K-edge subtraction imaging for iodine and calcium separation at a compact synchrotron x-ray source

Abstract

Purpose: About one third of all deaths worldwide can be traced to some form of cardiovascular disease. The gold standard for the diagnosis and interventional treatment of blood vessels is digital subtraction angiography (DSA). An alternative to DSA is K-edge subtraction (KES) imaging, which has been shown to be advantageous for moving organs and for eliminating image artifacts caused by patient movement. As highly brilliant, monochromatic x-rays are required for this method, it has been limited to synchrotron facilities so far, restraining the applicability in the clinical routine. Over the past decades, compact synchrotron x-ray sources based on inverse Compton scattering have been evolving; these provide x-rays with sufficient brilliance and meet spatial and financial requirements for laboratory settings or university hospitals.

Approach: We demonstrate a proof-of-principle KES imaging experiment using the Munich Compact Light Source (MuCLS), the first user-dedicated installation of a compact synchrotron x-ray source worldwide. A series of experiments were performed both on a phantom and an excised human carotid to demonstrate the ability of the proposed KES technique to separate the iodine contrast agent and calcifications.

Results: It is shown that the proposed filter-based KES method allows for the iodine-contrast agent and calcium to be clearly separated, thereby providing x-ray images only showing one of the two materials.

Conclusions: The results show that the quasimonochromatic spectrum of the MuCLS enables filter-based KES imaging and can become an important tool in preclinical research and possible future clinical diagnostics.

1. Introduction

Digital subtraction angiography (DSA) is one of the most important techniques in the diagnosis of cardiovascular disease.^1–3 It is effective in enhancing contrast between vascular structures and surrounding bones and tissue structures. However, motion artifacts limit the application of this method.⁴

Cardiovascular disease is the leading cause of death globally and accounts for 17.3 million deaths per year.⁵ The primary cause for cardiovascular diseases is atherosclerosis,⁶ which causes structural degradation of arterial blood vessels. The accumulation of white blood cells and fat, cell proliferation, and inflammatory responses cause stiffening and thickening due to the formation of plaque on the vessel walls.⁷ The plaque can also contain calcifications that build up at the base of older lesions of the vessel.⁶ The formation of plaque is a process of several years or even decades and can remain asymptomatic.⁸ Often, atherosclerosis is only diagnosed after a cardiac infarction or a stroke. It is known that atherosclerotic processes tend to occur more often at branchings, curvatures, and constrictions of arteries⁶ because the blood flow is laminar and turbulence occurs at the outer wall of the vessel. This turbulence leads to a reduction of shear stress acting onto the arterial wall, which further leads to a disorder of the endothelial cells, making the arterial wall more permeable for other cells. Accumulation of cholesterol in the arterial wall and formation of plaque that can eventually rupture and cause a total closure of the vessel are the long-term consequences.⁷

In conventional x-ray imaging, the image contrast arises from variations in absorption of different materials or tissues. The absorption is dependent on the elemental composition and density of the material, leading to a strong contrast between bone and tissue structures. However, the elemental composition of soft tissues is very similar, leading to a weak attenuation contrast.⁹ Therefore, blood vessels are usually imaged with subtraction x-ray imaging in which structures or organs are visualized using a contrast medium that changes the attenuation between the contrasted structure and the surrounding tissue. In K-edge subtraction (KES) imaging, which was first proposed by Jacobson in 1953,¹⁰ two x-ray images are taken at mean energies below and above the K-edge of the contrast medium. The subtraction of the two leads to an image showing only the contrasted structure while other anatomical structures or bones are eliminated. This can solve issues of the attenuation of a tissue being enhanced by a contrast agent such that it becomes inseparable from another structure, as is commonly the case with an iodine-based contrast agent and calcium in calcifications. KES imaging relies on a monochromatic x-ray beam, and has therefore been performed at synchrotron sources in the past. It has been shown that intravenous injection of a contrast agent provides good image quality for coronary angiography.¹¹ Also KES imaging with monochromatic x-rays improves the image quality of the neurovascular system in comparison with dual-energy imaging with conventional x-ray spectra.¹² This is particularly interesting for applications in which there is the risk of blocking small vessels by a catheter used for injection. KES can also be performed by changing the energy of the x-ray spectrum using an x-ray filter^13,14 that absorbs part of the spectrum and therefore changes its mean energy.

