Search NASA⌕ Search

SEARCH · Search NASA

Results for “Image Processing”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 613 records · Page 34
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

SU-E-T-12: A Comparative Dosimetric Study of Pre and Post Prostate Iodine-125 Permanent Seed Implants

Purpose: Post-implant dosimetry has become the gold standard for prostate implant evaluation. The goal of this research is to compare the dosimetry between pre-plan and post-plan in permanent prostate seed implant brachytherapy. Methods: A retrospective study of 91 patients treated with Iodine-125 prostate seed implant between year 2012∼2014 were performed. All plans were created using a VariSeed 8.0 planning system. Pre-plan ultrasound images were acquired using 0.5 cm slice thickness. Post-plan CT images acquired about 1–4 weeks after implant, fused with the preplan ultrasound images. The prostate and urethra contours were generated using the fusion of ultrasound and CT images. Iodine-125 seed source activities varied between 0.382 to 0.414 mCi per seed. The loading patterns varied slightly between patients depending on the prostate size. Statistical analysis of pre and post plans for prostate and urethra volumes, V100%, V150% and D90, and urethra D10 were performed and reported. Results: The pre and post implant average prostate size was 36.90cc vs. 38.58cc; V100% was 98.33% vs. 96.89%; V150% was 47.09% vs. 56.95%; D90 was 116.35Gy vs. 116.12Gy, urethra volume was 1.72cc vs. 1.85cc, urethra D10% was 122.0% vs. 135.35%, respectively. There was no statistically significant difference between the pre and post-plan values for D90(p-value=0.43). However, there are significant differences between other parameters most likely due to post surgical edema; prostate size (p-value= 0.00015); V100% (p-value=3.7803E-07); V150% (p-value=1.49E-09); urethra volume (p-value= 2.77E-06); Urethra D10 (p-value=7.37E-11). Conclusion: The post-plan dosimetry using CT image set showed similar D90 dose coverage to the pre-plan using the ultrasound image dataset. The study showed that our prostate seed implants have consistently delivered adequate therapeutic dose to the prostate while sparing urethra. Future studies to correlate dose versus biochemical response using patients’ PSA values as well as patients’ survival are warranted.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-62: A Preliminary Experience of Using EPID Transit Dosimetry for Monitoring Daily Dose Variations in Radiation Treatment Delivery

Purpose: To investigate the use of EPID transit dosimetry for monitoring daily dose variations in radiation treatment delivery. Methods: A patient with head and neck cancer treated using nine field IMRT beams was used in this study. The prescription was 45 Gy in 25 fractions. A KV CBCT was acquired before each treatment on a Varian NTX linear accelerator. Integrated images using MV EPID were acquired for each treatment beam. Planning CT images, treatment plan, and daily integrated images were imported into a commercial QA software Dosimetry Check (v4r4 Math Resolutions, LLC, Columbia, MD) to calculate 3D dose of the day assuming 25 fractions treatment. Planning CT images were deformed and registered to each daily CBCT using Varian SmartAdapt (v11.MR2). ROIs were then propagated from planning CT to daily CBCT. The correlation between maximum, average dose of ROIs and ROI volume, center of mass shift, Dice Similarity Coefficient (DSC) were investigated. Results: Not all parameters investigated showed strong correlations. For PTV and CTV, the average dose has inverse correlation with their volume change (correlation coefficient −0.52, −0.50, respectively) and DSC (−0.59, −0.59, respectively). The average dose of right parotid has correlation with its volume change (0.56). The maximum dose of spinal cord has correlation with the center of mass superior-inferior shift (0.52) and inverse correlation with the center of mass anterior-posterior shift (−0.73). Conclusion: Transit dosimetry using EPID images collected during treatment delivery offers great potential to monitor daily dose variations due to patient anatomy change, motion, and setup errors in radiation treatment delivery. It can provide a patient-specific QA tool valuable for adaptive radiation therapy. Further work is needed to validate the technique.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-151: Breathing Synchronized Delivery (BSD) Planning for RapicArc Treatment

