Volume 34, Issue 6, June 2007
Index of content:
- Imaging Scientific Session: Room L100J
- Imaging Registration, Fusion, Segmentation, and Visualization
34(2007); http://dx.doi.org/10.1118/1.2761210View Description Hide Description
Purpose: To make anatomy contouring more efficient by integrating contours drawn in any combination of transverse/sagittal/coronal (T/S/C) views and reconstructing the T contours from the surface defined by the drawn contours so that the user may draw/edit until the object is completed. Method and Materials: Our program performs the surface reconstruction by thin‐plate spline interpolation of points in the drawn contours. The surface is reconstructed as a variational implicit function, and runs in real‐time because we abstract the most shape‐informative points from the input contours. These points correspond to maxima in one of several functions of the second derivatives of the contour functions. We demonstrate contouring results for one of these methods—the scalar second derivative of the curve function. Before abstracting the shape salient points, contours are re‐computed to have consistent rotational directions, and to have origins placed consistently in the patient coordinate system. Since the contours' points are non‐uniform samples of the curve, each contour is reconstructed, using all input points, by B‐spline interpolation to resample the curve uniformly along the its length to compute the derivatives by finite differences. Results: We demonstrate in the supporting document: 1) the operation of the shape abstraction algorithm, and 2) the production of contours approximating the prostate from a single S and C contours, and 3) the production of contours in 270 T sections of a spinal canal from a single S and 4 C contours. Conclusion: We have developed a contouring program that integrates drawn contours in any combination of T/S/C views to create a coherent surface that is clipped to produce T contours at the appropriate image planes, so the user can and assess how much more drawing or editing is needed to make a satisfactory match to object boundaries visible in the image.
Sponsored by CMS, Inc.
MO‐D‐L100J‐02: Automated Propagation of Region‐Of‐Interest Contours Between 4DCT Images Using a Regional Deformable Model34(2007); http://dx.doi.org/10.1118/1.2761212View Description Hide Description
Purpose: To develop a regional algorithm for contour propagation between difference phases of 4D CT and show its impact on 4D radiation therapy simulation. Methods and materials: The ROI contours were manually delineated on a selected phase of 4D CT on one of the breathing phases by a physician. A narrow band encompassing the ROI boundary was created on the image and used as a compact representation of the ROI boundary. A free form B‐Spline deformable registration was carried out to map the band to other phases of 4D CT. A Mattes mutual information function was used as the metric function and the limited memory Broyden‐Fletcher‐Goldfarb‐Shanno algorithm (L‐BFGS) was employed to optimize the multidimensional metric function. Upon successful registration, the deformation field was extracted and utilized to transform the manual contours to other phases. A bi‐directional contour mapping method was introduced to evaluate the success of the proposed technique. Three thoracic patients with 4D CT scans were used to test the proposed algorithm. Results: A formalism of automated contour mapping has been developed based on a regional narrow band deformable registration. Application of the algorithm to three thoracic patients indicates that clinically satisfactory results are achievable with a spatial accuracy better than 2mm for ROI mappings between adjacent phase, and 3mm between opposite phase mapping. As compared to the conventional whole image‐based registration, the computation here was found to be two orders of magnitude more efficient, in addition to the much reduced requirements in computer memory. Conclusions: A regional deformable registration is an efficient and accurate way for contour mapping and should find widespread application in 4D simulation and treatment planning in the future to maximally utilize the available spatial‐tempo information.
