Anatomical and functional imaging of myocardial infarction in mice using micro-CT and eXIA 160 contrast agent

Abstract

Non-invasive small animal imaging techniques are essential for evaluation of cardiac disease and potential therapeutics. A novel preclinical iodinated contrast agent called eXIA 160 has recently been developed, which has been evaluated for micro-CT cardiac imaging. eXIA 160 creates strong contrast between blood and tissue immediately after its injection and is subsequently taken up by the myocardium and other metabolically active tissues over time. We focus on these properties of eXIA and show its use in imaging myocardial infarction in mice. Five C57BL/6 mice were imaged ~ 2 weeks after LAD coronary artery ligation. Six C57BL/6 mice were used as controls. Immediately after injection of eXIA 160, an enhancement difference between blood and myocardium of ~340 HU enabled cardiac function estimation via 4D micro-CT scanning with retrospective gating. Four hours post-injection, the healthy perfused myocardium had a contrast difference of ~140 HU relative to blood while the infarcted myocardium showed no enhancement. These differences allowed quantification of infarct size via dual energy micro-CT. In vivo micro-SPECT imaging and ex vivo TTC staining provided validation for the micro-CT findings. Root mean squared error of infarct measurements was 2.7% between micro-CT and SPECT, and 4.7% between micro-CT and TTC. Thus, micro-CT with eXIA 160 can be used to provide both morphological and functional data for preclinical studies evaluating myocardial infarction and potential therapies. Further studies are warranted to study the potential use of eXIA 160 as a CT molecular imaging tool for other metabolically active tissues in the mouse.

Introduction

An ongoing need exists to develop non-invasive small animal imaging techniques for evaluation of cardiac disease and potential therapeutics. The mouse is a challenging model system for these studies. The mouse heart is nearly 3000 times smaller than a human’s and has an R-R interval 1/10th as long, which demands both high spatial and temporal resolution for effective imaging. Despite these challenges, several groups have demonstrated successful cardio-respiratory gated 4D micro-CT imaging of the mouse heart[1–3]. In addition to high resolution scanning, cardiac micro-CT imaging requires contrast agents that remain in the bloodpool. High relative cardiac output in the mouse [4] leads to rapid renal clearance of traditional contrast agents. As a result, cardiac micro-CT has been performed mostly using blood pool contrast agents with prolonged circulation times relative to clinical contrast agents. Various formulations of blood pool agents have been developed, including iodinated liposomes [5], iodine-containing micelles [6] and a pegylated iodinated triglyceride emulsion marketed as Fenestra VC [7, 8]. More recently, a new contrast agent called eXIA 160 (Binitio Biomedical, Inc. Canada) has become commercially available. eXIA 160 is an aqueous colloidal poly-disperse contrast agent containing a large concentration of iodine (160 mg/ml), which allows sufficient enhancement with very low injection volumes (0.125 ml/25 g mouse). The kinetics and biodistribution of eXIA 160 have been characterized and compared with Fenestra [9].

In a recent study, the murine cardiac-related enhancement caused by eXIA 160 was evaluated [10]. The authors showed that eXIA 160 provides blood enhancement sufficient for cardiac micro-CT approximately 30 minutes post-injection. More interestingly, they report that this contrast agent was taken up by the myocardium and brown adipose tissue (BAT) over time and provided continued enhancement in these tissues even after elimination from the blood pool. The reason for uptake in these tissues is not clear, but may be related to their high metabolic activity.

In this study, we focus on these unique properties of eXIA 160 and show its use in imaging the myocardium after coronary artery ligation. Based on the hypothesis that eXIA is taken up by tissues with high metabolic activity, we expected that infarcted tissues would show little to no uptake of the contrast agent, while the healthy myocardium would show significant uptake and retention of the agent. This would allow for easy differentiation of infarcted and healthy myocardium by micro-CT. We have previously developed a method using micro-CT to quantify both infarct size and cardiac function after coronary artery ligation using a continuous infusion of a traditional iodinated contrast agent [11]; however, this method required a large volume of contrast agent that was not tolerated well by some mice because of their susceptibility to acute cardiac decompensation after myocardial infarction. The objective of this study is to show how the unique uptake of eXIA 160 can provide an alternate method for determining the three-dimensional location and size of myocardial infarcts in mice using a dual energy micro-CT scan while also producing sufficient contrast for measuring cardiac function using 4D micro-CT after a single bolus injection.

