Dual-Energy CT Based Monochromatic Imaging

Dual-Energy CT Review Yu et al. Monochromatic Imaging Based on Dual-Energy CT Dual-Energy CT Review Lifeng Yu 1 Shuai Leng Cynthia H. McCollough Yu L, Leng S, McCollough CH Keywords: CT, dual-energy CT, virtual monochromatic imaging DOI:1.2214/AJR.12.9121 Received April 24, 212; accepted without revision April 3, 212. Publication of this supplement to the American Journal of Roentgenology is made possible by an unrestricted grant from Siemens Healthcare. 1 All authors: Department of Radiology, Mayo Clinic, 2 First St SW, Rochester, MN 5595. Address correspondence to C. H. McCollough (mccollough.cynthia@mayo.edu). AJR 212; 199:S9 S15 361 83X/12/1995 S9 American Roentgen Ray Society Dual-Energy CT Based Monochromatic Imaging OBJECTIVE. We summarize how virtual monochromatic images are synthesized from dual-energy CT using image-domain and projection-domain methods. The quality of virtual monochromatic images is compared with that of polychromatic single-energy images acquired at different tube potentials and the same radiation dose. Clinical applications of dual-energy CT based virtual monochromatic imaging are reviewed, including beam-hardening correction, contrast and noise optimization, metal artifact reduction, and material differentiation. CONCLUSION. Virtual monochromatic images synthesized from dual-energy CT data have the potential to reduce beam-hardening artifacts and to provide quantitative measurements. If there is no desire to obtain material-specific information or to correct for metal or beam-hardening artifacts from the dual-energy data, however, it is better to perform a conventional single-energy scan at the optimal tube potential. S ince the introduction of dual-source CT scanners and scanners capable of fast tube potential (kilovoltage) switching, dual-energy CT has been applied in many clinical areas, including automatic bone removal [1], stone composition characterization [2 4], virtual unenhanced imaging [5, 6], diagnosis of gout [7, 8], and assessment of myocardial blood supply [9]. All of these clinical applications require dual-energy processing to obtain materialspecific information [1, 11]. For example, the iodine and soft-tissue content of each voxel need to be quantified through a process known as basis material decomposition to generate iodine map images and virtual unenhanced images. In addition to the material-specific images, a single set of images is usually generated to be used for routine diagnosis, similar to conventional polychromatic single-energy CT images. These images use all the radiation dose in the dual-energy acquisition. A simple approach to generating these images is to perform a linear or nonlinear combination of the CT images reconstructed from the low- and high-energy data [12 16]. Another approach, which may improve the image quality of this single set of images, is to synthesize virtual monochromatic images from the dual-energy scans [1]. As early as the 197s, when the dual-energy CT technique was first discussed, it was wide- ly recognized that CT images obtained at different monochromatic x-ray beam energies could be synthesized from dual-energy scans. In principle, these images would be free of beam-hardening artifacts [1]. Virtual monochromatic images can be created in the projection (raw data) domain or in the image domain [17]. Reducing beam-hardening artifacts and providing more quantitatively accurate attenuation measurements are considered to be the main benefits of virtual monochromatic dualenergy CT images [1, 18]. The approach to synthesizing monochromatic images is usually based on basis material decomposition and the knowledge of mass attenuation of the basis materials. Depending on the CT system used to acquire the dual-energy data, the approach of basis material decomposition differs slightly. For dual-energy CT data acquired with a single-source fast-kilovoltage-switching technique, the basis material decomposition is performed in the projection domain. For dualenergy CT data acquired with dual-source CT systems, image domain basis material decomposition is typically used. This is primarily because the projection data from the low- and high-energy scans collected by the dual-source system in helical mode are not coincident with each other [19]. Iterative methods have been developed to enable projection-domain dual-energy processing of dual-source CT data, but have not yet been implemented in clinical practice [2]. AJR:199, November 212 S9

