Index of content:
Volume 36, Issue 6, June 2009
- Imaging Scientific Session: Room 304A
- Breast Imaging and CAD
36(2009); http://dx.doi.org/10.1118/1.3182537View Description Hide Description
Purpose: The goal of this work is to prove theoretically, and demonstrate in an example application, that 3‐class classification in a 2D decision space results in an equal or higher data concordance value than 2‐class classification in a 1D decision space. Material and Methods: Most medicaldiagnostic tests are interpreted using binary classification methods, i.e., patients are classified into normal vs. abnormal. Most of these tests, however, are not strictly binary since both normal and abnormal classes often contain distinct sub‐groups. Here, we will provide theoretical proof that, for a 3‐class problem, 3‐class classification in a 2D decision space results in an equal or higher data concordance value than 2‐class classification in a 1D decision space, and that, in order to take advantage of the 3‐class classification, the decision must be made in a 2D decision space. We applied the theory to computerized lesion classification in breast ultrasound imaging. The database consisted of sonographic images of 128 cancerous, 168 benign solid and 74 cystic lesions. We used 4 lesion descriptors as inputs for 3‐class and 2‐class Bayesian neural network (BNN) classifiers. The 3‐ and 2‐class BNNs were trained and tested in leave‐one‐case‐out analyses for the task of differentiating among 3 (cancerous, benign solid, and cystic) and 2 classes (cancerous vs. non‐cancerous), respectively. We performed a 2‐alternative‐forced‐choice (2AFC) experiment to assess the performance of both classifiers for the task of classifying cancerous vs. non‐cancerous. Results: In the 2AFC experiment, the percent correct was 92% using 3‐class classification in a 2D decision space, which was significantly higher than the 85% obtained using binary classification in a 1D decision space. Conclusion: This work demonstrates that 3‐class classification in a 2D decision space has the potential to improve computer‐aided diagnosis.
WE‐D‐304A‐02: Exploring Non‐Linear Feature Space Dimension Reduction and Data Representation in Breast CADx36(2009); http://dx.doi.org/10.1118/1.3182538View Description Hide Description
Purpose: Preliminary study to explore potential of recently developed non‐linear dimension reduction and data representation techniques as applied to breast lesion CADx generated feature spaces. Method and Materials: Two new methods were explored: Laplacian eignenmaps of (Belkin and Niyogi) and t‐distributed stochastic neighbor embedding (t‐SNE) (van der Maaten and Hinton). The properties of these methods were evaluated in the context of malignancy classification performance as well as visual inspection of the sparseness for two and three dimensional mapping representations. The robustness of the proposed techniques were tested against four separate imaging modality feature databases, including 2956 ultrasound(US), 735 full‐field digital mammography (FFDM), 850 screen‐film mammography (SFM), and 356 DCE‐MRI biopsy proven breast mass lesions. Using the reduced dimension mapped feature output as input, as opposed to feature selection, two classifiers, linear and non‐linear, were tested: Markov Chain Monte Carlo based Bayesian artificial neural network (MCMC‐BANN) and linear discriminate analysis (LDA). To evaluate performance, the AUC was estimated for each classifier using ROC analysis and 0.632+ bootstrap validation on over 500 samples. In addition, performance was compared to a non‐linear Automatic Relevance Determination (ARD), linear step‐wise feature selection method, as well as a linear reduction Principle Component Analysis based method. Results: The new methods were found to match or exceed performance of current state‐of‐the‐art breast lesion CADx algorithms for feature selection and classification across all modalities. Additionally, the new techniques possess the added benefit of naturally delivering sparse lower dimensional representations for visual interpretation, thus, revealing intricate data structure of the feature space. Conclusion: These methods maintain predictive power while also preserving both local and global structural information present in the original high dimensional feature space for breast lesion feature data used in CADx.Conflict of Interest: M. Giger is a stockholder and receives royalties from Hologic.
