Engineering Papers⌕ Search

Engineering topics

Xing, L

Publications and source records attributed to Xing, L.

At least 37 records · Page 2

MO-FG-CAMPUS-TeP2-05: Optimizing Stereotactic Radiosurgery Treatment of Multiple Brain Metastasis Lesions with Individualized Rotational Arc Trajectories

Purpose: Radiosurgery of multiple (n>4) brain metastasis lesions requires 3–4 noncoplanar VMAT arcs with excessively high monitor units and long delivery time. We investigated whether an improved optimization technique would decrease the needed arc numbers and increase the delivery efficiency, while improving or maintaining the plan quality. Methods: The proposed 4pi arc space optimization algorithm consists of two steps: automatic couch angle selection followed by aperture generation for each arc with optimized control points distribution. We use a greedy algorithm to select the couch angles. Starting from a single coplanar arc plan we search through the candidate noncoplanar arcs to pick a single noncoplanar arc that will bring the best plan quality when added into the existing treatment plan. Each time, only one additional noncoplanar arc is considered making the calculation time tractable. This process repeats itself until desired number of arc is reached. The technique is first evaluated in coplanar arc delivery scheme with testing cases and then applied to noncoplanar treatments of a case with 12 brain metastasis lesions. Results: Clinically acceptable plans are created within minutes. For the coplanar testing cases the algorithm yields singlearc plans with better dose distributions than that of two-arc VMAT, simultaneously with a 12–17% reduction in the delivery time and a 14–21% reduction in MUs. For the treatment of 12 brain mets while Paddick conformity indexes of the two plans were comparable the SCG-optimization with 2 arcs (1 noncoplanar and 1 coplanar) significantly improved the conventional VMAT with 3 arcs (2 noncoplanar and 1 coplanar). Specifically V16 V10 and V5 of the brain were reduced by 11%, 11% and 12% respectively. The beam delivery time was shortened by approximately 30%. Conclusion: The proposed 4pi arc space optimization technique promises to significantly reduce the brain toxicity while greatly improving the treatment efficiency.

60 APPLIED LIFE SCIENCES↗

TU-FG-201-04: Computer Vision in Autonomous Quality Assurance of Linear Accelerators

Purpose: Routine quality assurance (QA) of linear accelerators represents a critical and costly element of a radiation oncology center. Recently, a system was developed to autonomously perform routine quality assurance on linear accelerators. The purpose of this work is to extend this system and contribute computer vision techniques for obtaining quantitative measurements for a monthly multi-leaf collimator (MLC) QA test specified by TG-142, namely leaf position accuracy, and demonstrate extensibility for additional routines. Methods: Grayscale images of a picket fence delivery on a radioluminescent phosphor coated phantom are captured using a CMOS camera. Collected images are processed to correct for camera distortions, rotation and alignment, reduce noise, and enhance contrast. The location of each MLC leaf is determined through logistic fitting and a priori modeling based on knowledge of the delivered beams. Using the data collected and the criteria from TG-142, a decision is made on whether or not the leaf position accuracy of the MLC passes or fails. Results: The locations of all MLC leaf edges are found for three different picket fence images in a picket fence routine to 0.1mm/1pixel precision. The program to correct for image alignment and determination of leaf positions requires a runtime of 21– 25 seconds for a single picket, and 44 – 46 seconds for a group of three pickets on a standard workstation CPU, 2.2 GHz Intel Core i7. Conclusion: MLC leaf edges were successfully found using techniques in computer vision. With the addition of computer vision techniques to the previously described autonomous QA system, the system is able to quickly perform complete QA routines with minimal human contribution.