Yu et al. In this article, we summarize how virtual monochromatic images are synthesized from dual-energy CT, using either image-domain or projection-domain methods. We compare the iodine contrast, image noise, and iodine contrast-to-noise ratio (CNR) between virtual monochromatic images and polychromatic single-energy images acquired at different tube potentials and the same radiation dose. Finally, we will discuss clinical applications of virtual monochromatic imaging, including beam-hardening reduction, contrast and noise optimization, metal artifact reduction, and material differentiation. Synthesizing Virtual Monochromatic Images From Dual-Energy CT Projection-Based Method Figure 1 illustrates a projection-based method to synthesize virtual monochromatic images. In principle, virtual monochromatic CT images synthesized in the projection domain can fully eliminate beam-hardening artifacts. Figure 2 shows that typical beam-hardening correction methods in single-energy CT cannot completely remove the artifacts caused by use of polychromatic x-ray beams, whereas the monochromatic image generated from dual-energy CT does not have any beam-hardening artifacts. Projection domain basis material decomposition is briefly described below. In the diagnostic x-ray energy range, the linear attenuation coefficient of a voxel at energy E can be expressed as a linear combination of the mass attenuation coefficient of two basis materials [17]: µ µ µ(e) = ρ (E) ρ + 1 ρ (E) ρ 2 (1) where 1 µ ρ (E), i i = 1, 2 denote the mass attenuation coefficients of the two basis materials, and ρ i, i = 1, 2 denote the mass densities of the two basis materials. In the original work by Alvarez and Macovski [1], the attenuation coefficient was decomposed into Compton and photoelectric interactions. The basis material decomposition in equation 1 is fundamentally equivalent to the Compton and photoelectric decomposition because two basis materials with distinct atomic numbers can span the same linear space as that by the two types of interactions [17]. Two nonlinear equations are then solved to determine the mass density sinogram of each basis material. In practice, solution of the two 2 Fig. 1 Illustration of method used to synthesize virtual monochromatic images from dual-energy CT using projection-based method, which completely eliminates beam-hardening artifacts in this simulation. A and B, Sinogram images of bone (A) and water (B) content of simulated object after dual-energy basis material decomposition. C and D, Reconstructed images of bone (C) and water (D) sinograms provide map of mass density of each material. E G, Monochromatic images generated by synthesizing attenuation levels of bone and water at 2 kev (E), 4 kev (F), and 7 kev (G). nonlinear equations is obtained from calibration measurements [1, 21, 22]. The calibration establishes a relationship between the line integral of the mass density of each basis material and the measured data at low and high energies. This process inherently takes into account the effect of energy spectra and detector energy response. The mass density map for each basis material can then be reconstructed using typical reconstruction methods. From the knowledge of the mass density maps of the two basis materials and the mass attenuation coefficients of the two basis materials at each energy, monochromatic images can be synthesized using the linear combination in equation 1. This method is theoretically accurate for materials spanning a wide range of atomic numbers. Image-Based Method One can also create virtual monochromatic images based on reconstructed low- and highenergy images. The process is similar to the projection-based method, except that solving for the density maps of the two basis materials is performed using images reconstructed from low- and high-energy scans. Similar to equation 1, the linear attenuation coefficients derived from reconstructed images at low- and high-energy scans can be expressed as a linear combination of the mass attenuation coefficients of the two basis materials: k k µ µ µ k = ρ ρ + 1 ρ ρ 2, k = L, H (2). 1 Solving the two linear equations, one can obtain the mass density of the two basis materials, which can be used to calculate the monochromatic image at energy E by using equation 1. In practice, image-domain virtual monochromatic images also use a series of calibration measurements of CT numbers for different concentrations of the two basis materials in low- and high-energy scans [23]. To accommodate the effect of different patient sizes, this calibration measurement can be performed at 2 S1 AJR:199, November 212

