Introduction: The Heterogeneity of Alzheimer’s Disease
The clinical presentation of Alzheimer’s disease (AD) is remarkably heterogeneous. While the amyloid-tau-neurodegeneration (ATN) framework has provided a biological backbone for diagnosis, clinicians frequently encounter patients whose cognitive performance does not align with their biomarker profile. Some individuals remain cognitively stable despite high pathological burdens—a phenomenon termed cognitive resilience—while others exhibit rapid decline even with relatively low levels of classic AD pathology, often due to the presence of non-AD copathologies. Understanding these discrepancies is critical in the current era of disease-modifying therapies (DMTs), where predicting a patient’s individual trajectory is essential for risk-benefit stratification.
A landmark study by Brown et al. (2025), published in JAMA Neurology, explores this ‘tau-clinical mismatch.’ By evaluating individuals who are amyloid-positive (Aβ+), the researchers sought to determine whether the discordance between tau burden (measured via PET or plasma p-tau217) and clinical symptoms (measured by the Clinical Dementia Rating Sum of Boxes, CDR-SB) could serve as a proxy for identifying underlying copathologies or inherent resilience.
Study Design and Methodology
This longitudinal, observational cohort study integrated data from two major sources: the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Penn Alzheimer’s Disease Research Center (Penn-ADRC). The study spanned two decades (2004–2024), analyzing 998 Aβ+ individuals in the ADNI cohort and 248 in the Penn-ADRC cohort.
The researchers focused on two primary exposures: tau burden and clinical assessment. Tau burden was quantified using either tau positron emission tomography (tau-PET) or phosphorylated tau 217 (p-tau217), a highly specific blood-based biomarker. Clinical severity was determined using the CDR-SB score.
Participants were categorized into three distinct ‘mismatch’ groups based on the residuals of a regression model correlating tau levels with clinical symptoms:
1. Canonical Group
Individuals whose clinical symptoms were commensurate with their tau burden (approximately 55-57% of the cohort).
2. Resilient Group
Individuals whose clinical symptoms were significantly milder than predicted by their tau burden (approximately 24-25%).
3. Vulnerable Group
Individuals whose clinical symptoms were significantly more severe than predicted by their tau burden (approximately 19-20%).
Outcome measures included longitudinal changes in CDR-SB, neuroimaging signatures of neurodegeneration (such as medial temporal lobe volume), and biomarkers for copathologies, specifically TAR DNA-binding protein 43 (TDP-43) and α-synuclein (via cerebrospinal fluid seed-amplification assay).
Key Findings: Defining the Vulnerable and Resilient Phenotypes
The study’s results provide compelling evidence that tau-clinical mismatch is a robust indicator of underlying biological complexity.
Copathology in the Vulnerable Group
One of the most significant findings was the association between the ‘vulnerable’ phenotype and non-AD copathologies. Individuals classified as vulnerable showed neuroimaging patterns consistent with TDP-43-related neurodegeneration and were more likely to test positive for α-synuclein. This suggests that the ‘excess’ clinical impairment in these patients is not due to AD pathology alone but is driven by the synergistic effects of multiple proteinopathies. From a clinical perspective, this explains why some patients may appear to have ‘aggressive’ Alzheimer’s when, in reality, they are suffering from a mixed-etiology dementia.
Cognitive Resilience and Slower Trajectories
Conversely, the resilient group demonstrated a significantly slower rate of cognitive decline. Despite having substantial tau pathology, these individuals maintained better functional status over time compared to both the canonical and vulnerable groups. This resilience was associated with greater cortical thickness and larger medial temporal lobe volumes, suggesting a biological ‘buffer’ against the neurotoxic effects of tau aggregates.
Longitudinal Clinical Trajectories
Across both the ADNI and Penn-ADRC datasets, the mismatch classification successfully predicted future decline. The vulnerable group reached milestones of severe cognitive impairment much earlier than the canonical group, while the resilient group showed delayed progression. These trajectories remained consistent whether tau was measured via expensive PET scans or via the more accessible p-tau217 blood test, highlighting the potential for widespread clinical implementation.
Clinical Implications for Disease-Modifying Therapy
The emergence of antiamyloid monoclonal antibodies, such as lecanemab and donanemab, has shifted the focus toward precision medicine in neurology. The Brown et al. study applied their mismatch models to a cohort receiving antiamyloid therapy and found that the classification could predict individual cognitive trajectories during treatment.
For clinicians, this means that tau-clinical mismatch could be used to manage patient expectations. A patient in the ‘vulnerable’ group might see less clinical benefit from an antiamyloid drug because their symptoms are being driven by TDP-43 or α-synuclein—pathologies that are not targeted by current DMTs. On the other hand, identifying ‘resilient’ patients may help clinicians understand why some individuals remain stable for longer periods, potentially influencing the duration and intensity of therapeutic monitoring.
Expert Commentary: A New Tool for Prognosis
The ability to identify copathology through clinical and biomarker mismatch represents a significant leap forward. Traditionally, TDP-43 and α-synuclein could only be confirmed post-mortem. While seed-amplification assays for α-synuclein are becoming more common, we still lack a validated in vivo biomarker for TDP-43. The ‘vulnerable’ mismatch phenotype serves as a crucial clinical surrogate, alerting the physician to the likely presence of these ‘hidden’ pathologies.
However, the study is not without limitations. The cohorts involved, particularly ADNI, tend to represent a more highly educated and less diverse population than the general public. Cognitive reserve, often linked to education and socioeconomic factors, likely plays a role in the ‘resilient’ group, and further research is needed to see how these models perform in more diverse, community-based settings. Additionally, while the models are predictive, they are not yet diagnostic for specific copathologies; they indicate the *likelihood* of mixed pathology rather than confirming it.
Conclusion
The tau-clinical mismatch model provides a sophisticated yet practical framework for understanding the individual variability in Alzheimer’s disease. By recognizing that symptoms are the result of a complex interplay between AD pathology, non-AD copathology, and individual resilience, clinicians can move beyond a ‘one-size-fits-all’ approach. This study underscores the importance of integrating biomarkers—specifically tau PET or p-tau217—with rigorous clinical assessment to provide patients with an accurate, individualized prognosis in the era of modern dementia care.
References
Brown, C. A., Mundada, N. S., Cousins, K. A. Q., et al. (2025). Evaluation of Copathology and Clinical Trajectories in Individuals With Tau-Clinical Mismatch. JAMA Neurology. doi:10.1001/jamaneurol.2025.4974.
Jack, C. R., Jr, et al. (2018). NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimer’s & Dementia, 14(4), 535-562.
Stern, Y. (2012). Cognitive reserve in ageing and Alzheimer’s disease. The Lancet Neurology, 11(11), 1006-1012.
