Index of content:
Volume 34, Issue 6, June 2007
- Imaging Scientific Session: Room L100J
- Breast Imaging — Ultrasound and Tomosynthesis
TH‐E‐L100J‐01: Ultrasound Reflectivity Imaging with a Split‐Step Fourier Propagator for Cancer Detection and Diagnosis in Heterogeneous Breasts34(2007); http://dx.doi.org/10.1118/1.2761737View Description Hide Description
Purpose: To improve resolution and reduce speckle in ultrasound breast images by accounting for ultrasound scattering from breast heterogeneities during reflectivityimage reconstruction.Method and Materials: X‐ray mammography often fails to detect cancers in dense breasts, while breast ultrasound has the potential to detect them. Breast heterogeneities, particularly in dense breasts, generate significant ultrasound scattering. Properly handling ultrasound scattering is critical for reliable cancer detection and diagnosis in dense breasts. Ultrasoundwave propagation in the breast is governed by the acoustic‐wave equation in heterogeneous media, which can be decomposed into two one‐way wave equations describing wave propagation in opposite directions. A split‐step Fourier solution of a one‐way wave equation is used for backpropagation of reflected ultrasoundwaves. The backpropagation consists of two steps: one phase‐shift step in the frequency‐wavenumber domain, and another phaseshift step in the frequency‐space domain. During the backpropagation of ultrasound wavefields, heterogeneous breast sound‐speed models obtained from transmission ultrasound tomography are used to approximately account for ultrasoundwavescattering. The reflectivityimaging method based on the split‐step Fourier propagator is applied to computer‐generated ultrasound data and in‐vivo ultrasound breast data acquired using a ring transducer array. The ultrasoundimages are compared with those obtained using a uniform sound‐speed model.Results: Comparison of ultrasoundreflectivityimages obtained using heterogeneous breast sound‐speed models with those obtained with a uniform model shows that ultrasound scattering of breast heterogeneities needs to be taken into account to obtain high‐resolution and high‐quality breast images.Conclusion: Using heterogeneous sound‐speed models for ultrasoundwave backpropagation during reflectivityimage reconstruction significantly improves image resolution and reduces speckle. The resolution and quality of ultrasoundreflectivityimages are further enhanced with increasing accuracy and resolution of transmission ultrasound tomography.
34(2007); http://dx.doi.org/10.1118/1.2761738View Description Hide Description
Purpose: To develop a high precision geometry calibration method and an efficient image reconstruction algorithm for digital tomosynthesissystem.Method and Materials: A geometry calibration phantom was constructed with 40 markers arranged in two planes parallel to the breast holder. A 3×4 projection matrix which maps the coordinates (x,y,z) of a point in the object to the coordinates (u,v) of the correspondent projection point on the detector was constructed for each projection angle based on the projection images of the calibration phantom. All information for the system geometry, such as source‐to‐detector distance, source to ISO distance, central ray offset on the detector (u0, v0), and detector angle offsets, can be extracted from the projection matrices. The projection matrices, not explicit geometry parameters, were used in a modified Feldkamp algorithm to reconstruct the imaged object. A prototype tomosynthesissystem and a CIRS anthropomorphic breast phantom with multiple embedded structures were used to test the geometry calibration accuracy and the reconstruction algorithm. Results: 3‐D image of the breast phantom was reconstructed using the projection matrices. 4 fibers, 6 masses, and all 12 speck groups were visible in the focal plane.Conclusion: Geometry calibration based on the projection matrices is accurate and reconstruction using the projection matrices is efficient.
