Doctoral researcher Marta Arbizu Gómez presents how blood biomarkers and neuroimaging can stratify Alzheimer’s progression and what this means for clinical practice and cognitive rehabilitation.
Executive summary of the key points in this article:
1. Blood biomarkers such as GFAP and p-tau217 can predict the clinical progression of Alzheimer’s disease.
2. Dominant prognostic factors change throughout the course of Alzheimer’s disease.
3. Integrating biomarkers and neuroimaging in Alzheimer’s disease enables more personalized clinical monitoring.
Why is it important to improve prognosis in Alzheimer’s disease?
Alzheimer’s disease (AD) does not progress in the same way for everyone. Two patients with the same clinical diagnosis may show very different trajectories: while some remain stable for years, others progress rapidly to more advanced stages of dementia. This heterogeneity poses a major challenge for clinical practice, research, and the planning of therapeutic and cognitive rehabilitation interventions.
Traditionally, Alzheimer’s disease has been classified according to broad clinical stages—cognitively unimpaired individuals, mild cognitive impairment (MCI), and dementia—or biological models such as the A/T/N framework. However, these approaches do not always make it possible to accurately predict the rate of progression for each patient.
In this context, a study published in Nature Communications in 2026 proposes a new approach: an integrated prognostic staging system that combines clinical information, blood biomarkers, and neuroimaging to estimate the risk of progression across the entire Alzheimer’s disease continuum.
How was this Alzheimer’s disease research conducted?
The work was conducted in two major phases and was based on high-quality longitudinal data:
- K-ROAD cohort (South Korea): 1,263 participants, including cognitively unimpaired (CU) individuals, people with mild cognitive impairment (MCI), and people with dementia.
- ADNI cohort (external validation): 290 participants from an international consortium widely used in Alzheimer’s disease research.
All participants had:
- Longitudinal cognitive assessments using CDR-SB and MMSE. CDR-SB measures the functional and cognitive severity of dementia over time, while MMSE is a brief, standardized test that assesses global cognitive status.
- Plasma biomarkers, including p-tau217, GFAP, NfL, and Aβ42/40.
- Neuroimaging, particularly hippocampal volume on magnetic resonance imaging and amyloid PET.
In the first phase, the researchers used machine learning–based survival models (random survival forests) to identify risk subgroups within each initial cognitive stage. In the second phase, these subgroups were integrated into a unified system of six prognostic stages, ranging from stage 0 to stage IVB.
Subscribe
to our
Newsletter
What do the key findings of this Alzheimer’s study reveal?
One of the most relevant findings of this Alzheimer’s disease study is that dominant prognostic factors change throughout the course of the disease. In other words, the most informative biomarker is not the same in the earliest phases as in more advanced stages.
In cognitively unimpaired individuals, the primary predictor of progression is plasma GFAP, a marker of astroglial activation and early inflammation.
In the mild cognitive impairment (MCI) phase, the most informative factor becomes hippocampal volume, reflecting structural neurodegeneration.
In the dementia phase, age emerges as the main determinant of the rate of progression, with more aggressive courses observed in people with earlier onset.
In addition, consistently across all phases, plasma p-tau217 acts as a cross-stage marker, providing additional prognostic information regardless of clinical stage.
This dynamic shift in predictors can be summarized simply in the following table:
| Clinical phase | Primary predictor of progression | What it reflects biologically |
|---|---|---|
| Cognitively unimpaired (CU) | Plasma GFAP | Astroglial activation and early inflammation |
| Mild cognitive impairment (MCI) | Hippocampal volume (MRI) | Structural neurodegeneration |
| Dementia | Age | More aggressive course with earlier onset |
| All phases | Plasma p-tau217 | Cross-stage prognostic information associated with amyloid and tau pathology |
By integrating these factors, the proposed system defines six prognostic stages (0–IVB) that show stepwise increases in the risk of progression, progressively worsening CDR-SB and MMSE trajectories, and clear inflection points, especially in intermediate stages, where the risk of decline accelerates significantly.
What does this approach add compared with traditional Alzheimer’s disease models?
