What is it about?
Alzheimer’s disease (AD) is not just damaging but devastating to patients and their families. Once diagnosed with AD, patients and their loved ones are often anxiously concerned about how soon the disease will deteriorate. A reliable prediction of the severity of the disease in the future would help them relieve this concern and better prepare for the battle with AD, both mentally and materially. However, unlike in cancer treatment, where the individualized prognosis for personalized medicine is readily available, there is no such capability to perform a reliable prediction for a patient in the treatment, clinical trials, or research studies of AD nowadays. The main obstacle to this capability is partially due to the notoriously complex heterogeneity of AD and partially the limited ability of traditional methods. To overcome such an obstacle, this paper uses two validated strategies that are innovative in current AD research. One is the stage-specific prediction for mitigating the complex heterogeneity in AD stages, where we build a specific predictive model for each AD severity stage rather than use a one-fits-all model for all stages. The other is to leverage cutting-edge artificial intelligence techniques to remedy the methodological limitations of traditional methods. By doing so, this paper establishes a deep learning-based computational way to reliably and stage-specifically predict the time to future AD severity stages for a patient.
Photo by David Matos on Unsplash
Why is it important?
The prediction only uses the commonly available clinical data, including demographic variables, comorbidity, historical clinical records for this patient, and clinical/psychological examination results such as mini-mental state examination for dementia. The prediction results are of clinical grade, with the accuracy being at least 86% for two stages and about or above 90% for many other stages.
Read the Original
This page is a summary of: Machine Learning Approach Predicts Probability of Time to Stage-Specific Conversion of Alzheimer’s Disease, Journal of Alzheimer s Disease, November 2022, IOS Press, DOI: 10.3233/jad-220590.
You can read the full text:
The following have contributed to this page