WE‐D‐304A‐03: Quantitative Image Analysis for Prognosis in Newly Diagnosed Breast Cancer Patients with Sonographically Normal Appearing Lymph Nodes36(2009); http://dx.doi.org/10.1118/1.3182539View Description Hide Description
Purpose: To investigate whether a computerized quantitative image analysis (QIA) system for diagnostic breast ultrasound can predict whether a breast cancer lesion has metastasized in patients with sonographically normal appearing axillary lymph nodes. Method and Materials: This study used a database of sonographic images of breast cancers in patients who presented with a single cancerous lesion and axillary lymph nodes that appeared normal on diagnostic breast ultrasound. The dataset contained images of 50 patients. In 20 patients, the breast cancer had metastasized and in 30 patients the cancer remained localized at the time of surgery. The truth regarding disease grade/pathology was obtained from the final breast surgery and lymphadectomy. Features including both lesion and parenchymal characteristics were calculated for each lesion. A leave‐one‐case‐out analysis was performed and each iteration stepwise feature selection was combined with the training of an LDA classifier for the task of distinguishing between cancerous lesions that metastasized and ones that remained localized. Results: In the leave‐one‐case‐out analysis, the most frequently selected combination consisted of 3 features. In this set, a single feature pertained to the lesion itself and the other 2 features described a difference between the lesion and the surrounding parenchyma. The features were the lesion echogenicity, the difference in echogenicity, and the difference in entropy. The LDA classifier yielded an area under the ROC curve (AUC value) of 0.80 (standard error 0.06) for the task of distinguishing between metastasized breast cancer index lesions and localized breast cancers.Conclusion: The QIA scheme obtained promising performance in this preliminary study and shows potential to diagnose metastatic disease. It is important to note that this study was performed using images only of patients for whom pre‐surgery imaging studies suggested that the breast cancer had remained localized.
Conflict of Interest: Grants NIH and DOE. Stockholder, grant, royalties Hologic.
WE‐D‐304A‐04: Performance of a Method to Aid Breast Ultrasound Interpretation Using Image Processing and Case‐Based Reasoning36(2009); http://dx.doi.org/10.1118/1.3182540View Description Hide Description
Purpose: To examine factors that may impact accuracy, reproducibility of Breast UltrasoundCADxsoftware. Goal for CADx is standardized reporting of findings using well‐defined descriptors and to aid accuracy and reproducibility of interpretation by radiologists.Method and Materials: Breast Companion® (BC, Almen Laboratories and UCSD) provides tools to analyze breast masses. It computes 9 features of the mass, compares these to images in Reference Library of masses with known findings. Using Relative Similarity it retrieves instantaneously a cluster of the most similar cases and outputs numerical data including Computerized Lesion Assessment (CLA) following the ACR BI‐RADS assessment category (1 through 5). Results: On 596 cases with known findings BC achieved ROC Area 0.98±0.02, Sensitivity 99.7%, Specificity 96.1%, significantly higher than four experienced radiologists who read the same cases (ROC Areas 0.88 to 0.90±0.02). Increasing the number of cases in the Reference Library from 41–331 and the number Test Set cases from 20–924 keeping mix same, ROC varied from 0.90±0.07 to 0.98±0.02 (NS). In 55 cases (40% malignant) with confirmed findings BC calculated CLAs for radial and anti‐radial images that were highly correlated, not significantly different by two‐tailed t‐test (p>.5). CLA for simple cysts was 2.0, mean for complicated cysts was 2.5, 3.0 for solid benign and 4.5 for malignants. 28 subjects (40% malignant) were analyzed who had exams of the mass on two different occasions within 3 weeks on GE and Siemens (NS difference). Simulated masses of were imaged over 3 months at 9, 10, 13 MHz with varying gain, dynamic range, focal depth post‐processing. CLA was stable and independent of or linearly scaled to machine parameters. Conclusion: The BC system accuracy is high, stable over range of conditions. Measured parameters are dimensionless, relative to their backgrounds so it appears normalization achieves mitigation in machine effects.
WE‐D‐304A‐05: Analysis of Tumor Grade Breast Carcinoma Using Computer‐Extracted Morphological and Kinetic Features in DCE‐MRI36(2009); http://dx.doi.org/10.1118/1.3182541View Description Hide Description
Purpose: An important prognostic marker in breast cancer is the histological grade of the breast tumor. Thus, the purpose of our study is to investigate the performance of computer‐extracted morphological and kinetic features in DCE‐MRI in distinguishing breast tumors with varying histological grades: Grade 1, Grade 2, and Grade 3. Materials and Methods: Breast MR images were obtained with a T1‐weighted SPGR sequence using Gd‐DTPA on a 1.5T GEMRI scanner. Each case has one precontrast and three to five postcontrast series at intervals of 68 seconds, and each series contains 60 coronal slices. The database contains 135 invasive ductal carcinoma (IDC) breast lesions including 21 Grade 1 lesions, 72 Grade 2 lesions, 42 Grade 3 lesions, and 133 benign lesions (no cysts). All lesions were verified pathologically. Each lesion was segmented and its characteristic kinetic curve was extracted using the fuzzy c‐means method. Textural, morphological, kinetic, and spatial enhancement variance features were extracted, and stepwise linear discriminant analysis using a Wilks lambda cost function in a round‐robin fashion was used for feature selection. The selected features were merged using Bayesianneural network, and the classification performance was evaluated using receiver‐operating characteristics (ROC) analysis.Results: We achieved AUC values of 0.80±0.05, 0.85±0.02, 0.75±0.06, 0.74±0.05, 0.75±0.06, and 0.60±0.05 for the classification tasks of Grade 1 vs. benign, malignant (Grades 1–3 lesions) vs. benign, Grade 1 vs. Grade 3 lesions, Grade 1 vs. Grades 2 and 3 lesions, Grade 1 vs. Grade 2, and Grade 2 vs. Grade 3, respectively. Conclusions: Computerized analyses of breast MR images have the potential to yield image‐based prognostic markers for breast carcinoma, specifically in distinguishing different histological grades of breast tumors.