multiple attenuation levels. It has been shown that the monochromatic image generated from image space data is simply a linear combination of the two CT images at low and high energies, with the weighting factor being a function of monochromatic energy [19]. Monochromatic Imaging Based on Dual-Energy CT Image Quality of Virtual Monochromatic Images To evaluate the image quality of virtual monochromatic images synthesized from dualenergy scans, two questions need to be answered. One question is how well the beam-hardening artifact can be corrected and how accurate the CT numbers can be. The other question is how the quality of virtual monochromatic images compares with that of conventional singleenergy CT images acquired with polychromatic x-ray beams and the same radiation dose. Beam-Hardening Artifact Correction and Quantitative Accuracy of CT Number The primary difference between projectionand image-based methods lies in how to determine the mass density of the two basis materials. The projection-based method solves for mass density integration at each projection first and then reconstructs the mass density images of the two basis materials; the image-based method directly solves for mass density using the low- and high-energy images that are already reconstructed. The two approaches are not equivalent because of the nonlinear processes involved in solving for density. In principle, projection-based methods should be more effective in terms of beam-hardening artifact correction because beam hardening occurs in each x-ray projection. Although beam-hardening correction techniques are usually applied before image reconstruction, these correction techniques are not always accurate, especially when iodine and other dense structures are present. Even with iterative beam-hardening correction methods, residual artifacts may still exist after correction when exact knowledge of the physical model (i.e., spectrum, detector, and imaging materials) is not available [24]. Therefore, virtual monochromatic images created in the image domain may still contain beam-hardening artifacts propagated from the low- and high-energy images, especially in the presence of dense bone or iodine. Despite the theoretic benefit of the projection-based method over the image-based method, existing studies have not confirmed this benefit in practical implementations. Goodsitt et al. [18] evaluated the accuracy of CT number and effective atomic number measured in virtual monochromatic images obtained with Fig. 2 Computer simulation of effect of beam-hardening effect from polychromatic CT (width, 1 HU; level, HU). A and B, Images show simulation at 12 kv before (A) and after (B) beam-hardening correction. Cupping artifacts in A were corrected in B, but dark banding artifacts originating from small dense lateral structures remain. C, Complete elimination of beam-hardening artifacts using virtual monochromatic images at 7 kev synthesized from dual-energy scans is theoretically possible. Fig. 3 Polychromatic and virtual monochromatic CT images of electron density phantom (Gammex 467, Gammex) that contains samples of various atomic numbers and densities. Images were acquired using scanner capable of fast kilovoltage switching (HD75, GE Healthcare). Virtual monochromatic images were reconstructed using projection-based technique. Significant streaking and shadowing can be seen in 8- kvp polychromatic and 4- and 7-keV virtual monochromatic images, but not in 1- and 12-keV virtual monochromatic images. Display window level and width were 5 and 5 HU, respectively. (Reprinted with permission from [18]) a projection-based fast-kilovoltage-switching dual-energy technique. Their results showed that although the computed effective atomic numbers were reasonably accurate (within 15%), CT number inaccuracies remained, especially for dense materials at low energies, and there was still a dependency on