It is important to emphasize that this system is not intended to replace classic biological models or current diagnostic criteria. Instead, it offers a complementary perspective:
- It does not classify patients according to the presence or absence of pathology, but rather according to their risk of clinical progression.
- It is based on scalable biomarkers, particularly blood tests and magnetic resonance imaging, which are more accessible than highly specialized techniques.
- It facilitates intuitive clinical interpretation by translating complex data into clearly differentiated prognostic stages.
Ultimately, this is a prognosis-oriented framework, not a treatment decision-making tool, but it has great potential for clinical research and longitudinal monitoring.
What are the implications of this Alzheimer’s study for clinical practice and research?
This approach opens up new possibilities at several levels:
- Personalized monitoring: It makes it possible to identify patients at greater risk of rapid progression, even within the same clinical diagnosis, which may help prioritize monitoring and intervention.
- Clinical trial design: Prognostic stratification can improve participant selection and the interpretation of results, reducing heterogeneity.
- Longitudinal monitoring: By relying on blood biomarkers, it facilitates repeated assessments over time without resorting to invasive procedures.
How does this advance relate to NeuronUP?
At NeuronUP, we work with longitudinal cognitive performance data and personalized intervention programs. A prognostic stratification system such as the one proposed in this study naturally aligns with this philosophy, as it makes it possible to:
- Contextualize cognitive performance within an individual prognostic trajectory.
- Adjust the goals and intensity of cognitive rehabilitation according to the estimated risk of progression.
- Integrate biological and functional information to move toward a truly personalized approach for patients with Alzheimer’s disease.
Combining biomarkers, prognostic models, and digital rehabilitation platforms represents a key step toward more precise, dynamic, and person-centered care.
Conclusion
This study proposes a new integrated prognostic framework for Alzheimer’s disease that can capture the true heterogeneity of clinical progression across the entire disease continuum. By combining blood biomarkers, neuroimaging, and clinical data, it offers a powerful tool for research and longitudinal monitoring.
In the context of NeuronUP, these advances reinforce the importance of bringing together biological diagnosis, individualized prognosis, and cognitive rehabilitation, moving toward a more comprehensive and personalized care model.
References
- Shin D, Lee S, Kim JP, et al. Biomarker-integrated prognostic stagings for Alzheimer’s disease. Nature Communications. 2026. doi:10.1038/s41467-026-68732-6.
Frequently asked questions about biomarkers and prognosis in Alzheimer’s disease
1. Which blood biomarkers can predict Alzheimer’s disease progression?
The most relevant blood biomarkers currently are p-tau217, GFAP, NfL, and the Aβ42/40 ratio.
Plasma GFAP is particularly useful in preclinical phases, while p-tau217 provides prognostic information across all phases of the disease. These markers make it possible to estimate the risk of cognitive decline and the rate of clinical progression.
2. Which biomarker is most useful in the early stages of Alzheimer’s disease?
In cognitively unimpaired individuals with biological risk, plasma GFAP has been identified as one of the main predictors of progression. It reflects astroglial activation and early inflammation, processes that precede structural neurodegeneration.
3. How does the predictor of progression change according to clinical phase?
The dominant prognostic factor varies across the disease continuum:
- Preclinical phase: plasma GFAP.
- Mild cognitive impairment (MCI): hippocampal volume.
- Dementia: age at onset.
- All phases: p-tau217 as a cross-stage marker.
4. How does prognostic stratification differ from traditional diagnostic models?
Classical models (such as the A/T/N framework) classify the presence of biological pathology.
Prognostic stratification, in contrast, focuses on estimating the rate of clinical progression and the risk of functional decline, which is particularly useful for monitoring and treatment planning.
5. Can these biomarkers be used in routine clinical practice?
Blood biomarkers have high scalability potential because they are less invasive and more accessible than amyloid PET. Although their implementation depends on the healthcare setting, they represent a step toward precision medicine that is more applicable in real-world clinical environments.
6. What implications does this model have for neurorehabilitation?
This approach opens up new possibilities by making it possible to identify patients at greater risk of rapid progression, improve participant selection and the interpretation of results, and facilitate assessment over time without having to resort to invasive procedures.
Hospital Neuropsychological Assessment: Cognitive Impact of Hospitalization and Clinical Recommendations
Leave a Reply