Skip to content

BleuRadience/gbm-cart-spatial-model

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Spatial Fractional Reaction-Diffusion Model for Optimizing CAR-T Therapy in Glioblastoma

License: CC BY-NC-SA 4.0 Python 3.8+ DOI

Overview

This repository contains a comprehensive computational framework for modeling and optimizing CAR-T therapy in glioblastoma multiforme (GBM). The model integrates:

  • Fractional diffusion (α = 1.8) to capture anomalous invasion patterns
  • Tumor microenvironment barriers: Extracellular matrix (ECM), myeloid-derived suppressor cells (MDSCs), and acidic pH
  • Switchable CAR-T killing dynamics (inspired by UCSF E-SYNC platform)
  • IL-21 proliferation enhancement (based on MD Anderson trials)
  • Spatial dose optimization with entropy regularization

Key Features

  • Validated against clinical data: BraTS dataset with R² > 0.75
  • Comprehensive sensitivity analysis: Sobol indices for parameter importance
  • Statistical rigor: Bootstrap CI, hypothesis testing, effect size calculation
  • Numerical validation: Grid convergence study and mass conservation
  • Publication-ready figures: 7 figures with professional formatting
  • Fast execution: ~2 seconds per simulation on standard hardware

Installation

Requirements

  • Python 3.8+
  • NumPy ≥ 1.21.0
  • SciPy ≥ 1.7.0
  • Matplotlib ≥ 3.4.0

Quick Install

git clone https://github.com/bleuradience/gbm-cart-spatial-model.git
cd gbm-cart-spatial-model
pip install -r requirements.txt

Quick Start

Run basic simulation

python gbm_cart_model_fixed.py

Run complete publication analysis suite

# Quick test (15 minutes)
python publication_suite.py --mode quick

# Full analysis for publication (2-4 hours)
python publication_suite.py --mode full

Run individual analyses

python brats_validation.py        # Validate against clinical data
python sensitivity_analysis.py    # Parameter sensitivity
python statistical_analysis.py    # Virtual cohort statistics
python convergence_study.py       # Numerical convergence

Repository Structure

.
├── gbm_cart_model_fixed.py          # Main model implementation
├── swanson_baseline.py              # Baseline Swanson model
├── sensitivity_analysis.py          # Sobol sensitivity analysis
├── brats_validation.py              # Clinical data validation
├── statistical_analysis.py          # Cohort statistics
├── convergence_study.py             # Numerical validation
├── publication_suite.py             # Master analysis pipeline
├── references.bib                   # Complete bibliography
├── requirements.txt                 # Python dependencies
├── README.md                        # This file
├── LICENSE                          # CC BY-NC-SA 4.0
├── CITATION.cff                     # Citation metadata
│
├── figures/                         # Generated figures
│   ├── fig1_model_schematic.png
│   ├── fig2_validation_comparison.png
│   ├── fig3_tornado_plot.png
│   ├── fig4_statistical_cohorts.png
│   ├── fig5_convergence_study.png
│   └── ...
│
└── results/                         # Numerical results (.npz)
    ├── validation_results.npz
    ├── sensitivity_results.npz
    ├── statistical_results.npz
    └── convergence_results.npz

Model Description

Governing Equations

The model consists of five coupled partial differential equations:

Tumor cells (T):

∂T/∂t = D_T ∇^α T + r_T T(1-T) - k_CT C T (1-γT) η - h_T T

CAR-T cells (C):

∂C/∂t = D_C(x) ∇^α C + r_C C T/(T+0.5) - h_C C + δ(t) I(x)

Extracellular matrix (E):

∂E/∂t = D_E ∇² E + β_E - d_E E (1 + 0.1T)

MDSCs (M):

∂M/∂t = D_M ∇² M + T - 0.1M - d_M M

pH:

∂pH/∂t = D_pH ∇² pH - 0.1T + b_pH (7.4 - pH)

Where:

  • α = 1.8 (fractional diffusion exponent)
  • η = efficacy penalty from TME barriers
  • ∇^α = fractional Laplacian operator

Parameters

All parameters derived from published literature:

Parameter Value Source
D_T 0.001 cm²/day Swanson et al. 2008
r_T 0.012 day⁻¹ Rockne et al. 2010
k_CT 1.5 Tunable (E-SYNC)
r_C 0.27 day⁻¹ MD Anderson IL-21

See references.bib for complete citations.

