Predicting the Regulatory Dynamics of AML Disease Progression from Longitudinal Multi-Modal Clinical Data

Area of Science:

• Computational Biology
• Systems Biology
• Oncology

Background:

• Acute Myeloid Leukemia (AML) is a complex cancer with high mortality.
• Predictive models require longitudinal, multimodal patient data for accuracy.
• Understanding disease dynamics is crucial for effective treatment strategies.

Purpose of the Study:

• To develop a robust methodology for discovering disease progression dynamics in AML.
• To create predictive mathematical models of AML progression using a novel dataset.
• To identify key clinical, genetic, and treatment features influencing AML progression.

Main Methods:

• Analysis of a novel longitudinal, multimodal clinical dataset of AML patients.
• Development of a de novo inference algorithm based on evolutionary computation.
• Discovery of dynamic mathematical models including regulatory interactions and disease drivers.

Main Results:

• The methodology accurately estimated AML progression drivers and clinical dynamics (blast percentages).
• Predictions were validated on both training and novel patient data.
• The approach successfully leveraged heterogeneous and longitudinal patient data.

Conclusions:

• The developed methodology accurately predicts AML drivers and progression dynamics.
• This approach offers a flexible framework for modeling acute disease progression.
• Significant potential exists for advancing clinical and translational research in oncology.