TH‐E‐L100J‐03: Digital Breast Tomosynthesis Mammography: Computerized Classification of Malignant and Benign Masses34(2007); http://dx.doi.org/10.1118/1.2761739View Description Hide Description
Purpose: To develop a computer‐aided diagnosis system to assist radiologists in classification of malignant and benign masses on digital breast tomosynthesis (DBT) mammograms.Method and Materials: We used a data set of 60 cases (30 malignant and 30 benign) acquired with a GE Gen1 prototype DBT system located at the Massachusetts General Hospital. The DBT system acquired 11 projections in 5‐degree increments over a total arc of 50 degrees. The DBTs were reconstructed at a slice spacing of 1 mm using a simultaneous algebraic reconstruction technique. In this preliminary study, the volume of interest (VOI) containing the mass and its central slice were manually identified. The mass on each slice was segmented by an active contour model initialized with adaptive k‐means clustering. A band of pixels around the segmented mass boundary underwent the rubber band straightening transform (RBST) and the RBST image was enhanced by Sobel filtering. Morphological, spiculation, and texture features were extracted from the segmented mass and the enhanced RBST image. Two approaches for analysis of the masses were compared, one used the features from the central slice alone and the other used the average of the corresponding features over five slices centered at the central slice. Using leave‐one‐case‐out resampling, a linear discriminant analysis (LDA) classifier with stepwise feature selection was trained and tested in each feature space to classify the masses. Results: The LDA classifier achieved an area under the test receiver operating characteristic (ROC) curve of 0.93±0.03 using the 5‐slice approach, compared to 0.82±0.05 using the single‐slice approach. The difference was statistically significant with p<0.001. Conclusion: The preliminary results demonstrate the feasibility of computerized classification of masses in DBT. Further studies are underway to compare various approaches for using DBT volumetric information to characterize masses. Conflict of Interest: MGH group received support from GE Medical Systems.
TH‐E‐L100J‐04: Calibration of Bent‐Ray Ultrasound Tomography and Its Application to Breast Cancer Detection and Diagnosis34(2007); http://dx.doi.org/10.1118/1.2761740View Description Hide Description
Purpose: Travel‐time tomography plays a central role in our currentringtransducer array (Computerized Ultrasound Risk Evaluation — CURE) study of patients with suspicious breast masses. To properly calibrate an algorithm used for bent‐ray tomographic inversion of clinical ultrasound data, the following issues need to be addressed: balancing the data contribution and an a priorimodel, determining the stopping criteria for the iteration of the inversion, and estimating the resolving power of the tomography algorithm, that is, determining the reliability of the inversion. Method and Materials: In this study, quantitative assessments of the above issues are made based on synthetic models. The simulated ultrasonic data are generated using a high‐order finite‐difference time‐domain acoustic‐wave equation to model wave propagation through different numerical breast phantoms. The breast models are obtained by digitizing the reconstructed images of patient data and phantom data scanned by the CURE device. Checkerboard and perturbation tests are also performed to provide a quantitative estimate of the resolving power of the bent‐ray tomography algorithm. Guided by the synthetic study, the tomography algorithm with optimum parameters is applied to in‐vivo patient data acquired using CURE for breast cancer detection and diagnosis. Results: Our series of synthetic simulations show that an optimum trade‐off parameter can be chosen effectively using L‐Curve analysis. In addition, the stopping criteria can be chosen based on convergence rates of the iteration. Checkerboard and perturbation tests set a quantitative upper limit of the resolving power of the bent‐ray tomography for ultrasound data acquired with a ring array. Conclusion:Calibration of the bent‐ray tomography algorithm provides a high degree of confidence in tomographic inversion results, and helps in obtaining optimal tomographic images from in‐vivoultrasound breast data.
TH‐E‐L100J‐05: A Three‐Dimensional Linear Model for Investigating the Image Quality of Breast Tomosynthesis34(2007); http://dx.doi.org/10.1118/1.2761741View Description Hide Description
Purpose: To build a three dimensional (3D) linear model for breast tomosynthesis in order to investigate the effects of different imagingsystem design parameters on the reconstructedtomosynthesisimage quality. Method and Materials: The 3D model incorporates the detector performance, imaging geometry and reconstruction filters for the filtered backprojection (FBP) method. To validate the model, experiments were performed on a prototype breast tomosynthesissystem equipped with an amorphous selenium (a‐Se) digital mammographydetector. The detector can be operated in full resolution with 85 micron pixel size or binning mode to reduce acquisition time. Twenty—five projection images were acquired with an angular range of ±20°. The images were reconstructed using a slice thickness of 1 mm with 0.085 mm × 0.085 mm in‐plane pixel dimension. The 3D noise power spectrum (NPS) was computed using reconstructed scatter free uniform images and an edge phantom was imaged to measure the in‐plane modulation transfer function(MTF). An ACR mammography accreditation phantom was imaged to demonstrate the effects of detector operation modes and image acquisition geometry, i.e., angular range, on reconstructed image quality. Results: The measured in‐plane MTF and 3D NPS are both in good agreement with the model. Our results showed that both NPS and in‐plane MTF depend on the reconstruction filters and angular range. Increasing angular range helps improve the MTF at low frequencies, resulting in better detection of the large‐area, low‐contrast mass lesions in the phantom. Pixel binning cuts the acquisition time in half at the cost of decreased high frequency response. This results in reduced visibility of small calcification specks. Conclusion: A 3D linear system model for breast tomosynthesis was developed and validated with experimental measurements. The model can be used to predict the effects of system design parameters on reconstructed image quality.