60 APPLIED LIFE SCIENCES↗

TU-FG-BRB-08: Challenges, Limitations and Future Outlook Towards Clinical Translation of Proton Acoustic Range Verification

Purpose: To report our investigations of proton acoustic imaging, including computer simulations and preliminary experimental studies at clinical facilities. The ultimate achievable accuracy, sensitivity and clinical translation challenges are discussed. Methods: The acoustic pulse due to pressure rise was estimated using finite element model. Since the ionoacoustic pulse is highly dependent on the proton pulse width and energy, multiple pulse widths were studied. Based on the received signal spectrum at piezoelectric ultrasound transducer with consideration of random thermal noise, maximum spatial resolution of the proton-acoustic imaging modality was calculated. The simulation studies defined the design specifications of the system to detect proton acoustic signal from Hitachi and Mevion clinical machines. A 500 KHz hydrophone with 100 dB amplification was set up in a water tank placed in front of the proton nozzle A 40 MHz data acquisition was synchronized by a trigger signal provided by the machine. Results: Given 30–800 mGy dose per pulse at the Bragg peak, the minimum number of protons detectable by the proton acoustic technique was on the order of 10×10^6 per pulse. The broader pulse widths produce signal with lower acoustic frequencies, with 10 µs pulses producing signals with frequency less than 100 kHz. As the proton beam pulse width increases, a higher dose rate is required to measure the acoustic signal. Conclusion: We have established the minimal detection limit for protonacoustic range validation for a variety of pulse parameters. Our study indicated practical proton-acoustic range verification can be feasible with a pulse shorter than 10 µs, 5×10^6 protons/pulse, 50 nA beam current and a highly sensitive ultrasonic transducer. The translational challenges into current clinical machines include proper magnetic shielding of the measurement equipment, providing a clean trigger signal from the proton machine, providing a shorter proton beam pulse and higher dose per pulse.

60 APPLIED LIFE SCIENCES↗

TU-FG-BRB-12: Real-Time Visualization of Discrete Spot Scanning Proton Therapy Beam for Quality Assurance

Purpose: With the growing adoption of proton beam therapy there is an increasing need for effective and user-friendly tools for performing quality assurance (QA) measurements. The speed and versatility of spot-scanning proton beam (PB) therapy systems present unique challenges for traditional QA tools. To address these challenges a proof-of-concept system was developed to visualize, in real-time, the delivery of individual spots from a spot-scanning PB in order to perform QA measurements. Methods: The PB is directed toward a custom phantom with planar faces coated with a radioluminescent phosphor (Gd2O2s:Tb). As the proton beam passes through the phantom visible light is emitted from the coating and collected by a nearby CMOS camera. The images are processed to determine the locations at which the beam impinges on each face of the phantom. By so doing, the location of each beam can be determined relative to the phantom. The cameras are also used to capture images of the laser alignment system. The phantom contains x-ray fiducials so that it can be easily located with kV imagers. Using this data several quality assurance parameters can be evaluated. Results: The proof-of-concept system was able to visualize discrete PB spots with energies ranging from 70 MeV to 220 MeV. Images were obtained with integration times ranging from 20 to 0.019 milliseconds. If not limited by data transmission, this would correspond to a frame rate of 52,000 fps. Such frame rates enabled visualization of individual spots in real time. Spot locations were found to be highly correlated (R{sup 2}=0.99) with the nozzle-mounted spot position monitor indicating excellent spot positioning accuracy Conclusion: The system was shown to be capable of imaging individual spots for all clinical beam energies. Future development will focus on extending the image processing software to provide automated results for a variety of QA tests.

60 APPLIED LIFE SCIENCES↗

TU-H-CAMPUS-TeP3-03: Dose Enhancement by Gold Nanoparticles Around the Bragg Peak of Proton Beams

