Jeffrey L. Cummings | Department of Brain Health, Kirk Kerkorian School of Medicine, University of Nevada; Director, Chambers-Grundy Center for Transformative Neuroscience, Las Vegas, Nevada; Scientific Advisory Board Member, ALZpath, USA
Citation: EMJ Neurol. 2026; https://doi.org/10.33590/emjneurol/77IX5Z12
You’ve spent your career tracking how Alzheimer’s disease is understood, diagnosed, and treated. What has changed most fundamentally in our understanding of the disease since you started in this field?
The most influential change has been the evolution of biomarkers, beginning around 2012 with the approval of amyloid PET, so that we could see the plaque in the brain for the very first time. Now, we have these fantastic blood biomarkers across a whole range of changes that are going on in the brain.
Ultimately, we want treatments. That will be the big breakthrough. We have some progress in that regard, but the tools that are going to make all the drugs possible are the biomarkers. That’s why I give them precedence in terms of their importance.
You’ve argued that the biological complexity of Alzheimer’s disease makes combination approaches an important direction for treatment. What is the rationale for targeting multiple biological processes rather than focusing on a single disease mechanism?
It’s something that we’ve thought a lot about. If you look at the drugs that have been approved, the monoclonal antibodies, they slow the disease by 30% on cognitive measures, and maybe 40% on functional measures. That means we have 60–70% unresolved. The disease is still progressing, but more slowly. They are a great gift, but it’s not enough. We would like to arrest the disease, prevent the disease, improve the disease, but this is all well beyond what the monoclonals can accomplish. So, we’re going to have to add other drugs to them.
As Alzheimer’s treatment becomes increasingly biomarker-driven, what does precision medicine actually mean at the level of an individual patient? How close are we to using a patient’s biological profile to guide treatment selection in routine practice?
Precision medicine aims to individualise and provide optimal therapy for each patient. I think we’re pretty far from that. We only have one set of drugs that are approved that are disease-targeted, disease-modifying therapies; everyone who qualifies is going to get the same drug at the same dose.
We have a long way to go to be where we want to be, but having that goal is the way we define how we’re going to get there. The biomarkers are tremendous for that, starting at the level of the diagnosis. We use phosphorylated tau at threonine 217 (p-tau217) to get a diagnosis, and then we can refine that and find out the stage of the disease. We’re coming to a place where we have increasingly precise decision frameworks, but I would say in the grand scheme of precision medicine, we have a long way to go yet. Ultimately, I think we would need genetic architecture to get to precision medicine, and then the biomarkers are actually an expression of what’s being driven by the gene changes.
We’re making progress. I’m flabbergasted by where we are now compared to where we were just a few years ago; I would never have guessed it. But we still cannot deliver the kind of care to patients that we want.
Can you give a brief overview of where we currently are with biomarkers in clinical practice?
The blood biomarkers are the ones that are most scalable. They’re easiest to use. Patients and doctors already have some understanding of them, compared with lumbar puncture and studies of cerebrospinal fluid (CSF), which understandably people don’t want to have unless they absolutely need to. So, blood puts us in a whole different circumstance in terms of being able to apply biomarkers in a clinical setting.
In the USA there are now four approved blood tests. Most of the work so far has been on p-tau217, but we need a larger repertoire of biomarkers that would reflect more of the picture of the biology of the Alzheimer brain.
P-tau217 is a great reporter of amyloid in the brain, but there’s a lot else going on. We need better inflammatory markers, in particular. We have quite a few anti-inflammatory drugs in trials, we can see inflammation in the brain, and yet our biomarkers for inflammation in CSF or blood are pretty poor at this time.
Co-pathology is increasingly understood to be common in dementia. How should clinicians think about conditions such as vascular disease, Lewy body pathology, or TAR DNA-binding protein 43 (TDP-43) pathology when assessing a patient with Alzheimer’s pathology?
This is one of the top issues in the field right now. It’s actually unusual for a patient to have only amyloid in the brain. Nearly all of them have plaques and tangles, and a reasonable number also have TDP-43, synuclein, and vascular pathology. All of these are playing roles that we haven’t figured out yet. For example, monoclonal antibodies can slow the disease by 30%, but if you treat earlier with lower levels of tau, you get much more than 30%. Some of those patients are actually improving. It is quite amazing to look at those very early patients. So, the co-pathology may be putting a ceiling on the effects we can get from the monoclonal antibodies, but we need to better understand this.
We have one promising biomarker for TDP-43 in the CSF, but really very little else. The synuclein biomarker that’s most widely used is also in the CSF. On the other hand, we can assess vascular pathology with MRI. That’s widely available, and the skill set to interpret those scans is available, too. Although we have some tools for vascular disease, we need much better tools for TDP-43 and synuclein.
Also, we still don’t know if co-pathology is limiting the effect of the monoclonal antibodies. There is an accumulating body of evidence about these co-pathologies that suggests that they exaggerate the rate of decline and that they make the patient less responsive. We can see that particularly in the donanemab studies, where the patients who had amyloid and high levels of tau responded less well to treatment than patients who had amyloid with low levels of tau. This is an example where we could actually see the co-pathology limiting the response to treatment, and that tells us that we have to treat early, particularly with the anti-amyloid drugs. It may be that affecting other targets will work later into the disease process: inflammation and tau drugs might not be limited to this very early window of therapy.
Looking ahead, what do you think are the most important unanswered questions that need to be addressed for Alzheimer’s treatment to have a greater impact on disease progression?
I think there are several areas that need to be approached simultaneously. One is biomarkers, and that’s been the focus of our discussion. We have some co-pathology biomarkers, but most of them are very early in development and limited. That would be one area where we currently don’t have what we want. I’ve also mentioned inflammatory biomarkers as a place where we just don’t have enough guidance.
Another area is earlier treatment. Not only do we want to treat patients early because that’s targeting the biological window, but we also want to have dignified ageing with cognitive integrity for the longest possible time. That means intervening when patients have either no deficits or very minimal deficits, so that they can have those additional years with their grandchildren or those last travel adventures. Cognitive integrity later in life is a key goal.
We need a general raising of consciousness about the importance of clinical trials, and to make people aware that clinical trials are the way to get new therapies. There’s still a huge educational challenge out there to allow patients to understand that they can contribute to the science and be citizen scientists by participating in clinical trials. They can help develop drugs that their children and their grandchildren will benefit from, but we just haven’t gotten that message out there yet.
How can patients access and enrol in clinical trials? Is this typically facilitated through referral by their clinician, or are there opportunities for patients to identify and enrol in trials independently?
There are a variety of ways to get into a trial. The Alzheimer’s Association has TrialMatch, where you can find a trial on the Alzheimer’s Association website. In the USA, you can find a list on a website of all the trials and where they’re being done.1 Researchers advertise locally so that patients know they can participate in trials in the local area. The newspaper and social media are also widely used now to inform people about clinical trials.
By and large, I would say it’s still not enough. In our annual report, we calculated that just to populate the trials happening right now requires 55,000 people. However, there’s a screening failure rate of 7:1, which means we need to screen 350,000 people to get 55,000 people into trials. That is a huge number. We can do that only if people are widely aware of the trials and their ability to make an important contribution to solving Alzheimer’s disease.