Although KES has been successfully performed at synchrotrons in the past, the integration of synchrotrons into the clinical routine is difficult.¹⁵ Although synchrotrons provide highly brilliant, monochromatic x-rays, they not only are expensive both in installation and maintenance but also have large spatial requirements. Additionally, the beam size provided by a synchrotron is far below the beam size required for clinical imaging. Compact synchrotron sources have been developed over the past decades to provide high-brilliance x-ray beams with reduced financial and spatial requirements.^16,17 The lower investment costs of these compact x-ray sources enable the transfer of techniques that have been limited to synchrotrons so far, like KES-imaging, into a laboratory environment. One of these sources is the Munich Compact Light Source (MuCLS). It has been shown previously that coronary angiography at the MuCLS results in higher contrast-to-noise ratio between contrast enhanced blood vessels and surrounding tissue than at a conventional x-ray source because the spectrum at the MuCLS can be tuned to lie directly at the K-edge of the contrast agent.¹⁸

Here we present a proof-of-principle of KES at the MuCLS demonstrating the differentiation of the iodine-based contrast agent and calcium in the calcified plaque of a carotid vessel. It is shown that filter-based KES overcomes the clinically faced problem of indistinguishability of iodine and calcium.

2. Materials and Methods

2.1. Filter-Based K-Edge Subtraction

In KES imaging, images are acquired using x-ray energies directly below and above the K-edge of a contrast agent injected into the desired structures. The intensities of the x-ray beams at the low energy $I_{\text{low}}$ and at the high energy $I_{\text{high}}$ behind the imaged tissue can be described with

$$I_{\text{low}}=I_{0,\text{low}}{e}^{-[{\mu }_{T,\text{low}}(D-G)+{\mu }_{c,\text{low}}G]}$$ (1)

and

$$I_{\text{high}}=I_{0,\text{high}}{e}^{-[{\mu }_{T,\text{high}}(D-G)+{\mu }_{c,\text{high}}G]},$$ (2)

where ${\mu }_{T,\text{low}}$ and ${\mu }_{T,\mathrm{high}}$ are the mean attenuation coefficients of the surrounding tissue with thickness $D$ at low and high energy, respectively, and ${\mu }_{c,\text{low}}$ and ${\mu }_{c,\mathrm{high}}$ are the attenuation coefficients of the contrast agent in the stained structure with thickness $G$ at low and high energy, respectively. These radiographs are then subtracted logarithmically after flatfield correction:

$$\begin{gathered} \ln \frac{I_{\mathrm{high}}}{I_{0,\mathrm{high}}}-\ln \frac{I_{\text{low}}}{I_{0,\text{low}}}=\ln \frac{I_{0,\mathrm{high}}}{I_{0,\mathrm{high}}}-{\mu }_{T,\mathrm{high}}(D-G)-{\mu }_{c,\mathrm{high}}G-\ln \frac{I_{0,\text{low}}}{I_{0,\text{low}}}+{\mu }_{T,\text{low}}(D-G)+{\mu }_{c,\text{low}}G \\ \approx ({\mu }_{c,\text{low}}-{\mu }_{c,\mathrm{high}})G \end{gathered}$$ (3)

as the attenuation coefficient of most tissues hardly changes over the range of a few keV [${\mu }_{T,\text{low}}(D-G)-{\mu }_{T,\mathrm{high}}(D-G)\approx 0$]. The contrast in the image is, therefore, only dependent on the change in attenuation of the contrast agent, which has an absorption discontinuity between the two energies.

For KES imaging, one needs an x-ray source with the possibility of a fast change of energy and desirably a narrow energy bandwidth. The MuCLS has this desired almost monochromatic, narrow spectrum, but the tuning of the x-ray energy takes a considerable amount of time. Therefore, a filter-based approach that employs a filter containing the element also used in the contrast agent (in this case iodine) is implemented. This filter absorbs most of the spectrum above the iodine K-edge, thereby reducing the mean energy of the incident spectrum. From the radiographs taken with and without the iodine filter, it is possible to calculate images with energies below and above the iodine K-edge and obtain a difference image identical to one taken with two different energies without any filtering.^13,14 The work flow of calculation can be seen in Fig. 1. The weighting factors used to calculate the images were determined using x-ray spectra acquired with and without iodine filtering.

Fig. 1.