Purpose: To make clear the spatial distribution of dose enhancement around gold nanoparticles (GNPs) located near the proton Bragg peak, and to evaluate the potential of GNPs as a radio sensitizer. Methods: The dose enhancement by electrons emitted from GNPs under proton irradiation was estimated by Geant4 Monte Carlo simulation toolkit in two steps. In an initial macroscopic step, 100 and 195 MeV proton beams were incident on a water cube, 30 cm on a side. Energy distributions of protons were calculated at four depths, 50% and 75% proximal to the Bragg peak, 100% peak, and 75% distal to the peak (P50, P75, Peak, and D75, respectively). In a subsequent microscopic step, protons with the energy distribution calculated above were incident on a 20 nm diameter GNP in a nanometer-size water box and the spatial distribution of dose around the GNP was determined for each energy distribution. The dose enhancement factor (DEF) was also deduced. Results: The dose enhancement effect was spread to several tens of nanometers in the both depth and radial directions. The enhancement area increased in the order of P50, P75, Peak, and D75 for both cases with 100 and 195 MeV protons. In every position around the Bragg peak, the 100 MeV beam resulted in a higher dose enhancement than the 195 MeV beam. At P75, the average and maximum DEF were 3.9 and 17.0 for 100 MeV, and 3.5 and 16.2 for 195 MeV, respectively. These results indicate that lower energy protons caused higher dose enhancement in this incident proton energy range. Conclusion: The dose enhancement around GNPs spread as the position in the Bragg peak region becomes deeper and depends on proton energy. It is expected that GNPs can be used as a radio sensitizer with consideration of the location and proton beam energy.

60 APPLIED LIFE SCIENCES↗

WE-AB-209-02: A New Inverse Planning Framework with Principle-Based Modeling of Inter-Structural Dosimetric Tradeoffs

Purpose: Traditional radiotherapy inverse planning relies on the weighting factors to phenomenologically balance the conflicting criteria for different structures. The resulting manual trial-and-error determination of the weights has long been recognized as the most time-consuming part of treatment planning. The purpose of this work is to develop an inverse planning framework that parameterizes the inter-structural dosimetric tradeoff among with physically more meaningful quantities to simplify the search for a clinically sensible plan. Methods: A permissible dosimetric uncertainty is introduced for each of the structures to balance their conflicting dosimetric requirements. The inverse planning is then formulated as a convex feasibility problem, which aims to generate plans with acceptable dosimetric uncertainties. A sequential procedure (SP) is derived to decompose the model into three submodels to constrain the uncertainty in the planning target volume (PTV), the critical structures, and all other structures to spare, sequentially. The proposed technique is applied to plan a liver case and a head-and-neck case and compared with a conventional approach. Results: Our results show that the strategy is able to generate clinically sensible plans with little trial-and-error. In the case of liver IMRT, the fractional volumes to liver and heart above 20Gy are found to be 22% and 10%, respectively, which are 15.1% and 33.3% lower than that of the counterpart conventional plan while maintaining the same PTV coverage. The planning of the head and neck IMRT show the same level of success, with the DVHs for all organs at risk and PTV very competitive to a counterpart plan. Conclusion: A new inverse planning framework has been established. With physically more meaningful modeling of the inter-structural tradeoff, the technique enables us to substantially reduce the need for trial-and-error adjustment of the model parameters and opens new opportunities of incorporating prior knowledge to facilitate the treatment planning process.

60 APPLIED LIFE SCIENCES↗

WE-AB-209-09: Optimization of Rotational Arc Station Parameter Optimized Radiation Therapy

Purpose: To develop a fast optimization method for station parameter optimized radiation therapy (SPORT) and show that SPORT is capable of improving VMAT in both plan quality and delivery efficiency. Methods: The angular space from 0° to 360° was divided into 180 station points (SPs). A candidate aperture was assigned to each of the SPs based on the calculation results using a column generation algorithm. The weights of the apertures were then obtained by optimizing the objective function using a state-of-the-art GPU based Proximal Operator Graph Solver (POGS) within seconds. Apertures with zero or low weight were thrown out. To avoid being trapped in a local minimum, a stochastic gradient descent method was employed which also greatly increased the convergence rate of the objective function. The above procedure repeated until the plan could not be improved any further. A weighting factor associated with the total plan MU also indirectly controlled the complexities of aperture shapes. The number of apertures for VMAT and SPORT was confined to 180. The SPORT allowed the coexistence of multiple apertures in a single SP. The optimization technique was assessed by using three clinical cases (prostate, H&N and brain). Results: Marked dosimetric quality improvement was demonstrated in the SPORT plans for all three studied cases. Prostate case: the volume of the 50% prescription dose was decreased by 22% for the rectum. H&N case: SPORT improved the mean dose for the left and right parotids by 15% each. Brain case: the doses to the eyes, chiasm and inner ears were all improved. SPORT shortened the treatment time by ∼1 min for the prostate case, ∼0.5 min for brain case, and ∼0.2 min for the H&N case. Conclusion: The superior dosimetric quality and delivery efficiency presented here indicates that SPORT is an intriguing alternative treatment modality.