Results

Biodistribution

Two C57BL/6 control mice were serially imaged using a dual energy micro-CT system for five hours after injection of eXIA 160 to determine the biodistribution and kinetics of the contrast agent. At each time point, CT number (in HU) was measured in an ROI within the blood in the left ventricle (LV), myocardial wall, liver, spleen, and BAT. The average CT number for the two mice at each time point is plotted in Figure 1. These results are similar to those seen in previous studies [9, 10]. The spleen showed the greatest enhancement, peaking at 1440 HU at 30 minutes and slowly decreasing until 5 hours (not shown in figure). The subscapular BAT showed significant enhancement that continually increased throughout the 5 hour study. Liver enhancement peaked at 15 minutes and remained fairly constant (~130 HU above unenhanced blood) for the remainder of the study. Blood contrast peaked at 15 minutes with a 360 HU enhancement relative to the myocardium. This enhancement returned to baseline by 3 hours and then remained stable. The myocardium showed mild enhancement at 15 minutes (~100 HU over unenhanced blood) and then gradually increased in enhancement over the duration of the study with a maximum enhancement of 140 HU relative to the blood at 4 hours. Figure 2 shows representative standard micro-CT and dual energy images of the heart at 15 minutes (high blood contrast) and at 4 hours (high myocardial contrast). The myocardium is easily differentiated from the blood and surrounding structures at both time points. The brown adipose tissue (white arrow) is not visible at 15 minutes, but shows significant enhancement at 4 hours.

Figure 1.

Mean CT enhancement at 80 kVp in various organs over 5 hours post injection of eXIA 160 in two control mice. The blood is initially strongly enhancing, allowing for cardiac function estimation. At 4 hours the difference between the myocardium and the blood is reversed (brighter myocardium). This delayed timepoint imaging has been used to quantify the size of the myocardial infarction.

Figure 2.

A comparison of differential enhancement at 15 mins and 4 hours post injection of eXIA 160 in a control mouse for both filtered micro-CT and dual energy (DE) calcium/iodine decomposed maps with iodine in red and calcium in green. Axial (A) and coronal (B) slices through the center of the left ventricle are shown. The CT images are windowed between −500 to 750 HU and the DE maps have iodine windowed between 1.5 and 10 mg/ml and calcium between 0 and 50mg/ml. The arrows show brown adipose tissue.

4D micro-CT cardiac function imaging

Control (n=6) and surgically-induced MI (n=5) C57BL/6 mice were imaged 10 minutes after eXIA injection when contrast between the blood and myocardium was maximal. Figure 3 shows a series of 10 axial heart slices for both a control and an MI mouse. Each of the 10 images corresponds to one of the ten phases of the cardiac cycle. Phase 1 corresponds to the end diastolic volume (EDV), while phase 5 corresponds to the end systolic volume (ESV). From these values, ejection fraction (EF), stroke volume (SV), and cardiac output (CO) were determined. Figure 4 summarizes these data for both the control and the MI groups. There was a statistically significant difference between the two groups for EF, SV, and CO (p<0.05). There was not a statistically significant difference in EDV (p=0.19) and ESV (p=0.09) due to large variation in heart volumes in the MI group.

Figure 3.

A comparison of cine sequences in a mouse with myocardial infarction (A) and a control mouse (B) for similar axial micro-CT slices over the 10 phases of the cardiac cycle acquired immediately after injection of eXIA 160. Phases corresponding to end-diastolic volume (EDV) and end-systolic volume (ESV) are labeled. Note the enlarged ventricular size and lack of noticeable wall motion in the infarcted heart. The images are windowed between −500 and 700 HU.

Figure 4.

A comparison of cardiac function between the control and MI mice. Note the chamber enlargement of the LV and decreased ejection fraction, stroke volume and cardiac output in the MI mice. Error bars represent standard error of the mean, asterisks represent statistical significance (p<0.05).

Measurement of infarct size

MI mice were re-imaged 4 hours after eXIA injection when myocardial enhancement relative to blood was maximal. At this time point, perfused myocardium in the MI mice showed enhancement similar to the controls, but the infarcted myocardium showed no enhancement. Images of the infarct from micro-CT were then used to calculate infarct size.

