Cancer Nanoparticle Therapy Optimization via PhysiCell Agent-Based Simulation

Note on Codebase & Modifications

This project builds upon the pc4nanobio framework, an agent-based modeling extension of PhysiCell.

Each therapy strategy (baseline, metronomic, single_high_dose, adaptive_therapy) retains the original repository structure.

The primary modifications for each experiment were limited to the following core files:

• src/custom_modules/nanobio.cpp
  → Implemented therapy scheduling logic, dosing strategies, and adaptive ON/OFF control

• data/PhysiCell_settings.xml
  → Tuned simulation parameters including EC50, Hill coefficient, and nanoparticle boundary conditions

All other files follow the standard pc4nanobio implementation.

This project uses pc4nanobio, an agent-based modeling framework built on PhysiCell, to simulate four distinct NP therapy strategies on a 3D tumor spheroid over 30 days. The core simulation logic is including therapy scheduling, pharmacodynamic modeling, and adaptive dosing triggers . This was implemented by directly modifying nanobio.cpp and PhysiCell_Settings.xml in the pc4nanobio source and adding the therapy logic in the source code.

Question: Which NP delivery schedule achieves the greatest tumor suppression while minimizing regrowth?

Figure 1. Tumor spheroid snapshots across four nanoparticle therapy strategies simulated over 30 days using pc4nanobio (PhysiCell v1.10.4). Each row represents one therapeutic strategy; columns correspond to Days 3, 10, 15, 18, 21, and 30. Blue: viable tumor cells. Orange/Brown: drug-stressed or apoptotic cells. Pill icons indicate active dosing events. In the Adaptive strategy (bottom row), ON/OFF labels denote dynamic therapy activation based on live cell count thresholds (activate at >850 cells, deactivate at <600 cells). Metronomic therapy achieves the greatest tumor reduction by Day 30; Single High Dose shows rapid early kill followed by strong regrowth.

Drug-Induced Apoptosis Modeling

Drug-induced apoptosis was modeled using two formulations depending on the strategy:

AUC-based model (Baseline, Metronomic, Adaptive)

$$ \text{Drug Effect} = \frac{AUC^{\text{HillPower}}}{AUC^{\text{HillPower}} + EC50^{\text{HillPower}}} $$

Instantaneous concentration model (Single High Dose)

$$ \text{Drug Effect} = \frac{C^{\text{HillPower}}}{C^{\text{HillPower}} + EC50^{\text{HillPower}}} $$

Key Parameters by Strategy

Parameter	Baseline	Metronomic	Single High Dose	Adaptive
EC50	0.1	0.15	0.15	0.025
Hill Power	1.5	1.5	1.8	1.5
Initial Tumor Size	~571 cells	~571 cells	~571 cells	~1,285 cells
Simulation Duration	25 days	25 days	25 days	40 days

Microenvironment

The tumor microenvironment tracked two diffusible substrates governing cell viability and drug response:

Substrate	Role	Diffusion Coeff. (μm²/min)	Decay Rate (/min)	Boundary Condition
Oxygen (O₂)	Supports viability	1×10⁵	0.1	Constant Dirichlet
Nanoparticle (NP1)	Triggers apoptosis via internalization	1×10³	0.001	Dirichlet during therapy events

Results

Strategy	Outcome
Metronomic	Lowest final tumor burden. Sustained low-dose exposure prevented regrowth most effectively.
Adaptive	Largest initial tumor (~1,285 cells), greatest % reduction, but stabilized at a higher residual size.
Baseline	Moderate suppression with some regrowth between doses.
Single High Dose	Rapid initial kill followed by strong tumor regrowth — worst long-term outcome.

Key finding: Frequent low-dose metronomic scheduling outperformed a single aggressive dose, consistent with the hypothesis that sustained NP circulation maintains higher intracellular drug accumulation over time.

Repository Structure

Cancer-nanoparticle-optimization/
├── baseline/               # Baseline therapy simulation files & output
├── metronomic/             # Metronomic therapy setup & outputs
├── single_high_dose/       # Single high-dose simulation & outputs
├── adaptive_therapy/       # Adaptive therapy simulation code & data
├── POSTER.pptx             # Academic poster presented at IU Bloomington
├── Therapy_Progress.pdf    # Simulation output analysis report
├── Macklin_pc4nanobio_reference.pdf  # Reference paper (Wang et al., 2024)
└── README.md