Purpose: To propose a workflow for breathing synchronized delivery (BSD) planning for RapicArc treatment. Methods: The workflow includes three stages: screening/simulation, planning, and delivery. In the screening/simulation stage, a 4D CT with the corresponding breathing pattern is acquired for each of the selected patients, who are able to follow their own breathing pattern. In the planning stage, one breathing phase is chosen as the reference, and contours are delineated on the reference image. Deformation maps to other phases are performed along with contour propagation. Based on the control points of the initial 3D plan for the reference phase and the respiration trace, the correlation with respiration phases, the leaf sequence and gantry angles is determined. The beamlet matrices are calculated with the corresponding breathing phase and deformed to the reference phase. Using the 4D dose evaluation tool and the original 3D plan DVHs criteria, the leaf sequence is further optimized to meet the planning objectives and the machine constraints. In the delivery stage, the patients are instructed to follow the programmed breathing patterns of their own, and all other parts are the same as the conventional Rapid-Arc delivery. Results: Our plan analysis is based on comparison of the 3D plan with a static target (SD), 3D plan with motion delivery (MD), and the BSD plan. Cyclic motion of range 0 cm to 3 cm was simulated for phantoms and lung CT. The gain of the BSD plan over MD is significant and concordant for both simulation and lung 4DCT, indicating the benefits of 4D planning. Conclusion: Our study shows that the BSD plan can approach the SD plan quality. However, such BSD scheme relies on the patient being able to follow the same breathing curve that is used in the planning stage during radiation delivery. Funded by Varian Medical Systems.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-222: Computational Optimization of Monte Carlo Simulation On 4D Treatment Planning Using the Cloud Computing Technology

Purpose: This study evaluated the efficiency of 4D lung radiation treatment planning using Monte Carlo simulation on the cloud. The EGSnrc Monte Carlo code was used in dose calculation on the 4D-CT image set. Methods: 4D lung radiation treatment plan was created by the DOSCTP linked to the cloud, based on the Amazon elastic compute cloud platform. Dose calculation was carried out by Monte Carlo simulation on the 4D-CT image set on the cloud, and results were sent to the FFD4D image deformation program for dose reconstruction. The dependence of computing time for treatment plan on the number of compute node was optimized with variations of the number of CT image set in the breathing cycle and dose reconstruction time of the FFD4D. Results: It is found that the dependence of computing time on the number of compute node was affected by the diminishing return of the number of node used in Monte Carlo simulation. Moreover, the performance of the 4D treatment planning could be optimized by using smaller than 10 compute nodes on the cloud. The effects of the number of image set and dose reconstruction time on the dependence of computing time on the number of node were not significant, as more than 15 compute nodes were used in Monte Carlo simulations. Conclusion: The issue of long computing time in 4D treatment plan, requiring Monte Carlo dose calculations in all CT image sets in the breathing cycle, can be solved using the cloud computing technology. It is concluded that the optimized number of compute node selected in simulation should be between 5 and 15, as the dependence of computing time on the number of node is significant.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-287: Dose Verification On the Variation of Target Volume and Organ at Risk in Preradiation Chemotherapy IMRT for Nasopharyngeal Cancer

Purpose: To quantify the target volume and organ at risk of nasopharyngeal carcinoma (NPC) patients with preradiation chemotherapy based on CT scanned during intensity-modulated radiotherapy (IMRT), and recalculate the dose distribution. Methods: Seven patients with NPC and preradiation chemotherapy, treated with IMRT (35 to 37 fractions) were reviewed. Repeat CT scanning was required to all of the patients during the radiotherapy, and the number of repeat CTs varies from 2 to 6. The plan CT and repeat CT were generated by different CT scanner. To ensure crespectively on the same IMPT plan. The real dose distribution was calculated by deformable registration and weighted method in Raystation (v 4.5.1). The fraction of each dose is based on radiotherapy record. The volumetric and dose differences among these images were calculated for nascIpharyngeal tumor and retro-pharyngeal lymph nodes (GTV-NX), neck lymph nodes(GTV-ND), and parotid glands. Results: The volume variation in GTV-NX from CT1 to CT2 was 1.15±3.79%, and in GTV-LN −0.23±4.93%. The volume variation in left parotid from CT1 to CT2 was −6.79±11.91%, and in right parotid −3.92±8.80%. In patient 2, the left parotid volume were decreased remarkably, as a Result, the V30 and V40 of it were increased as well. Conclusion: The target volume of patients with NPC varied lightly during IMRT. It shows that preradiation chemotherapy can control the target volume variation and perform a good dose repeatability. Also, the decreasing volume of parotid in some patient might increase the dose of it, which might course potential complications.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-428: Feasibility Study of 4D Image Reconstruction by Organ Motion Vector Extension Based On Portal Images