For direct comparison, high resolution single-photon emission computed tomography (SPECT) images of the same group of MI mice were acquired using Tc99m-tetrofosmin to identify viable cardiac tissue. Individual slices (both axial and coronal) from SPECT and CT were compared side-by-side for determination of infarct location and size. Figure 5 shows a representative comparison between SPECT and CT images. The two modalities showed a very high level of agreement for both location and size of infarct in all mice studied (Figure 6). The total size of each infarct (% of total LV wall infarcted) was determined for both SPECT and CT by manually segmenting the healthy and infarcted regions of the myocardium.

Figure 5.

A comparison of similar axial (A) and coronal (B) slices in an MI mouse for the bilateral filtered CT at 80 kVP, the dual energy decomposition of iodine (red) and calcium (green) and SPECT data. Note the lack of myocardial enhancement at the place of MI (arrows). The CT images are windowed between −80 and 350 HU, the DE maps have iodine windowed between 1.0 and 3.0 mg/ml and calcium between 0 and 70 mg/ml, and the SPECT images are windowed between 245 and 1824 7Ci/ml.

Figure 6.

Comparison of similar slices in SPECT (A), CT (B), and TTC stain (C) shown with manual segmentation of healthy myocardium (yellow line) for infarct size calculation. Plots of CT measurement of infarct size vs. measurements by SPECT (D) and TTC staining (E).

Following SPECT and CT imaging, infarct size was also determined histologically using triphenyl tetrazolium chloride (TTC) staining of 1 mm heart slices. Figure 6 shows an example of the manual segmentation in CT, SPECT and TTC images as well as a plot directly comparing the infarct size as measured by each method. Linear regression of the data is shown, as well as the identity line (y=x) for comparison. In both plots, the CT method was consistent with the traditional method, so the linear regression equation is very close to the identity line. The root mean squared error between the two measurement techniques was 2.7% in the CT/SPECT comparison and 4.7% in CT/TTC comparison. A paired t-test for equality of the CT and SPECT measurements could find no statistically significant difference between the two techniques (t=−0.49 and p=0.65). A paired t-test for equality of the CT and TTC measurements could also not find a statistically significant difference between the two techniques (t=-0.16 and p=0.88).

Discussion

eXIA 160 shows a unique biodistribution. It not only enhances the liver and spleen, as expected with clearance via the reticuloendothelial system, but it also enhances the myocardium and BAT. Uptake by these tissues may be related to their high metabolic activity [10]. Binitio Biomedical, Inc. scientists have explained that eXIA is a colloidal formulation that is fully metabolized by catabolic pathways in living animals. eXIA consists of components that can serve as energy substrates and, thus, can be utilized by the animal body. Since the heart is continuously metabolically active and requires a constant supply of energy, it is not unreasonable to assume that eXIA is consumed by cardiac myocytes as an energy source. As such, eXIA can be used to quantitatively determine the metabolic integrity of myocardium analogous to FDG-based positron emission tomography [Personal communication from E. Rizhevskaya to C. Badea, unreferenced]. In this experiment, contrast in both the myocardium and the BAT continued to increase even after the contrast agent was eliminated from the blood. Iodine levels in the blood decreased below the levels in both the myocardium and BAT about 1.5 hours after injection, but enhancement in these tissues continued to increase until a peak was reached in the myocardium at 4 hours post injection. For the BAT, the peak enhancement was never determined because it continually increased through the duration of the experiment. This shows that uptake of eXIA 160 is occurring in these tissues not only by a passive transport mechanism (which would require a gradient from the blood into the tissues for uptake), but also by an active transport that concentrates the iodine in the myocardium and BAT against the gradient. Although specific details about the formulation of eXIA 160 are not publicly available, we postulate that eXIA is recognized as an energy source by cellular receptors on metabolically active tissues and is taken up by those tissues in a specific manner via active transport.

