Alzheimer’s disease (AD) is the most common cause of dementia, affecting over 55 million people globally, with nearly 10 million new cases reported each year (World Health Organisation, 2023). It is a progressive neurodegenerative condition marked by memory loss, language difficulties, disorientation, and functional decline. The average age of diagnosis is around 65 years, although changes in the brain often begin years or even decades earlier. Currently, diagnosis often occurs at a relatively late stage, once cognitive impairment becomes apparent and irreversible neuronal damage has already occurred (Jack et al., 2018). This delay significantly limits the effectiveness of potential therapeutic interventions.
Traditionally, the amyloid cascade hypothesis has been the leading explanation for AD pathogenesis. This theory posits that the accumulation of amyloid-beta (Aβ) peptides in the brain triggers a cascade of events culminating in tau hyperphosphorylation, neuronal dysfunction, and ultimately, cognitive decline (Busche and Hyman, 2020). Detection of amyloid-beta plaques through cerebrospinal fluid analysis or Positron Emission Tomography (PET) imaging can precede symptom onset by up to 20 years, making it a valuable early diagnostic biomarker (Palmqvist et al., 2017). However, the relationship between amyloid burden and symptom severity remains inconsistent, leading researchers to consider other pathological features.
Tau pathology, specifically the abnormal phosphorylation and aggregation of tau protein into neurofibrillary tangles, is now considered a more direct correlate of cognitive impairment (Franzmeier et al., 2020). Tau buildup disrupts intracellular transport and neuronal stability, directly impairing brain function. Importantly, tau-related changes are typically visible on neuroimaging only after significant neuronal damage has occurred, thereby limiting their use-fulness in the very early stages of diagnosis.
Given the limitations of current biomarkers, researchers are increasingly examining systemic biomarkers, particularly those associated with metabolic abnormalities, as alternative diagnostic tools. Insulin resistance, a hallmark of Type 2 diabetes, has been linked to neuroinflammation, mitochondrial dysfunction, and synaptic damage processes that unify several theories of AD pathogenesis (Arnold et al., 2018; Alves et al., 2021). Moreover, disrupted energy metabolism in the brain has been linked to impaired neural circuit function, further supporting the metabolic connection to AD (Kapogiannis and Mattson, 2018). These findings suggest that individuals with metabolic disorders may exhibit AD-related neuropathology years before traditional biomarkers become detectable.
Another promising avenue for early diagnosis of AD is the use of artificial intelligence (AI). AI-powered tools, including deep neural networks and machine learning algorithms, can analyse large-scale clinical, genetic, and imaging datasets to detect subtle patterns indicative of early AD (Kumar et al., 2024). Predictive modelling using AI enhances early diagnostic accuracy and holds promise for personalised treatment planning in at-risk individuals.
People with AD and their families bear a heavy financial burden, primarily from lost wages, medical costs, and long-term care. Among the significant cost projections are:
In the US:
As stated by the Alzheimer’s Association (2023), the average lifetime cost of care for a person with Alzheimer’s is $392,874.
Without successful therapies, the $345 billion in payments for AD and other dementias in 2023 is expected to rise to almost $1 trillion by 2050 (Alzheimer’s Association, 2023).
Depending on the stage of the disease, families may incur annual out-of-pocket expenses ranging from $10,000 to $50,000.
In Europe:
According to Prince et al. (2015), the annual cost of dementia care in Europe is predicted to be €250 billion.
Costs for individual patients vary by country, but they may exceed €30,000 per year, with long-term care being the most expensive aspect.
Worldwide:
Dementia is predicted to cost the world $2.8 trillion by 2050, up from an anticipated $1.3 trillion in 2019 (World Alzheimer Report, 2019).
It is becoming more widely acknowledged that insulin resistance, a defining feature of Type 2 diabetes, may be a contributing factor to neurodegeneration. Important brain processes like cell survival, neurotransmitter release, and synaptic plasticity are all regulated by insulin. A series of pathological alterations may arise when insulin signalling is compromised, as is the case with diabetes.
One example is glucose hypometabolism. The brain requires glucose for energy, and AD patients have been shown to have decreased glucose absorption even before cognitive symptoms appear. This metabolic deficit may exacerbate neuronal mortality and synaptic dysfunction (Alves et al., 2021; Arnold et al., 2018).
