Search

Purpose: Radiation therapy treatment planning is a trial-and-error, often time-consuming process. An optimal dose distribution based on a specific anatomy can be predicted by pre-trained deep learning (DL) models. However, dose distributions are often optimized based on not only patient-specific anatomy but also physician preferred trade-offs between planning target volume (PTV) coverage and organ at risk (OAR) sparing. Therefore, it is desirable to allow physicians to fine-tune the dose distribution predicted based on patient anatomy. In this work, we developed a DL model to predict the individualized 3D dose distributions by using not only the anatomy but also the desired PTV/OAR trade-offs, as represented by a dose volume histogram (DVH), as inputs. Methods: The desired DVH, fine-tuned by physicians from the initially predicted DVH, is first projected onto the Pareto surface, then converted into a vector, and then concatenated with mask feature maps. The network output for training is the dose distribution corresponding to the Pareto optimal DVH. The training/validation datasets contain 77 prostate cancer patients, and the testing dataset has 20 patients. Results: The trained model can predict a 3D dose distribution that is approximately Pareto optimal. We calculated the difference between the predicted and the optimized dose distribution for the PTV and all OARs as a quantitative evaluation. The largest average error in mean dose was about 1.6% of the prescription dose, and the largest average error in the maximum dose was about 1.8%. Conclusions: In this feasibility study, we have developed a 3D U-Net model with the anatomy and desired DVH as inputs to predict an individualized 3D dose distribution. The predicted dose distributions can be used as references for dosimetrists and physicians to rapidly develop a clinically acceptable treatment plan.

We recently built an analytical source model for GPU-based MC dose engine. In this paper, we present a sampling strategy to efficiently utilize this source model in GPU-based dose calculation. Our source model was based on a concept of phase-space-ring (PSR). This ring structure makes it effective to account for beam rotational symmetry, but not suitable for dose calculations due to rectangular jaw settings. Hence, we first convert PSR source model to its phase-space let (PSL) representation. Then in dose calculation, different types of sub-sources were separately sampled. Source sampling and particle transport were iterated. So that the particles being sampled and transported simultaneously are of same type and close in energy to alleviate GPU thread divergence. We also present an automatic commissioning approach to adjust the model for a good representation of a clinical linear accelerator . Weighting factors were introduced to adjust relative weights of PSRs, determined by solving a quadratic minimization problem with a non-negativity constraint. We tested the efficiency gain of our model over a previous source model using PSL files. The efficiency was improved by 1.70 ~ 4.41, due to the avoidance of long data reading and transferring. The commissioning problem can be solved in ~20 sec. Its efficacy was tested by comparing the doses computed using the commissioned model and the uncommissioned one, with measurements in different open fields in a water phantom under a clinical Varian Truebeam 6MV beam. For the depth dose curves, the average distance-to-agreement was improved from 0.04~0.28 cm to 0.04~0.12 cm for build-up region and the root-mean-square (RMS) dose difference after build-up region was reduced from 0.32%~0.67% to 0.21%~0.48%. For lateral dose profiles, RMS difference was reduced from 0.31%~2.0% to 0.06%~0.78% at inner beam and from 0.20%~1.25% to 0.10%~0.51% at outer beam., Comment: page line has been removed

VMAT optimization is a computationally challenging problem due to its large data size, high degrees of freedom, and many hardware constraints. High-performance graphics processing units have been used to speed up the computations. However, its small memory size cannot handle cases with a large dose-deposition coefficient (DDC) matrix. This paper is to report an implementation of our column-generation based VMAT algorithm on a multi-GPU platform to solve the memory limitation problem. The column-generation approach generates apertures sequentially by solving a pricing problem (PP) and a master problem (MP) iteratively. The DDC matrix is split into four sub-matrices according to beam angles, stored on four GPUs in compressed sparse row format. Computation of beamlet price is accomplished using multi-GPU. While the remaining steps of PP and MP problems are implemented on a single GPU due to their modest computational loads. A H&N patient case was used to validate our method. We compare our multi-GPU implementation with three single GPU implementation strategies: truncating DDC matrix (S1), repeatedly transferring DDC matrix between CPU and GPU (S2), and porting computations involving DDC matrix to CPU (S3). Two more H&N patient cases and three prostate cases were also used to demonstrate the advantages of our method. Our multi-GPU implementation can finish the optimization within ~1 minute for the H&N patient case. S1 leads to an inferior plan quality although its total time was 10 seconds shorter than the multi-GPU implementation. S2 and S3 yield same plan quality as the multi-GPU implementation but take ~4 minutes and ~6 minutes, respectively. High computational efficiency was consistently achieved for the other 5 cases. The results demonstrate that the multi-GPU implementation can handle the large-scale VMAT optimization problem efficiently without sacrificing plan quality., Comment: 23 pages, 3 figures, 3 tables

