A continuous-time Markov model approach for modeling myelodysplastic syndromes progression from cross-sectional data

Area of Science:

• Computational biology
• Bioinformatics
• Medical informatics

Background:

• Integrating genomics and clinical data enables disease progression modeling.
• Longitudinal data is crucial but often difficult to obtain for disease progression studies.
• Novel decision support tools are needed for precise patient risk stratification and tailored therapies.

Purpose of the Study:

• To develop a framework for simulating disease progression using cross-sectional data.
• To create patient trajectories and analyze disease evolution using Markov models.
• To investigate genomic factors associated with disease progression in Myelodysplastic syndromes (MDS).

Main Methods:

• Combined disease progression simulation from cross-sectional data with a continuous-time Markov model.
• Derived transition probabilities using Cox regression and patient similarity via matrix tri-factorization.
• Applied the framework to cross-sectional genomic and clinical data from MDS patients to predict Acute Myeloid Leukemia (AML) development.

Main Results:

• Generated patient trajectories across increasing AML risk stages.
• Identified genomic characteristics associated with disease progression probability using a Cox model.
• Validated findings against previous studies, confirming the utility of simulated longitudinal data.

Conclusions:

• Simulated longitudinal data is a valuable resource for studying disease progression in MDS.
• The developed framework supports the creation of decision support tools for personalized medicine.
• This approach enhances understanding of disease evolution and therapeutic intervention strategies.