36(2009); http://dx.doi.org/10.1118/1.3182542View Description Hide Description
Purpose: To demonstrate the effectiveness of contrast‐enhanced dual energy mammography (CEDEM) in characterizing the contrast kinetic curve. Method and Materials: The acquisition of CEDEM was performed using a dedicated cone‐beam breast CT scanner in our laboratory, with the system operating in stationary imaging mode. CEDEM images were acquired with a filter combination of 0.2 mm Cu and 0.2 mm Sn at the same kVp. Previous studies in our group have shown that 55 kVp is the optimal technique factor to acquire high‐quality CEDEM images when imaging a 6‐cm, anatomical‐complex breast phantom. As an addition to the static breast phantom, a tumor perfusion chamber was designed and fabricated to simulate the contrast kinetics of breast tumors. The chamber is composed of a “leaky” vessel (soaker hose) and the tumor interstitial space (foam). The contrast kinetics was controlled by two peristaltic pumps, the “leakiness” of the soaker hose, and the permeability of the foam. When imaged experimentally, the tumor perfusion chamber was fixed in the middle of the static breast phantom. A sequence of CEDEM images were acquired with an initial bolus injection of 100 ml Visipaque®. The contrast kinetics was determined by evaluating the signal‐to‐noise ratio of each dual‐energy subtracted image as a function of time. Results: The CEDEM image sequence of the dynamic breast phantom was acquired at 55 kVp. The contrast kinetic curve was found to follow closely that of human subjects. The influence of the phantom parameters on the kinetic curve will be presented. The contrast kinetic curve modeling benign and malignant tumors using this phantom will be presented. Conclusion: The contrast kinetic curve was evaluated using a dynamic breast phantom. CEDEM was found to be effective in characterizing the contrast kinetic curve around breast tumors.
WE‐D‐304A‐07: A Novel Stereo‐Scanning Tomosynthesis Scheme and Its Application in Image Guidance for Breast Radiotherapy36(2009); http://dx.doi.org/10.1118/1.3182543View Description Hide Description
Purpose: In conventional digital tomosynthesis (DTS), images are generated by scanning an x‐ray source and flat‐panel detector pair on a one‐dimensional trajectory. A novel stereo‐scanning tomosynthesis (STS) is proposed in which the x‐ray source is allowed to scan on a two‐dimensional (2D) surface. By allowing the x‐ray source to project from more stereo angles, the scanning scheme of STS provides more coverage in the spatial‐frequency domain of the object, consequently better image quality. Method and Materials: The feasibility and effectiveness of STS is corroborated by computer simulations, using a three‐dimensional (3D) numerical breast phantom in its natural shape. Iterative algorithm in the form of total variation regulated expectation maximization (TV‐EM) is developed for image reconstructions. A STS scheme is proposed as the image guidance method for prone‐position breast radiotherapy.Results: The STS images possess excellent image quality for online tumor targeting, superior to conventional DTS. STS outperforms DTS in restoring certain features of the object and reducing artifacts, due to its more frequency‐domain coverage. Conclusion: By allowing more general scanning trajectories than conventional DTS, STS may provide more flexible imaging geometry or/and better image quality, in various clinical tasks of image guided radiotherapy and diagnostic imaging.