Monte Carlo (MC) simulation is considered as the most accurate method for radiation dose calculations. Accuracy of a source model for a linear accelerator is critical for the overall dose calculation accuracy. In this paper, we presented an analytical source model that we recently developed for GPU-based MC dose calculations. A key concept called phase-space-ring (PSR) was proposed. It contained a group of particles that are of the same type and close in energy and radial distance to the center of the phase-space plane. The model parameterized probability densities of particle location, direction and energy for each primary photon PSR, scattered photon PSR and electron PSR. For a primary photon PSRs, the particle direction is assumed to be from the beam spot. A finite spot size is modeled with a 2D Gaussian distribution. For a scattered photon PSR, multiple Gaussian components were used to model the particle direction. The direction distribution of an electron PSRs was also modeled as a 2D Gaussian distribution with a large standard deviation. We also developed a method to analyze a phase-space file and derive corresponding model parameters. To test the accuracy of our linac source model, dose distributions of different open fields in a water phantom were calculated using our source model and compared to those directly calculated using the reference phase-space file. The average distance-to-agreement (DTA) was within 1 mm for the depth dose in the build-up region and beam penumbra regions. The root-mean-square (RMS) dose difference was within 1.1% for dose profiles at inner and outer beam regions. The maximal relative difference of output factors was within 0.5%. Good agreements were also found in an IMRT prostate patient case and an IMRT head-and-neck case. These results demonstrated the efficacy of our source model in terms of accurately representing a reference phase-space file., Comment: page line has been removed

Monte Carlo (MC) method has been recognized the most accurate dose calculation method for radiotherapy. However, its extremely long computation time impedes clinical applications. Recently, a lot of efforts have been made to realize fast MC dose calculation on GPUs. Nonetheless, most of the GPU-based MC dose engines were developed in NVidia CUDA environment. This limits the code portability to other platforms, hindering the introduction of GPU-based MC simulations to clinical practice. The objective of this paper is to develop a fast cross-platform MC dose engine oclMC using OpenCL environment for external beam photon and electron radiotherapy in MeV energy range. Coupled photon-electron MC simulation was implemented with analogue simulations for photon transports and a Class II condensed history scheme for electron transports. To test the accuracy and efficiency of our dose engine oclMC, we compared dose calculation results of oclMC and gDPM, our previously developed GPU-based MC code, for a 15 MeV electron beam and a 6 MV photon beam on a homogenous water phantom, one slab phantom and one half-slab phantom. Satisfactory agreement was observed in all the cases. The average dose differences within 10% isodose line of the maximum dose were 0.48-0.53% for the electron beam cases and 0.15-0.17% for the photon beam cases. In terms of efficiency, our dose engine oclMC was 6-17% slower than gDPM when running both codes on the same NVidia TITAN card due to both different physics particle transport models and different computational environments between CUDA and OpenCL. The cross-platform portability was also validated by successfully running our new dose engine on a set of different compute devices including an Nvidia GPU card, two AMD GPU cards and an Intel CPU card using one or four cores. Computational efficiency among these platforms was compared., Comment: page line has been removed

Simulation of x-ray projection images plays an important role in cone beam CT (CBCT) related research projects. A projection image contains primary signal, scatter signal, and noise. It is computationally demanding to perform accurate and realistic computations for all of these components. In this work, we develop a package on GPU, called gDRR, for the accurate and efficient computations of x-ray projection images in CBCT under clinically realistic conditions. The primary signal is computed by a tri-linear ray-tracing algorithm. A Monte Carlo (MC) simulation is then performed, yielding the primary signal and the scatter signal, both with noise. A denoising process is applied to obtain a smooth scatter signal. The noise component is then obtained by combining the difference between the MC primary and the ray-tracing primary signals, and the difference between the MC simulated scatter and the denoised scatter signals. Finally, a calibration step converts the calculated noise signal into a realistic one by scaling its amplitude. For a typical CBCT projection with a poly-energetic spectrum, the calculation time for the primary signal is 1.2~2.3 sec, while the MC simulations take 28.1~95.3 sec. Computation time for all other steps is negligible. The ray-tracing primary signal matches well with the primary part of the MC simulation result. The MC simulated scatter signal using gDRR is in agreement with EGSnrc results with a relative difference of 3.8%. A noise calibration process is conducted to calibrate gDRR against a real CBCT scanner. The calculated projections are accurate and realistic, such that beam-hardening artifacts and scatter artifacts can be reproduced using the simulated projections. The noise amplitudes in the CBCT images reconstructed from the simulated projections also agree with those in the measured images at corresponding mAs levels., Comment: 21 pages, 11 figures, 1 table

