Development of a GPU-accelerated Radiation Dose Calculation System, and its Application towards MRI-guided Radiation Therapy

Yuhe Wang, Department of Physics, Washington University
March 10, 2017 at 4:00 pm
204 Crow
Event Description 

Monte Carlo radiation transport simulation is generally considered to be the most accurate method for dose calculation in radiation therapy. However, it typically requires long simulation time (hours to days) to achieve adequate statistical uncertainty. In this talk, I’ll introduce how to rewrite a well-known dose calculation engine called PENELOPE on GPU platform (namely gPENELOPE), and show you some tricks to maximize the overall efficiency (10 times faster than CPU). In addition, a fancy graphic user interface will be introduced to greatly simplify users’ operation. Followed by are three important applications:

(1) gPENELOPE is incorporated to a MRI-guided radiation therapy system to verify that its treatment planning system is accurate and reliable through both experiment and simulation comparisons.

(2) gPENELOPE is exploited to evaluate the dose distortion caused by magnetic field in this system, and prove that the distortion is negligible with the default 3mm*3mm*3mm grid size and a weak (0.35T) magnetic field.

(3) gPENELOPE can serve as a quick verification tool for adaptive treatment planning in which experimental verification is not immediately applicable.