patient or phantom size. Beam-hardening artifacts still existed, particularly for images at lower monochromatic energies (Fig. 3). Their evaluation shows that the synthesized virtual monochromatic images generated from the current fastkilovoltage-switching dual-energy technique are not truly monochromatic, even though they are processed in the projection domain. Further research is required to fully realize the theoretic benefit of the projection-based method. Currently, it remains unclear which implementation, fast-kilovoltage-switching projectionbased or dual-source dual-energy image-based method, has an advantage over the other regarding beam-hardening artifact reduction and quantitative CT number accuracy. Image Quality Comparison Between Monochromatic and Polychromatic Images at the Same Radiation Dose In addition to the material-specific information provided by a dual-energy CT examination, interpreting physicians also need a single AJR:199, November 212 S11

Yu et al. Iodine Contrast (HU) Noise (HU) 4 6 8 1 12 6 5 4 3 2 1 4 5 3 25 2 15 1 5 4 5 Small (monochromatic) Medium (monochromatic) Large (monochromatic) X-Large (monochromatic) Polychromatic Energy (kv) 6 7 8 9 1 11 Monochromatic Energy (kev) Small (monochromatic) Medium (monochromatic) Large (monochromatic) X-Large (monochromatic) Small (polychromatic) Medium (polychromatic) Large (polychromatic) X-Large (polychromatic) 14 16 18 Polychromatic Energy (kv) 4 6 8 1 12 14 16 18 35 6 7 8 9 1 11 Monochromatic Energy (kev) Small (polychromatic) Medium (polychromatic) Large (polychromatic) X-Large (polychromatic) Fig. 4 Iodine contrast as function of monochromatic energy for four different phantom sizes. Four curves are very close to each other, indicating that beam hardening in larger phantoms minimally affects CT numbers in virtual monochromatic images. These data were acquired using dual-source scanner (Definition, Siemens Healthcare). Virtual monochromatic images were synthesized using image-based technique. CT numbers of iodine measured from different polychromatic single-energy scans show larger effect of beam hardening due to phantom size. Fig. 5 Noise as function of monochromatic energy for four different phantom sizes. For comparison, noise levels at different polychromatic beam energies are also displayed. Radiation doses, in terms of volume CT dose index, were matched for each phantom size. Radiation dose allocated to 8 kv in dual-source dual-energy scan was 5%. Data were acquired and reconstructed in same manner as in Figure 4. set of images for routine diagnosis. Virtual monochromatic images appear to be ideal for this purpose. An important question to be answered is how the image quality of virtual monochromatic images compares with that of conventional single-energy CT images acquired with polychromatic x-ray beams and the same radiation dose. If the image quality is the same or better, then dual-energy CT virtual monochromatic images can be used as the basis for routine diagnosis without requiring additional radiation dose. The image quality of virtual monochromatic images was studied in an early work by Alvarez and Seppi [25]. They found that a minimum noise exists, corresponding to a certain monochromatic energy, and similar noise levels could be achieved by monochromatic images or conventional polychromatic images for the same radiation dose. The image quality of virtual monochromatic images generated from the latest dual-energy CT techniques has also been evaluated by several investigators [22, 26 28]. Matsumoto et al. [27] evaluated the noise and iodine CNR in virtual monochromatic images acquired with a fast-kilovoltage-switching dual-energy technique. They compared monochromatic images to 12-kV single-energy images and evaluated one setting of dose partitioning between the low- and high-energy scans and one phantom size. They claimed that the virtual monochromatic image at approximately 7 kev had a better image quality and had the potential to replace 12 kv as the standard image modality. Because of the lack of a systematic evaluation of tube potentials other than 12 kv, this conclusion is not definitive [29]. On the same fast-kilovoltage-switching dualenergy CT platform, equivalent performance was also observed by Zhang et al. [26] between virtual monochromatic images acquired at 65 kev and images acquired with routine abdomen protocols at 12 kv for a matched volume CT dose index (CTDI vol ). Li et al. [22] compared the radiation dose in terms of CTDI vol between virtual monochromatic images and single-energy images when the low-contrast detectability was matched. They found that the CTDI vol in dual-energy CT was roughly 22% and 14% higher than that in routine single-energy CT for head and body examinations, respectively, to yield comparable lowcontrast detectability. In another recent work, Yu et al. [19] evaluated the image quality of virtual monochromatic images generated from dual-source dual-energy S12 AJR:199, November 212