36(2009); http://dx.doi.org/10.1118/1.3182544View Description Hide Description
Purpose: Traditionally, brachytherapytreatment planning relies on two‐dimensional imaging using orthogonal radiographs to provide basic positioning. True organ volumetric 3D information would improve planning and the ability to better conform dose delivery. Modern brachytherapy suites include imaging devices capable of producing cone beam CT(CBCT)images; however collision and patient dose may limit the viability of complete axial rotation. Digital tomosynthesis (DTS) or limited arc techniques overcome these problems but have reduced edge information in two planes. In this study, we compare the efficacy of different combinations of short‐arc techniques to CBCT and CT for volumetric fidelity. Method and Materials: The Digital Integrated Brachytherapy Unit (Nucletron, Veenendaal, The Netherlands), with combined L‐arm and C‐arm construction, is capable of multiple axes of isocentric rotation and allows imaging with non‐axial arcs. For this study, a comparison is made between CT and DTS using multiple image‐sweep combinations: 190‐degree L‐arm sweep (190‐L), 50‐degree L‐arm sweep (50‐L), combination of 50‐degree L‐arm sweep and 90‐degree offset orthogonal 50‐degree C‐arm sweep (L+C), and combination of two orthogonally spaced 50‐degree L‐arm sweeps (L+L). Our phantom consists of a contrast‐filled plastic sphere immersed in water. Data from CT reconstructions are considered the standard for which comparisons are made. For each of the four reconstructions and each orthogonal plane, central slice diameters were measured and compared for fidelity to the CT dataset. Results: As expected, the 190‐L results matched closest with CT. The 50‐L images (standard DTS) showed the lowest fidelity. The two combined‐arc techniques gave improved results over a single arc. Of the limited arc techniques, the combined L+C reconstruction gave superior overall results and compared well with the 190‐L volume. Conclusion: Short‐angle orthogonal‐sweep imaging may give adequate volumetric data versus CBCT and improvement versus single‐arc DTS.
Conflict of Interest: Research supported by Nucletron, B.V. and T32‐CA113267.
WE‐D‐304A‐09: Characterization of Multi‐Beam Field Emission X‐Ray Source for Stationary Digital Breast Tomosynthesis36(2009); http://dx.doi.org/10.1118/1.3182545View Description Hide Description
Purpose: The current prototype digital breast tomosynthesis (DBT) scanners are based on the regular full‐field digital mammography systems and require partial isocentric motion of a mammography x‐ray tube over certain angular range to record the projection views needed for reconstruction. This prolongs the scanning time and in turn degrades the imaging quality due to motion blur. We are developing a stationary DBT (s‐DBT) scanner to mitigate the above limitations. Method and Materials: The proposed s‐DBT system is based on the carbon nanotube multi‐pixel field emission x‐ray (MBFEX) technology demonstrated by our group. The pixilated and spatially distributed MBFEX source can generate x‐ray radiation from multiple views without any mechanical motion of the source, detector, or object. This enables the design of tomography systems with great flexibility in source configuration and imaging sequence. It further enables multiplexing imaging — simultaneously collection of multiple images using one detector.Results and Conclusions: To demonstrate the feasibility of the s‐DBT scanner, we have designed and constructed a proof‐of‐concept full‐field s‐DBT system. The configuration of the system, in terms of angular coverage, number of views, and dose, etc., closely resembles the Siemens DBT scanner for a more realistic comparison. In the Siemens DBT scanner, the x‐ray tube moves along an arc with the exposure points evenly distributed along the rotation route. Our s‐DBT scanner is designed with the MBFEX pixels positioned in a straight line parallel to the detector plane. We will report a detailed study on the design and the performance characteristics of the MBFEX source. In particular evaluation on key parameters including the x‐ray flux, the lifetime of the x‐ray source, effective x‐ray focal spot size, variation between different source, system spatial resolution, and x‐ray energy spectrum will be reported. These preliminary results demonstrate the feasibility of the proposed s‐DBT scanner.
- Computed Tomography and Radiation Dose
TU‐C‐304A‐01: The Need and Feasibility of a Modern Software for Reporting Patient Doses From CT Scans36(2009); http://dx.doi.org/10.1118/1.3182348View Description Hide Description
Purpose: To demonstrate the need and feasibility to develop a modern software tool for reporting the organ dose and effective dose for patients undergoing CT examinations. Method and Materials: Existing CT dose reporting software do not meet the need because of the simplified anatomical phantoms, updated ICRP data and scanner information. A new software is designed with original dose data derived from Monte Carlo simulations involving CTscanner models from various companies and anatomically realistic phantoms. X‐ray sources and protocols are modeled. The Pregnant Women, Adult Male and Adult Female phantoms are utilized. Organ doses and effective doses are computed using both the ICRP‐60 and the latest ICRP‐103 recommendations. The software is developed using the Visual C♯.NET with a modern graphical user interface (GUI) design to allow a user to specify the patient type, body scan region, and scanner operating parameters. Object‐oriented programming technology allows the phantoms to be displayed in 3D interactively. Results: Compared to values reported by the existing software, the organ dose estimates can be different by a ratio as 0.77 to 1.24 for the organ or tissue covered in the scan range, and 0.13 for the organs out of the scan region between calculations using the anatomically realistic phantoms. In addition to the improved dose accuracy, the new program offers a number pf modern GUI features through which 3D phantoms are vividly inspected for organs that receive a high doses. Based on the user‐specified scanning parameters, organ and effective doses are rapidly reported. Conclusion: Preliminary results have demonstrated the aim of the new software design in addressing the needs for new CTscanners, ICRP recommendations and anatomically realistic phantoms. When fully developed, this new tool is expected to improve both the accuracy and usability in reporting CT doses in the future.