Chronic low-grade inflammation in the brain, characterised by the increased production of pro-inflammatory cytokines, also plays a significant role in the progression of AD. This neuroinflammatory response is often observed in individuals with insulin resistance, where elevated levels of cytokines such as interleukin-6 (IL-6), tumour necrosis factor-alpha (TNF-α), and interleukin-1 beta (IL-1β) are present (De Felice and Ferreira, 2014). These inflammatory mediators contribute to neuronal damage by promoting oxidative stress, disrupting synaptic signalling, and accelerating tau hyperphosphorylation (Heneka et al., 2015). Furthermore, microglial activation in response to metabolic dysfunction enhances the release of these cytokines, establishing a self-perpetuating cycle of inflammation and neurodegeneration. As a result, neuroinflammation serves as both a marker and a mediator of cognitive decline in AD, further reinforcing the pathological link between metabolic disease and neurodegeneration.
Insulin resistance impairs mitochondrial activity, which results in oxidative stress and damage to neurons (Kapogiannis and Mattson, 2018). According to research, insulin resistance may increase tau phosphorylation, which in turn accelerates the development of neurofibrillary tangles, a defining feature of AD pathology (Busche and Hyman, 2020).
Based on these results, tracking metabolic biomarkers, including insulin sensitivity, blood glucose variations, and inflammatory markers, may enable the identification of AD before conventional tau-based diagnostics become practical.
With its substantial benefits in processing massive datasets and recognising intricate illness patterns, AI technologies are transforming medical diagnostics. AI is being used in AD in a variety of ways:
Machine learning for biomarker analysis: AI systems can examine genetic information, metabolic profiles, and electronic medical records to identify early indicators of AD. According to Kumar et al. (2024), machine learning models have proven to recognise high-risk patients based on minute alterations in glucose metabolism, inflammatory markers, and cognitive performance.
Analysis of neuroimaging: AI-powered image processing can identify subtle structural and functional brain alterations that occur before symptoms appear. Deep learning models trained on MRI and PET images can accurately classify early-stage AD (Franzmeier et al., 2020).
The role of retinal imaging in brain health: AI-assisted retinal imaging has been investigated as a non-invasive technique for identifying early neurodegenerative alterations, as the retina and the brain share anatomical similarities (Kumar et al., 2024). AI models that analyse retinal scans have shown promise in identifying early vascular and neuronal abnormalities associated with AD. Retinal imaging offers a non-invasive window into cerebral pathology, as the retina shares embryological and anatomical features with the brain (Cheung et al., 2021). AI-driven analysis has detected changes, including retinal nerve fibre layer thinning, microvascular alterations, and amyloid-related deposits, all of which correlate with AD-related neurodegeneration (Sánchez et al., 2020; Jo et al., 2022). These findings highlight the potential of AI-enhanced retinal screening as a cost-effective, accessible method for early detection of AD, particularly in at-risk populations.
There are numerous encouraging clinical ramifications when metabolic indicators and AI are used in early Alzheimer’s diagnosis. Early therapeutic interventions, including lifestyle modifications, pharmacological treatment, and cognitive training, can be implemented when individuals at risk of AD are identified through the presence of metabolic biomarkers, prior to the accumulation of tau pathology. Detecting markers such as insulin resistance, dysregulated glucose metabolism, or mitochondrial dysfunction allows for proactive strategies aimed at slowing disease progression and preserving cognitive function (Arnold et al., 2018; Kapogiannis and Mattson, 2018). AI in personalised medicine can assist in customising treatment plans according to a patient’s genetic risk factors and metabolic profile.
Additionally, non-invasive screening tools, such as AI-assisted metabolic biomarker analysis and retinal imaging, may offer affordable and easily accessible alternatives to costly, invasive procedures such as CSF analysis and PET scans.
Future research should aim to validate these approaches on a large scale in various populations to guarantee their efficacy in clinical settings. Developing and incorporating predictive models into standard medical care requires interdisciplinary partnerships involving neurologists, data scientists, and endocrinologists.