MO‐D‐L100J‐03: Generating Patient Specific Motion Models Using a Navigator Channel and a Liver Population FEM Motion Model34(2007); http://dx.doi.org/10.1118/1.2761213View Description Hide Description
Purpose: To incorporate motion information from 2D images using a navigator technique and finite element modelling (FEM) to generate patient‐specific 3D motion models.Methods and Materials: A liver population motion model was created using a FEM deformable image registration platform, MORFEUS, to simulate liver deformation during a respiratory cycle. Twenty patient's exhale livers were constructed from the population FEM, and then deformed into their inhale livers. Deformation maps were generated and used to check the accuracy of the navigator‐updated patient models. NavigatorView, an in‐house algorithm, places a rectangular navigator channel to define a region of interest on 2D image slices. An operator chosen navigator is placed on the exhale image while another navigator is automatically placed at the corresponding location on the inhale image. Motion was calculated as the shift required to align the intensity profiles within the channels at the superior dome and the inferior tip of the liver on coronal CT slices and simulated radiographimages. The navigator shifts were used in a weighting equation to generate patient‐specific motion models from the population motion model.Results: The average accuracy ± standard deviation of the navigator channel at the superior and inferior edges is 0.12 ± 0.12cm and 0.25 ± 0.25cm respectively. The navigator‐updated patient‐specific models was 100% and 80% successful of the 20 patient cases for the coronal CT and simulated radiograph slices respectively, where a successful case achieved an accuracy error less than the image voxel size (0.25cm). Conclusions: The navigator technique allows for updating patient‐specific 3D motion models from a liver population model using more easily acquired 2D images. This can be a useful tool for image‐guidedtherapeutics, such as intra‐fraction image guided tracking during radiotherapy, where 2D data may be more rapidly acquired.
Research sponsored by the National Cancer Institute of Canada — Terry Fox Foundation.
MO‐D‐L100J‐04: Non‐Rigid Registration Based Respiratory Motion Models of the Lung Using Two Parameters34(2007); http://dx.doi.org/10.1118/1.2761214View Description Hide Description
Purpose: Single parameter motion models based on the phase of a respiratory signal can model the motion over an average respiratory cycle. Two parameter models based on amplitude and gradient may also be able to model some of the inter‐cycle variation. We present a method of constructing two parameter motion models and evaluate different functions for the model. Method and Materials: A reference CT volume is non‐rigidly registered to free breathing CT data. A function is then fitted to each of the control point displacements that define the registrations, relating them to the respiratory parameter(s). Three different functions were evaluated on data from an example patient: a 1D cyclic b‐spline function relating the displacements to phase, a 2D linear function, and a 2D 3rd order polynomial function, both relating the displacements to amplitude and gradient. Models built from each of the functions were used to produce transformations at the same parameter values as the registration results. Models were built both leaving out the target registration and including it. Sample points covering the entire region of interest were deformed using the model results and the registration results, and the differences in the displacements of the points were calculated. Results: The mean differences, when using all registrations and when leaving out the target respectively, were 0.53 mm and 0.67 mm for the 1D cyclic b‐spline function, 0.59 mm and 0.71 mm for the 2D linear function, and 0.43 mm and 0.96 mm for the 2D polynomial function. Conclusion: These results suggest that the performance of the 2D linear function is comparable to the 1D b‐spline function. The 2D polynomial function models the data more accurately but would appear to ‘over‐fit’ the data.
34(2007); http://dx.doi.org/10.1118/1.2761215View Description Hide Description
Purpose: Contouring plays an important role in radiation therapy. The traditional slice‐by‐slice manual contouring method is very time‐consuming. There have been several automatic and semi‐automatic segmentation methods proposed to make contouring easier. However, those methods are limited by the image quality and image features. In particular, all those methods fail in the prostate region where both prostate and rectum are difficult to segment. Methods:The contouring method we propose here only requires a few manual contours drawn in TCS (transversal, coronal, and sagittal) views. Our algorithm will then generate a 3D surface based on those manual contours. This method consists of two steps. The first step is a surface‐fitting process (the same principle as curve‐fitting). It determines a surface that approximates the manually drawn contour/control points. Different model is used for different organ (prostate, rectum) to create fit surface. The second step is a surface deformation process. A smooth deformation field is created automatically to deform the surface so that it passes through the manual contour points. This makes sure that the resulting surface conforms to the given contours. Results: This algorithm is very fast and user interactions can be easily incorporated. The algorithm was tested on 39 prostate and rectum cases and the results match manual contour results very well. Five manual contours are used. The false‐negative rate is around 0.07–0.08 and the speed‐up is around 3–5. Conclusions: We have presented a semi‐automatic method that only require human operator to draw a few manual contours to delineate a 3D structure. This method will produce a smooth contour and it is very robust.