60 APPLIED LIFE SCIENCES↗

TH-AB-209-07: High Resolution X-Ray-Induced Acoustic Computed Tomography

Purpose: X-ray radiographic absorption imaging is an invaluable tool in medical diagnostics, biology and materials science. However, the use of conventional CT is limited by two factors: the detection sensitivity to weak absorption material and the radiation dose from CT scanning. The purpose of this study is to explore X-ray induced acoustic computed tomography (XACT), a new imaging modality, which combines X-ray absorption contrast and high ultrasonic resolution to address these challenges. Methods: First, theoretical models was built to analyze the XACT sensitivity to X-ray absorption and calculate the minimal radiation dose in XACT imaging. Then, an XACT system comprised of an ultrashort X-ray pulse, a low noise ultrasound detector and a signal acquisition system was built to evaluate the X-ray induced acoustic signal generation. A piece of chicken bone and a phantom with two golden fiducial markers were exposed to 270 kVp X-ray source with 60 ns exposure time, and the X-ray induced acoustic signal was received by a 2.25MHz ultrasound transducer in 200 positions. XACT images were reconstructed by a filtered back-projection algorithm. Results: The theoretical analysis shows that X-ray induced acoustic signals have 100% relative sensitivity to X-ray absorption, but not to X-ray scattering. Applying this innovative technology to breast imaging, we can reduce radiation dose by a factor of 50 compared with newly FDA approved breast CT. The reconstructed images of chicken bone and golden fiducial marker phantom reveal that the spatial resolution of the built XACT system is 350µm. Conclusion: In XACT, the imaging sensitivity to X-ray absorption is improved and the imaging dose is dramatically reduced by using ultrashort pulsed X-ray. Taking advantage of the high ultrasonic resolution, we can also perform 3D imaging with a single X-ray pulse. This new modality has the potential to revolutionize x-ray imaging applications in medicine and biology.

60 APPLIED LIFE SCIENCES↗

TH-CD-206-01: Expectation-Maximization Algorithm-Based Tissue Mixture Quantification for Perfusion MRI

Purpose: To investigate the feasibility of estimating the tissue mixture perfusions and quantifying cerebral blood flow change in arterial spin labeled (ASL) perfusion MR images. Methods: The proposed perfusion MR image analysis framework consists of 5 steps: (1) Inhomogeneity correction was performed on the T1- and T2-weighted images, which are available for each studied perfusion MR dataset. (2) We used the publicly available FSL toolbox to strip off the non-brain structures from the T1- and T2-weighted MR images. (3) We applied a multi-spectral tissue-mixture segmentation algorithm on both T1- and T2-structural MR images to roughly estimate the fraction of each tissue type - white matter, grey matter and cerebral spinal fluid inside each image voxel. (4) The distributions of the three tissue types or tissue mixture across the structural image array are down-sampled and mapped onto the ASL voxel array via a co-registration operation. (5) The presented 4-dimensional expectation-maximization (4D-EM) algorithm takes the down-sampled three tissue type distributions on perfusion image data to generate the perfusion mean, variance and percentage images for each tissue type of interest. Results: Experimental results on three volunteer datasets demonstrated that the multi-spectral tissue-mixture segmentation algorithm was effective to initialize tissue mixtures from T1- and T2-weighted MR images. Compared with the conventional ASL image processing toolbox, the proposed 4D-EM algorithm not only generated comparable perfusion mean images, but also produced perfusion variance and percentage images, which the ASL toolbox cannot obtain. It is observed that the perfusion contribution percentages may not be the same as the corresponding tissue mixture volume fractions estimated in the structural images. Conclusion: A specific application to brain ASL images showed that the presented perfusion image analysis method is promising for detecting subtle changes in tissue perfusions, which is valuable for the early diagnosis of certain brain diseases, e.g. multiple sclerosis.