*Research sponsored by Siemens AG.
34(2007); http://dx.doi.org/10.1118/1.2761742View Description Hide Description
Purpose: To investigate the use of operator independent ultrasoundtomography for breast imaging with the goal of differentiating breast masses. Method and Materials: A series of in‐vitro and in‐vivo experiments were carried out using a recently developed prototype based on the principles of ultrasoundtomography. To characterize the performance of the prototype, an anthropomorphic breast tissue phantom with embedded inclusions was imaged to yield reflection and transmission (i.e. sound speed and attenuation) tomograms. The in‐vivo performance of the prototype was assessed by imaging 125 patients with suspicious mammograms. Each data set yielded ∼45 to ∼75 tomograms covering the whole breast volume. Masses were identified by biopsy and their locations inferred from conventional mammography and ultrasound exams. These data were compared with the ultrasound tomograms to evaluate the in vivo detection capabilities of the prototype. Results: Our techniques successfully demonstrated tomographic imaging of breast architecture in both reflection and transmission imaging. Furthermore, phantom studies indicated that masses as small as 6 mm in size were detected in all imaging tomograms. In addition, statistically significant differences were observed between the sound speed and attenuationmeasurements of the benign inclusions and cancerous masses. The in vivo data suggested that ∼90% of masses >15mm in size were routinely detected including cancers as small as 8 mm in size. Overall, reflection, sound speed, and attenuationimaging of breast masses were demonstrated both in vitro and in vivo. Conclusion: Operator‐independent whole‐breast imaging and the detection of breast masses are feasible using ultrasoundtomography techniques. Our approach has the potential to provide a cost‐effective, non‐invasive, and non‐ionizing means of evaluating breast masses, leading to more routine analysis and evaluation of treatment response.
TH‐E‐L100J‐07: Comparison of Mammographic Percent Density and Volumetric Percent Density Determined From Ultrasound Tomography Images34(2007); http://dx.doi.org/10.1118/1.2761743View Description Hide Description
Purpose: Previous ultrasoundtomography work conducted by our group showed a direct correlation between measuredsound speed and physical density in vitro, and increased in vivosound speed with increasing mammographic density, a known risk factor for breast cancer. Building on these empirical results, we investigated the use of volumetric ultrasound percent density (USPD) for breast density estimation. Method and Materials: A breast phantom comprised of fat inclusions embedded in fibroglandular tissue was scanned four times with both ultrasoundtomography and CT. The coronal transmission tomograms and corresponding CTimages were analyzed using a semi‐automatic segmentation routine. Next, a cohort of ∼100 patients was imaged encompassing the entire breast volume (50–75 tomograms/patient). USPD was determined by segmenting high sound speed areas from each slice using a k‐means clustering routine, integrating these results over the entire breast, and dividing by total breast area. USPD was evaluated using two mammographic density measures: (1) qualitative, as determined by a radiologist's visual assessment using BI‐RADS Criteria and (2) quantitative, via digitization and semi‐automatic segmentation of craniocaudal and medio‐lateral oblique mammograms.Results: The integrated areas of the phantom's fat inclusions were compared using both transmission ultrasound and CT for four repeated scans. The average variability for inclusion segmentation was ∼2% and ∼10%, respectively, and a close correlation was observed in the integrated areas between the two modalities. A strong positive association between BI‐RADS category and USPD was demonstrated. Furthermore, comparing USPD to calculated mammographic density yielded moderate to strong positive associations (Pearson r = 0.76–0.84) for MLO and CC views, respectively. Conclusion: These results support the hypothesis that utilizing USPD as an analogue to mammographic breast density is feasible. USPD has the potential to provide a non‐ionizing, whole‐breast analysis of breast density, which may better elucidate the relationship between breast density and breast cancer risk.