Scheme for calculating a KES image at the MuCLS. The unfiltered and the iodine filtered images are dark current, flatfield, and flux corrected. The iodine filtered image is weighted to correct for the absorption of the part of the spectrum below the K-edge in the filter. After the weighting, the low-energy part of the weighted iodine image (d) has the same intensity as the unfiltered image (a). The unfiltered imaged is weighted in such a way that it corresponds to the remaining high-energy part of the weighted iodine filtered image. The high-energy image (e) is calculated by subtracting the unfiltered (a) and weighted iodine images (d). By subtracting the weighted iodine (d) and weighted unfiltered (c), the low-energy image (f) is obtained. The KES image is obtained by logarithmically subtracting the two.

First, the unfiltered and the iodine filtered images are dark current, flatfield, and flux corrected. To obtain two images only containing the high- or low-energy part of the spectrum, respectively, with precise separation at the K-edge, the high-energy part in the iodine filtered image and the low-energy part in the unfiltered image are removed. For the calculation of the high-energy image, the iodine filtered image [Fig. 1(b)] is weighted to correct for the absorption of the part of the spectrum below the K-edge in the filter. After the weighting, the low-energy part of the weighted iodine image [Fig. 1(d)] has the same intensity as the unfiltered image [Fig. 1(a)]. By then subtracting the weighted iodine filtered image (d) from the unfiltered image (a), the high-energy image [Fig. 1(e)] is obtained as the low-energy parts of the spectra are identical and only the high-energy part of the spectrum remains. To obtain the low-energy image, the high-energy part of the weighted iodine filtered image (d) is eliminated. Therefore, the unfiltered image (a) is weighted in such a way that it corresponds to the remaining high-energy part of the weighted iodine filtered image. By then subtracting the weighted unfiltered image [Fig. 1(c)] from the weighted iodine filtered image (d), one obtains a low-energy image [Fig. 1(f)] where the high-energy part of the spectrum is completely canceled out. For more information on the work flow and the equations for the calculations, see our previous work.¹⁹ Finally, the subtracted image is obtained by logarithmically subtracting the high- and low-energy image. Using an empirical energy correction factor, the high- and the low-energy image can also be subtracted such that the calcium is highlighted while the iodine is removed from the resulting image.

2.2. Samples

When examining arterial vessels, an iodine contrast agent is injected into the blood to visualize them more clearly. Yet, iodine absorbs x-ray radiation similarly to calcium in plaques as their attenuation coefficients are very similar when using the iodine concentrations common for the clinical routine (compare Fig. 2). This makes the discrimination between the two very difficult. When imaging around the K-edge of iodine, the attenuation coefficient $\mu /\rho$ of iodine will change clearly, whereas the attenuation coefficient of calcium will remain almost unchanged. By subtracting the two images, the jump in the attenuation coefficient of iodine will become visible and make the discrimination of iodine and calcium possible.

Fig. 2.

The total attenuation coefficients of calcium and iodine are similar at clinical concentrations. At 33.17 keV, the iodine K-edge can be seen as a jump in the attenuation coefficient of iodine. At this energy, the calcium attenuation coefficient changes only slightly. By doing KES imaging around the iodine K-edge, the two substances can be distinguished.

For a proof of principle, a phantom was constructed with three microcentrifuge tubes [Fig. 3(a)], filled with a sodium iodide solution with an iodine concentration of $12\mathrm{mg}/\mathrm{ml}$, a $75\mathrm{mg}/\mathrm{ml}$ calcium chloride solution, and as a reference, water. These were fixed in a plastic container to ensure a stable position during the measurements. The concentrations of the solutions were chosen in such a way that the attenuation $\mu d$ of the iodine and the calcium solution are identical, to simulate a case in which the two materials are inseparable in normal x-ray imaging. Additionally, a human carotid artery with calcification and a sodium iodide solution with an iodine concentration of $50\mathrm{mg}/\mathrm{ml}$ inside a microcentrifuge tube was imaged [Fig. 3(b)]. Unfortunately, the human carotid sample we used was not allowed to come into contact with the iodine contrast agent directly due to the requirements of the histological investigations. However, the separation of the contrast agent and calcification is also achievable when there is contact of the two materials (similar to the iodine contrast agent and bone in our previous work on coronary angiography¹⁹). The project was approved by the local ethics committee and complies with the Declaration of Helsinki of 1975, as revised in 2008. Written informed consent to provide tissue for research purposes had been obtained from the donor or the donor’s relatives.

Fig. 3.