MO‐D‐L100J‐06: Evaluation of An Automatic Algorithm Based On Kernel Principal Component Analysis for Segmentation of the Bladder and Prostate in CT Scans of Prostate Radiotherapy Patients34(2007); http://dx.doi.org/10.1118/1.2761216View Description Hide Description
Purpose: to evaluate the performance of an automatic segmentation algorithm used to segment the prostate and bladder in CT scans of patients undergoing external beam radiotherapy.Method and Materials: A nonlinear kernel principal component analysis (KPCA) based algorithm has been developed to model the surface shapes of soft tissueorgans in the pelvis of prostate patients. A library of 25 patients, each with approximately 13 CT scans acquired during their treatment, has been manually segmented by a physician. The library is used to both train and evaluate the algorithm. In this preliminary analysis, the training sets consisted of the contours from 8 to 10 scans of a single patient. Evaluation was performed using 3 CT studies previously unseen by the system from each patient, from the beginning, middle and end of treatment. Performance was measured by comparing the volumes and center‐of‐gravity positions of the generated surfaces with those of the physician‐drawn surfaces. The prostate shapes were also compared by generating maps of the surface separation after center‐of‐mass alignment. Three models were evaluated: a prostate‐only model, a bladder‐only model, and a joint prostate‐bladder model. Results: All models demonstrate the ability to successfully segment the soft tissue structures in the majority of cases. The average ratios of the prostate and bladder overlap volumes to the references volumes for the first seven patients were found to be 0.829 and 0.868 for the prostate only and bladder only models. The joint model results are similar, the corresponding ratios being 0.853 and 0.844. Surface separations for the two models were similar, the models generally being within 1 mm of the physician‐drawn shapes at mid‐organ, and along the right and left sides. Conclusion: Preliminary results from the proposed automatic segmentation method indicate it is sufficiently accurate for radiotherapy applications.
MO‐D‐L100J‐07: Improved Deformable Image Registration Using Hybrid Models: An Application to Solitary Pulmonary Nodule34(2007); http://dx.doi.org/10.1118/1.2761217View Description Hide Description
Purpose: To present a fast and robust hybrid deformable image registration algorithm (Juggler) for target motion estimation, automated segmentation and internal target volume (ITV) generation in 4DCT lungimaging.Method and Materials: The Juggler algorithm utilizes two alternating separately optimized diffusion models: one for low gradient features (soft tissues), the other for high gradient features (bony anatomy). Clinical lungimaging, acquired from 4DCT, consisted of free breathing CT and 10 phased CT sets. Comparison with three other conventional algorithms (demons, accelerated demons, and free‐form deformation method) was carried out using both simulated data and clinical 4DCT. Algorithm validation was carried out using comparison of 1) displacement vectors (for simulated data), 2) correlation coefficient, 3) difference imaging, and 4) vector streamlines. Clinical efficacy of Juggler was evaluated by visual inspection of anatomical structures mapped from the reference onto target imaging using the deformation map. Tumor motion/ deformation was determined using image moments. Results: For simulated data, Juggler achieved the highest correlation coefficient and minimum error. Its unique alternating mechanism greatly improved the convergence rate, resulting in at least 40% reduction in iterations. For 256×256×80 imaging, Juggler required <3 min for 50 iterations using Pentium‐4 3.2 GHz PC. For clinical 4DCT, Juggler performed best based on correlation coefficient and visual inspection. We used Juggler for generating ITV, and autosegmenting normal structures onto the phased CTs based on the initial segmentation on the reference CT.Conclusion: Juggler demonstrated the benefits of utilizing multiple diffusion models into a single unified model. Initial experiments indicated Juggler achieved the fastest rate of convergence per iteration, and overall superior registration based on correlation coefficient, difference imaging and error analysis. Its success in autosegmentation and ITV generation make this highly promising in implementation of adaptive radiation therapy. With new computer hardware, <30sec computation times are expected.