Monochromatic Imaging Based on Dual-Energy CT Fig. 6 Iodine contrast-to-noise ratio (CNR) as function of monochromatic energy for four different phantom sizes. For comparison, iodine CNR at different polychromatic beam energies are also displayed. Note that radiation doses, in terms of volume CT dose index, were matched for each phantom size. Radiation dose allocated to 8 kv in dual-source dual-energy scan was 5%. Data were acquired and reconstructed in same manner as in Figure 4. Fig. 7 Reduction of pseudoenhancement using virtual monochromatic images acquired using fast-kilovoltage-switching approach (HD75, GE Healthcare) and created using projection-domain technique. Phantom containing vial of water was placed in iodine solution (2 mg/ml) and was scanned using conventional single-energy CT at 8 and 14 kv. Phantom was scanned again with dual-energy CT with fast-kilovoltage-switching technique, and virtual monochromatic images at 7 and 1 kev were synthesized. CT numbers in water were measured on each image. Significant pseudoenhancement was observed on 8- and 14-kV images, with CT numbers much higher than expected CT number of water ( HU). Virtual monochromatic images at both 7 and 1 kev showed negligible pseudoenhancement. (Reprinted with permission from [6]) scans and compared it with that from singleenergy CT scans acquired at various tube potentials but using the same radiation dose. The effects of several important technical factors, including patient size, dose partitioning, and image quality metric to be optimized, were considered in this evaluation. They found that monochromatic images provided a consistent iodine CT number across phantom sizes, whereas polychromatic single-energy images showed larger differences, primarily because of the beam-hardening effect (Fig. 4). With optimal monochromatic energies, the noise level was similar to that of single-energy scans at 12 kv for all phantom sizes included in this study (Fig. 5). With respect to iodine CNR, monochromatic images were similar to or better than single-energy images at 12 kv for the same radiation dose (Fig. 6). Therefore, if dual-energy CT is performed to obtain material-specific information, monochromatic images synthesized from the dual-energy scan can provide image quality equivalent to or better Attenuation Value of Water (HU) Iodine CNR 4 6 8 1 12 45 4 35 3 25 2 15 1 5 4 5 35 33.7 3 25 2 15 1 5 5 Small (monochromatic) Medium (monochromatic) Large (monochromatic) X-Large (monochromatic) than at 12 kv with no increase in radiation dose. However, comparing virtual monochromatic images from dual-energy CT to polychromatic single-energy CT at other tube potentials, the conclusions regarding optimal Polychromatic Energy (kv) 6 7 8 9 1 11 Monochromatic Energy (kev) 2.9 Single-energy CT 8 kvp 14 kvp Small (polychromatic) Medium (polychromatic) Large (polychromatic) X-Large (polychromatic) Iodine solution 14 16 18 Water.6 1.9 Dual-energy CT 7 kev 1 kev Fig. 8 Metal artifact reduction using virtual monochromatic images synthesized from dual-source dualenergy CT data using image-space techniques. A, Image shows pedicle screw in water phantom, acquired with single-energy scan at 12 kv. B, Monochromatic image at 127 kev was generated from dual-energy scan with same scanner output (volume CT dose index) as used in single-energy scan. Streaking caused by metal was almost completely eliminated. settings are dependent on patient size. Iodine CNR in virtual monochromatic images from dual-energy CT was lower than that for singleenergy 8-kV images for small, medium, and large phantom sizes. Therefore, if there is no AJR:199, November 212 S13