- Breast Imaging‐Mammography and CT
WE‐E‐L100J‐01: Simulation of Tomosynthesis Mammograms with Cone Beam CT Images of Mastectomy Breast Specimens34(2007); http://dx.doi.org/10.1118/1.2761576View Description Hide Description
Purpose: To describe and discuss about a technique to simulate digital mammograms and tomosynthesisimages with cone beam CTimages of mastectomy breast specimens. Method and Materials: To allow x‐ray images to be simulated, cone beam breast CTimages of a mastectomy breast specimen were acquired at 80 kVp and used to generate a 3‐D map of linear attenuation coefficients for mammographic x‐rays. The map was then deformed to generate a new map for a compressed breast. A software re‐projector was developed and used to compute projection images for a stationary detector and a linearly shifted x‐ray source. A special re‐projector was developed to preserve the contrast of microcalcifications (MCs). The resulting images were then used to reconstructtomosynthesismammograms using various reconstruction algorithms. Results: We have successively used cone beam CTimages of mastectomy breast specimens to generate 3‐D maps of attenuation coefficients to model compressed breasts and to use them to simulate digital mammograms and tomosynthesisimages. It was found that MCs clearly visible in cone beam CTimages were not visible in regular mammograms but faintly visible in tomosynthesisimages. The MCs were clearly visible in both MC enhanced mammograms and the tomosynthesisimagesreconstructed from them. Conclusions: Cone beam CTimages may be used to model a compressed breast for simulating mammograms and tomosynthesisimages. Cone beam CTimages are superior to regular or tomosynthesismammograms in depicting MCs within their resolution limit. MC enhanced re‐projection, however, could be useful in displaying the cone beam CTimages with the more conventional appearances of regular and tomosynthesismammograms while maintaining the contrast and visibility of MCs in the CTimages.
This work was supported in part by a research grant EB00117 from the NIH‐NIBIB and a research grant CA104759 from the NIH‐NCI.
34(2007); http://dx.doi.org/10.1118/1.2761577View Description Hide Description
Purpose: To optimize and study the effectiveness of CEDEM using simple breast‐equivalent phantoms in the experimental setting. Method and Materials: The experimental evaluation of CEDEM was performed using a dedicated cone‐bream breast CT scanner in our laboratory, with the system operating in stationary imaging mode. A series of physical phantoms approximating the anatomical complexity of the breast were fabricated. Polyethylene and water were used for the adipose and glandular aspects of the breast phantom, respectively, due to their similar x‐ray attenuation characteristics. Holes of various sizes were randomly drilled on the polyethylene slabs to create an “anatomical” noise pattern. Rectangular cells (cuvettes) filled with various iodine concentrations were placed in a polyethylene slab to maintain a homogeneous projected iodine thickness (mg/cm2). When imaged experimentally, the polyethylene slab with five cuvettes of various iodine concentrations was overlaid with a number of the slabs with pseudo‐anatomical noise distributions while immersed in a plastic container filled with water. Data were acquired at 55 kVp and 100 kVp (suggested by computer optimization) with additional elemental filters (0.1 – 0.3 mm copper filtration), and dual‐energy subtracted images were produced. Results: Preliminary results were acquired at 55 kVp and 100 kVp (0.3 mm additional copper filtration). The sensitivity to iodine contrast agent in the dual‐energy subtracted image (signal‐difference‐to‐noise ratio (SDNR) per mg/cm2) was evaluated and found to be linear with concentration. The influence of the breast phantom thickness, composition, and dose on contrast enhancement at different kVp combinations will be presented. Conclusion: The effectiveness of CEDEM with an anatomical‐complex phantom was evaluated. The results indicate that the dual‐energy subtracted image can enhance the iodine contrast agent with background anatomical structures. This approach may be useful for the evaluation of the kinetic curve in contrast enhanced breast imaging procedures.
34(2007); http://dx.doi.org/10.1118/1.2761578View Description Hide Description
Purpose: To compare and validate Monte Carlo simulation using the GATE software with experimental measurements for scattered radiation in X‐ray cone beam CT breast imaging . In addition, to determine the scatter distribution for single‐order, and multiple‐order incoherent and coherent scatter using validated Monte Carlo simulation.Methods: This study used a bench‐top cone‐beam CT breast imagingsystem using a flat‐panel detector. A cylindrical phantom with equivalent composition of 50% fibro‐glandular and 50% adipose tissues was used. Scatter distributions were measured by beam stop and aperture methods. A lead strip was positioned between the X‐ray source and the breast phantom in the beam stop method. Its inverse version was used in the aperture method. We computed the scatter contribution in the breast phantom by subtracting background scatter. The background scatter fraction was measured without the phantom. The Geant4‐based simulation package GATE was used to model X‐ray photon interactions in the phantom and detector. Two implemented electromagnetic interactions packages, standard and low energy models, were compared for computing efficiency and physics ingredients. A structured breast phantom was used in GATE simulations to determine the characteristics of various scatter components which cannot be separated in measurements. Results: Measurements by the two methods are consistent within 5% after background subtraction, and agree with the low energy model simulations. A hybrid model in GATE, with photoelectric process from the standard model and Compton and Rayleigh scattering processes from the low energy model, is computationally efficient while maintaining physics accuracy. The GATE simulations in the Bakic phantom show that the multiple‐order scatter distribution, as well as single‐order Compton scatter, has predominantly low‐frequency characteristics. Single‐order Rayleigh scatter was observed to be the primary contribution to the spatially variant scatter component.
