increases a clinician's confidence and an accurate diagnosis provides the foundation for excellent medical care.”

“The data are compelling and this should encourage CMS to revisit its decision and consider coverage for amyloid PET scans,” added Dr. Salloway, who helped in the first stages of the study design and enrolled over 90 patients in the study. “It's frustrating that you have a test that can lead you closer to an answer and you can't order it. These types of tests modernize the evaluation of patients with dementia. It reduces the guesswork.”

He added that these tests are even more important because “we can see the pathology 15 years before the clinical signs and it is possible that treatments will work better in the earliest stages.”

“I am not surprised by the good outcome,” said Oskar Hansson, MD, professor of neurology at Lund University. He and his colleagues in Sweden have been ordering cerebrospinal fluid (CSF) amyloid biomarkers in clinical practice for a decade. “Our clinical experience is that amyloid biomarkers lead to a more accurate diagnosis and better patient management. However, this impressive study now proves how important this really is when a doctor is uncertain about a clinical diagnosis.” In Sweden, they mostly order CSF amyloid assays. He said they do amyloid PET testing also but it is more expensive than ordering a lumbar puncture to look for amyloid pathology.

“The scale and scope of this study is enormous,” added Sudha Seshadri, MD, FAAN, founding director of the Glenn Biggs Institute for Alzheimer's & Neurodegenerative Diseases at UT Health San Antonio. “People are hungry for certainty when faced with a diagnosis of Alzheimer's. The breadth of this study shows scientists can do pragmatic studies quickly.” She said that it was a large “but convenient sample because clinicians wanted the test for their patients,” but it is not representative of the whole population of dementia patients. Only four percent of the sample were African-American or Hispanic, she noted.

She also said that the few drugs approved for AD were tested in patients well before the availability of amyloid-PET scans and “only time will tell whether this increased use of anti-cholinesterase inhibitors in amyloid-beta positive MCI and dementia patients will be good, bad, or indifferent on patient outcome measures. Also, many people who were amyloid-beta negative were taken off their medications. We don't know if that is good or bad, either.”

The price of these scans is between $3,000 and $5,000. “This test is very expensive,” Dr. Seshadri said. “There are other biomarkers that measure amyloid (and tau) that may be less expensive. We need to be sure that this is the best way to structure a diagnostic decision and a management plan.”

“If this was a relatively cheap test would we even be discussing this?” asked Neill R. Graff-Radford, MD, FAAN, the David Eisenberg professor at Mayo Clinic in Jacksonville, FL. “Patients and families deserve an accurate diagnosis. This is a good tool to help with that. Why wouldn't we want to give patients an accurate diagnosis. It informs what they do to plan for their future.”

Finally, Norman L. Foster, MD, FAAN, director of the Center for Alzheimer's care, imaging and research at the University of Utah, said: “I don't think anyone can argue that this is an accurate method of determining Alzheimer's pathology. No imaging scan will ever diagnose an illness. You have to know the clinical context.

“To think it is not a problem that we would otherwise misdiagnosis a large group of patients reflects a tremendous degree of nihilism on the part of the payers and the health care system.”


Dr. Rabinovici reported receiving grants from the American College of Radiology, the Alzheimer's Association, Avid Radiopharmaceuticals Inc., GE Healthcare, Life Molecular Imaging, and Eli Lilly. He has received personal fees from Eisai, Piramal Imaging, Merck, and Genentech. Dr. Petersen reported receiving personal fees from Hoffman-La Roche, Merck, Genentech, Biogen, GE Healthcare, and Eisai.

Link Up for More Information

•. Rabinovici GD, Gatsonis C, Apgar C, et al. Association of amyloid positron emission tomography with subsequent change in clinical management among Medicare beneficiaries with mild cognitive impairment or dementia JAMA 2019; 321:13 1286–1294.