Carlos Cruchaga, PhD, director of the NeuroGenomics and Informatics Center at Washington University, discusses emerging blood-based biomarkers for monitoring biological and cognitive responses to lecanemab treatment in Alzheimer disease.
As anti-amyloid therapies become part of routine care for Alzheimer disease (AD), researchers are investigating how blood-based biomarkers can help characterize biological responses to treatment and explain differences in clinical outcomes. Although lecanemab reduces amyloid burden and slows cognitive decline, patients do not experience the same degree of clinical benefit. Understanding how biomarkers change during treatment could help clarify the biological processes associated with amyloid clearance and cognitive outcomes.
Published in The Lancet Neurology on September 24, 2026, a longitudinal study led by investigators at Washington University School of Medicine in St. Louis evaluated plasma proteomic changes in patients receiving lecanemab in real-world clinical practice. Using Alamar Biosciences’ NULISAseq CNS 120 panel, researchers measured 130 plasma proteins spanning amyloid and tau pathology, inflammation, neurodegeneration, and synaptic function. Among 197 patients treated with lecanemab, 34 biomarkers changed significantly with the number of infusions received, with some moving toward levels observed in amyloid-negative individuals without cognitive impairment and others changing in the opposite direction.
In this Q&A with NeurologyLive®, Carlos Cruchaga, PhD, professor of psychiatry and director of the NeuroGenomics and Informatics Center at Washington University School of Medicine in St. Louis, discusses the study’s findings and their implications for understanding treatment response. He also explores how candidate blood-based biomarkers associated with amyloid clearance and cognitive decline could inform future approaches to treatment monitoring, while highlighting the need for further validation before these markers can be incorporated into clinical practice.
Carlos Couchaga, PhD: Currently, there are a couple of new therapies for AD, including lecanemab, which was among the first disease-modifying treatments approved for the disease. These treatments focus on removing amyloid plaques from the brain. Clinical trials have demonstrated that these therapies can remove amyloid plaques and slow cognitive decline, but there have been relatively few studies examining proteomic changes in people receiving lecanemab in real-world settings as part of standard clinical care.
In this study, we followed approximately 200 people receiving lecanemab at a memory clinic. We collected blood samples over nearly 2 years and generated proteomic data using the NULISAseq CNS 120 panel. This panel includes AD biomarkers, such as phosphorylated tau 217 (p-tau217), as well as proteins that capture other biological processes, including neurodegeneration, microglial activity, and inflammation.
Through this approach, we identified a substantial number of proteomic changes associated with treatment. Some of these changes were unexpected. For example, we observed changes in several proteins associated with immune responses and inflammation that moved in the opposite direction from what we would have anticipated.
One possible explanation is that, as amyloid plaques are removed, microglia undergo reprogramming as they adapt to the brain's changing biological environment. In other words, we may be capturing aspects of brain plasticity and microglial reprogramming in response to treatment.
We identified several groups of proteins based on how their levels changed during treatment.The first group consisted of proteins that we described as normalizing. As patients received treatment, the levels of these proteins moved closer to those observed in people without AD. These included proteins related to tau metabolism and neurodegeneration more broadly. This was what we initially expected to see across most of the proteins.
However, we also identified proteins that appeared to overcorrect. Their levels moved toward those observed in controls but then went beyond the levels we would expect in people without AD. This group included some established biomarkers and proteins associated with inflammation.
The third group, which I think is particularly interesting, consisted of proteins that changed in the opposite direction from what we expected, moving further away from the levels observed in controls. Neurofilament light (NfL) was among the proteins that changed in this unexpected manner.
Importantly, we did not observe an association between these changes in the opposite direction and cognitive outcomes, even among patients who became amyloid-negative. Becoming amyloid-negative is an important milestone in treatment with these therapies.
We interpret these findings as potentially reflecting normal biological changes associated with amyloid removal rather than necessarily indicating a negative treatment response. Clinicians should not automatically interpret these unexpected biomarker changes as a reason for alarm. Further research will be needed to better understand the underlying mechanisms and clinical significance.
Another important finding was the identification of potential biomarkers that could help monitor treatment and identify patients who may benefit most. Patients receiving lecanemab typically undergo amyloid PET imaging at around 18 months to determine whether they have become amyloid-negative. Those who remain amyloid-positive may need additional imaging, which can be expensive and burdensome.
We identified several proteins that may help indicate when a patient is transitioning from amyloid-positive to amyloid-negative status. If these findings are validated, blood-based biomarkers could potentially help clinicians determine the most appropriate time to perform amyloid PET imaging rather than relying exclusively on a predetermined time point.
We also identified proteins associated with greater cognitive stability and less memory decline during treatment. These findings could eventually help provide patients with more individualized information about their expected treatment response.