TU‐C‐304A‐02: The Impact of the New ICRP‐103 Recommendations On the Assessment of Effective Doses From CT Procedures36(2009); http://dx.doi.org/10.1118/1.3182349View Description Hide Description
Purpose: To apply a pair of adult phantoms representing ICRP‐89 50th‐percentile adult males and females to the study of impact of the new ICRP‐103 recommendations on the assessment of Effective Dose from CT procedures. Method and Materials: a pair of mesh‐based computational phantoms, RPI Adult Male (RPI‐AM) and RPI Adult Female (RPI‐AF) that were recently developed to represent the ICRP‐89 50th‐percentile adult males and adult females. This pair of phantoms has the detailed bone structures, including the spongiosa which contains the red bone marrow. The detailed RBM distribution was adjusted according to ICRP Publication 70. The CT scanner model used in this study is an MDCT scanner which includes the source geometry and movement, the source energy spectrum, the bow‐tie filter as well and the beam shape. CT scan protocols including whole body scan were carefully modeled in this study, and tube potential of 120 kVp were considered. All simulations were performed using the Monte Carlo code, MCNPX 2.5.0. The three‐correction factor method was used to calculate the RBM dose. Effective Dose results were calculated following the algorithm from ICRP 103. Results: A new set of organ absorbed dose results has been presented using this pair of new developed reference adult phantoms from CT procedures, as well as the effective dose results. Also the new results of red bone marrow dose have been provided. The recently published ICRP 103 updated the radio‐sensitive organ list; also it improved the algorithm of effective dose calculations. Conclusion: Advanced red bone marrow dose calculation method has been used in this study due to the detailed bone structures of this pair of RPI‐AM and RPI‐AF phantoms. This new set of effective dose dataset based on the new ICRP‐103 recommendations could be used to provide latest information for clinical diagnostic dosimetry area.
TU‐C‐304A‐03: Stylized MIRD Phantoms Should Be Replaced by Anatomically Realistic Phantoms: Discrepancies In Red Bone Marrow Doses From CT Scans36(2009); http://dx.doi.org/10.1118/1.3182350View Description Hide Description
Purpose: To test the hypothesis that the stylized MIRD phantoms would cause significant error in the estimated red bone marrow (RBM) dose from CT scans in comparison with anatomically realistic phantoms. Method and Materials: The MC model of the CT scanner include the source geometry, movement, source energy spectrum, bow‐tie filter, as well and the beam shape. MCNPX 2.5.0 was used to simulate the RBM dose from various CT scanning procedures. To calculate the absorbed dose to the RBM as a function of photon fluence in the spongiosa and the photon energy, an F4 tally together with a set of DE/DF cards in MCNPX were used to score the photon fluence in MCNPX. The stylized MIRD phantom and the anatomically realistic RPI Adult Male and Adult Female phantoms were implemented in the MCNPX to determine organ doses using the same dose algorithm. Results: For all the cases studied, the RBM doses calculated using RPI adult phantoms were gearter than those obtained from MIRD‐ORNL phantoms. For the chest CT scan, the RBM dose ratio (RPI‐AM to MIRD‐ORNL) is about 1.50 (1.48–1.51), and RBM dose ratio of female phantoms is about 1.28 (1.24–1.32). For the abdominal‐pelvis CT scan, the RBM dose ratios are 1.30 (1.28–1.31) and 1.32 (1.28–1.38) for male and female phantom, respectively. These differences are mainly from the anatomical differences in the phantoms. Conclusion: As the RBM is not uniformly distributed in the human body, the homogeneous bone mixtures definition by MIRD phantoms underestimated the dose by as much as 50% in certain cases. This test concludes that the simplified MIRD phantoms used in existing CT dose software should and can be replaced by realistic phantoms. This is an opportunity to improve the anatomical realism and therefore the associated dose and risk assessments for patients who undergo CT examinations.