By enabling earlier interventions that slow the progression of the disease, early diagnosis through metabolic screening and AI-based techniques could significantly reduce healthcare expenses. One possible cost-saving advantages are:
Delaying the advancement of the disease by only five years: Studies have shown that delaying the onset of AD by five years may result in a roughly 50% decrease in cases and a significant reduction in healthcare expenses (Brookmeyer et al., 2018). As a result, medical and long-term care costs could be reduced by hundreds of billions of dollars a year.
Cutting down on emergency room visits and hospitalisations: Hospitalisation rates are higher for AD patients due to infections, falls, and the consequences of the disease's progression. According to the Alzheimer’s Association (2023), early intervention might save $7,000 to $15,000 annually per patient by preventing needless hospitalisations.
Reducing the cost of long-term care: Most AD patients eventually require full-time care support or nursing home care, which costs between $80,000 and $100,000 annually in the United States. By extending independent living, slowing progress might result in a 20–30% reduction in these costs (Gaugler et al., 2022).
Although the implementation of AI-based metabolic screenings and early diagnostic tools comes with some upfront expenses, research suggests that they could result in significant long-term savings:
Costs associated with screening for AI: Compared to PET scans, which cost between $5,000 and $7,000 per scan, AI-driven blood tests and retinal imaging are expected to cost between $500 and $1,000 per individual (Kumar et al., 2024). Implementing AI widely could lower overall diagnostic costs by 40–50%, increasing accessibility to screening.
Economic Rationale for Investing: According to a 2009 study by Weimer and Sager, investing in early detection techniques could result in a 10:1 return on investment by lowering hospitalisation and long-term care expenses.
By funding AI-based early detection initiatives, governments could save billions of dollars in public healthcare costs over time.
Amyloid-beta and tau detection are the mainstays of traditional Alzheimer’s diagnostics, which frequently diagnose the illness too late for effective treatment. Nonetheless, mounting data points to insulin resistance and metabolic dysfunction as potential early markers of AD, offering a significant chance for an earlier diagnosis. AI-driven approaches significantly improve diagnostic accuracy by examining large datasets and identifying minor illness signals before tau clumps appear in scans. Moreover, mounting evidence highlights insulin resistance and metabolic dysfunction as potential early markers of AD, offering a critical window for earlier diagnosis. AI-driven approaches significantly enhance diagnostic accuracy by analysing large-scale datasets and detecting subtle disease signals well before tau clumps appear on neuroimaging. This suggests that integrating AI into diagnostic systems has the potential to transform early detection of AD, enabling timely interventions. In addition to improving patient outcomes, early diagnosis also presents a compelling financial advantage by reducing the long-term costs associated with advanced-stage care. In addition to improving patient outcomes, early diagnosis presents a compelling financial argument by significantly reducing medical expenses. Future research will be necessary to confirm these techniques and apply them in clinical settings, which could revolutionise the diagnosis and treatment of AD.
Alzheimer’s Association (2023) Alzheimer’s disease facts and figures. Alzheimer’s & Dementia, 19(4), pp.1–120. Available at: https://doi.org/10.1002/alz.13016
Alves, S.S., Silva-Junior, R.M.P.D., Servilha-Menezes, G., Homolak, J., Šalković-Petrišić, M. and Garcia-Cairasco, N. (2021) ‘Insulin resistance as a common link between current Alzheimer’s disease hypotheses’, Journal of Alzheimer’s Disease, 82(1), pp.71–105.
Arnold, S.E., Arvanitakis, Z., Macauley-Rambach, S.L., Koenig, A.M., Wang, H.Y., Ahima, R.S., Craft, S., Gandy, S., Buettner, C., Stoeckel, L.E., Holtzman, D.M. and Nathan, D.M. (2018) ‘Brain insulin resistance in type 2 diabetes and Alzheimer disease: concepts and conundrums’, Nature Reviews Neurology, 14(3), pp.168–181. Available at: https://doi.org/10.1038/nrneurol.2017.185
Brookmeyer, R., Abdalla, N., Kawas, C.H. and Corrada, M.M. (2018) ‘Forecasting the prevalence of preclinical and clinical Alzheimer's disease in the United States’, Alzheimer’s & Dementia, 14(2), pp.121–129. Available at: https://doi.org/10.1016/j.jalz.2017.10.009
Busche, M.A. and Hyman, B.T. (2020) ‘Synergy between amyloid-β and tau in Alzheimer's disease’, Nature Neuroscience, 23(10), pp.1183–1193. Available at: https://doi.org/10.1038/s41593-020-0687-6
Cheung, C.Y., Xu, D., Cheng, C.Y. and Wong, T.Y. (2021) ‘Retinal imaging and its potential role in dementia: A review of current research and recommendations’, Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring, 13(1), e12204.
