A simple blood test that can predict the onset of Alzheimer's disease before recognisable symptoms appear could be available in two years, say scientists.
The test, likely to cost £100 - £300, can show with almost 90% accuracy which individuals suffering from mild memory loss are going to develop Alzheimer's within a year.
It is expected to transform the search for treatments for the devastating brain illness that affects around 600,000 people a year in the UK.
To date, trials of drugs to halt or reverse Alzheimer's have all ended in failure. Therapies exist that can reduce its symptoms, but they only work for a short period of time and are not very effective.
Scientists believe a major reason for the lack of progress is that trial patients are being recruited too late, when their disease is already far advanced.
The new blood test, based on 10 "biomarker" proteins, will make it possible to test new treatments at an early stage of Alzheimer's progression.
It could also help families plan ahead and adjust to the likely prospect of one of their members being stricken by the disease.
One of the test's inventors Professor Simon Lovestone, from King's College London, said: "People come to me at the clinic because they want to know what's happening to them, and I currently can't tell them. I tell them you've got symptoms of memory loss, come back in a year's time.
"That's grim. It's horrible. You can only imagine what it's like for the patient."
Writing in the journal Alzheimer's & Dementia, the researchers describe how they investigated 26 proteins previously associated with Alzheimer's disease.
The scientists analysed blood samples from 476 confirmed Alzheimer's patients, 220 individuals with "mild cognitive impairment" (MCI) who experienced occasional memory loss, and 450 healthy elderly people.
In the vast majority of cases, memory lapses do not lead to Alzheimer's. But the researchers identified 10 blood proteins that appeared in 87% of MCI patients diagnosed with the disease within a year.
Prof Lovestone said: "Alzheimer's begins to affect the brain many years before patients are diagnosed with the disease. Many of our drug trials fail because by the time patients are given the drugs, the brain has already been too severely affected.
"A simple blood test could help us identify patients at a much earlier stage to take part in new trials and hopefully develop treatments which could prevent the progression of the disease.
"The next step will be to validate our findings in further sample sets, to see if we can improve accuracy and reduce the risk of misdiagnosis, and to develop a reliable test suitable to be used by doctors."
Working with the leading UK biotech company Proteome Sciences, the scientists hope to get the test on the market in two to five years.
Talks are under way with several potential partners who might be interested in producing a commercial product.
Dr Ian Pike, a director of Proteome Sciences and co-author of the research, said: "By linking the best British academic and commercial research, this landmark study in Alzheimer's disease is a major advance in the development of a simple blood test to identify the disease before clinical symptoms appear. This is the window that will offer the best chance of successful treatment."
Speaking at a news briefing in London, he revealed that he witnessed the destruction of his own mother by Alzheimer's.
"My family would have appreciated a test like this so we could have made the most of the time when she was interacting with us fully," he said. "A degree of clarity allows you to plan and prepare for the inevitable."
Dr Eric Karran, science director at the charity Alzheimer's Research UK, said a test identifying those at risk of Alzheimer's at an early stage would be of "real value" but warned that it would have to be used responsibly.
"Alzheimer's disease is now the most feared diagnosis," he said. "We have to be very careful about how we use these tests, especially in the absence of effective therapy."
In practice the blood test is not expected to be used on its own, but as a first step in a diagnostic process that could also involve brain scans and taking samples of spinal fluid.
Dr James Pickett, head of research at the Alzheimer's Society, said: "Finding a way to detect dementia before symptoms develop would revolutionise research into the condition."
But he pointed out that with an accuracy of less than 90%, one in 10 people who took the test would get a wrong result.
He added: "Accuracy would need to be improved before it could be a useful diagnostic test. Only through further research will we find answers to the biggest questions around dementia, so we will watch the progress of this study with interest."
Gordon Wilcox, Emeritus Professor of Geratology at Oxford University, said: "This is great news. It is important confirmation that it will probably be feasible to use a blood protein test to help identify people with memory problems who will go on to develop dementia. It will also assist in measuring the severity of their disease.
"A blood test is much simpler and less invasive than testing cerebrospinal fluid obtained at lumbar puncture, and less troublesome to patients than having a brain scan."