Purpose: To develop and to test a method to generate a new 4D CT images of the treatment day from the old 4D CT and the portal images of the day when the motion extent exceeded from that represented by plan CTs. Methods: A motion vector of a moving tumor in a patient may be extended to reconstruct the tumor position when the motion extent exceeded from that represented by plan CTs. To test this, 1. a phantom that consists of a polystyrene cylinder (tumor) embedded in cork (lung) was placed on a moving platform with 4 sec/cycle and amplitudes of 1 cm and 2 cm, and was 4D-scanned. 2. A 6MV photon beam was irradiated on the moving phantoms and cineEPID images were obtained. 3. A motion vector of the tumor was acquired from 4D CT images of the phantom with 1 cm amplitude. 4. From cine EPID images of the phantom with the 2 cm amplitude, various motion extents (0.3 cm, 0.5 cm, etc) were acquired and programmed into the motion vector, producing CT images at each position. 5. The reconstructed CT images were then compared with pre-acquired “reference” 4D CT images at each position (i.e. phase). Results: The CT image was reconstructed and compared with the reference image, showing a slight mismatch in the transition direction limited by voxel size (slice thickness) in CT image. Due to the rigid nature of the phantom studied, the modeling the displacement of the center of object was sufficient. When deformable tumors are to be modeled, more complex scheme is necessary, which utilize cine EPID and 4D CT images. Conclusion: The new idea of CT image reconstruction was demonstrated. Deformable tumor movements need to be considered in the future.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-431: Feasiblity of Using CT Scout Images for 2D LDR Brachytherpay Planning

Purpose: i) To show the feasibility of using CT scout images for 2D low-dose rate brachytherapy planning with BrachyVision (version 10.4); ii) to show their advantages and disadvantages over DRRs. Methods: A phantom was constructed to house a Fletcher-Suite applicator. The phantom is made of Styrofoam with metal BBs positioned at well-defined separations. These markers are used to assess the image distortion in the scout images. Unlike DRRs, scout images are distorted only in the direction normal to the couch direction; therefore, they needed to be scaled unidirectionally prior to importing into BrachyVision. In addition to confirming the scaling is performed correctly by measuring distances between well-positioned BB, we also compare a LDR plan using scout images to a 3D CT-based plan. Results: There is no distortion of the image along the couch direction due to the collimation of the CT scanner. The distortion in the transverse plane can be corrected by multiplying by the ratio of distances between source-to-isocenter and source-to-detector. The results show the distance separations between BBs as measured in scout images and by a caliber are within a few millimeters. Dosimetrically, the difference between the dose rates to points A and B based on scout images and on 3D CT are less than a few percents. The accuracy can be improved by correcting for the distortion on the transverse plane. Conclusion: It is possible to use CT scout images for 2D planning in BrachyVision. This is an advantage because scout images have no metal artifacts often present in CT images or DRRs. Another advantage is the lack of distortion in the couch direction. One major disadvantage is that the image distortion due to beam divergence can be large. This is due to the inherent short distance between source-to-isocenter and source-to-detector on a CT scanner.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-457: Impact of Interfractional Variations On Anterior Vs. Lateral-Field Proton Therapy of Prostate Cancer

Purpose: To investigate the effects of interfractional anatomy and setup variations on plans with anterior-oblique vs. lateral beams for prostate cancer pencil beam scanning (PBS) and passive scattered (PS) proton therapy. Methods: Six patients with low/intermediate risk prostate cancer treated with PS proton therapy at our institution were selected. All patients underwent weekly verification CT scans. Implanted fiducials were used for localization, and endorectal balloons for prostate immobilization. New PBS plans with lateral beams, as well as PBS and PS plans with anterior-oblique beams (±35 deg) were created. PBS plans used two different spot sizes: ∼10mm (large) and ∼5mm (medium) sigma at 25cm range and optimized as single-field-uniform-dose with ∼8% non-uniformity. No range uncertainty margins were applied in PBS plans to maximize rectal sparing. Field-specific apertures were used when planning with large spots to sharpen the penumbrae. The planned dose was recomputed on each weekly CT with fiducials aligned to the simulation CT, scaled and accumulated via deformable image registration. Results: The dose volume analysis showed that although difference between planned and accumulated dose remains negligible for plans with conventional lateral beams using both PS and PBS, this is not the case for plans with anterior beams. The target coverage in anterior plans was largely degraded due to the variations in the beam path length and the absence of range margins. The average prostate D95 was reduced by 7.5/15.9% (using PS/PBS) after accumulation for anterior plans, compared with 0/0.4% for lateral plans. The average mean dose in organs-at-risk decreased by 1% for lateral and 2% for anterior plans, similarly for PS and PBS. Spot size did not affect the dose changes. Conclusion: Prostate plans using anterior beams may undergo clinically relevant interfractional dose degradation. Corrective strategies guided by in-vivo range measurements should be studied before clinical application of this technique.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-496: Influence of the Intrafraction Patient Repositioning On Spot Scanning Proton Therapy for Prostate Cancer

