Tae Suk Suh

Compressed sensing based cone-beam computed tomography reconstruction with a first-order method

PURPOSE This article considers the problem of reconstructing cone-beam computed tomography (CBCT) images from a set of undersampled and potentially noisy projection measurements. METHODS The authors cast the reconstruction as a compressed sensing problem based onℓ1 norm minimization constrained by statistically weighted least-squares of CBCT projection data. For accurate modeling, the noise characteristics of the CBCT projection data are used to determine the relative importance of each projection measurement. To solve the compressed sensing problem, the authors employ a method minimizing total-variation norm, satisfying a prespecified level of measurement consistency using a first-order method developed by Nesterov. RESULTS The method converges fast to the optimal solution without excessive memory requirement, thanks to the method of iterative forward and back-projections. The performance of the proposed algorithm is demonstrated through a series of digital and experimental phantom studies. It is found a that high quality CBCT image can be reconstructed from undersampled and potentially noisy projection data by using the proposed method. Both sparse sampling and decreasing x-ray tube current (i.e., noisy projection data) lead to the reduction of radiation dose in CBCT imaging. CONCLUSIONS It is demonstrated that compressed sensing outperforms the traditional algorithm when dealing with sparse, and potentially noisy, CBCT projection views.

Fast compressed sensing-based CBCT reconstruction using Barzilai-Borwein formulation for application to on-line IGRT

PURPOSE Compressed sensing theory has enabled an accurate, low-dose cone-beam computed tomography (CBCT) reconstruction using a minimal number of noisy projections. However, the reconstruction time remains a significant challenge for practical implementation in the clinic. In this work, we propose a novel gradient projection algorithm, based on the Gradient-Projection-Barzilai-Borwein formulation (GP-BB), that handles the total variation (TV)-norm regularization-based least squares problem for the CBCT reconstruction in a highly efficient manner, with speed acceptable for routine use in the clinic. METHODS CBCT is reconstructed by minimizing an energy function consisting of a data fidelity term and a TV-norm regularization term. Both terms are simultaneously minimized by calculating the gradient projection of the energy function with the step size determined using an approximate Hessian calculation at each iteration, based on the Barzilai-Borwein formulation. To speed up the process, a multiresolution optimization is used. In addition, the entire algorithm was designed to run with a single graphics processing unit (GPU) card. To evaluate the performance, the Shepp-Logan numerical phantom, the CatPhan 600 physical phantom, and a clinically-treated head-and-neck patient were acquired from the TrueBeam™ system (Varian Medical Systems, Palo Alto, CA). For each scan, in total, 364 projections were acquired in a 200° rotation. The imager has 1024 × 768 pixels with 0.388 × 0.388-mm resolution. This was down-sampled to 512 × 384 pixels with 0.776 × 0.776-mm resolution for reconstruction. Evenly spaced angles were subsampled and used for varying the number of projections for the image reconstruction. To assess the performance of our GP-BB algorithm, we have implemented and compared with three compressed sensing-type algorithms, the two of which are popular and published (forward-backward splitting techniques), and the other one with a basic line-search technique. In addition, the conventional Feldkamp-Davis-Kress (FDK) reconstruction of the clinical patient data is compared as well. RESULTS In comparison with the other compressed sensing-type algorithms, our algorithm showed convergence in ≤30 iterations whereas other published algorithms need at least 50 iterations in order to reconstruct the Shepp-Logan phantom image. With the CatPhan phantom, the GP-BB algorithm achieved a clinically-reasonable image with 40 projections in 12 iterations, in less than 12.6 s. This is at least an order of magnitude faster in reconstruction time compared with the most recent reports utilizing GPU technology given the same input projections. For the head-and-neck clinical scan, clinically-reasonable images were obtained from 120 projections in 34-78 s converging in 12-30 iterations. In this reconstruction range (i.e., 120 projections) the image quality is visually similar to or better than the conventional FDK reconstructed images using 364 projections. This represents a dose reduction of nearly 67% (120∕364 projections) while maintaining a reasonable speed in clinical implementation. CONCLUSIONS In this paper, we proposed a novel, fast, low-dose CBCT reconstruction algorithm using the Barzilai-Borwein step-size calculation. A clinically viable head-and-neck image can be obtained within ∼34-78 s while simultaneously cutting the dose by approximately 67%. This makes our GP-BB algorithm potentially useful in an on-line image-guided radiation therapy (IGRT).

Evaluations of Poly(vinyl alcohol)/Alginate Hydrogels Cross-linked by γ-ray Irradiation Technique

In this work, we prepared hydrogels for wound dressing from a mixture of poly(vinyl alcohol) (PVA) and alginate using the60Co γ-ray irradiation technique. We examined the physical properties of these hydrogels, including gelation, water absorptivity, and gel strength, to evaluate the applicability of these hydrogels for wound dressings. The biocompatibility of these hydrogels was also evaluated in vitro, in cultures of mouse fibroblasts, and in vivo, by subcutaneous implantation studies in rats. The gel content and strength increased upon increasing the radiation dose and upon decreasing the concentration of alginate. The degree of swelling was inversely proportional to the gel content and strength. The degree of cytotoxicity of the γ-ray-treated hydrogels was ca. 60% compared to the (−) control (serum) after 1 day of incubation. When the incubations were prolonged up to 2 days, the toxicity of all the samples decreased remarkably and reached that of the control. Subcutaneous implantation studies in rats indicated that foreign body reactions occurring around the implanted hydrogels were moderate and became minimal upon increasing the implantation time.

