Bongyong Song

Network Duality for Multiuser MIMO Beamforming Networks and Applications

For any multiple-input multiple-output (MIMO) network with linear beamformers, it is shown that there exists a dual network that attains the same signal-to-interference-plus-noise ratio (SINR) performance. This network duality concept is a generalization of the virtual uplink concept investigated by Rashid-Farrokhi . We first develop the dual relation using a generic duality theory in linear programming. This approach naturally leads to the construction of a dual network, and provides machinery for handling a generalized cost function. We then consider the joint MIMO beamforming and power control problem with individual SINR constraints. We apply the network duality to this problem and propose a high-performance iterative algorithm which, compared with past centralized approaches, has improved convergence behavior. Finally, we propose a simpler distributed version of the algorithm tailored for a cellular downlink. This algorithm can be readily implemented in a distributed fashion, since it does not require information exchange between base stations. The algorithm achieves performance close to that of centralized ones, and outperforms other decentralized approaches available in the literature

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).

Liver motion during cone beam computed tomography guided stereotactic body radiation therapy

PURPOSE Understanding motion characteristics of liver such as, interfractional and intrafractional motion variability, difference in motion within different locations in the organ, and their complex relationship with the breathing cycles are particularly important for image-guided liver SBRT. The purpose of this study was to investigate such motion characteristics based on fiducial markers tracked with the x-ray projections of the CBCT scans, taken immediately prior to the treatments. METHODS Twenty liver SBRT patients were analyzed. Each patient had three fiducial markers (2 × 5-mm gold) percutaneously implanted around the gross tumor. The prescription ranged from 2 to 8 fractions per patient. The CBCT projections data for each fraction (∼650 projections∕scan), for each patient, were analyzed and the 2D positions of the markers were extracted using an in-house algorithm. In total, >55 000 x-ray projections were analyzed from 85 CBCT scans. From the 2D extracted positions, a 3D motion trajectory of the markers was constructed, from each CBCT scans, resulting in left-right (LR), anterior-posterior (AP), and cranio-caudal (CC) location information of the markers with >55 000 data points. The authors then analyzed the interfraction and intrafraction liver motion variability, within different locations in the organ, and as a function of the breathing cycle. The authors also compared the motion characteristics against the planning 4DCT and the RPM™ (Varian Medical Systems, Palo Alto, CA) breathing traces. Variations in the appropriate gating window (defined as the percent of the maximum range at which 50% of the marker positions are contained), between fractions were calculated as well. RESULTS The range of motion for the 20 patients were 3.0 ± 2.0 mm, 5.1 ± 3.1 mm, and 17.9 ± 5.1 mm in the planning 4DCT, and 2.8 ± 1.6 mm, 5.3 ± 3.1 mm, and 16.5 ± 5.7 mm in the treatment CBCT, for LR, AP, and CC directions, respectively. The range of respiratory period was 3.9 ± 0.7 and 4.2 ± 0.8 s during the 4DCT simulation and the CBCT scans, respectively. The authors found that breathing-induced AP and CC motions are highly correlated. That is, all markers moved cranially also moved posteriorly and vice versa, irrespective of the location. The LR motion had a more variable relationship with the AP∕CC motions, and appeared random with respect to the location. That is, when the markers moved toward cranial-posterior direction, 58% of the markers moved to the patient-right, 22% of the markers moved to the patient-left, and 20% of the markers had minimal∕none motion. The absolute difference in the motion magnitude between the markers, in different locations within the liver, had a positive correlation with the absolute distance between the markers (R(2) = 0.69, linear-fit). The interfractional gating window varied significantly for some patients, with the largest having 29.4%-56.4% range between fractions. CONCLUSIONS This study analyzed the liver motion characteristics of 20 patients undergoing SBRT. A large variation in motion was observed, interfractionally and intrafractionally, and that as the distance between the markers increased, the difference in the absolute range of motion also increased. This suggests that marker(s) in closest proximity to the target be used.

