MOQUI: an open-source GPU-based Monte Carlo code for proton dose calculation with efficient data structure

Abstract

Objective: Monte Carlo (MC) codes are increasingly used for accurate radiotherapy dose calculation. In proton therapy, the accuracy of the dose calculation algorithm is expected to have a more significant impact than in photon therapy due to the depth-dose characteristics of proton beams. However, MC simulations come at a considerable computational cost to achieve statistically sufficient accuracy. There have been efforts to improve computational efficiency while maintaining sufficient accuracy. Among those, parallelizing particle transportation using graphic processing units (GPU) achieved significant improvements. Contrary to the central processing unit (CPU), a GPU has limited memory capacity and is not expandable. It is therefore challenging to score quantities with large dimensions requiring extensive memory. The objective of this study is to develop an open-source GPU-based MC package capable of scoring those quantities. Approach: We employed a hash-table, one of the key-value pair data structures, to efficiently utilize the limited memory of the GPU and score the quantities requiring a large amount of memory. With the hash table, only voxels interacting with particles will occupy memory, and we can search the data efficiently to determine their address. The hash-table was integrated with a novel GPU-based MC code, moqui. Main results: The developed code was validated against an MC code widely used in proton therapy, TOPAS, with homogeneous and heterogeneous phantoms. We also compared the dose calculation results of clinical treatment plans. The developed code agreed with TOPAS within 2%, except for the fall-off and regions, and the gamma pass rates of the results were >99% for all cases with a 2mm/2% criteria. Significance: We can score dose-influence matrix and dose-rate on a GPU for a 3-field H&N case with 10 GB of memory using moqui, which would require more than 100 GB of memory with the conventionally used array data structure.

Low-Dose Computed Tomography Scanning Protocols for Online Adaptive Proton Therapy of Head-and-Neck Cancers

“…Our framework for online adaptive proton therapy [ 8 ] is based on fast GPU–accelerated Monte Carlo (GPU-MC) calculations. For this study, the gPMC code was used [ 24 , 25 , 26 , 27 ], while the current version of the adaptive framework is based on a newly developed GPU-MC code with an efficient data structure: Moqui [ 28 ]. Contour propagation is realized by applying DIR of the planning CT to daily image, which for this study was vCT.…”

A fast analytical dose calculation approach for MRI-guided proton therapy

“…While MC simulations are widely considered the most accurate method to calculate proton dose distributions, they require considerable computation times. Dedicated MC implementations using simplified physics and running on multiple central processing units (CPUs) or GPUs have led to considerable reductions in computation times (Jia et al 2012, Wan Chan Tseung et al 2015, Souris et al 2016, Schiavi et al 2017, Lee et al 2022. However, to our knowledge, only one such accelerated MC can also consider the impact of magnetic fields in patients (Lysakovski et al 2021), the speed of which would still limit its use for inverse plan optimisation for online plan adaptations.…”

A GPU‐based fast Monte Carlo code that supports proton transport in magnetic field for radiation therapy

“…But in an optimization system, one needs to score dose‐influence matrix, and the required memory is equal to the number of voxels in the scoring grid times the number of beamlets(about 100GB), which is much larger than GPU memory. Fortunately, Lee et al 37 proposed an efficient data structure method, which only scores the voxels interacting with particles, to solve this problem. Furthermore, our code takes into account magnetic field effects on beam transport and dose distortions of proton beams, so it can be applied to future MRI‐guided proton therapy.…”