60 APPLIED LIFE SCIENCES↗

TH-CD-206-05: Machine-Learning Based Segmentation of Organs at Risks for Head and Neck Radiotherapy Planning

Purpose: Accurate and efficient delineation of tumor target and organs-at-risks is essential for the success of radiotherapy. In reality, despite of decades of intense research efforts, auto-segmentation has not yet become clinical practice. In this study, we present, for the first time, a deep learning-based classification algorithm for autonomous segmentation in head and neck (HaN) treatment planning. Methods: Fifteen HN datasets of CT, MR and PET images with manual annotation of organs-at-risk (OARs) including spinal cord, brainstem, optic nerves, chiasm, eyes, mandible, tongue, parotid glands were collected and saved in a library of plans. We also have ten super-resolution MR images of the tongue area, where the genioglossus and inferior longitudinalis tongue muscles are defined as organs of interest. We applied the concepts of random forest- and deep learning-based object classification for automated image annotation with the aim of using machine learning to facilitate head and neck radiotherapy planning process. In this new paradigm of segmentation, random forests were used for landmark-assisted segmentation of super-resolution MR images. Alternatively to auto-segmentation with random forest-based landmark detection, deep convolutional neural networks were developed for voxel-wise segmentation of OARs in single and multi-modal images. The network consisted of three pairs of convolution and pooing layer, one RuLU layer and a softmax layer. Results: We present a comprehensive study on using machine learning concepts for auto-segmentation of OARs and tongue muscles for the HaN radiotherapy planning. An accuracy of 81.8% in terms of Dice coefficient was achieved for segmentation of genioglossus and inferior longitudinalis tongue muscles. Preliminary results of OARs regimentation also indicate that deep-learning afforded an unprecedented opportunities to improve the accuracy and robustness of radiotherapy planning. Conclusion: A novel machine learning framework has been developed for image annotation and structure segmentation. Our results indicate the great potential of deep learning in radiotherapy treatment planning.

60 APPLIED LIFE SCIENCES↗

TH-CD-207B-07: Noise Modeling of Single Photon Avalanche Diode (SPAD) for Photon Counting CT Applications

Purpose: Silicon photomultiplier (SiPM) has recently emerged as a promising photodetector for biomedical imaging applications. Due to its high multiplication gain (comparable to PMT), fast timing, low cost and compactness, it is considered a good candidate for photon counting CT. Dark noise is a limiting factor which impacts both energy resolution and detection dynamic range. Our goal is to develop a comprehensive model for noise sources for SiPM sensors. Methods: The physical parameters used in this work were based upon a test SPAD fabricated in 130nm CMOS process. The SPAD uses an n+/p-well junction, which is isolated from the p-substrate by a deep n-well junction. Inter-avalanche time measurement was used to record the time interval between two adjacent avalanche pulses. After collecting 1×106 counts, the histogram was obtained and multiple exponential fitting process was used to extract the lifetime associated with the traps within the bandgap. Results: At room temperature, the breakdown voltage of the SPAD is ∼11.4V and shows a temperature coefficient of 7.7mV/°C. The dark noise of SPAD increases with both the excess biasing voltage and temperature. The primary dark counts from the model were validated against the measurement results. A maximum relative error of 8.7% is observed at 20 °C with an excess voltage of 0.5V. The probabilities of after-pulsing are found to be dependent of both temperature and excess voltage. With 0.5V excess voltage, the after-pulsing probability is 63.5% at - 30 °C and drops to ∼6.6% at 40 °C. Conclusion: A comprehensive noise model for SPAD sensor was proposed. The model takes into account of static, dynamic and statistical behavior of SPADs. We believe that this is the first SPAD circuit simulation model that includes the band-to-band tunneling dark noise contribution and temporal dependence of the after-pulsing probability.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