Samples used for iodine/calcium separation and the experimental setup at the MuCLS. (a) Falcon tubes filled with $12\mathrm{mg}/\mathrm{ml}$ iodine solution, $75\mathrm{mg}/\mathrm{ml}$ calcium chloride solution, and as a reference, water. (b) Human carotid with calcifications and falcon tube filled with sodium iodide solution with $50\mathrm{mg}/\mathrm{ml}$ iodine. (c) Experimental setup at the MuCLS with the compact synchrotron x-ray source on the left emitting x-rays to the iodine filter and the sample. At 16 m from the interaction points, the x-rays are detected in the detector.

The experimental setup can be seen in Fig. 3(c). All samples were positioned at a source-to-sample distance of 15.6 m. The images were acquired with a flat panel detector (Dexela 1512, PerkinElmer, Inc.) placed at a distance of 16.4 m from the source point. The detector was equipped with a ${\mathrm{Gd}}_2{\mathrm{O}}_2\mathrm{S}$ scintillator and had a pixel size of $74.8\times 74.8\mu {\mathrm{m}}^2$. The effective pixel size of the acquired images was $70\times 70\mu {\mathrm{m}}^2$. Flatfield images as well as images of the samples were taken with and without iodine filter. Afterward, the images were darkfield and flatfield corrected. The subtraction was done according to the scheme presented in Fig. 1. Two images of each sample were taken: the first one with the native spectrum and the second one filtered with an iodine filter. The iodine filter was made of an acrylic glass container filled with an iodine solution exhibiting a thickness of 2 mm along the beam direction and an iodine content of $400\mathrm{mg}/\mathrm{ml}$ (IMERON 400 MCT, Bracco Imaging, Germany), amounting to an effective iodine thickness of around $150\mu \mathrm{m}$. The iodine filter was constructed such that a large part of the spectrum above the K-edge of iodine was absorbed (3% of the original intensity remaining) while maintaining a large part of the low-energy part of the spectrum (48% of the original intensity remaining). The switching of the filter was done manually.

2.3. Munich Compact Light Source

The x-ray beam at the MuCLS is produced by a compact synchrotron source that was developed and manufactured by Lyncean Technologies Inc. It is based on the process of inverse Compton scattering in which relativistic electrons collide with infrared laser photons, thereby providing a tunable, quasimonochromatic x-ray beam.^16,20 Assuming head-on collision of electrons and laser photons and backscattering of the produced x-ray photons, the energy of an x-ray photon is calculated with

$$E_x\approx 4{\gamma }^2{E}_L,$$ (4)

where $\gamma =E_e/(mc{)}^2$ is the ratio of electron energy to electron rest energy and $E_L$ is the laser photon energy.²¹ By adjusting the electron energy, the x-ray beam is tunable between 15 and 35 keV. To reach a high x-ray flux, the relativistic electrons circulate in a storage ring while the laser photons are amplified in a high finesse optical cavity. The produced x-rays are emitted into a solid angle of about 4 mrad. At the time of the measurements, the source produced a flux of up to $1.2\times {10}^{10}$ photons per second with a source size of $\sim 50\times 50\mu {\mathrm{m}}^2$. However, for the particular experiments presented here, the x-ray flux was $\sim 0.3\times {10}^{10}$ photons per second, leading to longer acquisition times. Since the time of the experiments, the x-ray source has been upgraded to produce an x-ray flux of up to $3.5\times {10}^{10}$ photons per second (at 35 keV).²²

The experiments were performed at the far end station of a dedicated imaging beam line, which was developed and installed by the Technical University of Munich and features two end stations. For all experiments, the sample was placed into the far end station at a source-to-sample distance of about 15.5 m and the detector was located 16 m from the interaction point, where the beam has an elliptic extent of $62\times 74{\mathrm{mm}}^2$. The photon rate at the detector position was $4.5\times {10}^5\mathrm{photons}/(\mathrm{s}{\mathrm{mm}}^2)$. The iodine filter was placed in front of the sample.

3. Results

The projection images of the iodine–calcium phantom can be seen in Fig. 4. In the unfiltered reference image (a), one can see that the tubes with iodine and calcium give the same absorption signal and cannot be distinguished. Their attenuation values $\mu d$ were found to be identical with a difference of 0.3%. In the iodine filtered image (b), a slight difference in attenuation can be seen. This is the case because the high-energy part of the spectrum, for which the absorption in the iodine solution is particularly high, is absorbed by the filter. This leads to a decrease in the mean absorption of the iodine solution. When calculating the high- and low-energy images and subtracting the two from each other, only the iodine tube remains visible in the resulting KES image (c) as the absorption of iodine changes drastically around the K-edge while the absorption of all other substances hardly changes.