TU‐C‐304A‐04: Monte Carlo Based Multidetector CT Modeling and Dose Calculations for Pregnant Patients36(2009); http://dx.doi.org/10.1118/1.3182352View Description Hide Description
Purpose: To model and validate the multidector CT (MDCT) scanner and to assess radiation dose to the fetus and pregnant patient in three different gestational periods. Method and Materials:Monte Carlo code, MCNPX, was used to simulate the x‐ray source including the energy spectrum, filter, and scan trajectory. Detailed CTscanner components were specified using an iterative trial‐and‐error procedure for a GE LightSpeed CTscanner. The scannermodel was validated by comparing simulated results against measured CTDI values and dose profiles reported in the literature. The source movement along the helical trajectory was simulated using the pitch of 0.9375 and 1.375, respectively. The validated scannermodel was then integrated with phantoms of a pregnant patient in three different gestational periods to calculate organ doses and fetal doses. Results: Comparison between simulated results and reported results in literature shows good agreement in terms of CTDI values as well as dose profiles. It was found that the dose to the fetus of the 3‐month pregnant patient phantom was 0.13 mGy/100mAs and 0.57 mGy/100mAs from the chest and kidney scan, respectively. For the chest scan of the 6‐month patient phantom and the 9‐month patient phantom, the fetal doses were 0.21 mGy/100mAs and 0.26 mGy/100mAs, respectively. All these scans were performed with protocols that did not contain the fetus directly in the x‐ray beam. The paper also discusses how these fetal dose values can be used to evaluate imaging procedures and to assess risk using recommendations of the report from AAPM Task Group 36. Conclusion: This work demonstrates the ability of modeling and validating MDCT scanner by Monte Carlo method, as well as rapidly and accurately assessing fetal dose and organ doses by combining the MDCT scannermodel and pregnant patient phantom.
TU‐C‐304A‐05: Dose Reductions of Bismuth Shields in Diagnostic Radiology: Measurements and Monte Carlo Simulations36(2009); http://dx.doi.org/10.1118/1.3182353View Description Hide Description
Purpose: To assess the dosimetric characteristics of bismuth breast shields for a CT beam with ion chamber measurements and Monte Carlo simulations.Method and Materials: Primary attenuation and backscattereffects of both adult and pediatric bismuth superficial organ (e.g. breast) shields were measured with a 0.18‐cc ion chamber and a rectangular slab phantom made of a tissue equivalent material. Simulated CT beams (120 kVp with 100∼400 mAs) were used to irradiate the ion chamber with and without bismuth shields. Both 2‐ply (pediatric) and 4‐ply (adult) bismuth shields (F&L Medical Products, Vandergrift, PA) were placed at the front of the chamber for primary attenuation measurements and at the back of the chamber to measure backscatterradiation.Radiationdoses were measured free‐in‐air and in the tissue‐equivalent slabs. Monte Carlo simulations for the same measurement settings were performed by using EGSnrc/BEAMnrc code. Results: Mean radiationdose reduction from primary attenuation was about 23% (2‐ply) and 40% (4‐ply) for the free‐in‐air and tissue slab measurements. The radiationdose increase from backscatter was around 2% for both the 2‐ply and 4‐ply shields. The Monte Carlo simulations produced the dose reduction from primary attenuation was about 20% (2‐ply) and 38% (4‐ply), and the dose increase from backscatter was around 6% for both shields.
Conclusion: Primary attenuation is the dominant factor that induces radiationdose changes in bismuth shields in CT examinations. The dose contribution from backscatteredradiation in bismuth shields is very small.
36(2009); http://dx.doi.org/10.1118/1.3182354View Description Hide Description
Purpose: Previous work demonstrated that there are significant dose variations on the peripheral, or surface of either a CTDI 32cm phantom or an anthropomorphic phantom when helical CT scanning is performed. The purpose of this work is to investigate the effectiveness of exploiting these variations to reduce dose to targeted radiosensitive organs solely by varying the tube start angle in CT scans.Method and Materials: Radiation dose to several radiosensitive organs (including breasts, thyroid, uterus, gonads, lens of eyes) from a MDCT CT scanner were estimated using Monte Carlo simulation methods on GSF Baby phantom. Whole body scans were simulated using 120kVp, 300mAs, 28.8 mm nominal collimation, pitch 1.5 under a wide range of start angles (0 to 340 degrees in 20 degree increments). The relationship between tube start angle and organ dose was examined for each organ and the potential dose reduction was calculated. Results: The organ dose shows obvious variation depending on the tube start angle. For small peripheral organs, (e.g. the lens of eyes), the minimum dose can be 35% lower than the maximum dose, depending on tube start angle. For pitch 1.5 scans, the dose is usually lowest when the tube start angle is such that the x‐ray tube is posterior to the patient when it passes the longitudinal location of the organ.Conclusion: Helical MDCT scanning results in “cold spots” and “hot spots” that are created both at surface and even in‐depth locations within patients. If organs have a relatively small longitudinal extent, their dose may be reduced by selecting the tube start angle such that the location of these “cold spots” may be manipulated by appropriately selecting the tube start angle. This dose reduction should not have any implications for image quality as there is no change in mAs or total mAs.