34(2007); http://dx.doi.org/10.1118/1.2761579View Description Hide Description
Purpose: To develop a new screening tool for the detection of breast cancer by the wide‐field collection of x‐ray coherent scatter from the human breast, to potentially improve the diagnosis of breast cancer.Method and Materials: Coherent scatter analysis is normally performed using a highly collimated narrow beam. This complicates adding coherent scatter analysis to screening mammography. A technique has been developed which allows broad field discrimination of coherent scatter using standard high ratio Bucky grids. Using the Kα line from a standard Mo tube, the coherent scattering angle (2θ) for normal breast tissue and cancertissue would be 9 and 13 degrees, respectively. For the first step of work, pork and beef fat (2θ = 9 degrees) were used as a phantom for normal human breast tissue, while graphite (2θ = 12.16 degrees) was used as a phantom for breast cancertissue. The coherent scatter images were recorded with CR plates. Tests were carried out with a 10:1 grid for a series of tilt angles and sample‐to‐grid distances. Results: The intensity distribution showed that the coherent scatter from the graphitecancer phantom where observable above the fat background at certain grid tilt angles and sample‐to‐grid distances. Theoretical calculations for the intensity distributions were compared to measurements. Conclusion: This work indicates promising potential for including coherent scatter analysis in screening mammography. The patient is not exposed to any additional x‐ray radiation. Coherent scatter is normally present but not collected. This technology provides an entirely new basis for diagnosis, which may improve the sensitivity and specificity of mammography.
34(2007); http://dx.doi.org/10.1118/1.2761580View Description Hide Description
Purpose: To compute the radiation dose to different tissues of the body from a standard mammogram.Method and Materials: Using the Geant4 Monte Carlo toolkit, a simulation was developed in which the human body was represented using a mathematical anthropomorphic phantom. A total of 66 different organs and tissues were simulated, including the uterus, which was used to compute the dose to the fetus during early‐stage pregnancy. The imaged breast was represented as compressed, and in both the cranio‐caudal and the medio‐lateral oblique views. At each energy level, the simulation emitted 60 million monochromatic x‐rays and recorded the location of each interaction and the amount of energy deposited. The dose to the red bone marrow and to the bone surfaces of each bone was estimated separately. The monochromatic data was used to compute the resulting dose for different typical polyenergetic mammographic spectra, and the values were normalized to the resulting glandular dose to the breast, resulting in the dose ratio. The effectiveness of using a lead shield between the patient and the x‐ray field was also investigated. Results: Only 16 organs and skeletal tissues received a dose higher than 0.10% of the glandular dose to the breast. Of the organs, the contralateral breast received the highest dose ratio with 0.62%. Of the skeletal tissues, the sternum received the highest dose, with a dose ratio of 2.36% to the bone surface and 0.56% to the bone marrow. The dose to the uterus and fetus received a maximum dose ratio of <10−5. The lead shield did not lower the dose values substantially. Conclusion: The radiation dose to all tissues outside the x‐ray field, including the fetus, during standard mammography is extremely low.
