We employ large-scale computer simulations to model the real world at different spatial-temporal scales. For radiation transport simulation, Monte Carlo (MC) simulation is commonly regarded as the most accurate method. However, the high computational burden limits its wide applications. Over the years, we have performed systematic developments of Graphics Processing Unit (GPU)-based ultra-fast MC simulation tools on different spatial scales (from nm to m), over various energy ranges (from eV to GeV), and of popular particle types (photon, electron, proton, and heavy ion) and phantom geometries (voxelized geometry and parameterized geometry). With the GPU platform and novel parallelization schemes, substantial acceleration factors (typically over 100x) over conventional CPU based MC simulations have been achieved. Complex simulations were performed for a variety of clinical applications, including but not limited to x-ray imaging, radiation dose calculation, cone-beam CT scattering, and radiotherapy treatment planning. We also perform simulations to facilitate the development, prototyping, evaluation, regulation, translation and refinement of AI-based medical practice, especially medical imaging-guided diagnosis and therapy.
Understanding the impacts of radiation in biological systems requires simulation of not only radiation transport, but also the interactions between radiation and the biological systems at multiple spatial and temporal scales. To this end, we have extended our GPU-based MC expertise to developing a microscopic radiation transport simulation tool gMicroMC. Coupled with a multi-scale DNA model, we study DNA damages caused by radiation under various spatial and temporal modulations.
Funding Support:
- “Adversarially Based Virtual CT Workflow for Evaluation of AI in Medical Imaging.” NIH/NIBIB, R01EB032716, 01/2021-12/2025.
- “Monte Carlo based biological treatment plan optimization,” Project#2 of Multi-investigator Research Awards “Towards Carbon Beam Stereotactic Body Radiation Therapy (C-SBRT) for Higher Risk Early Stage Lung Cancer,” Cancer Prevention and Research Institute of Texas RP160661, 08/2016 – 07/2021.
