A consistency evaluation of signal-to-noise ratio in the quality assessment of human brain magnetic resonance images.

BACKGROUND: Quality assessment of medical images is highly related to the quality assurance, image interpretation and decision making. As to magnetic resonance (MR) images, signal-to-noise ratio (SNR) is routinely used as a quality indicator, while little knowledge is known of its consistency regarding different observers. METHODS: In total, 192, 88, 76 and 55 brain images are acquired using T2*, T1, T2 and contrast-enhanced T1 (T1C) weighted MR imaging sequences, respectively. To each imaging protocol, the consistency of SNR measurement is verified between and within two observers, and white matter (WM) and cerebral spinal fluid (CSF) are alternately used as the tissue region of interest (TOI) for SNR measurement. The procedure is repeated on another day within 30 days. At first, overlapped voxels in TOIs are quantified with Dice index. Then, test-retest reliability is assessed in terms of intra-class correlation coefficient (ICC). After that, four models (BIQI, BLIINDS-II, BRISQUE and NIQE) primarily used for the quality assessment of natural images are borrowed to predict the quality of MR images. And in the end, the correlation between SNR values and predicted results is analyzed. RESULTS: To the same TOI in each MR imaging sequence, less than 6% voxels are overlapped between manual delineations. In the quality estimation of MR images, statistical analysis indicates no significant difference between observers (Wilcoxon rank sum test, p w ≥ 0.11; paired-sample t test, p p ≥ 0.26), and good to very good intra- and inter-observer reliability are found (ICC, p icc ≥ 0.74). Furthermore, Pearson correlation coefficient (r p ) suggests that SNRwm correlates strongly with BIQI, BLIINDS-II and BRISQUE in T2* (r p ≥ 0.78), BRISQUE and NIQE in T1 (r p ≥ 0.77), BLIINDS-II in T2 (r p ≥ 0.68) and BRISQUE and NIQE in T1C (r p ≥ 0.62) weighted MR images, while SNRcsf correlates strongly with BLIINDS-II in T2* (r p ≥ 0.63) and in T2 (r p ≥ 0.64) weighted MR images. CONCLUSIONS: The consistency of SNR measurement is validated regarding various observers and MR imaging protocols. When SNR measurement performs as the quality indicator of MR images, BRISQUE and BLIINDS-II can be conditionally used for the automated quality estimation of human brain MR images.

Active learning using adaptable task-based prioritisation.

Supervised machine learning-based medical image computing applications necessitate expert label curation, while unlabelled image data might be relatively abundant. Active learning methods aim to prioritise a subset of available image data for expert annotation, for label-efficient model training. We develop a controller neural network that measures priority of images in a sequence of batches, as in batch-mode active learning, for multi-class segmentation tasks. The controller is optimised by rewarding positive task-specific performance gain, within a Markov decision process (MDP) environment that also optimises the task predictor. In this work, the task predictor is a segmentation network. A meta-reinforcement learning algorithm is proposed with multiple MDPs, such that the pre-trained controller can be adapted to a new MDP that contains data from different institutes and/or requires segmentation of different organs or structures within the abdomen. We present experimental results using multiple CT datasets from more than one thousand patients, with segmentation tasks of nine different abdominal organs, to demonstrate the efficacy of the learnt prioritisation controller function and its cross-institute and cross-organ adaptability. We show that the proposed adaptable prioritisation metric yields converging segmentation accuracy for a new kidney segmentation task, unseen in training, using between approximately 40% to 60% of labels otherwise required with other heuristic or random prioritisation metrics. For clinical datasets of limited size, the proposed adaptable prioritisation offers a performance improvement of 22.6% and 10.2% in Dice score, for tasks of kidney and liver vessel segmentation, respectively, compared to random prioritisation and alternative active sampling strategies.

A comparative study of medical image enhancement algorithms and quality assessment metrics on COVID-19 CT images.

Medical imaging can help doctors in better diagnosis of several conditions. During the present COVID-19 pandemic, timely detection of novel coronavirus is crucial, which can help in curing the disease at an early stage. Image enhancement techniques can improve the visual appearance of COVID-19 CT scans and speed-up the process of diagnosis. In this study, we analyze some state-of-the-art image enhancement techniques for their suitability in enhancing the CT scans of COVID-19 patients. Six quantitative metrics, Entropy, SSIM, AMBE, PSNR, EME, and EMEE, are used to evaluate the enhanced images. Two experienced radiologists were involved in the study to evaluate the performance of the enhancement techniques and the quantitative metrics used to assess them.