Application of proton boron fusion reaction to radiation therapy: A Monte Carlo simulation study

Three alpha particles are emitted from the point of reaction between a proton and boron. The alpha particles are effective in inducing the death of a tumor cell. After boron is accumulated in the tumor region, the emitted from outside the body proton can react with the boron in the tumor region. An increase of the protons maximum dose level is caused by the boron and only the tumor cell is damaged more critically. In addition, a prompt gamma ray is emitted from the proton boron reaction point. Here, we show that the effectiveness of the proton boron fusion therapy was verified using Monte Carlo simulations. We found that a dramatic increase by more than half of the protons maximum dose level was induced by the boron in the tumor region. This increase occurred only when the protons maximum dose point was located within the boron uptake region. In addition, the 719 keV prompt gamma ray peak produced by the proton boron fusion reaction was positively detected. This therapy method features the advantages such as the application of Bragg-peak to the therapy, the accurate targeting of tumor, improved therapy effects, and the monitoring of the therapy region during treatment.

Tomographic image of prompt gamma ray from boron neutron capture therapy: A Monte Carlo simulation study

Purpose of paper is to confirm the feasibility of acquisition of three dimensional single photon emission computed tomography image from boron neutron capture therapy using Monte Carlo simulation. Prompt gamma ray (478 keV) was used to reconstruct image with ordered subsets expectation maximization method. From analysis of receiver operating characteristic curve, area under curve values of three boron regions were 0.738, 0.623, and 0.817. The differences between length of centers of two boron regions and distance of maximum count points were 0.3 cm, 1.6 cm, and 1.4 cm.

Optimisation of dose distribution for linear accelerator-based stereotactic radiosurgery

The work presented in the paper addresses a method for obtaining the optimal dose distribution for LINAC-based stereotactic radiosurgery. As many targets have nonspherical or irregular shapes and three-dimensional dose calculations included in dose optimisation, long computation times are required to determine the optimum isocentre separation and collimator sizes to shape the irregular target using the multiple-isocentre approach, by trial-and-error types of method. The simple approach, using a computer-aided design optimisation technique and a fast approximate dose model, has been developed to find the optimum isocentre positions and collimator sizes quickly and automatically. A spherical dose model has been developed to represent the dose for a standard arc system with a single isocentre. The implementation of computer-aided design algorithms with the spherical dose model and their application to several cases are discussed. It is shown that the spherical dose model gives dose distribution similar to that of the exact dose model, which makes this simple dose model more efficient, with computer-aided design optimisation, in finding optimum isocentre positions and collimator sizes used in stereotactic radiosurgery.

Dosimetry in an IMRT phantom designed for a remote monitoring program.

An accurate delivery of prescribed dose is essential to ensure the most successful outcome from advanced radiation treatments such as intensity modulated radiation therapy (IMRT). An anthropomorphic phantom was designed and constructed to conduct a remote-audit program for IMRT treatments. The accuracy of the dosimetry in the phantom was assessed by comparing the results obtained from different detectors with those from Monte Carlo calculations. The developed phantom has a shape of a cylinder with one target and three organs at risk (OARs) inside the unit. The target and OARs were shaped similar to those of nasopharyngeal cancer patients, and manufactured for their identification during computed tomography imaging. The phantom was designed with thermoluminescent dosimeter (TLD) holders that were inserted inside the target and the OARs for the measurements of absolute dose. In addition, the phantom allowed measurements with ionization chambers placed at the TLD locations. As a result, an inter-comparison between the two types of dosimeters was possible. For the measurement of the relative dose distribution across the target and OARs, two film slots were orthogonally placed near the center of the phantom, which also enabled the insertion of inhomogeneities near the target. Measurements with TLDs, provided by Korea Food and Drug Administration and Radiological Physics Center, and measurements with an ionization chamber (IC) were performed in four cases. The first case was one anterior field of 6 MV x rays delivered to the phantom; the second case used the same anterior field, but with inhomogeneities inserted into the phantom. The third case was three fields of 6 MV beams at an equi-gantry angle for the homogeneous phantom, and the fourth case was IMRT delivery to the phantom without inhomogeneities. For each case, measurements with both TLDs and IC were performed. For cases 1-3, theoretical predictions were obtained by using the Monte Carlo (MC) codes (BEAMnrc and DOSXYZnrc6.0). The TLD measurements were larger than the IC readings by 2.2% (1.3-2.5%), 2.2% (1.2-2.9%), and 2.1% (0-3.3%) on average for case 1, case 2 and case 3, respectively. The average deviation between TLDs and MC results was 0.97% (-0.13-2.07%) for the first case, 1.27% (0.34-2.18%) for the second case, and 1.13% (0.31-1.94%) for the third case. The IC reading was less than the MC results; the average deviations were -1.2% (-2.44--0.43%), -0.96% (-1.74 - -0.54%) and -0.94% (-1.53-0.27%) for the first, second, and third cases, respectively. For the IMRT case, the average deviation between IC readings and TLD measurements was 0.5% (-7.0-3.9%). In conclusion, the TLD measurements in the developed phantom agreed with IC and MC results with less than 3% of an average difference. The developed phantom with TLD dosimeters should be useful for remote monitoring of IMRT.