34(2007); http://dx.doi.org/10.1118/1.2761581View Description Hide Description
Purpose: Technique optimization for contrast‐enhanced dual energy mammography in a breast CT scanner. Method and Materials: Using kVp accuracy and HVL measurements physically performed on the x‐ray source of a dedicated cone beam breast CT system, a mathematical spectral model was calibrated to have the same spectral and output properties as the physical system. Mathematical breast phantoms were fabricated by generating random sized and randomly placed glandular spheres on an adipose background to approximate the anatomical noise of the breast. The simulated imaging geometry matched that of the breast CT system. Quantum noise was simulated using Poisson model and a validated random number generator. Contrast‐enhanced dual energy mammography was optimized for parameters including kVp, x‐ray spectral filtration, and mAs distribution between high and low energy images. Mean glandular dose to the breast was calculated using dose coefficients generated from previously published Monte Carlo simulations. A figure of merit of the contrast‐to‐noise ratio at constant mean glandular dose was chosen for the optimization task to maximize image quality at minimal radiation dose levels. Optimization was carried out over a range of breast diameters, breast compositions, and iodine concentrations. Results: Spectra study suggested the optimal kVp to be between 120kV and 140kV for the high energy image, and between 50kV and 60kV for the low energy image, for the studied ranges of breast diameters and compositions. Added copper filtration up to 1mm for the high kVp spectra continued to improve CNR at set dose levels. Iodine as diluted as 0.05% of the 350mg/mL standard solution could be detected with a 2.5mGy total dose. Optimized mAs distribution suggested a higher dose proportion for the low kVp image than for the high kVp image.Conclusion: An optimized contrast‐enhanced dual energy mammography can efficiently suppress anatomical noise and significantly improve contrast detection.
34(2007); http://dx.doi.org/10.1118/1.2761582View Description Hide Description
Purpose: To evaluate breast CT with energy resolved detection. Methods and Materials: A new breast CT system with energy resolved x‐ray detector and Scanning Multi‐Slit (SMS) data acquisition was evaluated. The SMS CT provides efficient scatter rejection, and uses linear array of photon counting detectors. A breast CT phantom with 14 cm diameter and 50/50 adipose/glandular composition was used. The phantom included carcinoma, adipose, blood, iodine, and CaCO3 as contrast elements. The x‐ray technique was 90 kVp with 660 mR entrance exposure. Photon counting, charge (energy) integrating, and photon energy weighting CTimages were generated. The CNR improvement with photon energy weighting was calculated. The CTimages with calcium/tissue and iodine/tissue decomposition were generated with dual energy subtraction. The x‐ray spectrum was electronically split into low and high energy parts by photon counting detector to perform dual energy subtraction. Results:CNR of the energy weighting CTimages of carcinoma, blood, iodine, and CaCO3 was higher by 19.3%, 19.9%, 29.7% and 40.4%, respectively, as compared to conventional CT.Photon counting dual energy subtraction provided 61% higher CNR as compared to conventional dual kVp dual energy subtraction. CNR of the photon counting dual energy subtracted CTimages of CaCO3 was increased by 69% when photon energy weighting was applied to CT projections prior subtraction. CNR improvement of SMS CT due to scatter rejection is 32% as compared to cone beam breast CT. Combination of photon energy weighting with scatter rejecting results in 50%‐85% CNR improvement in SMS CT as compared to conventional CT.Conclusion:Photon countingCT has a potential for improved sensitivity due to higher CNR with scatter rejection, quantum limited detection, and photon energy weighting. It has a potential for improved specificity due to single kVp dual energy subtraction that enables calcium/tissue and iodine/tissue decomposition.
- Computed Tomography — New Developments and Applications
34(2007); http://dx.doi.org/10.1118/1.2761365View Description Hide Description
Purpose: Partial voluming in PETimaging leads to underestimation in activity concentration. The aim of this abstract is to correct for partial volume artifacts in PET/CT scans. Method and Materials: A Jaszczak phantom with hollow spheres of varying sizes (4.95 – 31.27 mm inner diameter (ID)) was filled with F‐18 water using 3 different sphere‐to‐background ratios (SBR), ranging from 3:1 to 10:1. For each SBR, several acquisitions were conducted. All PET data was reconstructed using OSEM (2 iterations, 21 subsets). Regions were drawn on the CTimages to obtain accurate sphere volume and location. A software tool was written to correct for partial voluming by incorporating the sphere size and the non‐stationary response function of the scanner. The original maximum (OM), original average (OA), corrected maximum (CM), and corrected average (CA) activity concentrations (AC) were measured and compared. Results: For all SBRs, spheres larger than 19 mm the measured OM and OA AC were 111 and 77% of the true value, respectively. Following correction, these values changed to 128 (CM) and 102% (CA), respectively. For the smallest sphere size (4.95 mm), the measured OM and OA AC were both 20% of the true value. Following correction, these values changed to 121 (CM) and 104% (CA) of the true value. The CM, however, did vary between 83 and 176% of the true AC. An analytical relationship between the lesion size (obtained from CT) and the amount of correction needed to recover the true AC based on the multiple acquisitions was generated. Conclusions: To determine the true AC of a lesion from a PET/CT scan, the corrected average is more accurate than the original maximum, and should be used for clinical assessment.