Purpose To verify the effect of intrafraction patient repositioning based on real time image guidance on the proton dose distribution for prostate cancer. Methods Dose distributions were simulated using CT images of prostate cancer patients. Treatment plans were prepared using proton beam from left side of the patient with 3 and 6 mm lateral margins. During the beam delivery, prostate position was assumed to be shifted by 5 mm in anterior, posterior, superior and inferior directions. The dose distributions were calculated for two scenarios; (i) the patient position was readjusted by 5 mm just after the prostate shift occurs (ii) the patient position was not adjusted. Such a real time readjustment can be realized with real time image guidance. The differences of the values D{sub 5} and D{sub 95} in CTV (D{sub 5–95}) were compared between different sizes of the lateral margins and different scenarios. Results D{sub 5–95} values in the case of scenario (i) are smaller than scenario (ii). The differences between scenario (i) and (ii) were not depend on the directions of the prostate shift. In the case of scenario (ii), margin expansion makes D{sub 5–95} values small. In the case of scenario (i), margin size is uncorrelated to D{sub 5–95} values. Conclusion These results indicate that the intrafraction patient repositioning is effective to minimize dose deterioration when the prostate moves during the irradiation. In the case without intrafraction patient repositioning, margin expansion is effective to reduce dose deterioration. In the case with intrafraction patient repositioning, 3 mm margin is enough to minimize dose deterioration.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-509: Inter-Observer and Inter-Modality Contouring Analysis for Organs at Risk for HDR Gynecological Brachytherapy

Purpose This study quantifies errors associated with MR-guided High Dose Rate (HDR) gynecological brachytherapy. Uncertainties in this treatment results from contouring, organ motion between imaging and treatment delivery, dose calculation, and dose delivery. We focus on interobserver and inter-modality variability in contouring and the motion of organs at risk (OARs) in the time span between the MR and CT scans (∼1 hour). We report the change in organ volume and position of center of mass (CM) between the two imaging modalities. Methods A total of 8 patients treated with MR-guided HDR brachytherapy were included in this study. Two observers contoured the bladder and rectum on both MR and CT scans. The change in OAR volume and CM position between the MR and CT imaging sessions on both image sets were calculated. Results The absolute mean bladder volume change between the two imaging modalities is 67.1cc. The absolute mean inter-observer difference in bladder volume is much lower at 15.5cc (MR) and 11.0cc (CT). This higher inter-modality volume difference suggests a real change in the bladder filling between the two imaging sessions. Change in Rectum volume inter-observer standard error of means (SEM) is 3.18cc (MR) and 3.09cc (CT), while the inter-modality SEM is 3.65cc (observer 1), and 2.75cc (observer 2). The SEM for rectum CM position in the superior-inferior direction was approximately three times higher than in other directions for both the inter—observer (0.77 cm, 0.92 cm for observers 1 and 2, respectively) and inter-modality (0.91 cm, 0.95 cm for MR and CT, respectively) variability. Conclusion Bladder contours display good consistency between different observers on both CT and MR images. For rectum contouring the highest inconsistency stems from the observers’ choice of the superior-inferior borders. A complete analysis of a larger patient cohort will enable us to separate the true organ motion from the inter-observer variability.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-754: Three-Dimensional Patient Modeling Using Photogrammetry for Collision Avoidance