MO‐D‐L100J‐08: Construction of 4D‐CT Motion Model Using Deformable Registration: Comparison of Eulerian and Lagrangian Approaches34(2007); http://dx.doi.org/10.1118/1.2761218View Description Hide Description
Purpose: To analyze and compare two motion models constructed from 4D‐CT using deformable registration using two different computation approaches: Eulerian and Lagrangian.Method and Material: Accurate motion modeling within the lung is an important consideration in different clinical applications. We consider 4D‐CT scans for three patients treated in radiotherapy for lungcancer. 4D‐CTs were acquired using a 4‐slice fan‐beam CT scanner (GE Lightspeed QX/i; GE Heathcare Technologies, Waukesha, WI), and a respiratory surrogate (Real‐time Position Management; Varian Medical Systems, Palo Alto, CA). Two motion models were constructed from vector fields computed with demons algorithm with Gaussian regularization. We used also a image pre‐treatment method, a priori lung density modification, to handle this limitation of the demons algorithm. The first model, obtained with the Eulerian approach, uses small deformation estimations between successive phases of the 4D‐CT. The second model, obtained with the Lagrangian approach, was generated by estimation of larger deformations between the end‐exhale phase and all other states. The models were validated and compared using consistency (symmetry and transitivity) and accuracy (based on landmark points) metrics. Results: Mean values of accuracy were on the order of the image resolution and comparable to inter‐observer variability (1.9 mm), with slightly better results for the Lagrangian approach: 2.3 mm vs. 2.6 mm. The differences were not statistically significant for consistency.Conclusions: The results of this study suggest that the Lagrangian approach is more appropriate to use for generating a 4D‐CT motion model with deformable registration. In ongoing works, lung and GTV contours are used in order to conclude on the superiority of one motion model over the other for an automatic contour propagation tool, and for lung physiological information computation and analysis.
MO‐D‐L100J‐09: A Multi‐Resolution, Multi‐Scale, Mutual Information Technique for Registration of High‐ and Low‐KVp Projections in Dual‐Energy Imaging34(2007); http://dx.doi.org/10.1118/1.2761219View Description Hide Description
Purpose: In dual‐energy (DE) imaging, double‐shot acquisition provides superior DQE and detectability index compared to sandwiched detectors, but introduces the potential for misregistration artifacts (e.g., respiratory, cardiac, and bulk motion). This paper reports a projection registration scheme operating at various levels of scale and resolution to resolve misregistration errors prior to DE decomposition. Method and Materials: The method is based on joint histograms of the high‐ and low‐kVp images (or ROIs therein), with optimal image transformations computed to maximize the mutual information between the images. The image is subdivided into a series of ROIs, with an optimal transformation computed for each ROI. Large ROIs are downsampled to reduce the computational complexity of the optimization. The ROI transformations are smoothed and interpolated to determine a pixel‐wise transformation for the entire image. This is repeated with progressively smaller ROIs. Results: The results demonstrate that large scale ROIs (400×400 pixel) are effective in correcting bulk patient motion such as drift or relaxation. A second pass with a smaller ROI (200×200 pixel) corrects breathing and cardiac motion. A final pass with yet smaller ROIs (100×100 pixels) is effective at correcting the motion of fine bronchio‐vascular structure. The combination of these in an iterative multi‐resolution, multi‐scale method effectively registers the high‐ and low‐kVp projections such that DE images exhibit significantly reduced motion artifacts — particularly in the scapulae, aorta, heart,liver, and bronchioles. Expert radiologist readings suggest a significant improvement in image quality and diagnostic performance. Conclusion: The iterative, multi‐resolution, multi‐scale registration corrects misregistration progressively at scales ranging from bulk anatomical drift down to smaller scale motion such as that of fine pulmonary vasculature. The approach is a vital part of the DE image processing chain that has been implemented for a clinical DE imaging trial with 200 patients.
Research partly supported by Eastman Kodak Co.