SU-C-213-04: Application of Depth Sensing and 3D-Printing Technique for Total Body Irradiation (TBI) Patient Measurement and Treatment Planning

Purpose: To develop and validate an innovative method of using depth sensing cameras and 3D printing techniques for Total Body Irradiation (TBI) treatment planning and compensator fabrication. Methods: A tablet with motion tracking cameras and integrated depth sensing was used to scan a RANDOTM phantom arranged in a TBI treatment booth to detect and store the 3D surface in a point cloud (PC) format. The accuracy of the detected surface was evaluated by comparison to extracted measurements from CT scan images. The thickness, source to surface distance and off-axis distance of the phantom at different body section was measured for TBI treatment planning. A 2D map containing a detailed compensator design was calculated to achieve uniform dose distribution throughout the phantom. The compensator was fabricated using a 3D printer, silicone molding and tungsten powder. In vivo dosimetry measurements were performed using optically stimulated luminescent detectors (OSLDs). Results: The whole scan of the anthropomorphic phantom took approximately 30 seconds. The mean error for thickness measurements at each section of phantom compare to CT was 0.44 ± 0.268 cm. These errors resulted in approximately 2% dose error calculation and 0.4 mm tungsten thickness deviation for the compensator design. The accuracy of 3D compensator printing was within 0.2 mm. In vivo measurements for an end-to-end test showed the overall dose difference was within 3%. Conclusion: Motion cameras and depth sensing techniques proved to be an accurate and efficient tool for TBI patient measurement and treatment planning. 3D printing technique improved the efficiency and accuracy of the compensator production and ensured a more accurate treatment delivery.

60 APPLIED LIFE SCIENCES↗

SU-D-201-03: Imaging Cellular Pharmacokinetics of 18F-FDG in Inflammatory/Stem Cells

Purpose: Atherosclerosis is a progressive inflammatory condition that underlies coronary artery disease (CAD)—the leading cause of death in the USA. Thus, understating the metabolism of inflammatory cells can be a valuable tool for investigating CAD. To the best of our knowledge, we are the first to successfully investigate the pharmacokinetics of [18F]fluoro-deoxyglucose (18F-FDG) uptake in a single macrophages and compared with induced pluripotent stem cells (iPSCs) and mesenchymal stem cells (MSCs) with a novel imaging technique, radioluminescence microscopy, initially developed for cancer imaging. Methods: Live cells were cultured sparsely on Matrigel in a glass-bottom dish and starved for 1 hour before incubation with 250 microCi of 18F-FDG for 45 minutes. Excess radiotracer was removed using DMEM medium without glucose. Before imaging, DMEM (1 mL) was added to the cell culture and a 100 microm-thin CdWO4 scintillator plate was placed on top of the cells. Light produced following beta decay was imaged with a highly sensitive inverted microscope (LV200, Olympus) fitted with a 40x/1.3 high-NA oil objective, and an EM-CCD camera. The images were collected over 18,000 frames with 4×4 binning (1200 MHz EM Gain, 300ms exposure). Custom-written software was developed in MATLAB for image processing (Figure 1). For statistical analysis 10 different region-of-interests (ROIs) were selected for each cell type. Results: Figures 2A–2B show bright-field/fusion images for all three different cell types. The relationship between cell-to-cell comparisons was found to be linear for macrophages unlike iPSCs and MSCs, which were best fitted with moving or rolling average (Figure 2C). The average observed decay of 18F-FDG in a single cell of MSCs per second (0.067) was 20% and 36% higher compared to iPSCs (0.054) and macrophages (0.043), respectively (Figure 2D). Conclusion: MSCs was found to be 2–3x more sensitive to glucose molecule despite constant parameters for each cell type examined.