As these treatments become more common, clinicians will increasingly encounter patients undergoing biomarker testing, including measurements of proteins such as NfL.One important implication of our findings is that some biomarkers may move in an unexpected direction during treatment.
In our study, these changes were not associated with treatment outcomes. This suggests that they may reflect biological responses to amyloid removal rather than necessarilyindicating a problem with treatment. Clinicians should interpret these findings in theappropriate clinical context rather thanassuming that every unexpected biomarker change signals an adverse outcome.
Another potential implication relates to treatment monitoring. Currently, there is a relatively established approach to determining when patients should undergo amyloid PET imaging. I do not expect that protocol to change in the immediate future, but blood-based biomarkers could eventually help inform the timing of these scans.
NfL is one protein of interest because its levels changed as patients transitioned to amyloid-negative status. If these findings are validated, this type of biomarker could potentially help identify the appropriate time for amyloid PET imaging based on an individual's biological response to treatment.
We also found that certain protein levels were associated with greater cognitive stability. This information could be useful when discussing treatment expectations with patients and their families. However, I do not necessarily see these findings changing treatment eligibility criteria at this stage.
On average, these treatments have been shown to slow cognitive decline, although individual responses vary. Patients who may experience a smaller benefit could still benefit from treatment. We should not exclude someone from receiving therapy solely because a biomarker suggests that their response might be less pronounced.
Instead, these findings could eventually give clinicians additional information to support discussions with patients about what they might expect from treatment.
This was a relatively large real-world study of people receiving lecanemab. We followed approximately 200 participants for nearly 2 years, which required a substantial effort from physicians, research coordinators, nurses, and other team members at Washington University.
However, this was a single-site study conducted at Washington University. The findings need to be validated in additional studies involving different sites and patient populations. As these treatments become more widely used, we hope that other groups will conduct similar studies to replicate our findings.We are encouraged by the results, but independent validation remains important.
Another limitation is that we identified potential biomarkers for treatment monitoring and efficacy that have not yet completed the necessary validation process. These candidates will need to progress through the usual biomarker development pathway, including additional clinical studies and the appropriate regulatoryreview, before they can be incorporated into routine clinical practice.
I think the field is moving in this direction, and we may see new biomarkers for monitoring treatment response emerge as the evidence develops.
Looking ahead, where do you see the AD biomarker field heading, particularly regarding treatment monitoring?
It is important to recognize how much the field has changed. Five to seven years ago, or even more recently, many people did not expect that we would be able to develop reliable blood-based biomarkers for AD, Parkinson disease, and other neurodegenerative conditions.
Today, we have promising blood-based biomarkers for AD, including p-tau217, that can help identifyAlzheimer-related pathology. These biomarkers are also helping make disease-modifying treatments more accessible.
To receive these treatments, patients need confirmation of amyloid pathology in the brain, which has traditionally been established using amyloid PET imaging or cerebrospinal fluid testing through lumbar puncture. At Washington University, a substantial proportion of patients undergoing evaluation for treatment can now be assessed using a blood test for p-tau217. This is already helping streamline the process of identifying eligible patients.
Now that we have disease-modifying treatments available and additional therapies potentially on the horizon, we need to develop biomarkers that can do more than identifyAlzheimer-related pathology. We need biomarkers that can monitor treatment response, determine when patients become amyloid-negative, and potentially identify adverse effects, including amyloid-related imaging abnormalities (ARIA).
We did not specifically analyze ARIA in this study because the number of events in our sample was too low to support that analysis. However, we now have a larger dataset and are investigating ARIA as well.
Ultimately, I think we will see continued development of biomarkers for treatment monitoring, including biomarkers that can help track the transition to amyloid-negative status, identify potential adverse effects, and determine which patients are benefiting most from treatment. Identifying and validating these biomarkers will be an important next step for the field.
I want to highlight the effort required to conduct this study in a real-world clinical setting. We recruited patients receiving routine lecanemab treatment and collected their blood samples over time. Because this was part of standard clinical care rather than a dedicated clinical trial, the process required substantial coordination among neurologists, physicians, research coordinators, nurses, and other members of the clinical team.
It was a major institutional effort to bring everything together and make the study possible. Many people were instrumental in making this research happen, including Suzanne Schindler, the paper's senior author, as well as Randy Bateman, who helped organize much of the sample collection.
One reason I think this study is important is that collecting longitudinal samples from patients receiving routine treatment is challenging, and relatively little research has examined these proteomic changes in this setting. The findings reflect a collaborative effort to build a dataset that can help us better understand the biological effects of treatment.That collaborative environment was essential to completing the study, and I think it is important to recognize everyone who contributed.