Fig. 4.

Projection images of the iodine–calcium phantom: (a) (unfiltered) reference x-ray image (0.5-s exposure time); (b) x-ray image filtered with iodine filter (2-s exposure time); (c) KES image of (a) and (b) calculated according to the scheme shown in Fig. 1; and (d) KES image of (a) and (b) with an additional energy correction, so the calcium and not the iodine is visible.

Yet, this is only possible because the absorption around the iodine K-edge has a large influence on the resulting overall attenuation of the x-rays by the iodine solution as the spectrum of the MuCLS is comparably narrow. When imaging with a broad x-ray spectrum, the iodine K-edge would have less effect on the spectrum as a whole and the effective attenuation value $\mu d$ of iodine would change less. The difference of the attenuation values of iodine and calcium between the reference and iodine filtered images would not be that large, and thus the iodine would not be as clearly visible in the KES image. To visualize the calcium solution without the iodine, the low-energy image was multiplied with an empirical energy correction coefficient, calculated by the ratio of the mean attenuation of the iodine solution in the high-and low-energy images. With this correction, it is possible to subtract the images in such a way that the iodine vanishes in the difference image [see Fig. 4(d)]. This allows for the generation of two KES images showing only one of the two materials, iodine or calcium, respectively.

Next, the same measurements were performed on a biological sample. A carotid with an iodine filled microcentrifuge tube was measured with the same procedure as the phantom. In Fig. 5, the acquired images are shown. In the unfiltered reference image [Fig. 5(a)], the calcifications of the carotid can be seen in the right branch and on the left side of the vessel wall at the branching of the two blood vessels. The absorption of the iodine solution is identical to the absorption in the calcification at the branching. Because in a clinical setting the iodine contrast agent would be flowing through the carotid, the two materials would not be easily separable. In the iodine filtered image [Fig. 5(b)], the absorption of the iodine solution is reduced and the calcifications are still clearly visible. In the KES image [Fig. 5(c)], calculated from (a) and (b), only the iodine solution remains visible. The calcifications and the iodine can therefore be distinguished through KES imaging. With an additional energy correction, the subtraction can be done in such a way that only the calcifications remain visible in the image [Fig. 5(d)]. However, the identification of them is more difficult than in the case of the phantom due to the tissue structures and the high noise level. Because KES is a method that separates the contrast agent from surrounding materials, the vessel wall and the calcification cannot be separated if both are visible in the conventional unfiltered x-ray image. In other words, KES allows for the unambiguously identification of which of the highly absorbing structures in the unfiltered image are due to the iodine contrast agent. By elimination, this allows for the association of the remaining high-contrast features with calcifications. In its natural state, the vessel would be embedded inside other tissue and filled with blood, so x-ray contrast of the vessel walls would not be as prominent as in the ex situ images here and highly absorbing features would either be calcifications or iodine. Thus, after KES, calcifications could be more readily identified. This natural state could be better simulated by putting the sample in a water bath. Then there would be almost no contrast between water and vessel walls, so in the energy corrected KES image [Fig. 5(d)], only the calcification would be visible. Therefore, in a more realistic setting, the KES image [Fig. 5(c)] and energy corrected KES image [Fig. 5(d)] would only show iodine and calcifications, respectively.

Fig. 5.

Projection images of the human carotid: (a) reference x-ray image, exposure time: 0.5 s; (b) iodine-filtered x-ray image, exposure time: 2 s; (c) KES image of (a) and (b); and (d) KES of (a) and (b) with an additional energy correction, so the calcifications and not the iodine is visible. In this sample, it is more difficult to identify the calcifications than in the phantom.

4. Discussion

Although the separation of iodine from calcifications in blood vessels is a big issue in clinical diagnostics, it could be shown in a proof of principle experiment that the separation in projection images is possible when using an x-ray source with a narrow spectrum, as provided at the MuCLS, and the method of KES. Although the attenuation values $\mu d$ of iodine and calcium solutions were identical in the unfiltered reference image, only the iodine solution remained visible in the KES image. With further correction during the subtraction, it was also possible to highlight the calcium in the image and eliminate the iodine. With these experiments, it was shown that using KES imaging at a compact synchrotron source enables the differentiation of the iodine contrast agent from calcium. This allows for identification of calcifications in blood vessels via radiography.