Robotic Ultrasound Scanning With Real-Time Image-Based Force Adjustment: Quick Response for Enabling Physical Distancing During the COVID-19 Pandemic.

During an ultrasound (US) scan, the sonographer is in close contact with the patient, which puts them at risk of COVID-19 transmission. In this paper, we propose a robot-assisted system that automatically scans tissue, increasing sonographer/patient distance and decreasing contact duration between them. This method is developed as a quick response to the COVID-19 pandemic. It considers the preferences of the sonographers in terms of how US scanning is done and can be trained quickly for different applications. Our proposed system automatically scans the tissue using a dexterous robot arm that holds US probe. The system assesses the quality of the acquired US images in real-time. This US image feedback will be used to automatically adjust the US probe contact force based on the quality of the image frame. The quality assessment algorithm is based on three US image features: correlation, compression and noise characteristics. These US image features are input to the SVM classifier, and the robot arm will adjust the US scanning force based on the SVM output. The proposed system enables the sonographer to maintain a distance from the patient because the sonographer does not have to be holding the probe and pressing against the patient's body for any prolonged time. The SVM was trained using bovine and porcine biological tissue, the system was then tested experimentally on plastisol phantom tissue. The result of the experiments shows us that our proposed quality assessment algorithm successfully maintains US image quality and is fast enough for use in a robotic control loop.

Deep Learning Based Model Observer by U-Net.

Model Observers (MO) are algorithms designed to evaluate and optimize the parameters of new medical imaging reconstruction methodologies by providing a measure of human accuracy for a diagnostic task. In contrast with a computer-aided diagnosis system, MOs are not designed to outperform human diagnosis but only to find a defect if a radiologist would be able to detect it. These algorithms can economize and expedite the finding of optimal reconstruction parameters by reducing the number of sessions with expert radiologists, which are costly and prolonged. Convolutional Neural Networks (CNN or ConvNet) have been successfully used in the computer vision field for image classification, segmentation and video analytics. In this paper, we propose and test several U-Net configurations as MO for a defect localization task on synthetic images with different levels of correlated noisy backgrounds. Preliminary results show that the CNN based MO has potential and its accuracy correlates well with that of the human.

Temporal stability assessment in shear wave elasticity images validated by deep learning neural network for chronic liver disease fibrosis stage assessment.

PURPOSE: To automatically detect and isolate areas of low and high stiffness temporal stability in shear wave elastography (SWE) image sequences and define their impact in chronic liver disease (CLD) diagnosis improvement by means of clinical examination study and deep learning algorithm employing convolutional neural networks (CNNs). MATERIALS AND METHODS: Two hundred SWE image sequences from 88 healthy individuals (F0 fibrosis stage) and 112 CLD patients (46 with mild fibrosis (F1), 16 with significant fibrosis (F2), 22 with severe fibrosis (F3), and 28 with cirrhosis (F4)) were analyzed to detect temporal stiffness stability between frames. An inverse Red, Green, Blue (RGB) colormap-to-stiffness process was performed for each image sequence, followed by a wavelet transform and fuzzy c-means clustering algorithm. This resulted in a binary mask depicting areas of high and low stiffness temporal stability. The mask was then applied to the first image of the SWE sequence, and the derived, masked SWE image was used to estimate its impact in standard clinical examination and CNN classification. Regarding the impact of the masked SWE image in clinical examination, one measurement by two radiologists was performed in each SWE image and two in the corresponding masked image measuring areas with high and low stiffness temporal stability. Then, stiffness stability parameters, interobserver variability evaluation and diagnostic performance by means of ROC analysis were assessed. The masked and unmasked sets of SWE images were fed into a CNN scheme for comparison. RESULTS: The clinical impact evaluation study showed that the masked SWE images decreased the interobserver variability of the radiologists' measurements in the high stiffness temporal stability areas (interclass correlation coefficient (ICC) = 0.92) compared to the corresponding unmasked ones (ICC = 0.76). In terms of diagnostic accuracy, measurements in the high-stability areas of the masked SWE images (area-under-the-curve (AUC) ranging from 0.800 to 0.851) performed similarly to those in the unmasked SWE images (AUC ranging from 0.805 to 0.893). Regarding the measurements in the low stiffness temporal stability areas of the masked SWE images, results for interobserver variability (ICC = 0.63) and diagnostic accuracy (AUC ranging from 0.622 to 0.791) were poor. Regarding the CNN classification, the masked SWE images showed improved accuracy (ranging from 82.5% to 95.5%) compared to the unmasked ones (ranging from 79.5% to 93.2%) for various CLD stage combinations. CONCLUSION: Our detection algorithm excludes unreliable areas in SWE images, reduces interobserver variability, and augments CNN's accuracy scores for many combinations of fibrosis stages.