Our 4D CT dataset shows that eXIA is capable of providing sufficient contrast between the myocardium and the blood for determination of cardiac function. The results that we found from the 4D data were as we expected. The control mice had fairly uniform cardiac function, well within the expected range for healthy mice. The MI mice, however, had significant variability in their heart size. While they all had reduced ejection fractions, some mice had markedly dilated hearts due to post-infarction adverse LV remodeling, while other mice had relatively normal diastolic chamber size but reduced ejection fraction. Because of the large variation in heart size, there was no statistical significance between end-diastolic volume and end-systolic volume in the MI mice compared to control. However, the MI mice showed significantly lower ejection fraction, stroke volume, and cardiac output compared to the control mice.

With delayed imaging (four hours after eXIA injection), we were able to successfully identify, localize, and measure the extent of the infarcted myocardium. As expected, uptake of the contrast agent occurred strongly in the healthy myocardium, but was entirely absent in the infarcted myocardium. The dual energy decomposition, while not strictly necessary for this study, was useful in our analysis. Decomposing the image into an iodine map helped to support our assumption that the enhancement seen in the myocardium was specifically due to iodine.

Additionally, the iodine map had significantly better contrast between the myocardial wall and the lumen of the ventricle at 4 hours because there was no iodine left in the blood pool. This produced an image with sharp contrast between a completely dark lumen and a bright wall, as seen in Figure 5. The iodine map also improved accuracy of segmentation, as the bright bone surrounding the left ventricle in the regular CT image is not as intense in the iodine map, so the left ventricle is more easily differentiated.

In our validation of eXIA as a method of determining infarct size, we compared size measurements from the eXIA-contrasted CT method with measurements from SPECT and TTC staining, which are the traditional gold standards for in vivo and ex vivo determination of infarct size and location. We aimed to show that measurements from the eXIA-contrasted CT are equivalent to the measurements made by SPECT or TTC for any given mouse. In figure 6, in which CT data was plotted against both SPECT and TTC data, we compared the linear regression of our data to the identity line (y=x), which represents equality of the two compared techniques. The regression line in both cases is extremely close to the identity line, showing strong agreement between the CT method and the gold standards. The root mean square error between the CT measurements and the gold standard measurements was 2.7% for SPECT and 4.7% for TTC, which is well within the expected variation for manually segmented images.

We saw more variability in the CT vs TTC measurements than in the CT vs SPECT measurements. The most likely explanation for this increased variability is that in the CT vs. TTC measurements, we are only comparing areas of infarct from a single anatomical slice, rather than comparing the entire volume from ~70 slices (as in the CT vs. SPECT data). This increases the potential for error in the comparison. Despite the slightly increased variability in the TTC data compared to the SPECT data, we were satisfied with the small observed error between the different types of measurements. We concluded that the ability of CT to measure infarct size was roughly equivalent to both SPECT and TTC staining. The results of our paired t-test agreed with this conclusion (we could not reject the hypothesis that the two techniques were equivalent), but due to the small sample size in this study, we did not have sufficient statistical power to detect small differences in the techniques. A more rigorous study including more mice could be performed to confirm these findings, but our results are sufficient for a proof of concept that using micro-CT with the eXIA 160 contrast agent is comparable to the traditional gold standards for detecting and quantifying myocardial infarction.

Another potential source of error in this analysis is the manual segmentation of images for infarct measurement. Using manual segmentation introduced the possibility of both random user error and expectation bias. The segmentation of all the data sets was performed by a single person to minimize variation in technique, and trials of manually segmenting the same CT images multiple times (while being blind to the associated SPECT/TTC results) yielded consistent measurements for infarct size. This gives us confidence that the amount of introduced error in this technique is small. Manual segmentation was chosen because automatic segmentation with basic software was not successful in properly identifying the infarcted myocardial tissue, even though infarcted tissue was easily visualized for manual segmentation. More advanced segmentation protocols could be developed specific to this problem which would further improve the accuracy of our results, but this was beyond the scope of the current study.

The ability of micro-CT to accurately identify and quantify myocardial infarction is an important step in preclinical imaging. Because of its low cost, high spatial resolution, and ease of use relative to other modalities for cardiac imaging (MR, SPECT), micro-CT has been widely used in preclinical imaging for the evaluation of a variety of cardiac-related therapeutics. eXIA improves the usefulness of micro-CT by providing a simple methodology for imaging myocardial infarctions with high levels of contrast. This high level of contrast is also an improvement over the delayed hyperenhancement seen in clinical CT for myocardial infarction using traditional iodinated contrast agents, which suggests that eXIA may be useful in the realm of clinical imaging as well.

