Monte Carlo (MC) simulation is commonly considered to be the most accurate dose calculation method in radiotherapy. However, its efficiency still requires improvement for many routine clinical applications. In this paper, we present our recent progress toward the development of a graphics processing unit (GPU)-based MC dose calculation package, gDPM v2.0. It utilizes the parallel computation ability of a GPU to achieve high efficiency, while maintaining the same particle transport physics as in the original dose planning method (DPM) code and hence the same level of simulation accuracy. In GPU computing, divergence of execution paths between threads can considerably reduce the efficiency. Since photons and electrons undergo different physics and hence attain different execution paths, we use a simulation scheme where photon transport and electron transport are separated to partially relieve the thread divergence issue. A high-performance random number generator and a hardware linear interpolation are also utilized. We have also developed various components to handle the fluence map and linac geometry, so that gDPM can be used to compute dose distributions for realistic IMRT or VMAT treatment plans. Our gDPM package is tested for its accuracy and efficiency in both phantoms and realistic patient cases. In all cases, the average relative uncertainties are less than 1%. A statistical t-test is performed and the dose difference between the CPU and the GPU results is not found to be statistically significant in over 96% of the high dose region and over 97% of the entire region. Speed-up factors of 69.1 ∼ 87.2 have been observed using an NVIDIA Tesla C2050 GPU card against a 2.27 GHz Intel Xeon CPU processor. For realistic IMRT and VMAT plans, MC dose calculation can be completed with less than 1% standard deviation in 36.1 ∼ 39.6 s using gDPM.

A novel phase-space source implementation has been designed for graphics processing unit (GPU)-based Monte Carlo dose calculation engines. Short of full simulation of the linac head, using a phase-space source is the most accurate method to model a clinical radiation beam in dose calculations. However, in GPU-based Monte Carlo dose calculations where the computation efficiency is very high, the time required to read and process a large phase-space file becomes comparable to the particle transport time. Moreover, due to the parallelized nature of GPU hardware, it is essential to simultaneously transport particles of the same type and similar energies but separated spatially to yield a high efficiency. We present three methods for phase-space implementation that have been integrated into the most recent version of the GPU-based Monte Carlo radiotherapy dose calculation package gDPM v3.0. The first method is to sequentially read particles from a patient-dependent phase-space and sort them on-the-fly based on particle type and energy. The second method supplements this with a simple secondary collimator model and fluence map implementation so that patient-independent phase-space sources can be used. Finally, as the third method (called the phase-space-let, or PSL, method) we introduce a novel source implementation utilizing pre-processed patient-independent phase-spaces that are sorted by particle type, energy and position. Position bins located outside a rectangular region of interest enclosing the treatment field are ignored, substantially decreasing simulation time with little effect on the final dose distribution. The three methods were validated in absolute dose against BEAMnrc/DOSXYZnrc and compared using gamma-index tests (2%/2 mm above the 10% isodose). It was found that the PSL method has the optimal balance between accuracy and efficiency and thus is used as the default method in gDPM v3.0. Using the PSL method, open fields of 4 × 4, 10 × 10 and 30 × 30 cm(2) in water resulted in gamma passing rates of 99.96%, 99.92% and 98.66%, respectively. Relative output factors agreed within 1%. An intensity modulated radiation therapy patient plan using the PSL method resulted in a passing rate of 97%, and was calculated in 50 s (per GPU) compared to 8.4 h (per CPU) for BEAMnrc/DOSXYZnrc.