Series: Practical Evaluation of Clinical Image Quality (3): Subjective Evaluation of Image Quality in Digital Radiography Systems.

In the environment of picture archiving and communication system (PACS), the subjective evaluation of medical image quality is performed using an image display monitor. Commonly, PACS and its image viewer can handle digital imaging and communications in medicine (DICOM) format image files that are displayed as 8-bit grayscale images in the monitor. However, the original grayscale of DICOM image is between 10-bit to 16-bit depth. Therefore, in order for DICOM images to be displayed in the monitor by PACS image viewer, they need to be converted into 8-bit depth grayscale. This conversion indicates that the original characteristics of DICOM images are altered when they are displayed in the monitor by PACS image viewer. Moreover, image data interpolation to match the image matrix for the resolution of the monitor also alters the characteristics of images. For these reasons, it is a must to recognize the influential factors of image display settings for the subjective evaluation of medical image quality.

SU-E-I-36: A Study to Compare Operator Dosimeter Exposure versus Fluoroscopic Time and Air Kerma in Fluoroscopically Guided Interventional Procedures.

PURPOSE: The total effective dose equivalent limit for occupationally radiation exposed persons has remained at 50 millisieverts per year since the 1960s. There is ongoing discussion whether this limit should be lowered. Whether or not it is lowered, all facilities have to adhere to the principle of ALARA (as low as reasonably achievable). In a hospital environment, the workers who typically have exposures which Result in ALARA investigations having to be performed include interventional radiologists, cardiologists, and mobile C-arm operators. At our institution there are a limited number of interventional radiologists with one person being responsible for performing the majority of the procedures and sometimes exceeding the ALARA investigational levels. METHODS: A limited study was performed to compare the number, type of studies, cumulative air kerma, and fluoroscopic time versus the exposure measured by the personnel dosimeters worn by the radiologists. This was also related to the type of angiographic equipment in use. RESULTS: The length of fluoroscopic time and cumulative air kerma and cumulative DAP appeared to be related to the higher exposures received by the operator. The number of procedures performed did not correlate with the personnel dosimeter exposure; rather it was the type and complexity of a study and the length of fluoroscopic time that was related. CONCLUSION: The number of ALARA investigations required has decreased gradually. This could be attributed to several factors. These include increased use of available safety items. With an increase in the number of interventionalists the work load is distributed more equitably. The purchase of new angiographic equipment has resulted in better image quality and more shielding options for the operators. The operators are periodically reminded to keep fluoroscopic time to a minimum and use appropriate settings on the equipment. The equipment is calibrated to provide adequate image quality at mid-dose settings.

SU-E-J-104: Evaluation of Atlas-Based Auto-Segmentation on Daily In-Room CT for Prostate Cancer.

PURPOSE: Anatomy delineation is a major time consuming task to correct for inter-fractional changes in anatomy. Atlas-based auto-segmentation (ABAS) was developed to expedite this process. This study aims to evaluate the performance of ABAS applying to high quality verification CT-imaging acquired using a CT-on-rail system for prostate cancer. METHODS: The prostate, rectum and bladder were manually contoured for seven prostate cancer patients. For each patient, three patient specific atlases were generated consisting of one, four and seven prior image and contour sets. ABAS was applied using these atlases for the last seven daily CT images of each patient. The auto- and manual-contours were compared both geometrically and dosimetrically. The reproducibility of the observation was validated by an experienced radiation oncologist performing the same procedure. The performance of ABAS with patient and non-patient specific atlases were also evaluated on 21 image sets. P<0.05 was considered statistical significant for two-tailed paired student t-test. RESULTS: Contours obtained from ABAS agreed well with the manual ones. With 1-image set atlas, the OI and DSC for the bladder were greater than 96% and 91%, respectively. Both indices were above 81% for rectum and prostate. The consistencies significantly improved by including 4 image sets in the atlas, while a further increase of atlas size to 7 did not show obvious benefits. Dose coverage for the auto- and manual-contours was similar for all organs. Similar results were obtained by the second observer. Compared with non-patient specific atlas, patient specific atlas yielded more accurate contours. The time for ABAS and manual contouring was ∼2 min and ∼20 min per image set, respectively. CONCLUSIONS: With diagnostic quality verification images, ABAS can provide fast and accurate delineations for prostate cancer on a daily basis. The inclusion of more than one CT set in atlas improves the contouring results.