One limitation of micro-CT versus other imaging modalities is the use of ionizing radiation. The dose associated with micro-CT methods is not negligible but representes approximately 8 to 11 times less than the LD_50/30[12]. This dose is not expected to influence the results of this study.

This study did not consider the long-term kinetics of eXIA, but it would be interesting to continue to track the distribution of iodine on a longer time scale. Previous studies have shown that both the myocardium and BAT show strong enhancement at 24 hours [10], but those studies did not determine how long the iodine is retained in those tissues or at what point the BAT reaches its maximal enhancement. Further studies of the BAT enhancement are also warranted to better characterize the contrast agent. Imaging mice at different temperatures (which would vary the amount of thermogenesis in the BAT) would help to elucidate the direct relationship between metabolic activity and eXIA uptake.

This study has shown that eXIA 160 is a potentially useful molecular imaging tool for gathering functional information about the metabolic activity of tissues at the cellular level. With the resemblance of eXIA-CT to FDG-PET, the question arises: can eXIA-CT be used in other areas where PET imaging is traditionally dominant? For example, do cancer cells with high metabolic activity also show increased uptake of eXIA? And, if so, can eXIA be used as a molecular imaging tool for the detection of cancer by CT? Would using eXIA in addition to FDG in dual-modality PET/CT imaging improve the sensitivity and specificity of cancer imaging? Further studies will focus on addressing these important questions.

Conclusions

eXIA 160 is a novel blood-pool contrast agent which shows uptake in metabolically active tissues, including the myocardium and brown adipose tissue. Infarcted myocardium showed no uptake of contrast agent, making it easy to differentiate from non-infarcted perfused myocardium. Infarct size was successfully measured using micro-CT with results comparable to both SPECT imaging and ex vivo TTC staining. Further studies are warranted to further characterize eXIA 160 as a CT molecular imaging tool for metabolically active tissues.

Materials and Methods

In vivo experiments in mice

The institutional subcommittee on research animal care at Duke University Institutional Animal Care and Use Committee approved all animal surgery and imaging studies. A total of (n=5) C57BL/6 mice underwent MI by left anterior coronary artery ligation as described previously[11, 13] Mice were maintained under general anesthesia throughout the surgical procedure with1–2% isoflurane. A thoracotomy was performed in the fourth left intercostal space. The LV was visualized and the left anterior descending coronary artery was permanently ligated with a monofilament nylon 8-0 suture at the level of the left atrial appendage. This procedure resulted in variable MIs involving the anterolateral, posterior and apical regions. The MI mice were imaged approximately 2 weeks after surgery. A total of n=6 C57BL/6 mice were used as a control group for the infarction studies and a total of n=2 control mice were used for the eXIA biodistribution studies.

For all micro-CT scans, mice were injected with a dose of 0.125 ml/25 g mouse eXIA 160 via a tail vein catheter. The animals were anesthetized with isoflurane (1.5%) mixed with 50% oxygen and balanced with nitrogen, delivered via nose cone. ECG was monitored via electrodes (Blue Sensor, Medicotest, UK) taped to the foot pads, and body temperature was maintained with heat lamps, a rectal probe and a back controller (Digi-Sense®, Cole Parmer, Chicago, IL). A pneumatic pillow on the thorax was used to monitor respiration.

Micro-CT system

Image data were acquired with a dual source micro-CT system developed explicitly for dynamic and dual energy applications which has been described in detail elsewhere[14]. The x-ray tubes and the detectors are arranged orthogonally. The system contains two G-297 x-ray tubes (Varian Medical Systems, Palo Alto, CA) with 0.3/0.8 mm focal spot size, two Epsilon High Frequency X-ray generators by EMD Technologies (Quebec, Canada) and two CCD based detectors with a Gd2O2S phosphor (XDI-VHR 2 Photonic Science, East Sussex, UK) with 22 micron pixels which we typically bin to 88 microns. The data acquisition was controlled by sequencing applications written in LabVIEW.

