Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.


Improving diagnostics for Alzheimer’s research

Series of computed tomography (CT) scans of an axial section through the head of a 74-year-old patient with Alzheimer's disease. Credit: ZEPHYR/ SPL/ Alamy Stock Photo.

Studies on Alzheimer’s disease (AD) are increasing worldwide, with a prevalent focus on the preclinical and early prodromal population: patients whose symptoms are mild or yet to appear. There is a need for highly accurate and validated diagnostic measures, in order to avoid including in the pharmacological trials subjects who have a low risk of progression to AD or are actually affected by other neurodegenerative disorders.

Researchers are now starting to recognize that, for current clinical trials on AD (in particular of anti-amyloid monoclonal antibodies), one of the reasons for failures may actually be the inclusion of subjects without Alzheimer disease, that may account for up to 30% out of the target population in prodromal phase or early dementia trials1 2. A recently launched Italian research project aims to address this problem together with validating an organizational model for public health purposes.

Alzheimer’s patients are currently classified according to the ATN (amyloid, tau, neurodegeneration) system3. But that system lacks a definition of the risk profile for AD progression at the level of the single patient. It may also underestimate the contribution of alterations that are shared between AD and other neurodegenerative diseases, as well as physiological ageing4, again raising the fundamental issue of an accurate diagnosis for the inclusion of subjects in pharmacological trials. We should also consider that AD and related dementia mechanisms act progressively for decades before the onset of symptoms. In addition, during the early phase, there is the plastic brain reorganization (brain reserve) that counteracts the progressive loss of neurons, synapses and connections, therefore maintaining brain functions and main cognitive abilities for longer.

At the preclinical stage, the presence of subjective cognitive decline is a risk factor for AD, but may result in the inclusion of subjects whose symptoms have a different cause (for example, depression), and in the exclusion of individuals with defective awareness of cognitive decline5 6.

Several biomarkers are used in clinical research studies for AD diagnosis and risk of progression: clinical and cognitive criteria, the deposition in the brain of the beta-amyloid and tau proteins (measured in cerebrospinal fluids and by PET), and brain atrophy measured by magnetic-resonance imaging (MRI). But they have a limited sensitivity and specificity for AD, especially in the earliest phases, and may have resulted in using anti-amyloid and tau therapies in people without the disease.

The addition of biological markers increases the predictive value, but has several limitations. For example, the MRI atrophy patterns appear late in neurodegenerative diseases, and even the measure of hippocampal atrophy does not represent an accurate differential diagnostic marker for AD. Presence of amyloid load in the brain is a marker of AD pathology but is also found in other neurodegenerative conditions and even in healthy ageing. The same is true for tau pathology. Despite being a significant risk factor, amyloid (and tau) positivity is not associated with clinical progression in a large majority of subjects, either with normal cognition or with mild cognitive decline1 2.

A crucial biomarker of neurodegeneration is brain metabolism, reflecting synaptic activity and density, and is included in the ATN classification. A brain-scanning technique called [18F] FDPET identifies a typical altered metabolism in the temporo-parietal region that is indeed a highly accurate diagnostic marker for AD, and a negative result on this test strongly predicts clinical stability. Thus, [18F] FDG-PET brain metabolism or other neurodegeneration measures should be coupled to amyloid-PET to provide a more accurate AD case inclusion and to exclude clinically stable individuals from clinical trials.

It is imperative that clinical trials adhere to inclusion criteria where not only amyloid or tau PET scans are obtained for all subjects at the initial recruitment phase, but also more specific biomarkers of neurodegeneration. This would reduce inappropriate patient recruitment which adds noise and contributes to negative results.

The INTERCEPTOR project is a strategic project by the Italian Ministry of Health and the Italian Medicines Agency, aiming to evaluate the best predictors of progression of cognitive decline in MCI subjects. A large cohort of subject, recruited by a nationwide network of clinical centres, have been submitted to a harmonized assessment of the main biomarkers available in clinical practice, including cognitive measures, MRI, EEG and FDG-PET assessing specific neuronal/synaptic/network dysfunction in the prodromal phase of disease progression.

The selection of biomarkers based on scientific evidence (highly accurate, non-invasive, and financially sustainable), allows advanced data analysis. The project is looking for the most effective biomarker combination predicting disease progression towards dementia, and to contribute to early diagnosis and prognosis (for risk of progression to AD) in subjects with mild cognitive impairment7 8. From 20 recruiting centres distributed on the Italian territory, biomarkers and clinical data have been acquired from more than 350 prodromal MCI subjects. ‘Expert centres’ for individual biomarkers risk evaluation, and a common web-based platform have been developed and validated. We expect the first results in the second half of 2023.



  1. Iaccarino, L., Sala, A., Perani, D., & Annals of clinical and translational neurology, 6(6), 1113-1120 (2019).

    PubMed  Article  Google Scholar 

  2. Tondo, G., Carli, G., Santangelo, R., Mattoli, M. V., Presotto, L., Filippi, M., et al., European Journal of Neurology, 28(4), 1123-1133 (2020).

    PubMed  Article  Google Scholar 

  3. Jack Jr, C. R., Bennett, D. A., Blennow, K., Carrillo, M. C., Dunn, B., Haeberlein, S. B. et al. Alzheimer's & Dementia, 14(4), 535-562 (2018).

    PubMed  Article  Google Scholar 

  4. Perani, D., Iaccarino, L., Lammertsma, A. A., Windhorst, A. D., Edison, P., et al., Alzheimer's & Dementia, 15(8), 1081-1103 (2019).

    PubMed  Article  Google Scholar 

  5. Cacciamani, F., Tandetnik, C., Gagliardi, G., Bertin, H., Habert, et al. Journal of Alzheimer's Disease, 59(2), 753-762. 10.3233/JAD-170399 (2017).

    PubMed  Google Scholar 

  6. Dubois, B., Villain, N., Frisoni, G. B., Rabinovici, G. D., Sabbagh, M., et al. The Lancet Neurology, 20(6), 484-496 (2021).

    PubMed  Article  Google Scholar 

  7. Rossini, P. M., Cappa, S. F., Lattanzio, F., Perani, D., Spadin, P., et al. Journal of Alzheimer's Disease, 72(2), 373-388 (2019).

    PubMed  Article  Google Scholar 

  8. Rossini, P. M., Miraglia, F., & Vecchio, F. Alzheimer's & Dementia (2022).

Download references

Nature Careers


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing


Quick links