Purpose: To evaluate photogrammetry for creating a three-dimensional patient model. Methods: A mannequin was configured on the couch of a CT scanner to simulate a patient setup using an indexed positioning device. A CT fiducial was placed on the indexed CT table-overlay at the reference index position. Two dimensional photogrammetry targets were placed on the table in known positions. A digital SLR camera was used to obtain 27 images from different positions around the CT table. The images were imported into a commercial photogrammetry package and a 3D model constructed. Each photogrammetry target was identified on 2 to 5 images. The CT DICOM metadata and the position of the CT fiducial were used to calculate the coordinates of the photogrammetry targets in the CT image frame of reference. The coordinates were transferred to the photogrammetry software to orient the 3D model. The mannequin setup was transferred to the treatment couch of a linear accelerator and positioned at isocenter using in-room lasers. The treatment couch coordinates were noted and compared with prediction. The collision free regions were measured over the full range of gantry and table motion and were compared with predictions obtained using a general purpose polygon interference algorithm. Results: The reconstructed 3D model consisted of 180000 triangles. The difference between the predicted and measured couch positions were 5 mm, 1 mm, and 1 mm for longitudinal, lateral, and vertical, respectively. The collision prediction tested 64620 gantry table combinations in 11.1 seconds. The accuracy was 96.5%, with false positive and negative results occurring at the boundaries of the collision space. Conclusion: Photogrammetry can be used as a tool for collision avoidance during treatment planning. The results indicate that a buffer zone is necessary to avoid false negatives at the boundary of the collision-free zone. Testing with human patients is underway. Research partially supported by a grant from Varian Medical Systems.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-F-BRD-11: A Virtual Simulator Designed for Collision Prevention in Proton Therapy

Purpose: In proton therapy, collisions between patient and nozzle potentially occur in attaining minimal air gap due to the large nozzle structure. Thus, we developed software predicting the collisions of the nozzle and patient by simulating treatments. Methods: 3D modeling of a gantry inner-floor, nozzle and robotic-couch was done by using the SolidWorks based on the manufacturer’s machine data. To obtain patient body information, a 3D-scanner was utilized to scan a patient right before CT scanning. From the acquired images, a 3D-image of the patient’s body contour was reconstructed. The accuracy of the image was confirmed against the CT image for a humanoid phantom. The machine components and the virtual patient were combined on the treatment-room coordinate system, resulting in a virtual simulator. The simulator simulated the motion of its components such as rotation and translation of gantry, nozzle and couch, in real scale. Collision, if any, was examined both in static mode and dynamic mode. The static mode checks only at fixed positions of the machine’s components while dynamic mode examines while one component is in motion. Collision was notified if any voxel of two components, for example a nozzle and a patient or couch, overlapped when calculating volume locations. The event and collision point are visualized and colliding volumes are reported. Results: All components were successfully assembled and the motions could be accurately controlled. The 3D-shape of a phantom agreed with CT images within a deviation of 2 mm. Collision situations can be simulated within minutes and the results are displayed and reported. Conclusion: The developed software will be useful in improving patient safety and clinical efficiency for proton therapy. This work was supported by the National Research Foundation of Korea funded by Ministry of Science, ICT & Future Planning (2012M3A9B6055201, 2013M2A2A7043507), and Samsung Medical Center grant (GFO1130081)

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-F-BRA-04: Prostate HDR Brachytherapy with Multichannel Robotic System

Purpose: High-dose-rate (HDR) brachytherapy is gradually becoming popular in treating patients with prostate cancers. However, placement of the HDR needles at desired locations into the patient is challenging. Application of robotic system may improve the accuracy of the clinical procedure. This experimental study is to evaluate the feasibility of using a multichannel robotic system for prostate HDR brachytherapy. Methods: In this experimental study, the robotic system employed was a 6-DOF Multichannel Image-guided Robotic Assistant for Brachytherapy (MIRAB), which was designed and fabricated for prostate seed implantation. The MIRAB has the provision of rotating 16 needles while inserting them. Ten prostate HDR brachytherapy needles were simultaneously inserted using MIRAB into a commercially available prostate phantom. After inserting the needles into the prostate phantom at desired locations, 2mm thick CT slices were obtained for dosimetric planning. HDR plan was generated using Oncetra planning system with a total prescription dose of 34Gy in 4 fractions. Plan quality was evaluated considering dose coverage to prostate and planning target volume (PTV), with 3mm margin around prostate, as well as the dose limit to the organs at risk (OARs) following the American Brachytherapy Society (ABS) guidelines. Results: From the CT scan, it is observed that the needles were inserted straight into the desired locations and they were adequately spaced and distributed for a clinically acceptable HDR plan. Coverage to PTV and prostate were about 91% (V100= 91%) and 96% (V100=96%), respectively. Dose to 1cc of urethra, rectum, and bladder were within the ABS specified limits. Conclusion: The MIRAB was able to insert multiple needles simultaneously into the prostate precisely. By controlling the MIRAB to insert all the ten utilized needles into the prostate phantom, we could achieve the robotic HDR brachytherapy successfully. Further study for assessing the system’s performance and reliability is in progress.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-F-303-17: Real Time Dose Calculation of MRI Guided Co-60 Radiotherapy Treatments On Free Breathing Patients, Using a Motion Model and Fast Monte Carlo Dose Calculation

