Related Experiment Videos
Disease progression timeline estimation for Alzheimer's disease using discriminative event based modeling
Vikram Venkatraghavan1, Esther E Bron1, Wiro J Niessen2
1Biomedical Imaging Group Rotterdam, Departments of Medical Informatics & Radiology, Erasmus MC, University Medical Center Rotterdam, the Netherlands.
Neuroimage
|November 25, 2018
Summary
A new discriminative event-based modeling (EBM) approach accurately estimates Alzheimer's disease biomarker progression. This method improves patient staging and disease timeline estimation, offering a promising tool for early diagnosis and prognosis.
Area of Science:
- Biomedical Informatics
- Computational Biology
- Neuroscience
Background:
- Alzheimer's Disease (AD) progression involves complex biomarker changes.
- Understanding disease dynamics is crucial for early diagnosis and prognosis.
- Existing event-based modeling (EBM) methods have limitations in accuracy.
Purpose of the Study:
- To develop a novel discriminative approach to EBM for improved accuracy.
- To accurately estimate biomarker abnormality sequences in AD.
- To enhance patient staging and disease progression timeline creation.
Main Methods:
- Developed a discriminative EBM approach using generalized Mallows models.
- Introduced a novel probabilistic Kendall's Tau distance for ordering events.
- Created a disease progression timeline using relative event distances.
- Implemented a patient staging algorithm based on the timeline.
Main Results:
- The proposed discriminative EBM method demonstrated higher accuracy than state-of-the-art EBM.
- Event orderings on Alzheimer's Disease Neuroimaging Initiative (ADNI) data align with current AD understanding.
- Patient staging algorithm consistently outperformed existing methods.
- Simulations showed improved event ordering accuracy and correlation with actual disease progression.
Conclusions:
- Discriminative EBM is a highly accurate and promising approach for modeling disease progression.
- The method facilitates better understanding of AD biomarker dynamics.
- This approach can significantly aid in early AD diagnosis and prognosis.