60 APPLIED LIFE SCIENCES↗

SU-E-I-39: Combining Conventional Tomographic Imaging Strategy and Interior Tomography for Low Dose Dual-Energy CT (DECT)

Purpose: Dual-energy CT (DECT) affords quantitative information of tissue density and provides a new dimension for disease diagnosis and treatment planning. The technique, however, increases the imaging dose because of the doubled scans, and thus hinders its widespread clinical applications. The purpose of this work is to develop a novel hybrid DECT image acquisition and reconstruction strategy, in which one of the energies is dealt by interior tomography while the other one is obtained using conventional tomography approach. Methods: In the proposed hybrid imaging strategy, the projection data of one of the energies (e.g., high-energy) were acquired and processed in an interior scanning model, whereas the other energy in the conventional tomographic approach. It known that, if the ROI is piecewise constant or polynomial, the interior ROI can be reconstructed with TV or HOT minimization. Here we extend the TV based interior reconstruction method into dual-energy situation. The ROI images so obtained were overlaid in the context of conventional CT of the companion energy. A material based composition in ROI was used in the proposed reconstruction framework. Results: In the simulation experiment with a diagnostic DECT geometry and energies, we were able to derive the densities of soft-tissues and bones in the ROI with high fidelity. In the experimental CBCT study, both kV and MV data were collected using the on-board kV and MV imaging system. The MV data were truncated only across the ROI. Using the interior tomography reconstruction above, we were able to obtain the ROI images as that obtained using un-truncated MV data with known tissue densities. Conclusion: The proposed DECT imaging strategy provides an effective way to extract tissue density information in the ROI and in the context of anatomical images of CT imaging, with much reduced imaging dose.

60 APPLIED LIFE SCIENCES↗

SU-E-I-41: Dictionary Learning Based Quantitative Reconstruction for Low-Dose Dual-Energy CT (DECT)

Purpose: DECT collects two sets of projection data under higher and lower energies. With appropriates composition methods on linear attenuation coefficients, quantitative information about the object, such as density, can be obtained. In reality, one of the important problems in DECT is the radiation dose due to doubled scans. This work is aimed at establishing a dictionary learning based reconstruction framework for DECT for improved image quality while reducing the imaging dose. Methods: In our method, two dictionaries were learned respectively from the high-energy and lowenergy image datasets of similar objects under normal dose in advance. The linear attenuation coefficient was decomposed into two basis components with material based composition method. An iterative reconstruction framework was employed. Two basis components were alternately updated with DECT datasets and dictionary learning based sparse constraints. After one updating step under the dataset fidelity constraints, both high-energy and low-energy images can be obtained from the two basis components. Sparse constraints based on the learned dictionaries were applied to the high- and low-energy images to update the two basis components. The iterative calculation continues until a pre-set number of iteration was reached. Results: We evaluated the proposed dictionary learning method with dual energy images collected using a DECT scanner. We re-projected the projection data with added Poisson noise to reflect the low-dose situation. The results obtained by the proposed method were compared with that obtained using FBP based method and TV based method. It was found that the proposed approach yield better results than other methods with higher resolution and less noise. Conclusion: The use of dictionary learned from DECT images under normal dose is valuable and leads to improved results with much lower imaging dose.

60 APPLIED LIFE SCIENCES↗

SU-E-J-49: Design and Fabrication of Custom 3D Printed Phantoms for Radiation Therapy Research and Quality Assurance

Purpose The rapid proliferation of affordable 3D printing techniques has enabled the custom fabrication of items ranging from paper weights to medical implants. This study investigates the feasibility of utilizing the technology for developing novel phantoms for use in radiation therapy quality assurance (QA) procedures. Methods A phantom for measuring the geometric parameters of linear accelerator (LINAC) on-board imaging (OBI) systems was designed using SolidWorks. The design was transferred to a 3D printer and fabricated using a fused deposition modeling (FDM) technique. Fiducials were embedded in the phantom by placing 1.6 mm diameter steel balls in predefined holes and securing them with silicone. Several MV and kV images of the phantom were collected and the visibility and geometric accuracy were evaluated. A second phantom, for use in the experimental evaluation of a high dose rate (HDR) brachytherapy dosimeter, was designed to secure several applicator needles in water. The applicator was fabricated in the same 3D printer and used for experiments. Results The general accuracy of printed parts was determined to be 0.1 mm. The cost of materials for the imaging and QA phantoms were $22 and $5 respectively. Both the plastic structure and fiducial markers of the imaging phantom were visible in MV and kV images. Fiducial marker locations were determined to be within 1mm of desired locations, with the discrepancy being attributed to the fiducial attachment process. The HDR phantom secured the applicators within 0.5 mm of the desired locations. Conclusion 3D printing offers an inexpensive method for fabricating custom phantoms for use in radiation therapy quality assurance. While the geometric accuracy of such parts is limited compared to more expensive methods, the phantoms are still highly functional and provide a unique opportunity for rapid fabrication of custom phantoms for use in radiation therapy QA and research.