Purpose: To demonstrate real-time dose calculation of free-breathing MRI guided Co−60 treatments, using a motion model and Monte-Carlo dose calculation to accurately account for the interplay between irregular breathing motion and an IMRT delivery. Methods: ViewRay Co-60 dose distributions were optimized on ITVs contoured from free-breathing CT images of lung cancer patients. Each treatment plan was separated into 0.25s segments, accounting for the MLC positions and beam angles at each time point. A voxel-specific motion model derived from multiple fast-helical free-breathing CTs and deformable registration was calculated for each patient. 3D images for every 0.25s of a simulated treatment were generated in real time, here using a bellows signal as a surrogate to accurately account for breathing irregularities. Monte-Carlo dose calculation was performed every 0.25s of the treatment, with the number of histories in each calculation scaled to give an overall 1% statistical uncertainty. Each dose calculation was deformed back to the reference image using the motion model and accumulated. The static and real-time dose calculations were compared. Results: Image generation was performed in real time at 4 frames per second (GPU). Monte-Carlo dose calculation was performed at approximately 1frame per second (CPU), giving a total calculation time of approximately 30 minutes per treatment. Results show both cold- and hot-spots in and around the ITV, and increased dose to contralateral lung as the tumor moves in and out of the beam during treatment. Conclusion: An accurate motion model combined with a fast Monte-Carlo dose calculation allows almost real-time dose calculation of a free-breathing treatment. When combined with sagittal 2D-cine-mode MRI during treatment to update the motion model in real time, this will allow the true delivered dose of a treatment to be calculated, providing a useful tool for adaptive planning and assessing the effectiveness of gated treatments.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

MO-DE-210-06: Development of a Supercompounded 3D Volumetric Ultrasound Image Guidance System for Prone Accelerated Partial Breast Irradiation (APBI)

Purpose: For early-stage breast cancer, accelerated partial breast irradiation (APBI) is a cost-effective breast-conserving treatment. Irradiation in a prone position can mitigate respiratory induced breast movement and achieve maximal sparing of heart and lung tissues. However, accurate dose delivery is challenging due to breast deformation and lumpectomy cavity shrinkage. We propose a 3D volumetric ultrasound (US) image guidance system for accurate prone APBI Methods: The designed system, set beneath the prone breast board, consists of a water container, an US scanner, and a two-layer breast immobilization cup. The outer layer of the breast cup forms the inner wall of water container while the inner layer is attached to patient breast directly to immobilization. The US transducer scans is attached to the outer-layer of breast cup at the dent of water container. Rotational US scans in a transverse plane are achieved by simultaneously rotating water container and transducer, and multiple transverse scanning forms a 3D scan. A supercompounding-technique-based volumetric US reconstruction algorithm is developed for 3D image reconstruction. The performance of the designed system is evaluated with two custom-made gelatin phantoms containing several cylindrical inserts filled in with water (11% reflection coefficient between materials). One phantom is designed for positioning evaluation while the other is for scaling assessment. Results: In the positioning evaluation phantom, the central distances between the inserts are 15, 20, 30 and 40 mm. The distances on reconstructed images differ by −0.19, −0.65, −0.11 and −1.67 mm, respectively. In the scaling evaluation phantom, inserts are 12.7, 19.05, 25.40 and 31.75 mm in diameter. Measured inserts’ sizes on images differed by 0.23, 0.19, −0.1 and 0.22 mm, respectively. Conclusion: The phantom evaluation results show that the developed 3D volumetric US system can accurately localize target position and determine target volume, and is a promising image-guidance tool for prone APBI.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

