Check out this answer from Consensus:
Predicting the rate of progression in Alzheimer’s disease is a complex but achievable goal, thanks to advancements in biomarkers, imaging technologies, and machine learning models. Factors such as neurofibrillary tangle burden and early cognitive decline play crucial roles in forecasting the disease’s trajectory. These insights not only enhance clinical trial design but also pave the way for personalized care strategies, ultimately improving the quality of life for individuals with Alzheimer’s.
Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and memory loss. Predicting the rate at which an individual will progress through the stages of Alzheimer’s is crucial for clinical management and the design of effective clinical trials. Recent research has focused on identifying biomarkers and developing models to forecast the progression of AD. This article explores the current scientific understanding and methodologies used to predict the rate of Alzheimer’s progression.
Factors Influencing Alzheimer’s Progression
Neurofibrillary Tangle Burden
One significant factor influencing the progression of Alzheimer’s is the burden of neurofibrillary tangles. A study utilizing data from the National Alzheimer Coordinating Center found that the stage of neurofibrillary tangles, when combined with other factors such as neuritic plaques, age, sex, education, and APOE genotype, could predict the rate of cognitive decline. Higher neurofibrillary tangle stages were associated with faster deterioration in cognitive functions, such as memory and verbal fluency1.
Early Cognitive Decline
The early rate of cognitive decline can also serve as a predictor for later stages of Alzheimer’s. A study involving 91 patients assessed using the Milan Overall Dementia Assessment (MODA) scale demonstrated that patients with a slow progression rate in the early stages were unlikely to experience rapid progression later on, and vice versa. This finding suggests that early cognitive assessments can provide valuable insights into the future course of the disease2.
Predictive Models and Technologies
Structural MRI and Generative Adversarial Networks
Advancements in imaging technologies and machine learning have led to the development of predictive models for Alzheimer’s progression. One such approach involves using structural Magnetic Resonance Imaging (sMRI) to capture cortical atrophy associated with AD. A novel framework employing a 3D multi-information generative adversarial network (mi-GAN) has been proposed to predict future brain MRI images and determine the clinical stage of Alzheimer’s. This model has shown high accuracy in generating realistic future brain images and classifying the disease stage, thereby offering a promising tool for long-term progression prediction3.
Implications for Clinical Trials and Care
Enhancing Clinical Trial Design
Understanding the factors that influence Alzheimer’s progression can significantly enhance the design of clinical trials. For instance, selecting participants based on their neurofibrillary tangle stage can improve the power of clinical trials and reduce the required sample sizes. This approach ensures that trials are more efficient and have a higher likelihood of detecting the efficacy of disease-modifying agents1.
Personalized Clinical Care
Predictive models and early cognitive assessments can also inform personalized clinical care strategies. By identifying individuals at risk of rapid progression, healthcare providers can tailor interventions and support to better manage the disease and improve patient outcomes2.
Can scientists predict how quickly a person will go through the stages of Alzheimer’s?
Simon Duchesne has answered Likely
An expert from Université Laval in Neuroimaging, Alzheimer’s Disease, Neurology, Artificial Intelligence
Thank you for your question. It is timely, and important.
The short answer is, yes.
The long answer is also yes; but, while the same, it is a bit longer to explain.
Given the right tests, scientists can tell if one already has the hallmarks of the disease (e.g. accumulation of amyloid plaques and tau proteins in and around neurons; loss of brain tissue; diminished brain energy consumption in certain regions); when coupled with established, clear symptoms (e.g. loss in memory; loss of subtle language functions; loss of planning and executive abilities), then a diagnostic can be posed.
Predicting the future is also possible. At this point, the error in the prognostic is less than the error in the probability of the event occurring. This is like predicting the weather: we can predict it will rain around 20mm tomorrow. It may end up being 10mm, or 30mm, but rain there will be. So, given the symptoms and hallmarks (which we call biomarkers) stated above, scientists can make predictions with good accuracy as far as Alzheimer’s disease is concerned.
Now, the brain is an interesting organ, which is surprisingly robust. One consequence is that the stage one is at for Alzheimer’s disease is not the same as the functional stage of the brain. To put another way, one may have the hallmarks of the disease, but may not have all the symptoms; the reverse, while rarer, is also true. To use a different analogy, just because one has contracted the flu virus of the year (and therefore, has the influenza disease) does not mean that he/she will get a running nose, high fever and a cough right away (the symptoms); nor experience them as fully as someone else.
So predicting that one will progress in their level of Alzheimer’s disease does not equate directly with a similar progression in the rate of the symptoms – which is what people, internally and externally, notice. To come back to our weather analogy, a 10mm rain may not overflow one house’s gutters; but a 30mm would (and it would be different for a different house). For Alzheimer’s, this difference may be the difference between still being able to make one’s breakfast, or setting fire to the kitchen. And this difference, can and does make all the difference. In that sense, our predictions of the functional impact of Alzheimer’s disease are still very much incomplete.
Hope this clarifies. Do come back with more questions!
