The way doctors diagnose Alzheimer's disease may be starting to change. Traditionally clinicians have relied on tests of memory and reasoning skills and reports of social withdrawal to identify patients with Alzheimer's. Such assessments can, in expert hands, be fairly conclusive—but they are not infallible. Around one in five people who are told they have the neurodegenerative disorder actually have other forms of dementia or, sometimes, another problem altogether, such as depression. To know for certain that someone has Alzheimer's, doctors must remove small pieces of the brain, examine the cells under a microscope and count the number of protein clumps called amyloid plaques. An unusually high number of plaques is a key indicator of Alzheimer's. Because such a procedure risks further impairing a patient's mental abilities, it is almost always performed posthumously.

In the past 10 years, however, scientists have developed sophisticated brain scans that can estimate the amount of plaque in the brain while people are still alive. In the laboratory, these scans have been very useful in studying the earliest stages of Alzheimer's, before overt symptoms appear. The results are reliable enough that last year the Food and Drug Administration approved one such test called Amyvid to help evaluate patients with memory deficits or other cognitive difficulties.

Despite the FDA's approval, lingering doubts about the exact role of amyloid in Alzheimer's and ambivalence about the practical value of information provided by the scan have fueled debate about when to order an Amyvid test. Not everyone who has an excessive amount of amyloid plaque develops Alzheimer's, and at the moment, there is generally no way to predict whom the unlucky ones will be. Recent studies have shown that roughly one third of older citizens in good mental health have moderate to high levels of plaque, with no noticeable ill effects. And raising the specter of the disorder in the absence of symptoms may upset more people than it helps because no effective treatments exist—at least not yet.

A new clinical trial will investigate whether giving an experimental drug as soon as the scans detect the formation of plaques can slow or halt the development of Alzheimer's. Even if the results are encouraging, a new drug will take many years to reach the market, and doctors must still determine how to use the scan responsibly in the meantime. “The whole field is grappling with this,” says neurologist Reisa Sperling of Brigham and Women's Hospital. She thinks the scan could bring clarity to challenging diagnoses, but “as a predictive test, it's not ready.”

Telltale Trouble
Scientists first linked amyloid plaques to what is now called Alzheimer's disease more than 100 years ago. In 1906 German psychiatrist Alois Alzheimer documented unusual protein knots in the brain of a deceased dementia patient. By the mid-1980s scientists determined that such plaques are made up of a protein they named beta-amyloid.

Healthy neurons produce plenty of beta-amyloid, but its precise purpose remains a mystery. In the initial stages of Alzheimer's and other neurodegenerative disorders, the proteins begin to behave strangely, sticking together to form larger and larger clumps. Scientists are still unsure whether the resulting plaques are primarily responsible for the devastating loss of millions of neurons that characterizes Alzheimer's or whether they are, instead, a by-product of some other, as yet undetermined, cause.

Nevertheless, the clumps form long before any explicit signs of dementia, and numerous plaques remain one of the best indicators of the disorder. Neurologists John C. Morris, Randall Bateman and their colleagues at Washington University in St. Louis have been tracking the health of people with a rare genetic mutation that guarantees Alzheimer's will strike them at a young age. They have detected amyloid plaques in the brain 15 years before cognitive problems typically appear in such individuals.

Attempts to design a scan that could spot amyloid first began about a dozen years ago at the University of Pittsburgh. Researchers injected patients with a small, benign amount of a radioactive dye they named Pittsburgh compound B, or PIB. The dye traveled through the blood to the brain and clung exclusively to clusters of amyloid protein. Scanning the brain with a positron-emission tomography (PET) machine then produced images that highlighted any plaques by detecting radiation in the form of gamma rays emanating from the dye.

Scientists at Philadelphia-based Avid Radiopharmaceuticals built on the Pittsburgh approach by developing Amyvid, a longer-lasting dye that gave clinicians more time to scan their patients. Early results were so promising that Eli Lilly bought the company in 2010 for $300 million. Two years later the FDA approved the Amyvid test, which more than 450 imaging centers in the U.S. now offer, usually for $3,000 or more.

To Scan or Not to Scan
Together the FDA and the neurological community have approached Amyvid with a mix of enthusiasm and trepidation. Officially the test has been approved primarily to exclude Alzheimer's as a diagnosis for someone who already has cognitive impairment, which is particularly helpful when the causes are unclear. This past January an expert task force convened by the Alzheimer's Association and the Society of Nuclear Medicine and Molecular Imaging published guidelines that advised limiting the test's use to patients with unexplained, persisting mild cognitive impairment (MCI) and to those who either have developed dementia unusually early or have dementia with atypical symptoms, such as hallucinations or delirium. Determining whether Alzheimer's is the culprit in such cases has always been difficult, and Amyvid is exactly the kind of tool that can provide some assistance. If a scan finds few plaques, physicians can more confidently rule out Alzheimer's and explore other explanations, such as a rare and rapid degeneration of the brain's frontal lobes.

The task force also advised against amyloid scans for people with no cognitive impairment. Panelists worried not only that amyloid-positive results might send some vulnerable individuals spiraling into depression or perhaps even suicide but also that the information may make it harder for them to get long-term care insurance or renew a driver's license, says task force member Jason Karlawish, a professor of medicine and medical ethics at the University of Pennsylvania.

Despite these cautious guidelines, which Eli Lilly supports, many people in the medical community have reservations. Task force member and neurologist Norman Foster of the University of Utah worries that some doctors and patients will rely on amyloid scans as a diagnostic shortcut in lieu of a more thorough evaluation that includes a battery of memory and reasoning tests. And given the lack of effective treatment, a scan suggesting that someone with mild cognitive impairment has early Alzheimer's may yield more anxiety than practical guidance. “What do you do with that information?” asks neuroscientist William Jagust of the University of California, Berkeley.

Some experts think that if scans help patients and their families prepare for the changes in health and lifestyle that come with Alzheimer's, then they are beneficial. Others argue that the test is not meaningful if it does not alter how a patient is treated. The federal agency that oversees Medicare insurance essentially took the latter view in September, when, because of insufficient evidence that the scan improves patient health, it declined to cover Amyvid for now except in certain clinical trials. Though disappointed, Avid's CEO Daniel Skovronsky acknowledges that figuring out how best to use the scan will take considerable time and debate because it is such a new tool. “I think the discussion is actually a good thing,” he says.

The clinical trial slated to begin this month could help quell the controversy if its results are auspicious. In that study, Sperling, Karlawish and investigators at 60 U.S. medical centers aim to screen 3,000 healthy senior citizens to identify 1,000 amyloid-positive individuals who will receive either a drug therapy for Alzheimer's called solanezumab or a placebo for three years.

Before anybody slides into a PET scanner, however, participants will be prescreened for mood, depression and anxiety to ensure they are capable of “handling uncertainty and, potentially, what could be construed as bad news if they learn that they are amyloid-positive on imaging,” Karlawish says. “There will be some people who are not, frankly, allowed to go forward.”

With any luck, the trial's results, which are not expected until 2018, will confirm that solanezumab could become a viable treatment in the future, as well as help doctors decide whether it makes sense to get an early look at Alzheimer's today.