We used coherent light scattering in a multi-speckle detection scheme to investigate the mesoscale dynamics in aqueous foam. Time-resolved correlation of the scattered speckle intensities reveals the details of foam dynamics during aging. We introduce Temporal Contrast Analysis, a novel statistical tool that can be effective in characterizing structural rearrangements. Using Temporal Contrast Analysis we were able to detect two distinct dynamical components present during foam aging: spontaneous and intermittent, avalanche-like events and continuous, flow-like rearrangements in the foam structure. We were able to measure these contributions separately from the intrinsic statistical noise contribution, and thereby independently analyze the decay of each dynamical component during foam aging process.

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 towards the development a 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 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. High performance random number generator and hardware linear interpolation are also utilized. We have also developed various components to handle 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 found not 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 2.27GHz 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 sec using gDPM., Comment: 18 pages, 5 figures, and 3 tables

Recently we have developed an algorithm for reconstructing volumetric images and extracting 3D tumor motion information from a single x-ray projection. We have demonstrated its feasibility using a digital respiratory phantom with regular breathing patterns. In this work, we present a detailed description and a comprehensive evaluation of the improved algorithm. The algorithm was improved by incorporating respiratory motion prediction. The accuracy and efficiency were then evaluated on 1) a digital respiratory phantom, 2) a physical respiratory phantom, and 3) five lung cancer patients. These evaluation cases include both regular and irregular breathing patterns that are different from the training dataset. For the digital respiratory phantom with regular and irregular breathing, the average 3D tumor localization error is less than 1 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for 3D tumor localization from each projection ranges between 0.19 and 0.26 seconds, for both regular and irregular breathing, which is about a 10% improvement over previously reported results. For the physical respiratory phantom, an average tumor localization error below 1 mm was achieved with an average computation time of 0.13 and 0.16 seconds on the same GPU card, for regular and irregular breathing, respectively. For the five lung cancer patients, the average tumor localization error is below 2 mm in both the axial and tangential directions. The average computation time on the same GPU card ranges between 0.26 and 0.34 seconds.

Purpose: To develop an algorithm for real-time volumetric image reconstruction and 3D tumor localization based on a single x-ray projection image for lung cancer radiotherapy. Methods: Given a set of volumetric images of a patient at N breathing phases as the training data, we perform deformable image registration between a reference phase and the other N-1 phases, resulting in N-1 deformation vector fields (DVFs). These DVFs can be represented efficiently by a few eigenvectors and coefficients obtained from principal component analysis (PCA). By varying the PCA coefficients, we can generate new DVFs, which, when applied on the reference image, lead to new volumetric images. We then can reconstruct a volumetric image from a single projection image by optimizing the PCA coefficients such that its computed projection matches the measured one. The 3D location of the tumor can be derived by applying the inverted DVF on its position in the reference image. Our algorithm was implemented on graphics processing units (GPUs) to achieve real-time efficiency. We generated the training data using a realistic and dynamic mathematical phantom with 10 breathing phases. The testing data were 360 cone beam projections corresponding to one gantry rotation, simulated using the same phantom with a 50% increase in breathing amplitude. Results: The average relative image intensity error of the reconstructed volumetric images is 6.9% +/- 2.4%. The average 3D tumor localization error is 0.8 mm +/- 0.5 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for reconstructing a volumetric image from each projection is 0.24 seconds (range: 0.17 and 0.35 seconds). Conclusions: We have shown the feasibility of reconstructing volumetric images and localizing tumor positions in 3D in near real-time from a single x-ray image., Comment: 8 pages, 3 figures, submitted to Medical Physics Letter