- Dosimetry, Radiation Protection, and Quality Control
34(2007); http://dx.doi.org/10.1118/1.2761321View Description Hide Description
Purpose: Though equipment and software may claim to be DICOM compliant, interaction with a given electronic environment may result in unexpected problems from poor image quality to a disabled PACS. In order to prevent clinical disruption as a result of introducing a new modality into our electronic environment, we sought to institute a commissioning process that ensures rigor in testing and provides a record of outstanding issues. Methods: We developed a generic testing protocol which encompasses connectivity, workflow verification and image quality. The protocol is designed to be adaptable to the requirements of different modalities. First, analysis is done with images created in and sent through a “test” electronic environment which replicates the system used in production (RIS, PACS, worklist broker, clinical viewer and archive). Once the images have passed the “test” system, and we are confident that the modality will not disrupt clinical workflow, the same testing procedure is continued in the “production” system. In our case, testing is a coordinated effort amongst four groups (IT, Engineering, Radiology Systems Coordinators, and Medical Imaging Technology Specialists). A web based tool was set‐up to facilitate document sharing, multiple‐author‐editing of data collection forms and communication between groups. The testing protocol itself consists of five main categories each of which are detailed by each group according to their roles. Results: The testing protocol was used in the installation of a digital radiographic unit that was never before seen by our PACS. With this, problems were found and modifications to the modality and PACS were made. Conclusion: With the complex interdependent software systems involved in the practice of radiology, adding new systems to a network often results in unintended dire consequences. To mitigate clinical disruption, we've established a commissioning procedure that will be required for introducing new equipment or software into our electronic environment.
TU‐C‐L100J‐02: A Fiber‐Optic Coupled Point Dosimetry System for the Characterization of Multi‐Detector CT34(2007); http://dx.doi.org/10.1118/1.2761322View Description Hide Description
Purpose: Advances in CT acquisition techniques, primarily multidetector CT (MDCT) and cone beam CT(CBCT), make it highly desirable to develop measurement techniques that provide a more physically meaningful measurement of dose than the traditional CT dose index (CTDI). This study presents data based on a point dosimetrysystem utilizing fiber‐optic‐coupled (FOC) radioluminescent dosimeters to measure fundamental parameters associated with CTdosimetry. This point detector approach provides remote, real‐time dose measurements and allows direct recording of single‐scan dose profiles that contain the essential information required to determine dosimetric quantities for MDCT. Method and Materials: FOC dosimeters based on sensitive elements of either a copper‐doped quartz or coupled scintillation phosphor are characterized for their performance across the CT energy range based on energy dependence, dose linearity, and angular response. A custom Labview program provides a user‐friendly interface to control the system.Measurements were made using traditional CTDI as well as FOC dosimeter measurements for a variety of MDCT acquisition protocols. Results:Measurements along the central axis of a CTDI phantom provide a direct evaluation of the single scan dose profile. FOC peripheral point measurementsdetect intensity variation with tube rotation, a dependence on scanner pitch, and permit the correlation of scan parameters and dose profiles. While CTDI remains an accurate prediction of MSAD for axial CT, it is empirically demonstrated to fail for multiple scan dose profiles when pitch is not equal to one. The development of a small dosimeter that can directly measure the helical dose profile provides a useful characterization of MDCT scanning performance and an accurate prediction of MSAD. Conclusion: FOC dosimeters demonstrate high sensitivity, reproducibility, excellent dose linearity, and combined with their small physical size permit accurate point‐dose measurements. These properties provide a useful tool for the characterization of the dosimetry quantities fundamental to MDCT.
TU‐C‐L100J‐03: Radiation Dose Reduction in Interventional Cardiology Using a Fixed Circular Collimator Attached to Cardiovascular Angiography System with Flat Panel Detector34(2007); http://dx.doi.org/10.1118/1.2761323View Description Hide Description
Purpose: Fixed Circular Collimator employed in cardiovascular angiography system with flat panel detector was developed to reduce radiation dose to both patient and cardiologist in coronary intervention. This study was designed to assess the radiation dose reduction using this new collimator.Method and Materials: This new collimator makes the radiation field circular and has tapered aluminum to reduce radiation dose at peripheral field.
1. Entrance air kerma with backscatter (patient) and personnel dose equivalent (Hp) to cardiologist (scattered radiation dose) were measured to verify the radiation dose reduction using fixed circular collimator under a fixed geometric arrangement. The PMMA phantom was employed to simulate the patient thickness varying from 1.0 cm to 35.0 cm in increments of 1.0 cm. Upon completion of the data acquisition, entrance air kerma and Hp to cardiologist were plotted against the phantom thickness.