Adaptive filtering algorithms for promoting sparsity

We provide a mathematical framework for developing adaptive filtering algorithms for exploiting/enforcing sparsity. The approach is based on minimizing a regularized mean squared error criterion with sparsity being promoted by the regularizing term which consists of a diversity measure. A steepest descent algorithm (SDA) is developed to minimize the regularized cost function. Then we extend the algorithm to the adaptive environment and develop a class of algorithms, which we term the pLMS algorithm class and which incudes important variants - pLLMS (leaky pLMS) and pNLMS (normalized pLMS). The framework is quite general and encompasses a broad range of adaptive algorithms with the pNLMS having similarity with the proportionate normalized least-mean-squares (PNLMS) algorithm. Computer simulations have been conducted using the echo canceller application as an example of a sparse environment. The simulations clearly show the ability of the developed algorithms to exploit the inherent sparsity structure, thereby outperforming conventional algorithms like the NLMS algorithm in this application.

Ultra-Fast Digital Tomosynthesis Reconstruction Using General-Purpose GPU Programming for Image-Guided Radiation Therapy

The purpose of this work is to demonstrate an ultra-fast reconstruction technique for digital tomosynthesis (DTS) imaging based on the algorithm proposed by Feldkamp, Davis, and Kress (FDK) using standard general-purpose graphics processing unit (GPGPU) programming interface. To this end, the FDK-based DTS algorithm was programmed “in-house” with C language with utilization of 1) GPU and 2) central processing unit (CPU) cards. The GPU card consisted of 480 processing cores (2 × 240 dual chip) with 1,242 MHz processing clock speed and 1,792 MB memory space. In terms of CPU hardware, we used 2.68 GHz clock speed, 12.0 GB DDR3 RAM, on a 64-bit OS. The performance of proposed algorithm was tested on twenty-five patient cases (5 lung, 5 liver, 10 prostate, and 5 head-and-neck) scanned either with a full-fan or half-fan mode on our cone-beam computed tomography (CBCT) system. For the full-fan scans, the projections from 157.5°–202.5° (45°-scan) were used to reconstruct coronal DTS slices, whereas for the half-fan scans, the projections from both 157.5°–202.5° and 337.5°–22.5° (2 × 45°-scan) were used to reconstruct larger FOV coronal DTS slices. For this study, we chose 45°-scan angle that contained ~80 projections for the full-fan and ~160 projections with 2 × 45°-scan angle for the half-fan mode, each with 1024 × 768 pixels with 32-bit precision. Absolute pixel value differences, profiles, and contrast-to-noise ratio (CNR) calculations were performed to compare and evaluate the images reconstructed using GPU- and CPU-based implementations. The time dependence on the reconstruction volume was also tested with (512 × 512) × 16, 32, 64, 128, and 256 slices. In the end, the GPU-based implementation achieved, at most, 1.3 and 2.5 seconds to complete full reconstruction of 512 × 512 × 256 volume, for the full-fan and half-fan modes, respectively. In turn, this meant that our implementation can process > 13 projections-per-second (pps) and > 18 pps for the full-fan and half-fan modes, respectively. Since commercial CBCT system nominally acquires 11 pps (with 1 gantry-revolution-per-minute), our GPU-based implementation is sufficient to handle the incoming projections data as they are acquired and reconstruct the entire volume immediately after completing the scan. In addition, on increasing the number of slices (hence volume) to be reconstructed from 16 to 256, only minimal increases in reconstruction time were observed for the GPU-based implementation where from 0.73 to 1.27 seconds and 1.42 to 2.47 seconds increase were observed for the full-fan and half-fan modes, respectively. This resulted in speed improvement of up to 87 times compared with the CPU-based implementation (for 256 slices case), with visually identical images and small pixel-value discrepancies (< 6.3%), and CNR differences (< 2.3%). With this achievement, we have shown that time allocation for DTS image reconstruction is virtually eliminated and that clinical implementation of this approach has become quite appealing. In addition, with the speed achievement, further image processing and real-time applications that was prohibited prior due to time restrictions can now be tempered with.