Monte Carlo analyses of X-ray absorption, noise and detective quantum efficiency considering therapeutic X-ray spectrum in portal imaging detector

The bremsstrahlung spectrum from 6 MV linear accelerator (LINAC) was obtained and used as an input X-ray source in the simulation to estimate several important physical quantities of the detector in therapeutic X-ray portal imaging system, such as quantum and energy absorption efficiencies, Swank factor and the detective quantum efficiency (DQE) etc. In addition, the authors have obtained a spatial distribution of energy deposit within the detector, resulting in a spatial-frequency dependent DQE. From the simulation results, it is found that the use of metal plate largely enhances the energy absorption, leading to the large output signal. However, considering the noise properties, the presence of metal plate degrades the DQE at non-zero spatial frequencies because the detector absorption noise is dominated by the quantum absorption. For the verification of their simulation results, the authors have compared with experimental measurements and results of other literatures.

Prompt gamma ray imaging for verification of proton boron fusion therapy: A Monte Carlo study

PURPOSE The purpose of this study was to verify acquisition feasibility of a single photon emission computed tomography image using prompt gamma rays for proton boron fusion therapy (PBFT) and to confirm an enhanced therapeutic effect of PBFT by comparison with conventional proton therapy without use of boron. METHODS Monte Carlo simulation was performed to acquire reconstructed image during PBFT. We acquired percentage depth dose (PDD) of the proton beams in a water phantom, energy spectrum of the prompt gamma rays, and tomographic images, including the boron uptake region (BUR; target). The prompt gamma ray image was reconstructed using maximum likelihood expectation maximisation (MLEM) with 64 projection raw data. To verify the reconstructed image, both an image profile and contrast analysis according to the iteration number were conducted. In addition, the physical distance between two BURs in the region of interest of each BUR was measured. RESULTS The PDD of the proton beam from the water phantom including the BURs shows more efficient than that of conventional proton therapy on tumour region. A 719keV prompt gamma ray peak was clearly observed in the prompt gamma ray energy spectrum. The prompt gamma ray image was reconstructed successfully using 64 projections. Different image profiles including two BURs were acquired from the reconstructed image according to the iteration number. CONCLUSION We confirmed successful acquisition of a prompt gamma ray image during PBFT. In addition, the quantitative image analysis results showed relatively good performance for further study.

Experimental investigation of a moving averaging algorithm for motion perpendicular to the leaf travel direction in dynamic MLC target tracking

PURPOSE In dynamic multileaf collimator (MLC) motion tracking with complex intensity-modulated radiation therapy (IMRT) fields, target motion perpendicular to the MLC leaf travel direction can cause beam holds, which increase beam delivery time by up to a factor of 4. As a means to balance delivery efficiency and accuracy, a moving average algorithm was incorporated into a dynamic MLC motion tracking system (i.e., moving average tracking) to account for target motion perpendicular to the MLC leaf travel direction. The experimental investigation of the moving average algorithm compared with real-time tracking and no compensation beam delivery is described. METHODS The properties of the moving average algorithm were measured and compared with those of real-time tracking (dynamic MLC motion tracking accounting for both target motion parallel and perpendicular to the leaf travel direction) and no compensation beam delivery. The algorithm was investigated using a synthetic motion trace with a baseline drift and four patient-measured 3D tumor motion traces representing regular and irregular motions with varying baseline drifts. Each motion trace was reproduced by a moving platform. The delivery efficiency, geometric accuracy, and dosimetric accuracy were evaluated for conformal, step-and-shoot IMRT, and dynamic sliding window IMRT treatment plans using the synthetic and patient motion traces. The dosimetric accuracy was quantified via a tgamma-test with a 3%/3 mm criterion. RESULTS The delivery efficiency ranged from 89 to 100% for moving average tracking, 26%-100% for real-time tracking, and 100% (by definition) for no compensation. The root-mean-square geometric error ranged from 3.2 to 4.0 mm for moving average tracking, 0.7-1.1 mm for real-time tracking, and 3.7-7.2 mm for no compensation. The percentage of dosimetric points failing the gamma-test ranged from 4 to 30% for moving average tracking, 0%-23% for real-time tracking, and 10%-47% for no compensation. CONCLUSIONS The delivery efficiency of moving average tracking was up to four times higher than that of real-time tracking and approached the efficiency of no compensation for all cases. The geometric accuracy and dosimetric accuracy of the moving average algorithm was between real-time tracking and no compensation, approximately half the percentage of dosimetric points failing the gamma-test compared with no compensation.