2. Entrance skin regions (patient) were imaged to verify the overlap region reduction under ten routine angulations (RAO30, RAO10‐CAU30, RAO30‐CAU30, RAO30‐CRA30, LAO10‐CRA30, LAO20‐CRA30, LAO30‐CRA30, LAO45‐CRA30, LAO45, and LAO45‐CAU30). All tabletop position for every angulations were set up to image the heart center of chest phantom, subsequently, chest phantom was replaced by computed radiography and irradiated 1 sec. Results: Entrance air kerma and Hp to cardiologist were reduced by 8.0 % and 47.6 %, respectively. On the other hand, the overlapping region of skin was also reduced considerably. Conclusions: Due to the use of fixed circular collimator, entrance air kerma, the overlap region of exposed skin and Hp to cardiologist were decreased concurrently. It was considered that the use of fixed circular collimator makes the lower density of backscatter and region of interest optimize. Therefore, rectangular images should be reconsidered for safer coronary intervention. Although this study doesn't include image quality evaluation, cardiologists adopt this new collimator in clinical study currently.
TU‐C‐L100J‐04: Absorbed Dose Estimates Obtained Via PET Imaging for I‐124 Intact Antibody in Nude Mice34(2007); http://dx.doi.org/10.1118/1.2761324View Description Hide Description
Purpose: To estimate tumor targeting and radiation dose in mice receiving I‐124‐labeled anti‐CEA (Carcinoembryonic Antigen) intact monoclonal antibody. Use of I‐124 as a PET emitter is complicated by high‐energy photons (603 keV and 723 keV at 63% and 10% respectively) emitted along with the positrons (23%). Method and Materials: Intact monoclonal antibody cT84.66 was labeled with I‐124 and injected via tail vein into nude mice bearing LS174T human colon tumors (0.04 to 0.19 g). PETimaging was performed using the Siemens MicroPET RF scanner at seven time points out to 8 days post‐injection. Standard energy window of 350–750 keV was employed. Average voxel values from organ volumes of interest were used to determine relative magnitudes of activity in tumor,liver,heart (blood) and whole body. Absolute activities were calculated by normalizing these PET data to dissection results at the last time point. Bi‐exponential functions were fitted to resultant activity vs time curves and integrated to establish Ã values and residence times. Organ‐to‐organ S factors were determined by Monte Carlo analysis of I‐124 in a 20g digital mouse phantom. Organ dose was then estimated as the matrix product of SÃ. Results:Tumors were visualized within one day of injection. Average residence times for blood, tumor,liver and whole body were 45, 6.5, 7.0 and 86 hours respectively. Mean dose for the tumors was 3.3 Gy/MBq; associated liver and whole body average values were only 0.60 and 0.31 Gy/MBq respectively. Conclusions: Small colon tumors can be PET‐imaged in nude mice receiving the I‐124 anti‐CEA antibody cT84.66. By normalizing voxel results to those obtained at sacrifice, quantitative I‐124 uptakes can be measured and used to generate residence times. Ratios of tumor/liver and tumor/whole body dose were approximately 5 and 10 respectively.
34(2007); http://dx.doi.org/10.1118/1.2761325View Description Hide Description
Purpose: To develop a filmless technique for the measurement of CT beam width for QA. Method and Materials:Images of pre‐patient collimated CT beams were acquired using Fuji computed radiography(CR) cassettes and direct‐exposure film (DF). CR cassettes were exposed using 80kV, 10mAs to avoid signal saturation. The DF was exposed at 80kV, 30mAs to achieve a proper optical density. The DF image was digitized then linearized. To achieve a proper signal range for CR, various plate‐reading protocols were evaluated. An IDL program was developed to automatically handle both the digitized film and a variety of CRimage sizes. “Unprocessed” logarithmic CR data were linearized, which is critical to validate the FWHM. The program executes in an interactive fashion, whereby the user selects the beam to analyze. The program determines the background signal for subtraction then samples multiple cross‐beam profiles to form an averaged profile. The profile is smoothed and the FWHM is determined. Results: Under proper conditions of image acquisition, results show both DF and CR are appropriate techniques for beam width measurements. The CR technique demonstrated a high level of reproducibility with a coefficient of variation below 0.5%. Also, the CR results were within 0.3mm of specifications. Using various sizes of CR cassettes had a minimal effect on the results. However, if the properly exposed image is saturated due to an inappropriate CR plate reading protocol, the inaccuracy is substantial (>1mm). The technique worked well for annual tests of six GEscanners. Over a wide range of beam width configurations, all measurements were within 0.9mm of specifications. Conclusion: A technique for determination of CT beam width in a filmless environment was developed and tested. The results obtained using CR yield accurate measurements of CT beam profile width. The technique is effective, practical, and has proven to be robust.