Dynamic modulated brachytherapy (DMBT) for rectal cancer

PURPOSE All forms of past and current high-dose-rate brachytherapy utilize immobile applicators during treatment delivery. The only moving part is the source itself. This paradigm misses an important degree of freedom that, if explored, can in some instances produce previously unachievable dose conformality; that is, the dynamic motion of the applicator itself during treatment delivery. Monte Carlo and treatment planning simulations were used to illustrate the potential benefits of moving applicators for rectal cancer applications in particular. This concept is termed dynamic modulated brachytherapy (DMBT). METHODS The DMBT system uses a high-density, 18.0 g∕cm(3), 45 mm long tungsten alloy shield, cylindrical in shape, with a small window on one side to encapsulate a (192)Ir source, to create collimation that results in a highly directional beam profile. This shield can be dynamically translated and rotated, using an attached robotic arm, during treatment to create a volumetric modulated arc therapy-type delivery, but from inside the rectal cavity. Monte Carlo simulations and planning optimization algorithms were developed inhouse to evaluate the effectiveness of this new approach using 36 clinical treatment plans comprised of 13 patients each treated using the intracavitary mold applicator (ICMA, Nucletron, The Netherlands) to quantify the potential clinical benefit. The prescription dose was 10 Gy∕fx and the group had an average clinical target volume of 9.0 ± 3.5 cm(3). Ideal phantom geometries were used to evaluate the impact of various shield dimensions and designs on the resulting plan quality. RESULTS Simulations of ideal phantom geometries found that shields as small as 10 mm in diameter can produce high quality plans. For the clinical patient cases, compared to the ICMA, for equal prescription tumor coverage, the DMBT plans provided >30% decrease in D(5) (high dose volume) resulting in a ∼40% decrease in dose heterogeneity index. In addition, mean dose and D(98) showed a reduction (typically 40%-60%) on all critical structures evaluated. However, for a 10 Gy prescribed dose there was an increase in total treatment time on average from 7.6 to 20.8 min for a source with an air-kerma strength of 40.25 kU (10 Ci). CONCLUSIONS Dosimetric properties of a novel DMBT system have been described and evaluated. Comparison with the ICMA commercial applicator has shown it to be a prospective step forward in high-dose-rate brachytherapy (192)Ir technology. Dynamic motion of an applicator during treatment, for any applicator and site in general, can provide additional degrees of freedom that, if properly considered, can potentially increase the plan quality significantly.

Network duality and its application to multi-user MIMO wireless networks with SINR constraints

For any multiple-input and multiple-output (MIMO) network with linear beamformers, there exists a dual network that attains the same signal to interference plus noise ratio (SINR) performance. This network duality is a generalization of the virtual uplink concept investigated in R. Rashid-Farrokhi et al. (1998) in the context of cellular networks. In this paper, we develop network duality which is applicable to arbitrary multi-user MIMO networks with a generalized cost function. More importantly, we provide an optimization theoretic perspective of network duality which naturally leads to the construction of a dual network. We then consider the joint MIMO beamforming and power control problem with individual SINR constraints. We apply the network duality to this problem and propose a high performance algorithm which compared to past approaches has improved convergence behavior.

Direction-Modulated Brachytherapy for High-Dose-Rate Treatment of Cervical Cancer. I: Theoretical Design