34(2007); http://dx.doi.org/10.1118/1.2761366View Description Hide Description
Purpose: The properties of noise in CTimages are important in system design, algorithm development, and assessment of diagnostic observer performance, as well as for developing accurate dose reduction simulators. Conventional analysis assumes noise stationarity, even though it is well known that elements in a CT scanner, such as bowtie filters or tube‐current modulation, as well as filtered backprojection reconstruction, introduce variation in measurement statistics. In this presentation, two additional contributors to nonstationarity, hardware component variation and data interpolation, are analyzed, and the properties of noise in CTimages are characterized. Method and Materials: Raw sinogram data was collected from open gantry scans. Offline reconstruction software, with access to data at individual processing steps, such as linear interpolation, was used for image reconstruction. White noise data was injected at various stages of reconstruction and changes in statistical properties at subsequent processing steps measured. Multiple simulations were used to create data ensembles for analysis. Covariance measures were calculated in the sinogram and image domains. Results: Raw CT data is nonstationary during the acquisition cycle, with covariances as high as 20% in unattenuated signals. Image reconstruction steps, such as linear interpolation and filtered backprojection, introduce variations in noise power on the order of +/− 50% in sinogram data. The variation of noise power can be up to 600% across a uniform image. The off‐diagonal covariance coefficients are on the order of 10% with nearest neighbor pixels. Conclusion: The conventional assumption of noise stationarity in CTimages needs to be extended to include additional processes that introduce local variations in image statistics.
34(2007); http://dx.doi.org/10.1118/1.2761367View Description Hide Description
Purpose: Many external surrogates exist for 4‐D CT, but all require some device attached to the patient. We evaluated a markerless optical surfaceimagingsystem for use with 4‐D CT.Method and Materials: A prototype of the GateCT optical surfaceimagingsystem (VisionRT, London, UK) was installed with a Philips Brilliance 64 CT simulator (Philips Medical, Cleveland, OH). The system is similar to the AlignRT product, but only contains one camera pod which is focused on the bore of the CTscanner. It also monitors the x‐ray on signal and sends regular pulses to the scanner similar to other external respiratory surrogates. First, a static plate was imaged to detect any drift in the measured motion. Next, the 4D phantom was used to create sinusoidal motion with simultaneous longitudinal motion to model breathing whilst the patient was scanned, and the motion measured by GateCT was compared with the phantom motion. Next, the 4‐D phantom moved a 2 cm marble in a sinusoidal pattern, and the scanner was operated in retrospective helical mode. The respiratory waveform was exported from the GateCT program to the scanner, and the inspiration and exhalation image sets were reconstructed and compared. Results: When the phantom was still, the GateCT point track gave a stable reading with a noise level of 0.5 mm. When the phantom was moving in one or two dimensions, the GateCT point track accurately recorded the amplitude, phase and frequency of the motion. For the 4‐D CT scans, the reconstructedCTimages indicated that the amplitude, phase, and frequency of the marble motion were recorded accurately by GateCT. Conclusion: GateCT hold potential as a markerless respiratory surrogate. More work is necessary to determine the optimal patient selection for this technique.
34(2007); http://dx.doi.org/10.1118/1.2761368View Description Hide Description
Purpose: In helical tomotherapy, Megavoltage CT (MVCT) images are acquired at a rate of one transverse slice per five seconds. The slow acquisition of raw data results in CTimages with varying degrees of motion‐induced artifacts. In addition, the high energy of the x‐ray beam (4–6 MV) presents unique challenges for image reconstruction. The purpose of this work was to correct for motion in the superior‐inferior (S/I) direction using retrospective manipulation of the unprocessed sinogram data. Method and Materials: A respiratory motion platform with a S/I motion range of 3‐cm, and periods of 2, 4 and 6 seconds was used to move an ACR CT accreditation phantom and a anthropomorphic thorax phantom in the superior‐inferior direction. The motion of the respiratory platform was measured using the Varian respiratory gating system. Sinogram data was acquired axially with the couch stationary for 5 complete rotations of the gantry. The respiration cycle was synchronized to the projections in the raw sinogram so that the position of the moving object was known in each projection. Motion correction was performed by identifying and extracting projections occurring within a specified respiration phase, correcting for missing projections, and then reconstructing the CTimages.Results:Images acquired with the ACR CT phantom showed considerable improvement with motion correction using temporal re‐binning over the uncorrected images.Images acquired with the RANDO phantom suffered from greater image artifacts than the ACR phantom because of the larger variation in the phantom's anatomy over the 3‐cm range of S/I motion. Improvement was seen for each period, however only slight improvement was seen for the 2‐second period of motion. Conclusion: Temporal re‐binning can successfully reduce S/I motion artifacts in reconstructed MVCT images.Conflict of Interest: This research is supported by TomoTherapy, Inc.