34(2007); http://dx.doi.org/10.1118/1.2761326View Description Hide Description
Purpose: To determine radiation dose to specific organs in a pediatric phantom from a CT scan, and to determine if there is a significant difference between: axial (step and shoot) and helical acquisitions, when using very similar photon flux values (effective mAs). Method and Materials: An anthropomorphic pediatric phantom (5 year old equivalent) was used to investigate organ dose at the surface and internal to the phantom. The phantom contains four different tissue equivalent materials: soft tissue, bone, brain, and lung, and was imaged on a 64‐channel CTscanner with a head protocol [axial and two helical scans (pitch = 0.516 and 0.984)], and a chest protocol [axial and three helical scans (pitch = 0.516, 0.984, and 1.375)]. Effective mAs was kept constant (within 3%) for head and chest protocols. Dose measurements were acquired using thermoluminescent dosimeter(TLD)powder in capsules placed in the phantom sections in plug holes. The organs of interest for this study were: brain, both eyes, thyroid, sternum, both breasts, and both lungs.Results: All axial organ dose measurements were significantly higher (p<0.05) than all helical organ dose measurements. There was no significant difference (p>0.05) in organ dose values between the pitch values 0.516 and 0.984 for both head and chest scans. The chest organ dose measurements obtained using a pitch of 1.375 were significantly higher (p<0.05) than the other helical pitches used for chest scans (attributed to the automatic selection of the large focal spot). Conclusion: Communications with the vendor indicate that there are likely physical explanations for the difference observed in organ doses due to axial versus helical acquisitions. It is unclear if this difference in organ dose is unique to this scanner design, because dose measurements are not typically performed in helical scan mode and have not been investigated using other scanner models.
34(2007); http://dx.doi.org/10.1118/1.2761327View Description Hide Description
Background. NCRP Report 93 estimates that diagnostic x‐rays contributed an average effective dose of 0.39 mSv in 1980, and accounted for 11% of the total US population dose from all sources of man made and background radiation. The last 25 years has seen a dramatic increase of medical imaging, particularly CT, but corresponding data on US population doses are virtually non‐existent. Purpose. To estimate the average effective doses to the US population from diagnostic radiology. Method. We obtained data on the number of CT scans performed in the US in 2005 from M Bhargavan and JH Sunshine, PhD. Average effective doses for CT examinations were obtained from the 2000 UNSCEAR report. Assuming that CT accouts for about 70% of all medical exposures, as reported by Mettler et al we computed the average dose to the current US population of 300 million. Results. The average effective dose per CT examination is reported to be 8.8 mSv. The average dose from diagnostic radiology in the US in 2005, where 60 million exams were reported to be performed in 2005, would be approximately 2.5 mSv. This value is comparable to values reported in the literature, including 1.8 mSv for Belgium (1999), 2.0 mSv for Germany (1997) and 2.0 mSv for Luxembourg (2002). In the US, average natural background is 1 mSv from cosmic/terrestial/internal sources, and 2 mSv from Radon exposure. Medical exposures are becoming the dominant sources of population radiation exposure in industrialized countries. Conclusion. Population doses from medical imaging in the US may appear to have increased a factor of six or so in the last generation.
