We wrote in Beating the Dementia Monster about the search for a reliable test that would detect the development of Alzheimer's disease well before the appearance of the first symptoms. We reported on several promising blood tests, although none of them have yet found FDA approval.
Why should we want a test for Alzheimer's disease? Most people say they don't want to know that they have it. They've been told that there's no cure, so why go through the painful process of consciously confronting a monster you can't beat? This is an understandable attitude. (Although they should read Beating the Dementia Monster.)
But there are reasons we want to know. Alzheimer's is a disease with a very long incubation stage. It's perhaps 15 years before the first symptoms become evident and another five years before the advent of dementia. But if there will be a more effective treatment and, dare we say it, a cure, it will likely be most effective if applied during that pre-clinical (or "prodromal") 15-year stage. But that's before anyone knows anything is wrong.
Also, for testing new drugs and treatments, we want certainty that the test subjects actually have Alzheimer's disease and with as few comorbidities as possible. So we want them as young as possible to minimize the number of late-appearing comorbidities because these can confound test results.
While current research on a reliable test for Alzheimer's disease has focused on looking for beta amyloid in blood samples and plaques in the retina of the eye, there's a new idea building steam.
Spectral phenotyping using Fourier Transform Infrared spectromicroscopy. What's that about?
The idea is that, during both Huntington's and Alzheimer's diseases, there are changes in cells throughout the body that can be detected by analyzing the absorption of different frequencies of infrared radiation. These changes seem to follow a set pattern in both Huntington's disease and Alzheimer's disease. So (I guess) shine a heat lamp (broad spectrum of infrared radiation) on a skin cell and see which frequencies are absorbed.
The nerds among us will understand that the absorption will depend on the specific bonds between different atoms and molecules, so that the presence of certain proteins and lipds with characteristic bonds will have their own signatures. So just apply some artificial intelligence (AI). Show a bunch of easily accessible skin cells from people known to have Alzheimer's disease to the system, and have its AI compare them to cells from subjects who do not. Then let the AI figure out how to predict Alzheimer's disease from skin with unknown status. You don't have to know what elements and molecules are absorbing the radiation, just look for consistent patterns. The claim by scientists at the Lawrence Berkeley National Laboratory is that this method predicts Alzheimer's disease very reliably.
On reading their report, I'd say it's very promising, but they're at a relatively early stage of developing their testing method. They've had good success, but with a small number of test subjects. Also, most of the research on humans is with Huntington's disease and less with Alzheimer's disease. Huntington's disease is far more a product of genetics than Alzheimer's disease, and so, in Huntington's disease, you should find a very consistent pattern of proteins in the cells that are a consequence of the Huntington's gene. Of course, the rare young onset form of Alzheimer's disease is also a product of genetics, and so I would expect their method to work well there. But I'm having trouble seeing from what I read in their reporting how this is supposed to work with the much more common old onset form of Alzheimer's disease.
And I can't get past the damage a test will do to the cost/availability of long-term care insurance. Once a pre-clinical person is known to have Alzheimer's disease, how will that affect the willingness of an insurance company to take them on -- and at what price? (Of course, our systems of funding healthcare are in flux, so people may not see this as a problem in a few years.)
No comments:
Post a Comment