Alzheimer Biomarkers
So what are biomarkers and what’s the state of play in Alzheimer’s disease?
Biomarkers are essentially any biological feature that can be measured reliably and corresponds to an underlying disease state.
A simple example is that a cough is generally regarded as a clinical marker of an underlying respiratory illness. Experienced doctors (and some parents) can also identify when a certain kind of cough signals a specific respiratory disease, for example the ‘brassy’ cough of croup. By extension, if some clever researchers found a vocal signature for the COVID19 cough (if it existed), and then created a phone app to allow doctors and the public to accurately and easily detect the illness, then this would be a good example of a useful and powerful biomarker.
A second criteria for biomarkers that has emerged through convention is that they generally should represent a more convenient, efficient, inexpensive or accessible means of predicting the underlying illness than the gold standard. Alzheimer’s is a good example because the gold standard is post mortem brain histology - not so convenient or practical if you’re someone worried about their memory.
The first wave of Alzheimer biomarkers were therefore based on brain imaging. Using MRI, several different biomarkers were identified that accurately predicted Alzheimer pathology in the brain. These were generally structural, often measurements of the volume or size of different memory-related structures such as the hippocampus which shrinks greatly in Alzheimers. More accurately, these were therefore biomarkers of neurodegeneration (neuronal loss), of which Alzheimer’s is just one of a number of diseases that can account for it.
The next wave used PET-imaging based on the advent of tracing compounds that could “light up” amyloid plaques using a radioactive tracer (that quickly degrades and is safe under careful supervision). These have undergone extensive research and there is now consensus on what amount of PET-amyloid burden in the brain constitutes normal ageing, pre-clinical or definitive Alzheimer’s disease. The mathematical and statistical techniques for measuring amyloid PET have been agreed upon, as has how to resolve site-to-site and machine-to-machine differences, a vital step in biomarkers so results can be comparable around the world. Because of the maturity of this technology, it is now the default ‘gold standard’ biomarker for Alzheimer’s.
Tau PET-imaging ligands have also been produced in the last few years and the field is following the same path as for amyloid PET biomarkers, albeit a few years behind.
The problem with PET Alzheimer biomarkers is access and cost. PET markers need to be produced in highly specialised radiotracer facilities that are generally only located in major cities in advanced rich countries, such that many big cities will not be able to access this technology, let alone in developing countries. It also needs a lot of specialist technical expertise to carry out the scan and analyze the results, and the costs to researchers or patients is very high.
PET biomarkers of Alzheimers disease are therefore not going to go prime time and are confined to research use.
On the other hand, there are two other emerging and exciting types of biomarkers for Alzheimer’s disease.
Blood-based biomarkers aim to measure minute levels of Alzheimer-related proteins and protein fragments in the blood. Typically these fragments have been released or sequestered in the interstitial fluid in the brain (fluid in between brain cells), which then enter the cerebrospinal fluid (CSF) and transfer across into the blood.
This field has taken off in the last couple of years, essentially because the technology for detecting such tiny quantities of proteins and protein fragments has improved. A powerful example are next generation SIMOA assays.
There are now several groups around the world that have reported blood based assays that can predict PET amyloid status at around 80% accuracy. Similarly, neurofilament light is emerging as a useful biomarker of neurodegeneration, applicable in a range of brain disorders, from Alzheimer’s to Parkinson’s and multiple sclerosis.
This degree of accuracy (>80%) is sufficient for the clinical use, but what is yet to emerge are the preferred protein targets, assays, analysis methods, harmonization of standards and interpretative thresholds. This is a matter of time and it is predicted that blood-based biomarkers for Alzheimer disease will enter clinical use in the next few years. These will be relatively cheap and easily accessible, and should help clinicians around the world better diagnose and predict clinical trajectories in older people with cognitive deficits.
The second type of biomarker attracting a lot of attention are vocal biomarkers, the idea that there are sonic patterns in the way people speak that could be linked to underlying brain diseases. This is a topic that Prof Valenzuela and colleagues are pioneering at UNSW. Look out for a separate blog post about that.
Stepping back, however, a meta-issue that biomarkers are not addressing is the remarkable ‘promiscuity’ of pathology underlying cognitive impairment. We now understand that <10% of dementia in late life is due to Alzheimer’s disease alone, and that the most common basis is Alzheimer’s + 2 or 3 other degenerative diseases.
So if we only have a hammer the world is a nail. To truly understand and be able to predict and treat cognitive impairment in the elderly will mean a suite of cheap and easy biomarkers so that we can assess patients and research participants in an unbiased and realistic way.