TU‐C‐304A‐07: X‐Ray Tube Current Modulation and Effective Dose Per Unit Dose‐Length Product Conversion Factors in CT Dosimetry36(2009); http://dx.doi.org/10.1118/1.3182355View Description Hide Description
Purpose: To quantify how axial and longitudinal x‐ray tube current modulation influence effective dose per unit dose‐length product (E/DLP) conversion factors in chest CT.Method: We simulated a chest CT examination using a 4 cm beam width with projections obtained at every 15° x‐ray tube position at a constant tube output (120 kV). A radiographic patient dosimetry software package (PCXMC) was used to quantify relative patient effective dose as a function of the angular position and longitudinal location (z) of the x‐ray tube. Typical angular and longitudinal mA modulation schemes were obtained from the scientific literature. E/DLP conversion factors were generated for: (a) no mA modulation; (b) angular modulation alone; and (c) longitudinal modulation alone. Results: As the x‐ray tube rotates around the patient, the highest effective dose was at 285° (AP projection) and the lowest effective dose was at 195° (lateral projection), with the maximum to minimum ratio of 2.2. An angular mA modulation scheme with an AP/PA tube current one third of the lateral tube current reduces the E/DLP conversion factor in chest CT by 4.2%. For x‐ray tube movement along the z‐axis, the maximum to minimum ratio of patient effective dose was 3.3. In chest CTimaging, the longitudinal mA modulation changes the tube current approximately seven fold between the central lung area and the upper thorax region above the patient's lungs. Application of this longitudinal mA modulation scheme reduces the E/DLP conversion factor in chest CT by 9.2%. Conclusions: Use of longitudinal and angular mA modulation schemes in chest CT examinations could reduce E/DLP conversion factors by ∼13%.
36(2009); http://dx.doi.org/10.1118/1.3182356View Description Hide Description
Purpose:CT manufacturers are required to report dose using standardized CTDI phantoms. However, CTDI can be confusing when applied to pediatric patients due to variations in patient size. We propose a simple method to scale standard CTDI to better represent a pediatric patient's dose by providing universal conversion tables that are indexed by patient size ranges. Method and Materials: The Medical Imaging and Technology Alliance (MITA) and Image Gently Pediatric CT Physics Work Group have obtained CTDI ratios between the 32 cm, 16 cm and 10 cm phantoms for current 16 and 64 slice scanners. X‐ray attenuation was measured for the CTDI phantoms and retrospectively determined from pediatric and adult pre‐scan projection images. Regression relationships between patient dimensions, attenuation and dose ratios were used to create tables of CTDI scale factors as a function of patient dimensions. Results: The range of attenuation ratios of the 2nd and 3rd quartiles for pediatric heads compared to the standard CTDI phantoms was 0.85 to 0.95 (16 cm) and 0.46 to 0.51 (32 cm). The range for pediatric bodies was 1.03 to 1.35 (16 cm) and 0.85 to 0.95 (32 cm). Individual measurement errors relative to the mean CTDI ratio due to manufacturer, scanner generation and kVp or bowtie filter selections were minimal. Preliminary results show errors increased from 0 % to 14 % with increasing difference between the attenuation of the pediatric patient and CTDI phantom. Conclusion: Universal dose conversion tables can be provided based on dimensions or can be automatically calculated from a patient's attenuation determined from a pre‐scan projection image with reasonable accuracy to scale CTDI or its future replacement to better represent dose for pediatric patients.