Validation Results

Clinical Data Fit

  • Fractional model: R² = 0.78 ± 0.12
  • Swanson baseline: R² = 0.65 ± 0.15
  • Improvement: +20% better fit

Virtual Cohort (n=100)

  • CAR-T treatment: 48.3% ± 12.7% tumor reduction
  • Control: -5.2% ± 8.3% (tumor growth)
  • p-value: < 0.001
  • Cohen's d: 1.87 (large effect)

Computational Performance

  • Simulation time: 2.1 seconds (Nx=51, 1D)
  • Convergence: Verified at Nx=51
  • Mass conservation: < 0.05% error

Key Findings

  1. Fractional diffusion improves model accuracy by 20% over standard diffusion
  2. Tumor invasion rate (D_T) is most influential parameter (ΔEffect ~22%)
  3. CAR-T therapy shows significant benefit: p < 0.001, Cohen's d = 1.87
  4. TME barriers reduce efficacy by ~15% without interventions
  5. Optimal CAR-T dose: 0.45-0.55 (normalized units)

Citation

If you use this code in your research, please cite:

@software{harrison2026gbm,
  author    = {Harrison, Cassandra D.},
  title     = {Spatial Fractional Reaction-Diffusion Model for 
               Optimizing CAR-T Therapy in Glioblastoma},
  year      = {2026},
  publisher = {Zenodo},
  version   = {1.0.0},
  doi       = {10.5281/zenodo.18361371},
  url       = {https://github.com/bleuradience/gbm-cart-spatial-model}
}

License

CC BY-NC-SA 4.0 (Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International)

  • Allowed: Academic research, clinical research (non-commercial), education, personal use, non-profit use
  • Prohibited: Commercial use, selling this software or derivatives, pharmaceutical/biotech company use without separate license
  • 💼 Commercial licensing available - Contact for pharmaceutical partnerships, commercial integration, or consulting

This ensures the model remains freely available for research while protecting against unauthorized commercial exploitation.

See the LICENSE file for complete terms and commercial licensing contact information.

Contributing

Contributions are welcome! Please:

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/YourFeature)
  3. Commit your changes (git commit -m 'Add YourFeature')
  4. Push to the branch (git push origin feature/YourFeature)
  5. Open a Pull Request

See CONTRIBUTING.md for detailed guidelines.

Acknowledgments

  • BraTS Challenge for multi-institutional GBM imaging data
  • City of Hope (Dr. Christine Brown) for IL13Rα2 CAR-T clinical trials
  • UPenn (Dr. Donald O'Rourke) for EGFRvIII CAR-T trials
  • UCSF for E-SYNC switchable CAR-T platform
  • MD Anderson for IL-21 enhancement work

Contact

Cassandra D. Harrison, MBA, MPH
Principal Consultant, BleuConsult/AvaBleu Design LLC
bleuisresting@gmail.com

References

See references.bib for complete bibliography (30+ citations).

Key references:

  • Swanson et al. (2008). Cancer Research, 68(6), 1725–1731.
  • Brown et al. (2016). New England Journal of Medicine, 375(26), 2561–2563.
  • O'Rourke et al. (2017). Science Translational Medicine, 9(399), eaaa0984.
  • Quail & Joyce (2013). Nature Medicine, 19(11), 1423–1437.

Status: Ready for manuscript submission
Last Updated: January 2026
Version: 1.0.0