De Felice, F.G. and Ferreira, S.T. (2014) ‘Inflammation, defective insulin signaling, and mitochondrial dysfunction as common molecular denominators connecting Alzheimer's disease to diabetes’, Diabetes, 63(7), pp.2262–2272.
Gaugler, J., James, B., Johnson, T., Marin, A. and Weuve, J. (2022) ‘The costs and benefits of delaying Alzheimer's disease onset’, Health Affairs, 41(6), pp.823–830. Available at: https://doi.org/10.1377/hlthaff.2022.00246
Franzmeier, N., Rubinski, A., Neitzel, J., Kim, Y., Damm, A., Na, D.L. and Ewers, M. (2020) ‘Functional brain architecture is associated with the rate of tau accumulation in Alzheimer's disease’, Nature Communications, 11(1), p.347.
Heneka, M.T., Golenbock, D.T. and Latz, E. (2015) ‘Innate immunity in Alzheimer's disease’, Nature Immunology, 16(3), pp.229–236.
Jack, C.R., Bennett, D.A., Blennow, K., Carrillo, M.C., Dunn, B., Haeberlein, S.B. and Siemers, E. (2018) ‘NIA-AA research framework: Toward a biological definition of Alzheimer's disease’, Alzheimer’s & Dementia, 14(4), pp.535–562.
Jo, Y., Lee, J.Y., Kim, S., Park, K.H., Lee, J.Y. and Lee, W.J. (2022) ‘Deep learning-based identification of early Alzheimer's disease using optical coherence tomography’, Scientific Reports, 12, p.12134.
Kapogiannis, D. and Mattson, M.P. (2018) ‘Disrupted energy metabolism and neuronal circuit dysfunction in Alzheimer’s disease’, The Lancet Neurology, 17(4), pp.381–392.
Kumar, R., Waisberg, E., Ong, J., Paladugu, P., Amiri, D., Saintyl, J. and Tavakkoli, A. (2024) ‘Artificial Intelligence-Based Methodologies for Early Diagnostic Precision and Personalized Therapeutic Strategies in Neuro-Ophthalmic and Neurodegenerative Pathologies’, Brain Sciences, 14(12), p.1266.
Palmqvist, S., Schöll, M., Strandberg, O., Mattsson, N., Stomrud, E., Zetterberg, H., Blennow, K., Landau, S., Jagust, W. and Hansson, O. (2017) ‘Earliest accumulation of β-amyloid occurs within the default-mode network and concurrently affects brain connectivity’, Nature Communications, 8(1), pp.1–13.
Prince, M., Wimo, A., Guerchet, M., Ali, G.C., Wu, Y.T. and Prina, M. (2015) World Alzheimer Report (2015). The Global Impact of Dementia: An analysis of prevalence, incidence, cost and trends. Doctoral dissertation, Alzheimer's Disease International.
Reitz, C., Brayne, C. and Mayeux, R. (2020) ‘Epidemiology of Alzheimer disease’, Nature Reviews Neurology, 16(5), pp.245–260.
Sánchez, D., Castilla-Marti, M., Rodríguez-Gómez, O., Valero, S., Pifarré, L. and Martínez-Lage, P. (2020) ‘Retinal biomarkers for Alzheimer's disease: A review of the evidence’, Journal of Alzheimer's Disease, 77(1), pp.57–76.
Weimer, D.L. and Sager, M.A. (2009) ‘Early identification and treatment of Alzheimer's disease: social and fiscal outcomes’, Alzheimer's & Dementia, 5(3), pp.215–226.
World Health Organization (2023) Dementia. Available at: https://www.who.int/news-room/fact-sheets/detail/dementia (Accessed: 23 June 2025).
© Angela Ganea. This article is licensed under a Creative Commons Attribution 4.0 International Licence (CC BY).