Clinical DSA is prone to artifacts arising from the movement of the patient between the masking image, the injection of the contrast agent, and the second x-ray image. Correcting these artifacts can be very cumbersome^1,4 but can be avoided when reducing the time delay between the two acquired images. In filter-based KES imaging, the whole data acquisition can be performed after the injection of the contrast agent as demonstrated here. This makes the technique suitable for general angiography of only slowly moving body parts. In case of moving organs, such as the heart, a solution for the rapid movement during a heart cycle would have to be found. As the narrow spectrum of the MuCLS can be directly tuned to the K-edge of the contrast agent, filter-based KES imaging with an effective energy switching below and above the K-edge is possible. The manual switching of the liquid iodine filter was the time limiting factor in this study. For future studies, a solid iodine filter mounted on a motorized optomechanical component will be needed to enable filter changes in the order of milliseconds. The idea of oscillating the beam energy of the source itself, as mentioned in earlier work,¹⁸ could also be implemented but would require operating the machine in a nonstandard fashion and is not possible for subsecond energy shifts at the moment.

Although KES is a well-known imaging technique at synchrotron sources, KES with polychromatic spectra is not commonly applied in clinical imaging. One approach has been to use the technique for contrast-enhanced spectral mammography;^23,24 however the postprocessing procedure is unknown, so a direct comparison with other imaging methods is not possible. In clinical computed tomography (CT) angiography, the use of dual-energy CT systems is being investigated, for example, to calculate so-called virtual noncontrast data^25,26 or to facilitate detection of calcifications of arteries (e.g., aorta, coronary arteries).^27–29 In addition, dual-energy CT makes it possible to replace a true nonenhanced scan by a virtual nonenhanced scan. Thus radiation dose can be reduced in dual-energy CT angiography in comparison with conventional CT.^26,30 However, there are still some issues in conventional dual-energy CT that have to be solved such as the possible overestimation of calcification and the degree of stenosis.^31,32 In contrast to these imaging techniques already used in clinical radiology, KES with monochromatic x-rays is a methodically different approach for subtraction imaging. Several studies have shown that monochromatic KES CT provides improved image quality and quantitative accuracy compared with conventional polychromatic CT.^33–35 So far, KES with monochromatic x-rays has mainly been performed at synchrotron sources, with only limited integration into preclinical research. However, more and more compact synchrotron sources are emerging, which will fulfill the need for lab-size monochromatic x-ray sources and will allow KES imaging to become an important tool for biomedical research on potential clinical applications.⁹

The strong attenuation of x-rays in the human body at around 33 keV x-ray energy limits the clinical applicability of KES imaging at the iodine K-edge. This proof-of-principle study was performed with an iodine contrast agent as it is the standard contrast agent in clinical diagnostics, having the K-edge of iodine at 33.17 keV. In particular, for patients with severe iodine allergies, the shift to other contrast agents such as, e.g. gadolinium, would be an option. In the past, there have been encouraging results showing the possibility of using gadolinium-based contrast agents for angiography.^36–38 The gadolinium K-edge, located at 50.2 keV, is above the current maximum x-ray energy (35 keV) of the MuCLS; however, there is no general physical limit on the achievable x-ray energy of inverse Compton sources. Currently, many projects to build inverse Compton sources are ongoing; thereby some research projects aim at higher x-ray energies than 50 keV, e.g., ThomX³⁹ and STAR,⁴⁰ using higher electron energies and, in the case of ThomX, larger storage rings. Even with the same footprint as the MuCLS, higher x-ray energies are, in principle, accessible by decreasing the laser wavelength. This ongoing evolution of compact synchrotron sources will allow methods that are currently used at synchrotrons to be transferred to a laboratory environment¹⁷ and provide the basis for dose-compatible KES imaging in the future.

All in all, it has been shown that KES at a compact synchrotron x-ray source has the potential to be beneficial for cardiovascular imaging in the future, especially when it comes to the separation of the contrast agent and calcifications. Still, there is need for technical development of this method, which is limited in its applicability in the clinical routine by the maximum energy of the x-ray source and the availability of contrast agents with K-edges at higher energies. By further improvements of the x-ray filter and pushing the x-ray energy toward the clinical setting, a compact synchrotron source together with KES has the ability to become an important tool in preclinical research in the future.

Biographies

Stephanie Kulpe received her BSc degree in physics from the Technical University of Darmstadt, Germany, in 2014 and her MSc degree in physics from the Technical University of Munich in 2017. She is currently a PhD candidate at the same university.

Biographies of the other authors are not available.

Disclosures

The authors declare no conflicts of interest.