Sci-Sat AM: Brachy - 09: Permanent seed re-implantation using image guidance for composite dose planning.

Treatment outcome has been associated with dosimetric implant quality assessed in a postplan. Uncertainties during the implant and in the period before postplan evaluation lead to dosimetric deviations from the original treatment plan. The observed range of postplan results can include patients with dosimetry below recommended levels. Further treatment options may be considered to boost the delivered dose and a second seed implant is one way to accomplish this. This work describes the development of a procedure to plan and deliver a second seed implant and an evaluation of the dosimetric improvement. A patient with a post-plan D90 of 124 Gy was offered a second seed implant 21 weeks after the initial implant. A total dose of 163 Gy was prescribed due to radiobiological considerations for the time between implants. A volume study was performed 2 weeks before the implant and variable angle images of the seeds from the first implant were obtained to reconstruct the original seed coordinates. The second implant was planned considering the total composite physical dose distribution by forward planning additional seeds onto the original implant coordinates. The additional seeds were manually added to a new plan on the volume study images at the planned coordinates and delivered as a typical preplanned implant. Post-plan evaluation four weeks after the second implant showed a D90 of 177 Gy and a V100 of 92.7%. These results demonstrate the effective use of a second seed implant with image guided composite planning to improve dosimetric implant quality.

SU-D-BRA-03: Simultaneous MV-KV Imaging for Intra-Fractional Motion Management during Volume Modulated Arc Therapy (VMAT) Delivery on the Varian TrueBeam.

PURPOSE: To evaluate a MV-kV intra-fractional imaging technique for use during volume modulated arc therapy (VMAT) with the Varian TrueBeam. METHODS: MV-kV image pairs were acquired intra-fractionally during VMAT delivery. kV images (11 fps) were acquired throughout delivery using a standard pre-programmed imaging template. MV images (9.5 fps) were acquired simultaneously by deploying the EPID and passively collecting the resulting images using Varian proprietary software, iTools Capture. Localization accuracy was evaluated by imaging a Rando phantom implanted with 3 fiducials while moving the couch according to XML- programmed trajectories simulating typical prostate and respiratory motion. VMAT delivery was done using a single 360 degree arc in TrueBeam Developer mode. The effect on accuracy of total MU and gantry speed was studied. To improve image quality, MV frame averaging was performed and the MV and kV images were then registered to their corresponding DRRs using in-house registration software. From these 2D registrations, the 3D position at each MV-kV acquisition point was determined. RESULTS: Between 130 and 390 MV-kV pairs were acquired for each delivery. The mean difference between planned couch and measured fiducial 3D positions with prostate motion was less than 0.03 cm in each direction (SD 0.03 cm). Neither gantry speed nor MU significantly impacted accuracy. for respiratory motion, the mean difference between planned and measured position was less than 0.04 cm. Standard deviation averaged 0.06 cm but increased to 0.12 cm with large instantaneous motion and less MV dose per frame. MV frame averaging and inaccuracies in MV image gantry angle determination also affected accuracy, particularly with significant motion. CONCLUSIONS: With high quality MV imaging, MV-kV localization techniques can be highly accurate, even in the presence of significant motion. As clinical MV-kV methods become available, such techniques can provide an efficient and accurate method for monitoring intra-fractional motion. This work was partially supported through a research agreement with Varian Medical Systems, Palo Alto, CA.

SU-F-BRCD-07: Experimental Validation of Fluence Field Modulation for Noise and Dose Management in CT.