07 ISOTOPE AND RADIATION SOURCES↗

SU-E-J-71: Spatially Preserving Prior Knowledge-Based Treatment Planning

Purpose: Prior knowledge-based treatment planning is impeded by the use of a single dose volume histogram (DVH) curve. Critical spatial information is lost from collapsing the dose distribution into a histogram. Even similar patients possess geometric variations that becomes inaccessible in the form of a single DVH. We propose a simple prior knowledge-based planning scheme that extracts features from prior dose distribution while still preserving the spatial information. Methods: A prior patient plan is not used as a mere starting point for a new patient but rather stopping criteria are constructed. Each structure from the prior patient is partitioned into multiple shells. For instance, the PTV is partitioned into an inner, middle, and outer shell. Prior dose statistics are then extracted for each shell and translated into the appropriate Dmin and Dmax parameters for the new patient. Results: The partitioned dose information from a prior case has been applied onto 14 2-D prostate cases. Using prior case yielded final DVHs that was comparable to manual planning, even though the DVH for the prior case was different from the DVH for the 14 cases. Solely using a single DVH for the entire organ was also performed for comparison but showed a much poorer performance. Different ways of translating the prior dose statistics into parameters for the new patient was also tested. Conclusion: Prior knowledge-based treatment planning need to salvage the spatial information without transforming the patients on a voxel to voxel basis. An efficient balance between the anatomy and dose domain is gained through partitioning the organs into multiple shells. The use of prior knowledge not only serves as a starting point for a new case but the information extracted from the partitioned shells are also translated into stopping criteria for the optimization problem at hand.

60 APPLIED LIFE SCIENCES↗

SU-E-T-422: Fast Analytical Beamlet Optimization for Volumetric Intensity-Modulated Arc Therapy

Purpose: To implement a fast optimization algorithm on CPU/GPU heterogeneous computing platform and to obtain an optimal fluence for a given target dose distribution from the pre-calculated beamlets in an analytical approach. Methods: The 2D target dose distribution was modeled as an n-dimensional vector and estimated by a linear combination of independent basis vectors. The basis set was composed of the pre-calculated beamlet dose distributions at every 6 degrees of gantry angle and the cost function was set as the magnitude square of the vector difference between the target and the estimated dose distribution. The optimal weighting of the basis, which corresponds to the optimal fluence, was obtained analytically by the least square method. Those basis vectors with a positive weighting were selected for entering into the next level of optimization. Totally, 7 levels of optimization were implemented in the study.Ten head-and-neck and ten prostate carcinoma cases were selected for the study and mapped to a round water phantom with a diameter of 20cm. The Matlab computation was performed in a heterogeneous programming environment with Intel i7 CPU and NVIDIA Geforce 840M GPU. Results: In all selected cases, the estimated dose distribution was in a good agreement with the given target dose distribution and their correlation coefficients were found to be in the range of 0.9992 to 0.9997. Their root-mean-square error was monotonically decreasing and converging after 7 cycles of optimization. The computation took only about 10 seconds and the optimal fluence maps at each gantry angle throughout an arc were quickly obtained. Conclusion: An analytical approach is derived for finding the optimal fluence for a given target dose distribution and a fast optimization algorithm implemented on the CPU/GPU heterogeneous computing environment greatly reduces the optimization time.

60 APPLIED LIFE SCIENCES↗