Simulation of 3D centimeter-scale continuum tumor growth at sub-millimeter resolution via distributed computing

Area of Science:

• Computational biology
• Biophysics
• Medical imaging and simulation

Background:

• Simulating centimeter-scale tumor growth is computationally intensive, limiting current models to smaller scales.
• Previous models were insufficient for studying clinically relevant tumor sizes and vascularized tissue.

Purpose of the Study:

• To develop a computationally efficient method for simulating 3D cm-scale vascularized tumor growth at sub-millimeter resolution.
• To enable biologically relevant simulations of large tumors, addressing limitations of prior research.

Main Methods:

• A distributed computing (parallelized) implementation of a mixture model for tumor growth.
• Utilized a two-stage parallelization framework combining Message Passing Interface (MPI) and CUDA for GPU computation.
• Overcame single-system RAM and processing limitations through distributed computing.

Main Results:

• Successfully simulated 3D cm-scale vascularized tissue at sub-millimeter resolution.
• Demonstrated that the MPI-CUDA implementation enables continuum modeling of cm-scale tumors at reasonable computational cost.
• The parallelized approach significantly reduces computational expense compared to traditional methods.

Conclusions:

• The developed MPI-CUDA framework effectively simulates large-scale tumor growth.
• This computational approach paves the way for simulating patient-specific tumors for clinical applications.
• Further calibration of model parameters could enhance the clinical relevance of these simulations.