PURPOSE: Dose management in CT is a growing concern as the number of CT scans per capita rises. Fluence field modulated computed tomography (FFMCT) is a proposed approach for more optimal dose management, where the incident fluence pattern can change independently for each projection in order to minimize dose while meeting prescribed image quality metrics. This work studies the application of FFMCT for dose and noise management to a small cylindrical phantom. METHODS: Experiments were carried out on an experimental CT system using a small cylindrical phantom comprised of acrylic spheres, water and teflon rods as the object of interest. Modulation of the incident fluence was optimized using a simulated annealing algorithm that attempts to achieve user-prescribed, regionally varying SNR criteria while limiting dose. The modulated projections were synthesized from projection sets taken at various tube current settings. SNR patterns and predicted dose outcomes were compared to the prescribed values as well as against expected results using a bowtie filter. RESULTS: FFMCT resulted in SNR outcomes with greater agreement to the prescribed regionally varying SNR criteria than achieved by the bowtie filter. Dose reductions were observed for the majority of the object, and reached as high as 60%. Limited regions, coinciding with lower noise, observed dose increases less than 20%. Integral dose reduction over the entire object was greater than 40% when compared to a bowtie filter with comparable image quality in the region of interest. CONCLUSIONS: The results support the hypothesis that FFMCT has the potential to meet user-prescribed, regionally varying image quality objectives, while decreasing radiation dose to the patient. These results suggest that given a suitable collimator approach, fluence field modulated computed tomography could reap significant benefits in terms of reducing dose and optimizing image quality. NSERC.

SU-E-I-105: Optimizing Bomb Squad X-Ray Systems for Incidental Human Exposures.

PURPOSE: A current terrorist tactic involves strapping an explosive device to a live victim. Technicians wishing to x-ray the device need to use appropriate exposures to create clear images while mitigating incidental health problems for the live victim. This project involves characterizing and optimizing a popular radiographic system in the explosive investigation community, both for general use and, more specifically, for generating x-rays with a human subject in the beam path. This project seeks to minimize the dose to a victim incidentally located in the beam path, while maximizing image quality, thereby allowing appropriate image evaluation to neutralize the explosive threat. METHODS: Image quality and dose optimization begins with characterization of the x-ray source. The source in this study is an XRS-3, a portable device developed by Golden Engineering, which nominally produces a 270 kVp beam. Characterization of the beam involves verifying the actual kVp, as well as determining the average energy of the beam with various amounts of filtration added to the beam path. The kVp was verified using a noninvasive technique, utilizing the measurement of the dose attenuation through lead and then modeling the energy dependent attenuation curve of the lead using specific mass attenuation coefficients and parameters suggested by Joseph (Joseph, P.M., Med. Phys. Vol. 2, July/Aug1975; 201-207) in order to calculate the kVp through linear regression. RESULTS: The average energy of an unfiltered beam was 52 keV, and increases through adding various combinations of Cu and brass filters to a maximum value of 121 keV with 3 mm copper + 3 mm brass filtration. The calculated kVp was 290 kV. CONCLUSIONS: The average energies using filtration fall within the diagnostic range. The kVp fits well within the manufacturer's specifications of 15% deviation. Optimization can be accomplished with these parameters. University of Oklahoma Health Sciences Center- Department of Radiological Sciences.

MO-A-218-01: CT Protocol Review - Practical Tips for Imaging Physicists.

In the 1980's and 90's, when every mammography department had a wet film processor and a sundial to keep the schedule, medical physicists performing mammography surveys were primarily focused on measuring machine performance and image quality. As our professional experience matured, medical physicists began to learn that they were uniquely qualified to help to recommend technique factors that would balance dose and image quality. Technique charts using different kVp, target-filter combinations and AEC modes gradually became common and patients benefitted from our input. With the revolutionary change in CT Scanner technology and utilization, medical physicists have begun to contribute their expertise to developing and improving CT protocols. This presentation will present practical challenges and offer some directions for the practicing medical physicist who desires to participate in this critical and emerging aspect of imaging physics practice: CT Protocol Review.

MO-D-BRA-01: Limits of Dose Reduction in CT: Where are They and How Will We Know When We Get There?