Yu et al. Fig. 9 Metal artifact reduction using virtual monochromatic images synthesized from dual-source dualenergy CT is ineffective for dense metal objects. A, Image of dense metal implant in water phantom acquired with single-energy scan at 12 kv. B, Monochromatic image at 127 kev generated from dual-energy scan with same scanner output (volume CT dose index) as used in single-energy scan. Streaking by metal becomes worse on virtual monochromatic image. Fig. 1 Use of spectral attenuation curve to determine simple and hemorrhagic renal cyst. A, Virtual monochromatic image synthesized from fast-kilovoltage-switching dual-energy CT. Arrows point to hyperattenuating left renal lesion (green), hypoattenuating right renal lesion (light blue), gallbladder fluid (dark blue), and renal parenchymal enhanced by iodinated contrast material (red). Left renal lesion (green) is indeterminate and could be either enhancing mass or hyperdense cyst. B, Spectral attenuation curves measured from virtual monochromatic image for left renal lesion (green), right renal lesion (light blue), gallbladder bile (dark blue), and renal parenchyma (red) at various photon energies (kiloelectron volt, kev). Attenuation curves of both renal lesions are similar to that of gallbladder fluid, which remains relatively flat, indicating nonenhancing tissues. In contrast, attenuation curve of renal parenchyma shows sharp increase at low energies, which is characteristic of iodine-containing materials. Therefore, spectral attenuation curves obtained from virtual monochromatic images helped to determine that left hyperattenuating renal lesion is hemorrhagic or proteinaceous cyst. (Reprinted with permission from [6]) desire to obtain material-specific information from the dual-energy scan, it is better to simply perform a conventional single-energy scan at the optimal tube potential, which for a given patient size and diagnostic task, can reduce radiation dose or improve image quality [3, 31]. Approaches for automatic selection of the most dose-efficient tube potential have been developed and implemented in practice [28]. Clinical Applications of Virtual Monochromatic Imaging Beam-Hardening Correction and Optimal Contrast, Noise, and CNR Virtual monochromatic imaging has the potential to reduce the beam-hardening artifact and provide more quantitative attenuation information. One clinical application is to reduce the pseudoenhancement artifact observed in contrast-enhanced scans of renal cysts [6] (Fig. 7). Although this is promising, more research is necessary to improve the quantitative accuracy and beam-hardening correction [18]. In dual-energy CT, creation of virtual monochromatic images at different energies may allow users to freely select the optimal energy that yields the best contrast, noise, CNR, or lesion detectability. Although this feature is appealing when the dual-energy examination is performed for obtaining material-specific information, it is important to recognize that the optimized quality in monochromatic images in terms of contrast, noise, and CNR is not adequate justification for performing dualenergy CT. Although dual-energy CT can be performed without dose penalty compared with the commonly used 12 kv tube potential, the image quality of virtual monochromatic images is inferior to the quality of images obtained at lower tube potentials, especially for small and medium patients. Therefore, if there is no intention to perform material classification or quantification from the dual-energy scan, it is better to deliver all the radiation dose in a single-energy scan at an optimal tube potential [19]. Despite the inferior dose efficiency of dualenergy CT compared with single-energy CT at an optimal tube potential for traditional CT diagnostic tasks, if the advantage of beamhardening correction using virtual monochromatic imaging can be clearly shown in some diagnostic tasks, a dual-energy scan can still be justified. Metal Artifact Reduction Virtual monochromatic images at high energies have been found to be able to reduce artifacts caused by metal implants on dualsource dual-energy CT [23, 32, 33]. Optimal monochromatic energies vary between 95 and 15 kev, depending on the composition and size of the metal implant [32]. Figure 8 compares two images of a pedicle screw acquired by single-energy CT at 12 kv and monochromatic images at 127 kev synthesized from a dual-source dual-energy scan