MO-FG-204-01: Improved Noise Suppression for Dual-Energy CT Through Entropy Minimization

Purpose: In dual energy CT (DECT), noise amplification during signal decomposition significantly limits the utility of basis material images. Since clinically relevant objects contain a limited number of materials, we propose to suppress noise for DECT based on image entropy minimization. An adaptive weighting scheme is employed during noise suppression to improve decomposition accuracy with limited effect on spatial resolution and image texture preservation. Methods: From decomposed images, we first generate a 2D plot of scattered data points, using basis material densities as coordinates. Data points representing the same material generate a highly asymmetric cluster. We orient an axis by minimizing the entropy in a 1D histogram of these points projected onto the axis. To suppress noise, we replace pixel values of decomposed images with center-of-mass values in the direction perpendicular to the optimal axis. To limit errors due to cluster overlap, we weight each data point’s contribution based on its high and low energy CT values and location within the image. The proposed method’s performance is assessed on physical phantom studies. Electron density is used as the quality metric for decomposition accuracy. Our results are compared to those without noise suppression and with a recently developed iterative method. Results: The proposed method reduces noise standard deviations of the decomposed images by at least one order of magnitude. On the Catphan phantom, this method greatly preserves the spatial resolution and texture of the CT images and limits induced error in measured electron density to below 1.2%. In the head phantom study, the proposed method performs the best in retaining fine, intricate structures. Conclusion: The entropy minimization based algorithm with adaptive weighting substantially reduces DECT noise while preserving image spatial resolution and texture. Future investigations will include extensive investigations on material decomposition accuracy that go beyond the current electron density calculations. This work was supported in part by the National Institutes of Health (NIH) under Grant Number R21 EB012700.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

MO-FG-204-03: Using Edge-Preserving Algorithm for Significantly Improved Image-Domain Material Decomposition in Dual Energy CT

Purpose: To significantly improve dual energy CT (DECT) imaging by establishing a new theoretical framework of image-domain material decomposition with incorporation of edge-preserving techniques. Methods: The proposed algorithm, HYPR-NLM, combines the edge-preserving non-local mean filter (NLM) with the HYPR-LR (Local HighlY constrained backPRojection Reconstruction) framework. Image denoising using HYPR-LR framework depends on the noise level of the composite image which is the average of the different energy images. For DECT, the composite image is the average of high- and low-energy images. To further reduce noise, one may want to increase the window size of the filter of the HYPR-LR, leading resolution degradation. By incorporating the NLM filtering and the HYPR-LR framework, HYPR-NLM reduces the boost material decomposition noise using energy information redundancies as well as the non-local mean. We demonstrate the noise reduction and resolution preservation of the algorithm with both iodine concentration numerical phantom and clinical patient data by comparing the HYPR-NLM algorithm to the direct matrix inversion, HYPR-LR and iterative image-domain material decomposition (Iter-DECT). Results: The results show iterative material decomposition method reduces noise to the lowest level and provides improved DECT images. HYPR-NLM significantly reduces noise while preserving the accuracy of quantitative measurement and resolution. For the iodine concentration numerical phantom, the averaged noise levels are about 2.0, 0.7, 0.2 and 0.4 for direct inversion, HYPR-LR, Iter- DECT and HYPR-NLM, respectively. For the patient data, the noise levels of the water images are about 0.36, 0.16, 0.12 and 0.13 for direct inversion, HYPR-LR, Iter-DECT and HYPR-NLM, respectively. Difference images of both HYPR-LR and Iter-DECT show edge effect, while no significant edge effect is shown for HYPR-NLM, suggesting spatial resolution is well preserved for HYPR-NLM. Conclusion: HYPR-NLM provides an effective way to reduce the generic magnified image noise of dual–energy material decomposition while preserving resolution. This work is supported in part by NIH grants 7R01HL111141 and 1R01-EB016777. This work is also supported by the Natural Science Foundation of China (NSFC Grant No. 81201091), Fundamental Research Funds for the Central Universities in China, and Fund Project for Excellent Abroad Scholar Personnel in Science and Technology.

60 APPLIED LIFE SCIENCES↗