Radiation Dose continues to be a concern with respect to all diagnostic imaging using ionizing radiation, but especially so with CT imaging. We have always known how to reduce radiation dose in CT - for example, simply turning down the system output (e.g. reduce mAs). What we have not been able to do is to simultaneously reduce dose and maintain "diagnostic image quality". Many recent technical developments have appeared, and will continue to appear, that will allow users to reduce radiation dose in CT while "maintaining image quality". However, this last term is ill-defined and current metrics of image quality are not very applicable to actual clinical practice. The purpose of this symposium is to: (a) describe several current and possible future radiation dose reduction methods and the magnitude of their potential for dose reduction, (b) some description of what "diagnostic image quality" means, the effects that dose reductions methods have on this property, description of some metrics that may help us assess this property quantitatively and this information can be used to guide how low radiation doses can be reduced. LEARNING OBJECTIVES: 1. Understand both conventional and emerging radiation dose reduction methods in CT. 2. Understand the implications on diagnostic image quality for each radiation dose reduction method. 3. Understand some of the issues in evaluating how much radiation dose can be reduced and still accomplish a diagnostic imaging task.

SU-E-I-27: Effect of Focal Spot Choice on the System Performance of a High- Resolution Micro-Angiographic Fluoroscope (MAF) Detector.

PURPOSE: The generalized metrics of GMTF and GDQE were used to evaluate the effect of focal-spot choice on system performance of the Micro- Angiographic Fluoroscope (MAF) imager with a geometric magnification common for neuro-endovascular interventions. METHODS: The MAF, a newly developed high-resolution detector with very low instrumentation noise and large variable gain, was used for the study. It has 35-micron pixels and a 500-micron thick CsI HR-type phosphor. The detector MTF was measured using the slanted edge method and the focal spot MTF's were measured using a standard pin-hole assembly. For the comparison analysis, the GMTF and GDQE were determined for three different focal spots (0.3 mm, 0.5 mm and, 0.8 mm) with a magnification factor of 1.25. A stainless- steel stent was also imaged with the same magnification for all three focal spots. RESULTS: The MAF's performance is affected significantly by the choice of focal-spot size because of its very high resolution. The GMTF and GDQE comparison for three focal spots showed significant degradation, increasing from small to large focal spot. The GMTF values were found to be 0.09, 0.05 and 0.01 at 5 cycles/mm for the small, medium and large focal spot, respectively. The corresponding values for GDQE were 0.1, 0.03 and 0.001. These results demonstrating the effect of choice of focal spot on the image quality are supported by line profiles obtained across the stent images. CONCLUSIONS: The results of this study demonstrate the significance of focal-spot size on system performance when using a high-resolution detector with a realistic magnification and shows the importance of choosing the minimum focal-spot size. (Support: NIH-Grant R01EB002873) NIH Grants R01-EB008425, R01-EB002873 and an equipment grant from Toshiba Medical Systems Corp.

SU-E-I-46: Quantitative Effects of Clinical Practice on the Efficiency of CT Tube Current Modulation for Radiation Dose Reduction.

PURPOSE: CT tube current modulation (TCM) represents one of the most important and efficient methods for radiation dose reduction and has been well accepted. However, the possible influence of clinical practice is likely to be ignored although venders may provide clinical protocols and instructions. This study is designed to further investigate the quantitative effects of clinical operation, including radiograph direction and patient positioning, on the efficiency of TCM by measuring patient radiation dose and image quality. METHODS: An anthromorphologic chest phantom was scanned in a Sensation 40 and a lightSpeed 16 CT scanner, respectively, using routine chest protocols with TCM. We first investigated the effects of radiograph direction. Anterior-posterior (AP), lateral, and posterior-anterior (PA) directions were chosen. CTDIvol and dose-length-product (DLP) were recorded for analyses. Our second experiment studied the influence of patient position. First the phantom was positioned at the iso-center then scanned with AP direction. CTDIvols and DLPs were recorded as reference. Then the phantom was moved out of iso-center, up or down 2 and 4 cm respectively then scanned. CTDIvols and DLPs were recorded for comparison. For each setting, image noise was measured. RESULTS: CTDIvol increases approximately 20% for PA direction, compared to AP or lateral direction which generates similar CTDIvol and DLP with TCM. Image noise for the PA direction is less than those for the AP or lateral directions. CTDIvol increases approximately 9% for LightSpeed 16 and 13% for Sensation 40 when the phantom was moved up 4 cm, while CTDIvol decreases approximately 5% for LightSpeed 16 and 8% for Sensation 40 when the phantom was moved down 4 cm. CONCLUSIONS: Our quantitative study can direct clinical practice to improve the efficiency of CT tube current modulation and reduce patient radiation dose.