for the same radiation dose. Streaking artifacts caused by the metal implant were almost completely eliminated on the monochromatic image. However, for very dense metal implants, the correction is not effective (Fig. 9). This is mainly because the metal artifacts in the presence of dense metal are caused by factors in addition to beam hardening, such as photon starvation and nonlinear partial volume averaging, which cannot be corrected by synthesizing high-energy monochromatic images. Additional iterative correction has been applied to monochromatic images to further reduce metal artifacts [6, 34, 35]. The Use of Attenuation Values From Virtual Monochromatic Images to Characterize Materials By use of virtual monochromatic images at different energies, a spectral attenuation curve as a function of energy can be plotted. Different chemical compositions may be differentiated according to the characteristics of the spectral attenuation curve. For example, iodineenhancing solid masses may be differentiated from hyperattenuating renal cysts on the basis of the attenuation curves, because iodine attenuation greatly increases at lower energies, S14 AJR:199, November 212

Monochromatic Imaging Based on Dual-Energy CT whereas cyst attenuation remains relatively flat at all energies [6] (Fig. 1). This is a qualitative way to differentiate materials. A more quantitative approach to characterizing materials is to evaluate the effective atomic number, dual-energy index, or the density map of a specific basis material calculated from dualenergy scans [2, 5, 36]. Conclusion Virtual monochromatic images synthesized from dual-energy CT data have the potential to reduce beam-hardening artifacts and to provide quantitative measurements. Optimal monochromatic energies may be selected according to patient size, the partitioning of radiation dose between the low- and high-energy scans, and the image quality metric to be optimized. If there is no desire to obtain materialspecific information or to correct for metal or beam-hardening artifacts from the dual-energy data, however, it is better to perform a conventional single-energy scan at the optimal tube potential. References 1. Johnson TR, Krauss B, Sedlmair M, et al. Material differentiation by dual energy CT: initial experience. Eur Radiol 27; 17:151 1517 2. Primak AN, Fletcher JG, Vrtiska TJ, et al. Noninvasive differentiation of uric acid versus nonuric acid kidney stones using dual-energy CT. Acad Radiol 27; 14:1441 1447 3. Graser A, Johnson TR, Bader M, et al. Dual energy CT characterization of urinary calculi: initial in vitro and clinical experience. Invest Radiol 28; 43:112 119 4. Boll DT, Patil NA, Paulson EK, et al. Renal stone assessment with dual-energy multidetector CT and advanced postprocessing techniques: improved characterization of renal stone composition pilot study. Radiology 29; 25:813 82 5. Takahashi N, Hartman RP, Vrtiska TJ, et al. Dualenergy CT iodine-subtraction virtual unenhanced technique to detect urinary stones in an iodinefilled collecting system: a phantom study. AJR 28; 19:1169 1173 6. Silva AC, Morse BG, Hara AK, Paden RG, Hongo N, Pavlicek W. Dual-energy (spectral) CT: applications in abdominal imaging. RadioGraphics 211; 31:131 146 7. Nicolaou S, Yong-Hing CJ, Galea-Soler S, Hou DJ, Louis L, Munk P. Dual-energy CT as a potential new diagnostic tool in the management of gout in the acute setting. AJR 21; 194:172 178 8. Glazebrook KN, Guimaraes LS, Murthy NS, et al. Identification of intraarticular and periarticular uric acid crystals with dual-energy CT: initial evaluation. Radiology 211; 261:516 524 9. Ruzsics B, Schwarz F, Schoepf UJ, et al. Comparison of dual-energy computed tomography of the heart with single photon emission computed tomography for assessment of coronary artery stenosis and of the myocardial blood supply. Am J Cardiol 29; 14:318 326 1. Alvarez RE, Macovski A. Energy-selective reconstructions in x-ray computed tomography. Phys Med Biol 1976; 21:733 744 11. Kalender WA, Perman WH, Vetter JR, Klotz E. Evaluation of a prototype dual-energy computed tomographic apparatus. Part I. Phantom studies. Med Phys 1986; 13:334 339 12. Fletcher JG, Takahashi N, Hartman R, et al. Dualenergy and dual-source CT: is there a role in the abdomen and pelvis? Radiol Clin North Am 29; 47:41 57 13. Holmes DR 3rd, Fletcher JG, Apel