Bayesian networks in modeling leucocyte interplay following brain irradiation: A comprehensive framework

Area of Science:

• Immunology
• Radiation Oncology
• Computational Biology

Background:

• Understanding leucocyte subpopulation dynamics post-radiotherapy is vital for cancer research.
• Interest in advanced radiotherapy like proton therapy is growing.
• Rodent models are crucial for studying complex biological interactions.

Purpose of the Study:

• To develop a framework using Bayesian networks to uncover leucocyte subpopulation interactions after brain irradiation.
• To compare immune responses to X-ray and proton radiotherapy modalities.
• To identify causal relationships between radiation exposure and circulating leucocytes.

Main Methods:

• Utilized data from 96 mice subjected to X-ray or proton brain irradiation.
• Quantified leucocyte subpopulations in blood 12 hours post-irradiation.
• Employed Bayesian networks with Bayesian information criterion for causal structure learning and parameter estimation.

Main Results:

• Discovered novel interactions between NK-cells and neutrophils, and monocytes and T-CD4+ cells in the X-ray model.
• Identified interplay between T-CD4+ cells and neutrophils in the proton model.
• Observed heightened interactions between T-CD8+ cells and B cells for both radiation types, and strengthened T-CD4+/T-CD8+ interactions with protons.

Conclusions:

• Bayesian networks effectively revealed novel insights into immune response dynamics following radiation exposure.
• The study highlights the utility of causal discovery methods in understanding complex immunological changes.
• Findings contribute to advancing knowledge in radiation immunology and therapeutic strategies.