36(2009); http://dx.doi.org/10.1118/1.3182357View Description Hide Description
Purpose: The purpose of this study was to evaluate potential radiation dose reduction for pediatric CT.Materials and Methods: A dose‐reduction simulation tool, which adds synthetic noise to raw projection measurements and reconstructsimages at a simulated lower dose (Massoumzadeh, et al. Med. Phys. Vol. 36, pp. 174–189, 2009), was used to simulate low‐dose CTimages. Simulated low‐dose CTimages are created from full‐dose CTimages of normal and pathological slices (lung nodules and abdominal visceral lesions). The amount of added noise was task dependent, with 10 sets of simulated low dose ranged from 1% to 85% of original dose. Sixteen pediatric cases were selected, including eight normal, three patients with pulmonary nodules, two patients with abdominal visceral organ lesions, and three patients with appendicitis without perforation. All studies performed on a 16‐row scanner, with the effects of tube current modulation and bow tie filters included. Following a short training session, 19 volunteer radiologists, from various clinical centers in the world who were attending the International Pediatric Radiology conference in Montreal in 2006, used a 5‐point scale to rate a sequence of simulated, low‐to‐high exposure images for the presence or absence of lesions. They were total of 176 images with viewing sessions limited to 45 minutes. Diagnostic agreement between full‐dose and reduced‐dose images was assessed with a weighted Kappa statistic. Result: For detection of pulmonary nodules, a decrease in the average intra‐observer agreement (kappa= 0.90) was found at 80% dose reduction, while for the detection of abdominal lesions or appendicitis a 50% reduction was observed. Conclusion: There is potential for dose reduction in CT studies, which is task dependent and greater for pulmonary nodule than for abdominal visceral lesions and appendicitis. The noise simulation methodology is a powerful tool to help understand the relationship among dose,noise, and observer agreement.
36(2009); http://dx.doi.org/10.1118/1.3182358View Description Hide Description
Purpose: The purpose of this research was to systematically explore the change in contrast‐noise ratio (CNR) for a variety of substances as a function of kV in CT examinations for small body habitus under the condition of constant dose. Method and Materials: The head module of a Gammex RMI 461A Head/Body phantom (constructed from solid water) was suspended from the end of the table of a Philips Brilliance 6‐slice CTscanner. An ion chamber introduced into the center of the phantom was used to adjust the mAs at kVp values of 90, 120 and 140 so that the exposure at the center was the same at all kVp values. Fifteen materials were tested at the center of the phantom for each of the kVp values with the CT number mean and standard deviation (SD) measured for each of the resulting 45 scans. A control area in the solid water was also tested for each scan using an identical ROI. Related tests were performed on all samples at once using other scanners.
Results: For all 45 scans, the SD was independent of kVp demonstrating that the noise was a function of dose only and not kV. For high density materials, lowering the kV improved the CNR in a predictable way. For materials with CT numbers not far from water, however, in some cases the CNR became worse with decreasing kVp while in others, the CNR improved. Conclusion: While lowering the kVp vs. lowering the mAs in small body habitus may result in an improved CNR, it should not be assumed that this is always the case. Dose reduction strategies for particular situations may benefit more by reducing the beam intensity and keeping the kV high.
- Computed Tomography and Reconstruction
TU‐D‐304A‐01: Development and Testing of a Novel, 4D Maximum A Posteriori (MAP) Image Reconstruction Algorithm36(2009); http://dx.doi.org/10.1118/1.3182389View Description Hide Description
Purpose: Deformable image registration has been proven to be useful in tracking organ motion for dose calculation using artifact‐free 4D RCCT images. Such methods are challenged in the presence of image artifacts. We present an alternative method which avoids binning artifacts by directly estimating organ deformation during the reconstruction process. Method and Materials: We have developed a maximum a posteriori (MAP) algorithm for tracking organ motion that uses raw time‐stamped data to reconstruct the images and estimate deformations in anatomy simultaneously. Since the algorithm does not rely on a binning process, binning artifacts are avoided. Signal‐to‐noise ratio (SNR) is also increased since the algorithm uses all of the collected data. The increased SNR provides the opportunity to reduce dose to the patient during scanning. This framework also facilitates the incorporation of fundamental physical properties such as the conservation of local tissue volume during the estimation of organ motion. In order to validate the accuracy of the 4D reconstruction algorithm, a phantom study was performed using the CIRS anthropomorphic thorax phantom in a CT scanner. An improvement in image quality was also demonstrated by application of the algorithm to data from a real liver stereotactic body radiation therapy(SBRT) patient. Results: The algorithm accurately estimated the known motion of the anthropomorphic phantom. Additionally, a significant SNR increase was observed when using 4D reconstruction over binning, even for a scan with X‐ray tube current reduced to 10%. Conclusion: A novel method of fully 4D CTreconstruction was presented. The geometric accuracy of the estimated deformation was validated in phantom. A marked improvement in image quality was observed when applying the algorithm to image data from a real liverSBRT patient. The method allows reduction of X‐ray tube current during scanning while simultaneously improving motion estimates for use in dose calculation.