A, et al. Evaluation of non-linear blending in dual-energy computed tomography. Eur J Radiol 28; 68:49 413 14. Yu L, Primak AN, Liu X, McCollough CH. Image quality optimization and evaluation of linearly mixed images in dual-source, dual-energy CT. Med Phys 29; 36:119 124 15. Behrendt FF, Schmidt B, Plumhans C, et al. Image fusion in dual energy computed tomography effect on contrast enhancement, signal-to-noise ratio and image quality in computed tomography angiography. Invest Radiol 29; 44:1 6 16. Kim KS, Lee JM, Kim SH, et al. Image fusion in dual energy computed tomography for detection of hypervascular liver hepatocellular carcinoma phantom and preliminary studies. Invest Radiol 21; 45:149 157 17. Lehmann LA, Alvarez RE, Macovski A, et al. Generalized image combinations in dual KVP digital radiography. Med Phys 1981; 8:659 667 18. Goodsitt MM, Christodoulou EG, Larson SC. Accuracies of the synthesized monochromatic CT numbers and effective atomic numbers obtained with a rapid kvp switching dual energy CT scanner. Med Phys 211; 38:2222 2232 19. Yu L, Christner JA, Leng S, Wang J, Fletcher JG, McCollough CH. Virtual monochromatic imaging in dual-source dual-energy CT: radiation dose and image quality. Med Phys 211; 38:6371 6379 2. Maass C, Meyer E, Kachelriess M. Exact dual energy material decomposition from inconsistent rays (MDIR). Med Phys 211; 38:691 7 21. Chuang KS, Huang HK. A fast dual-energy computational method using isotransmission lines and table lookup. Med Phys 1987; 14:186 192 22. Li B, Yadava G, Hsieh J. Quantification of head and body CTDI(VOL) of dual-energy x-ray CT with fast-kvp switching. Med Phys 211; 38:2595 261 23. Raupach R, Becker CR, Johnson TR, Krauss B, Flohr T, Schmidt B. Metal artifact reduction for dual source dual energy computed tomography In: Hricak H, ed. 96th Scientific Assembly and Meeting of the Radiological Society of North America. Oak Brook, IL: Radiological Society of North America, 21 24. Kyriakou Y, Meyer E, Prell D, Kachelriess M. Empirical beam hardening correction (EBHC) for CT. Med Phys 21; 37:5179 5187 25. Alvarez RE, Seppi E. A comparison of noise and dose in conventional and energy selective computed tomography. IEEE Trans Nucl Sci 1979; 26:2853 2856 26. Zhang D, Li XH, Liu B. Objective characterization of GE Discovery CT75 HD scanner: gemstone spectral imaging mode. Med Phys 211; 38:1178 1188 27. Matsumoto K, Jinzaki M, Tanami Y, Ueno A, Yamada M, Kuribayashi S. Virtual monochromatic spectral imaging with fast kilovoltage switching: improved image quality as compared with that obtained with conventional 12-kVp CT. Radiology 211; 259:257 262 28. Yu L, Li H, Fletcher JG, McCollough CH. Automatic selection of tube potential for radiation dose reduction in CT: a general strategy. Med Phys 21; 37:234 243 29. Venema HW. Virtual monochromatic spectral imaging with fast kilovoltage switching should not be used as standard CT imaging modality. Radiology 211; 26:916 917; author reply, 917 3. Kalender WA, Deak P, Kellermeier M, van Straten M, Vollmar SV. Application- and patient size-dependent optimization of x-ray spectra for CT. Med Phys 29; 36:993 17 31. Guimarães LS, Fletcher JG, Harmsen WS, et al. Appropriate patient selection at abdominal dualenergy CT using 8 kv: relationship between patient size, image noise, and image quality. Radiology 21; 257:732 742 32. Bamberg F, Dierks A, Nikolaou K, Reiser MF, Becker CR, Johnson TR. Metal artifact reduction by dual energy computed tomography using monoenergetic extrapolation. Eur Radiol 211; 21:1424 1429 33. Zhou C, Zhao YE, Luo S, et al. Monoenergetic imaging of dual-energy CT reduces artifacts from implanted metal orthopedic devices in patients with factures. Acad Radiol 211; 18:1252 1257 34. Lee YH, Park KK, Song HT, Kim S, Suh JS. Metal artefact reduction in gemstone spectral imaging dual-energy CT with and without metal artefact reduction software. Eur Radiol 212; 22:1331 134 35. Brook OR, Gourtsoyianni S, Brook A, Mahadevan A, Wilcox C, Raptopoulos V. Spectral CT with metal artifacts reduction software for improvement of tumor visibility in the vicinity of gold fiducial markers. Radiology 212 [Epub ahead of print] 36. Joshi M, Aluri S, Procknow K, et al. Effective atomic number accuracy for kidney stone characterization using spectral CT. Proc SPIE 21; 7622:76223K AJR:199, November 212 S15