Beyond static biomarkers: systems biology and AI for decoding cancer dynamics

Area of Science:

• Quantitative Oncology
• Systems Biology
• Cancer Research

Background:

• Precision oncology relies on static biomarkers (mutations, receptor status) that fail to capture cancer's dynamic nature.
• Cancer exhibits complex behaviors like progression, metastasis, and relapse, driven by dynamic, multiscale processes.
• Current biomarkers are mismatched to the inherent stochasticity and adaptability of cancer.

Purpose of the Study:

• To propose a framework for understanding cancer as a dynamic, stochastic system.
• To advocate for biomarkers that predict cancer's future states, not just current status.
• To integrate diverse research areas into a unified model of cancer dynamics.

Main Methods:

• Developing a mathematically grounded framework using stochastic state-space models and layer-specific dynamical motifs.
• Distinguishing latent biological states from observable clinical data and addressing challenges of partial observability.
• Integrating six biological layers (molecular, cellular, spatial, ecosystem, metabolic-epigenetic, mechanobiological) and a clinical observation layer.

Main Results:

• Proposed a novel framework viewing cancer as a dynamic, coupled biological system.
• Identified key dynamic processes including cellular plasticity, spatial organization, and metabolic-epigenetic coupling.
• Highlighted the critical role of noise (transcriptional, ecological, treatment-induced, measurement) in cancer state dynamics.

Conclusions:

• Future biomarkers should estimate transition risk, system instability, and trajectory direction, moving beyond static classification.
• Artificial intelligence can aid inference in dynamic oncology, complementing mechanistic understanding.
• Translating dynamic oncology into a reproducible clinical discipline requires addressing regulatory and health-equity considerations.