By the time a person goes to the doctor with symptoms of Alzheimer’s disease, his or her brain is substantially damaged. So scientists are avidly hunting for a way to detect the neurological hallmarks of Alzheimer’s early on, in the hopes they can predict – and eventually prevent – the disease.

Results of a new study testing a brain-imaging technique based on positron emission tomography (PET) show that it can reliably distinguish Alzheimer’s patients from healthy people and those with mild cognitive impairment by detecting plaques and tangles, two types of protein clumps that occur in the brains of Alzheimer’s patients. It’s the first method that’s been able to directly image these neurological changes in living patients, and it could eventually be used to diagnose patients and test new drugs.

The technique is based on a compound, known as FDDNP, that can bind to both plaques and tangles. The compound is radioactively labeled, then injected into patients, where it travels to the brain. The PET scan detects an accumulation of the radioactive compound as it binds to plaques and tangles in different parts of the brain.

In the current study, presented this week at the American Academy of Neurology meeting in San Diego, CA, researchers from the University of California, Los Angeles studied 60 subjects: 20 with Alzheimer’s disease, 20 with mild cognitive impairment, a risk factor for Alzheimer’s disease, and 20 normal controls. They found that Alzheimer’s patients had the greatest buildup of FDDNP, implying the greatest number of plaques and tangles, followed by those with mild cognitive impairment.

The levels also correlate with a progression of the disease. When the scientists retested nine of the patients two years after the initial scans, they found that those whose condition had worsened – normal subjects who had developed a mild cognitive impairment or patients with mild cognitive impairment who had been diagnosed with Alzheimer’s – showed a 5 to 11 percent increase in FDDNP levels, compared with a 3 percent increase in those who remained stable. “This suggests the technology may be useful in early diagnosis and in tracking people over time,” says Gary Small, the UCLA scientist who led the research.

The technique is one of several in development for early diagnosis of Alzheimer’s. Currently, the disease is most commonly diagnosed with behavioral tests, rather than using brain imaging. More recently, doctors have begun diagnosing it with a different form of PET imaging – one that measures brain metabolism, which declines in a characteristic pattern in Alzheimer’s patients. Other methods also show promise, notably, testing spinal fluid for certain substances that build up in those in the early stage of disease.

However, the advantage of a compound such as FDDNP is that it gives a more direct measure of the neurological problems – plaques and tangles – associated with the disease, says Mony de Leon, a neuroscientist at New York University School of Medicine. A similar PET method using a different tracer compound is also under development, which measures amyloid plaques but not tangles. “This road is very promising. It’s extraordinary watching these technologies point out who will get sick,” says De Leon.

These types of diagnostics could be important in drug development. For example, new therapies being tested include antibodies designed to prevent the accumulation of plaques. But evaluating the effectiveness of such treatments means following patients over years. PET technologies to measure both plaques and tangles could more cheaply and quickly assess the efficiency of these therapies.

“It will be important for diagnosis and drug development to have both of these modalities in the armamentarium,” says Sam Gandy, chair of the Alzheimer’s Association Medical and Scientific Advisory Council and director of the Farber Institute for Neurosciences at Thomas Jefferson University in Philadelphia. “If we develop a medication specifically for amyloid, we’ll want to see just plaques. But we’ll also want to see that patients don’t go on to develop tangles.”

These techniques aren’t ready to be used yet in the doctor’s office. Researchers will need to demonstrate how sensitive these compounds are, which is an issue for any PET method using a new molecular tracer. “How well does it get into the brain and bind to the bad guys? That’s a very important component,” says de Leon.

Scientists will also need to perform large-scale clinical trials to show that the test can accurately predict who will develop the disease. According to Small, the FDDNP technology has been licensed to Siemens, a leading manufacturer of medical technology, which plans to run such clinical trials.