TU‐D‐L100J‐05: Assessment of Mesothelioma Tumor Response: Correlation of Tumor Thickness and Tumor Area34(2007); http://dx.doi.org/10.1118/1.2761369View Description Hide Description
Purpose: The quantification of pleural mesothelioma tumor extent is required to evaluate the efficacy of clinical trials. The manual acquisition of up to three linear tumor thickness measurements on each of three sections across a series of computed tomography(CT) scans is the current standard for tumor response assessment. The purpose of this study was to determine the correlation of response based on linear tumor thickness measurements and response based on tumor area. Method/MATERIALS: Two CT scans from each of 22 mesothelioma patients were collected. Using a computer interface, a radiologist acquired linear tumor thickness measurements on three sections of each patient's baseline scan and on the corresponding sections of each patient's follow‐up scan in accordance with our clinical protocol. These linear measurements across 132 CT sections (3 sections per scan, 2 scans per patient, 22 patients) provided the standard for comparison of area measurements. Another radiologist used a computer interface to delineate the tumor border in the same 132 CT sections to obtain tumor area and the changes in tumor area between the baseline and follow‐up scans of each patient. Results: A comparison of the sum of tumor thickness measurements and tumor area yielded a correlation coefficient of 0.59 across the 132 sections. With regard to tumor response, a comparison of change in the sum of tumor thickness measurements and change in the total tumor area between the baseline and follow‐up scans of the 22 patients yielded a correlation coefficient of 0.83. This relatively high correlation, however, does not capture the extent of variability in the data. For example, among patients with RECIST‐based “stable disease,” change in tumor area ranged from a decrease of 58% to an increase of 89%. Conclusions: Although measurements of tumor thickness and tumor area demonstrated moderate correlation, variability in this association requires further investigation.
34(2007); http://dx.doi.org/10.1118/1.2761370View Description Hide Description
Purpose: To evaluate existing and a new detector technologies for photon counting/energy resolving x‐ray and CTimaging.Materials and Methods:Digital mammography with photon counting Si and gaseous Xe detectors is now available. Although CZT and CdTe detectors are more attractive for photon counting x‐ray and CTimaging, they suffer from hole trapping and charge sharing between pixels, which decrease energy resolution, count rate, and spatial resolution. We have proposed using CZT/CdTe crystals in a tilted angle configuration when the x‐ray hits the crystal surface at a small angle. This allows decreasing crystal thickness significantly while maintaining high photon absorption. The tilted angle CZT/CdTe was simulated for 0.3–1 mm crystal thickness, 10°–90° tilting angles, and 50–150 keV photon energies. The results were compared to a crystal with 3 mm thickness used in normal irradiation. Experiments were performed with a tilted angle CZT with 2 mm crystal thickness using 59 keV and 122 keV photons and 110 kVp x‐ray. Results: The count rate of the tilted angle CZT detector was higher by 10–20 times for 0.3–1 mm crystal thickness and 50–150 keV photon energies compared to CZT with normal irradiation. The electron collection time was shorter by 3–6 times. The charge diffusion and charge sharing between pixels was decreased correspondingly. The experiments have shown significant decrease in tailing of the energy spectrum with tilted angle detector. The peak/tail ratio of the measured energy spectrum was increased by 2.4 and 2 times for 59 keV and 122 keV photons, respectively, when 20° tilting angle was used. Conclusion: The proposed tilted angle CZT/CdTe detector allows significant increase in count rate, energy resolution, and spatial resolution, as compared to currently used CZT/CdTe detectors. This potentially enables using tilted angle CZT/CdTe detectors for photon counting/energy resolving x‐ray and CTimaging.