Computational Oncology of Chemotaxis-Driven Tumour--Immune Spatial Patterning and Stability

Technical Summary: Computational Oncology of Chemotaxis-Driven Tumour–Immune Spatial Patterning and Stability

Problem Statement
Solid-tumour progression is characterized by spatial heterogeneity, including patterns of immune infiltration, immune exclusion, and necrosis. While reaction–diffusion models have long described tumour growth and immune dynamics, the specific role of chemokine-mediated chemotaxis in generating spatially heterogeneous patterns (such as the transition between "hot" and "cold" tumours) requires rigorous mathematical and numerical investigation. The core problem addressed is the formulation and analysis of a minimal tumour–immune–chemokine system that couples logistic tumour growth, immune-mediated killing, chemokine production, and directed immune migration. The challenge lies in identifying the stability thresholds that separate homogeneous coexistence from chemotaxis-driven pattern formation, while ensuring that numerical simulations preserve biological constraints such as non-negativity of cell densities and no-flux tissue boundaries.

Methodology
The authors develop a computational oncology framework based on a reaction–diffusion–chemotaxis system defined on a bounded tissue domain $\Omega \subset \mathbb{R}^d$ with no-flux boundary conditions.

1. Mathematical Formulation:

   • Variables: The system tracks tumour-cell density ($T$), immune effector-cell density ($E$), and chemokine concentration ($A$).
   • Mechanisms: Tumour cells grow logistically and are killed by immune cells via contact ($ET$). Chemokines are produced by tumour cells and $ET$ contacts, recruiting immune cells and guiding their migration via chemotaxis ($-\chi \nabla \cdot (E \nabla A)$).
   • Nondimensionalization: The system is nondimensionalized to reduce the parameter space, yielding a system with variables $u$ (tumour), $v$ (immune), and $w$ (chemokine). Key dimensionless parameters include chemotactic sensitivity ($\xi$), immune supply ($\sigma_0, \sigma_1$), and decay rates ($\delta, \ell$).

2. Numerical Scheme:

   • Discretization: A conservative finite-volume method is employed. Diffusive fluxes are approximated using centered differences.
   • Chemotaxis Handling: The chemotactic flux is discretized using an upwind scheme to ensure robustness and positivity in regimes with steep gradients.
   • Time Integration: An implicit Backward Differentiation Formula (BDF) scheme (first or second order) is used to handle stiffness arising from diffusion and reaction terms.
   • Diagnostics: The scheme incorporates strict positivity monitoring, mass-balance diagnostics (verifying discrete conservation laws), and post-hoc residual calculations to distinguish genuine pattern formation from numerical artifacts.

3. Analytical Approach:

   • Well-posedness: The authors establish local classical solvability and prove that solutions remain non-negative. They derive a priori estimates, including a uniform bound on tumour density and global mass estimates for immune and chemokine variables.
   • Equilibrium Analysis: The paper identifies spatially homogeneous equilibria, specifically the tumour-free state and coexistence states.
   • Linear Stability: A linear stability analysis is performed around the coexistence equilibrium. The authors derive a mode-wise dispersion relation to determine how chemotactic sensitivity affects the stability of finite-wavelength perturbations.

Key Results

1. Tumour-Free Threshold: The tumour-free equilibrium ($u^*=0$) is linearly stable if the baseline immune supply exceeds the immune loss rate ($\sigma_0 > \delta$). Crucially, the authors demonstrate that chemotaxis does not influence this specific threshold; the stability of the tumour-free state is determined solely by the balance of immune recruitment and decay.
2. Coexistence Equilibria: The existence of a spatially homogeneous coexistence state is reduced to finding a root of a scalar equation $F(u)=0$. The paper establishes conditions under which a positive coexistence equilibrium exists (e.g., when $\sigma_0 < \delta$ and specific recruitment parameters are positive).
3. Chemotaxis-Driven Instability: The linear stability analysis of the coexistence state reveals that chemotaxis acts as a wavenumber-amplified coupling. While the spatially uniform mode ($k=0$) remains unaffected by chemotaxis, finite-wavelength modes can become unstable if the chemotactic sensitivity $\xi$ exceeds a critical value $\xi_c$. This leads to a Turing-type instability where spatially heterogeneous patterns emerge from a homogeneous state.
4. Numerical Verification: The proposed finite-volume scheme successfully verifies the theoretical thresholds. Numerical experiments confirm the predicted dominant unstable modes, sensitivity maps, and the preservation of positivity and mass balance. The residual diagnostics confirm that the observed patterns are consistent with the PDE system.

Significance and Claims
The paper claims to contribute to computational oncology in three specific ways:

1. Model Formulation: It formulates a minimal, analytically tractable reaction–diffusion–chemotaxis model that explicitly couples immune-mediated tumour killing, chemokine-mediated recruitment, and directed migration.
2. Analytical Thresholds: It derives interpretable thresholds for tumour-free control and chemotaxis-driven finite-wavelength instability, linking biological feedback mechanisms directly to stability and pattern-formation criteria.
3. Numerical Reliability: It implements a conservative finite-volume scheme with specific features (upwind chemotactic flux, no-flux boundary treatment, and rigorous diagnostics) designed to preserve the biological constraints of non-negativity and mass conservation.

The authors position this work as a bridge between mathematical analysis and numerical simulation, providing a framework to investigate how chemokine-mediated recruitment and chemotactic migration shape the spatial organization of tumour–immune interactions. The study is framed as a "caricature" of the biological transition between homogeneous coexistence and chemotaxis-driven heterogeneity (the "hot/cold" tumour dichotomy), offering a mechanistic basis for understanding immune exclusion and infiltration patterns without claiming to model specific clinical outcomes or propose new experimental therapies.