Can scientists predict how quickly a person will go through the stages of Alzheimer’s?
John Hardy has answered Likely
An expert from University College London in Neurology, Neurodegenerative Disease, Alzheimer’s Disease
The average is about 8 years from diagnosis… but it is variable (4-16 years) and we don’t understand that variability… we are working on this now
Can scientists predict how quickly a person will go through the stages of Alzheimer’s?
Yaakov Stern has answered Likely
An expert from Columbia University in Neurodegenerative Disease, Cognitive Science, Neuroimaging
There is a lot of variability in the rate of decline in Alzheimer’s disease. Therefore it has always been difficult to make individual predictions for time from diagnosis to important endpoints such as need for nursing home care or death. There are some features of the disease that are associated with more rapid decline, and many years ago I created a prediction formula that incorporated some of these features (1). Still, I would not say that this model accurately predicts rate of decline for individual patients.
Over the past few years I’ve been collaborating with a statistician on a new model of the course of Alzheimer’s disease. This model is unique in that it integrates multiple features of the disease, including measures of cognition, function, motor signs, dependents, and psychiatric features. In addition, it is based on data from studies that I conducted which followed small groups of Alzheimer’s patients (~225 in each group) every six months for up to 12 years. Using these new analyses, we have developed a new model for the course of Alzheimer’s disease which I am hoping will be able to supply much more accurate predictions. An initial version of the prediction model seemed to work well. We tested this by applying the model to data from people who were not use in the development of the model. The results were very promising (2). Our newly improved model (3) is the currently being tested in separate data sets as well. Our hope is that we can eventually create a calculator that can assist a patient and their physician and making predictions about time until important endpoints. Some of the endpoints that we consider important are time till a person needs to be watched well in the home, time until they are receiving the collective nursing home care, and time to death. However, the model that we are working with is flexible and could be used for other important features of the disease as well. For example, one and point that is important to many family members is the point at which the patient needs to wear diapers.
In sum, I am optimistic that our prediction abilities will improve. Besides my approach, other investigators are using various forms of AI approaches to attempt to model data as well. I think that at the least these models will give a rough idea of time to an endpoint, and certainly differentiate slow from more rapid decliners. The absolute accuracy of the prediction for any individual still needs to be determined.
References
- Stern Y, Tang MX, Albert MS, et al. Predicting time to nursing home care and death in individuals with Alzheimer disease. Journal of the American Medical Association 1997;277:806-812.
- Razlighi QR, Stallard E, Brandt J, … Stern, Y. A new algorithm for predicting time to disease endpoints in Alzheimer’s disease patients. J Alzheimers Dis 2014;38:661-668.
- Stallard E, Kinosian B, Stern Y. Personalized predictive modeling for patients with Alzheimer’s disease using an extension of Sullivan’s life table model. Alzheimers Res Ther 2017;9:75.
Can scientists predict how quickly a person will go through the stages of Alzheimer’s?
Simon F Eskildsen has answered Unlikely
An expert from Aarhus University in Neuroimaging, Biostatistics, Alzheimer’s Disease
The current models cannot do this. Perhaps when we collect more data and models become more advanced, this is a possibility.
Can scientists predict how quickly a person will go through the stages of Alzheimer’s?
Daniel C Alexander has answered Likely
An expert from University College London in Alzheimer’s Disease, Neurodegenerative Disease, Biostatistics, Artificial Intelligence, Computational Mathematics
Yes — through a combination of the latest brain scanning methods, cognitive tests, and emerging computational techniques.
Predicting the future for someone with Alzheimer’s disease is challenging because the disease varies from person to person. For example, it is widely accepted that there are three subtypes of typical Alzheimer’s disease that affect the brain in different ways, although memory problems are always present.
In recent years, the emerging field of data-driven disease progression modelling (D3PM) has revealed new insight into progressive diseases. These models are a set of computational tools for building disease signatures that integrate different kinds of patient information, for example: the latest brain scanning methods, serum and spinal fluid, and cognitive tests.
These computational models enable finer precision in patient staging and stratification, prediction of progression rates, and earlier and better identification of at-risk individuals [Oxtoby, et al., Current Opinion in Neurology 30, 371 (2017)].
Adding machine learning into the mix has refined this further to include subtype and stage inference [Young et al., Nature Communications 9, 4273 (2018)].
There are strong efforts to test these ideas within the research community. For example, the TADPOLE challenge (https://tadpole.grand-challenge.org) aims to predict the future for elderly individuals who are at-risk of Alzheimer’s disease.
Disease progression modelling currently sits in the research domain, but there are strong efforts towards translating them for use in the clinic, and for improving clinical trials into putative therapies for Alzheimer’s disease. For example, the EuroPOND consortium: http://europond.eu
In summary, the research in this field holds great promise for predicting how quickly a patient will progress through the stages of Alzheimer’s disease. I expect that clinical tools emerging from data-driven disease progression modelling will become commonplace in the not-too-distant future.