34(2007); http://dx.doi.org/10.1118/1.2761328View Description Hide Description
Purpose: As large bore CTscanners are becoming more prevalent in the marketplace, so have questions about the image quality and dose associated with these scanners. The aim of this work is to compare the image quality and dose characteristics of a large bore 16‐slice CTscanner with a standard 16‐slice CTscanner.Methods: Two CTscanners, a large bore CT (Aquilion Large Bore, Toshiba) and a standard 16‐slice CT (Aquilion 16, Toshiba) were compared with respect to noise,spatial resolution,CT number accuracy, and low contrast. The technique factors on the large bore were adjusted to yield the same dose as a standard abdominal protocol on a standard 16‐slice scanner.Noise was measured on a 24cm water phantom, spatial resolution via the Fourier transform of bead, CT number via several known materials contained in a 32cm diameter phantom, and low contrast via a traditional low contrast phantom. Dose was measured via standard CTDI measurements. Results: The large bore required an increase in mAs of approximately 30% to yield the same dose as a standard 16‐slice scanner. After this adjustment was made, the image quality metrics for both scanners were identical. Conclusion: The large bore CT in this study yielded the same image quality in all four categories as the standard 16‐slice protocol.
K. Boedeker, R. Mather, and M. MacLeod are employees of Toshiba America Medical Systems.
34(2007); http://dx.doi.org/10.1118/1.2761329View Description Hide Description
Purpose. To quantify how CT patient doses, and the corresponding scan times, have changed over the last 25 years. Method. We analyzed data for five generations of CTscanners from one vendor (GE), ranging from the 9800 single slice scanner without slip ring technology used in the early 1980s to a 64 slice VCT, the current state of the art. We determined values of CTDIair and CTDIw at a constant kV/mAs. Effective doses were performed for a standard 32 cm chest CT examination performed at 120 kV using contiguous axial scanning on the 9800, and a pitch of 1 on the VCT. Scan times were computed assuming (contiguous) axial images with a 5 mm thickness.
Results. CTDIair of a VCT scanner (30 mGy/100 mAs) is 12% higher than that of a 9800 scanner, whereas the corresponding body VCT CTDIw (9.5 mGy/100 mAs) is 53% higher. The increased doses on modern scanners are a result of reduced focus‐isocenter distances to accommodate faster rotation speeds, changes in beam shaping filters, and overbeaming used on MSCT scanners. Patient effective doses for a routine chest CT examination are 5 mSv/100 mAs on a VCT and 3.2 mSv/100 mAs on a 9800. Acquisition of 64 axial images for a routine chest CT in the early 1980's took ∼5 minutes, whereas modern scanners cover the same volume in only 8 x‐ray tube rotations taking ∼3 seconds.
Conclusion. Over the last 25 years, normalized CT patient doses (i.e., per unit mAs) have increased by ∼50%, whereas image acquisition times have been reduced by two orders of magnitude.
TU‐C‐L100J‐10: Combining Measurement and Monte Carlo Methods for Dose Assessment in Flat‐Detector CT34(2007); http://dx.doi.org/10.1118/1.2761330View Description Hide Description
Purpose: Dose assessment in flat‐detector CT (FD‐CT) combining measurements and Monte Carlo(MC) simulations. Material and Methods: FD‐CT scanners provide large irradiation fields of typically 100 mm to 250 mm in the longitudinal direction. In consequence dose assessment according to the current definition of the CTDI would demand larger ionization chambers and phantoms which are not practical. We propose a method which includes a measurement in air or in a phantom with an integrating dosimeter to assess the dose at that point and to combine it with MC simulations to assess 3D dose distributions and integral dose for arbitrary objects and geometries. For validation purposes measurements were performed on a C‐arm system (Siemens Medical Solutions, Forchheim, Germany) equipped with a flat‐detector of 40 × 30 cm2. Dose was assessed for various tube voltages in cylindrical PMMA phantoms of 16 cm and 32 cm diameter with a varying z‐extent from 15 to 60 cm. The MC results were compared to the values obtained with calibrated ionization chambers of 100 mm and 250 mm length and to TLD dose profiles along the complete z‐extent of the phantoms. Additionally a comparison to measurements of dose distribution in an anthropomorphic phantom was performed. Results: The MC simulation was in agreement with the reference TLD measurements to within better than 10%. Standard CTDI phantoms with a z‐extent of 15 cm underestimate the dose at the center by up to 20% whereas a z‐extent of 300 mm appears to be sufficient for FD‐CT. The 100 mm chamber underestimates the measured CTDI value by over 40%. The MC tool can be used to calibrate the measurements with the ionization chamber of 100 mm. Conclusion: The combination of measurement and validated MC tool appears to be a flexible solution to assess arbitrary dose characteristics.
- 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.