PURPOSE To demonstrate that utilization of the direction-modulated brachytherapy (DMBT) concept can significantly improve treatment plan quality in the setting of high-dose-rate (HDR) brachytherapy for cervical cancer. METHODS AND MATERIALS The new, MRI-compatible, tandem design has 6 peripheral holes of 1.3-mm diameter, grooved along a nonmagnetic tungsten-alloy rod (ρ = 18.0 g/cm(3)), enclosed in Delrin tubing (polyoxymethylene, ρ = 1.41 g/cm(3)), with a total thickness of 6.4 mm. The Monte Carlo N-Particle code was used to calculate the anisotropic (192)Ir dose distributions. An in-house-developed inverse planning platform, geared with simulated annealing and constrained-gradient optimization algorithms, was used to replan 15 patient cases (total 75 plans) treated with a conventional tandem and ovoids (T&O) applicator. Prescription dose was 6 Gy. For replanning, we replaced the conventional tandem with that of the new DMBT tandem for optimization but left the ovoids in place and kept the dwell positions as originally planned. All DMBT plans were normalized to match the high-risk clinical target volume V100 coverage of the T&O plans. RESULTS In general there were marked improvements in plan quality for the DMBT plans. On average, D2cc for the bladder, rectum, and sigmoid were reduced by 0.59 ± 0.87 Gy (8.5% ± 28.7%), 0.48 ± 0.55 Gy (21.1% ± 27.2%), and 0.10 ± 0.38 Gy (40.6% ± 214.9%) among the 75 plans, with best single-plan reductions of 3.20 Gy (40.8%), 2.38 Gy (40.07%), and 1.26 Gy (27.5%), respectively. The high-risk clinical target volume D90 was similar, with 6.55 ± 0.96 Gy and 6.59 ± 1.06 Gy for T&O and DMBT, respectively. CONCLUSIONS Application of the DMBT concept to cervical cancer allowed for improved organ at risk sparing while achieving similar target coverage on a sizeable patient population, as intended, by maximally utilizing the anatomic information contained in 3-dimensional imaging. A series of mechanical and clinical validations are to be followed.

Downlink optimization of indoor wireless networks using multiple antenna systems

We compare the performance of two multiple antenna systems to be used in quality of service (QoS) supported indoor wireless networks. While a conventional array antenna system (AAS) has collocated, closely spaced antenna elements, a distributed antenna system (DAS) has largely spaced antennas over the entire area of radio coverage. To support multimedia applications requiring high bandwidth and on time delivery, we propose a set of highly spectrum efficient radio resource management algorithms. We focus on the optimization of downlink since many kinds of Internet traffic show the downlink dominance in their traffic asymmetry. To maximize the downlink throughput, we present a new transmit beamforming algorithm which can be equally applied to both DAS and AAS. The beamforming algorithm is integrated with a link scheduling algorithm that exploits the space division multiplexing (SDM) capability of multiple antenna systems to meet the QoS requirements of all terminals. Numerical examples conducted for a line of sight (LOS) environment demonstrate that a network with DAS outperforms one with AAS in terms of signal coverage and provides 40-160% higher capacity.

HDR brachytherapy of rectal cancer using a novel grooved‐shielding applicator design

PURPOSE The aim of this work was to design a novel high-dose rate (HDR) ((192)Ir) brachytherapy applicator for treatment of rectal carcinomas that uses tungsten shielding for possibly improved dosimetric results over commercial brachytherapy applicator(s). METHODS A set of 15 single-depth applicators and one dual-depth applicator were designed and simulated using Monte Carlo (MCNPX). All applicators simulated were high-density tungsten alloy cylinders, 16-mm in diameter, and 60-mm long, with longitudinal grooves within which an (192)Ir source can be placed. The single-depth designs varied regarding the number and depth of these grooves, ranging from 8 to 16 and 1-mm to 3-mm, respectively. The dual-depth design had ten channels, each of which had two depths at which the source could be placed. Optimized treatment plans were generated for each design on data from 13 treated patients (36 fractions) with asymmetrical clinical target volumes (CTVs). All results were compared against the clinically treated plans which used intracavitary mold applicator (ICMA), as well as a recently designed, highly automated, and collimated intensity modulation device named dynamic modulated brachytherapy (DMBT) device. RESULTS All applicator designs outperformed the ICMA in every calculated dosimetric criteria, except the total dwell times (∼30% increase). There were clear, but relative, tradeoffs regarding both the number of channels and the depth of each channel. Overall, the 12-channel, 1-mm depth, and 14-channel 2-mm depth designs had the best results of the simpler designs, sparing the healthy rectal tissues the most while achieving comparable CTV coverage with the dose heterogeneity index and lateral spill doses improving by over 10% and the contralateral healthy rectum dose dropping over 30% compared to ICMA. The ten-channel dual-depth design outperformed each single-depth design, yielding the best coverage and sparing. CONCLUSIONS New grooved tungsten HDR-brachytherapy devices have been designed and simulated. The results of this work attest to the capability of these new, highly anisotropic, intelligently shielded applicators to limit dose to healthy tissues while maintaining a conformal prescription dose to the CTV.