SU-E-J-89: Deformable Registration Method Using B-TPS in Radiotherapy.

PURPOSE: A novel deformable registration method for four-dimensional computed tomography (4DCT) images is developed in radiation therapy. METHODS: The proposed method combines the thin plate spline (TPS) and B-spline together to achieve high accuracy and high efficiency. The method consists of two steps. First, TPS is used as a global registration method to deform large unfit regions in the moving image to match counterpart in the reference image. Then B-spline is used for local registration, the previous deformed moving image is further deformed to match the reference image more accurately. RESULTS: Two clinical CT image sets, including one pair of lung and one pair of liver, are simulated using the proposed algorithm, which results in a tremendous improvement in both run-time and registration quality, compared with the conventional methods solely using either TPS or B-spline. CONCLUSIONS: The proposed method can combine the efficiency of TPS and the accuracy of B-spline, performing good adaptively and robust in registration of clinical 4DCT image.

WE-E-217A-01: CAD: Quality Assurance of CAD Systems Implemented in Clinical Use.

PURPOSE: We propose a new metric called Distance Discordance (DD), which is defined as the distance between two anatomic points from two moving images, which are co-located on some reference image, when deformed onto another reference image. METHODS: To demonstrate the concept of DD, we created a reference software phantom which contains two objects. The first object (1) consists of a hollow box with a fixed size core and variable wall thickness. The second object (2) consists of a solid box of fixed size and arbitrary location. 7 different variations of the fixed phantom were created. Each phantom was deformed onto every other phantom using two B-Spline DIR algorithms available in Elastix and Plastimatch. Voxels were sampled from the reference phantom [1], which were also deformed from moving phantoms [2…6], and we find the differences in their corresponding location on phantom [7]. Each voxel results in a distribution of DD values, which we call distance discordance histogram (DDH). We also demonstrate this concept in 8 Head & Neck patients. RESULTS: The two image registration algorithms produced two different DD results for the same phantom image set. The mean values of the DDH were slightly lower for Elastix (0-1.28 cm) as compared to the values produced by Plastimatch (0-1.43 cm). The combined DDH for the H&N patients followed a lognormal distribution with a mean of 0.45 cm and std. deviation of 0.42 cm. CONCLUSIONS: The proposed distance discordance (DD) metric is an easily interpretable, quantitative tool that can be used to evaluate the effect of inter-patient variability on the goodness of the registration in different parts of the patient anatomy. Therefore, it can be utilized to exclude certain images based on their DDH characteristics. In addition, this metric does not rely on 'ground truth' or the presence of contoured structures.

Poster - Thur Eve - 12: Dosimetric manifestation of harmonic mode imaging for seed implant brachytherapy.

PURPOSE: To demonstrate the dosimetric effects of observer variability in defining the prostate and critical organs, using Tissue Harmonic (H) ultrasound imaging mode for permanent seed implant brachytherapy. METHODS: Images were acquired using a B -K medical 8848 probe with Brightness (B) and H mode for ten prostate brachytherapy patients. The prostate, rectum and urethra were contoured independently by five observers. The clinically used treatment plans based on B mode imaging fulfilling the dosimetric criteria were applied on these contours. Dosimetric parameters (prostate: D90, V100 and V200; rectum: V100; urethra: V140, V150 and V160) were computed using SPOT PRO™ planning system. Interobserver variability in dosimetric parameters was tested using standard deviations as percentages of means. RESULTS: Two-factor analysis of variances showed significant (p<0.05) interobserver variability in all dosimetric parameters for both modes. Interobserver agreement in dosimetric parameters improves in H mode due to improved interobserver consistency in contouring these organs on H mode images compared to B mode. There is no significant difference observed (paired student t test, p>0.05) in the mean values of dosimetric parameters in H and B mode for prostate and critical organs. CONCLUSIONS: H mode due to its better image quality helped to improve the interobserver agreement in contouring the prostate and critical organs and hence better interobserver consistency in all dosimetric parameter. Because the difference in the mean value of dosimetric parameters between two imaging modes is not statistically significant, H mode does not appear to offer any clinical advantages in terms of improving the dosimetric outcome.