Cone beam CT (CBCT) has been widely used for patient setup in image guided radiation therapy (IGRT). Radiation dose from CBCT scans has become a clinical 20 concern. The purposes of this study are 1) to commission a GPU-based Monte Carlo (MC) dose calculation package gCTD for Varian On-Board Imaging (OBI) system and test the calculation accuracy, and 2) to quantitatively evaluate CBCT dose from the OBI system in typical IGRT scan protocols. We first conducted dose measurements in a water phantom. X-ray source model parameters used in gCTD 25 are obtained through a commissioning process. gCTD accuracy is demonstrated by comparing calculations with measurements in water and in CTDI phantoms. 25 brain cancer patients are used to study dose in a standard-dose head protocol, and 25 prostate cancer patients are used to study dose in pelvis protocol and pelvis spotlight protocol. Mean dose to each organ is calculated. Mean dose to 2% voxels 30 that have the highest dose is also computed to quantify the maximum dose. It is found that the mean dose value to an organ varies largely among patients. Moreover, dose distribution is highly non-homogeneous inside an organ. The maximum dose is found to be 1~3 times higher than the mean dose depending on the organ, and is up to 8 times higher for the entire body due to the very high dose 35 region in bony structures. High computational efficiency has also been observed in our studies, such that MC dose calculation time is less than 5 min for a typical case.

Using fiducial markers on the patient's body surface to predict the tumor location is a widely used approach in lung cancer radiotherapy. The purpose of this work is to propose an algorithm that automatically identifies a sparse set of locations on the patient's surface with the optimal prediction power for the tumor motion. In our algorithm, it is assumed that there is a linear relationship between the surface marker motion and the tumor motion. The sparse selection of markers on the external surface and the linear relationship between the marker motion and the internal tumor motion are represented by a prediction matrix. Such a matrix is determined by solving an optimization problem, where the objective function contains a sparsity term that penalizes the number of markers chosen on the patient's surface. Bregman iteration is used to solve the proposed optimization problem. The performance of our algorithm has been tested on realistic clinical data of four lung cancer patients. Thoracic 4DCT scans with ten phases are used for the study. On a reference phase, a grid of points are casted on the patient's surfaces (except for the patient's back) and propagated to other phases via deformable image registration of the corresponding CT images. Tumor locations at each phase are also manually delineated. We use nine out of ten phases of the 4DCT images to identify a small group of surface markers that are mostly correlated with the motion of the tumor and find the prediction matrix at the same time. The tenth phase is then used to test the accuracy of the prediction. It is found that on average six to seven surface markers are necessary to predict tumor locations with a 3D error of about 1 mm. It is also found that the selected marker locations lie closely in those areas where surface point motion has a large amplitude and a high correlation with the tumor motion. Our method can automatically select sparse locations on the patient's external surface and estimate a correlation matrix based on 4DCT, so that the selected surface locations can be used to place fiducial markers to optimally predict internal tumor motions.

Adaptive radiation therapy (ART) can reduce normal tissue toxicity and/or improve tumor control through treatment adaptations based on the current patient anatomy. Developing an efficient and effective re-planning algorithm is an important step toward the clinical realization of ART. For the re-planning process, manual trial-and-error approach to fine-tune planning parameters is time-consuming and is usually considered unpractical, especially for online ART. It is desirable to automate this step to yield a plan of acceptable quality with minimal interventions. In ART, prior information in the original plan is available, such as dose-volume histogram (DVH), which can be employed to facilitate the automatic re-planning process. The goal of this work is to develop an automatic re-planning algorithm to generate a plan with similar, or possibly better, DVH curves compared with the clinically delivered original plan. Specifically, our algorithm iterates the following two loops. An inner loop is the traditional fluence map optimization, in which we optimize a quadratic objective function penalizing the deviation of the dose received by each voxel from its prescribed or threshold dose with a set of fixed voxel weighting factors. In outer loop, the voxel weighting factors in the objective function are adjusted according to the deviation of the current DVH curves from those in the original plan. The process is repeated until the DVH curves are acceptable or maximum iteration step is reached. The whole algorithm is implemented on GPU for high efficiency. The feasibility of our algorithm has been demonstrated with three head-and-neck cancer IMRT cases, each having an initial planning CT scan and another treatment CT scan acquired in the middle of treatment course. Compared with the DVH curves in the original plan, the DVH curves in the resulting plan using our algorithm with 30 iterations are better for almost all structures. The re-optimization process takes about 30 s using our in-house optimization engine.