Biodistribution Image Acquisition and Analysis

We used n=2 controls to produce dual energy data sets at 15 mins, 30 mins, 1h, 2h, 3h, 4h, and 5h post injection of eXIA 160. These mice were used to assess the dynamic contrast enhancement in various organs. The enhancement over time was measured with ImageJ (National Institute of Mental Health) using the 80 kVp data and the iodine maps. Regions of interest of ~100 pixels were selected in the LV lumen, myocardium, brown fat, spleen and liver to calculate mean signal intensities and absolute iodine concentrations.

4D micro-CT image acquisition

Immediately after the injection of eXIA 160 (dose 0.125l/25 g mouse) in the MI group and control group, we performed a 4D micro-CT sampling using retrospective cardio-respiratory gating [15]. In retrospective gating, projection images are acquired at a rapid and constant rate without waiting for cardiac and respiratory coincidence. Respiratory and cardiac motion are monitored and saved in synchrony with the acquisition of the projections. The parameters for both x-ray tubes exposures were the same (i.e., 80 kVp, 100 mA, 10ms). A total number of 2250 projections were acquired over a sampling time of 5–8 minutes.

4D micro-CT data reconstruction and analysis

For retrospectively gated 4D micro-CT data required to assess cardiac function, the projection images are sorted and associated with angle, cardiac and respiratory phase information using a Matlab (The MathWorks, Natick, MA) script. This script detects the R peaks in the ECG signal and the maxima in the respiratory signal. Each cardiac cycle is divided into 10 intervals (each equal to 10% of the RR interval). Each projection is registered with the ECG signal by finding the temporal distance from the previous R peak that was closest to the projection sampling time and dividing the distance by the R-R period. This information is kept in lists that are used during reconstruction. The procedures for reconstruction using retrospective gating have been described in detail elsewhere[16]. Essentially, the projections are sorted into sets based on their phase in the cardiac cycle, and these sets are reconstructed separately. Reconstruction of each set is performed with filtered backprojection (FBP)[17]. However, because of the retrospective gating, the angular distribution of the projections may be irregular, resulting in streaking artifacts in the images reconstructed with FBP. To overcome this problem, we generate a synthetic set of projections with a regular angular distribution by interpolating the projections in each set. The reconstructions also use 4D bilateral filtration[18] to reduce noise and any remaining artifacts. The output of the reconstruction process was ten 3D sets corresponding to the 10 phases of cardiac cycle, each with a matrix size of 512×512×220 and voxel size of 88 microns.

Measures of cardiac function were obtained using the Cardiac Function Analysis application available on the Vitrea software package (Vital Images, Inc, MN). The software semi-automatically segments the left ventricle (LV) over all phases of the cardiac cycle and measures LV volumes. The LV end-diastolic and end-systolic volumes are used to compute stroke volume (SV), ejection fraction (EF) and cardiac output (CO), with SV = EDV − ESV; EF = SV/EDV, and CO= SV × Heart Rate.

Dual energy image acquisition

Dual energy micro-CT was performed on the MI mice (n=5) four hours after eXIA injection, when myocardial enhancement relative to blood was maximal. The dual energy scans were performed using prospective cardio-respiratory gating to minimize cardiac and respiratory motion as previously described[19]. The two x-ray tubes had different exposure parameters which allowed simultaneous acquisition of two sets of projections at different energies. The parameters of one x-ray tube were 40 kVp, 250 mA, 16 ms per exposure, while the other X-ray tube’s parameters were 80 kVp, 160 mA, 10 ms per exposure. For each energy set, the acquisition consisted of 300 projections. A single scan required approximately 5 minutes to complete.

Dual energy micro-CT processing

Affine registration was performed to improve registration between corresponding 40 and 80 kVp reconstructed volumes using ANTs, an open-source, ITK-based registration toolkit (Advanced Normalization Tools, http://picsl.upenn.edu/ANTS/; ITK rev. 4.0.0). Affine registration in ANTs optimized mutual information. Registration between each pair of 40 and 80 kVp data generally completed in less than two minutes using a Mac OS X workstation with dual 2.66 GHz, quadcore Xeon processors and 16 GB of RAM.

To improve the results of DE decomposition, each data set was de-noised using bilateral filtration (BF). BF is a well-characterized edge-preserving smoothing filter that considers both the distribution of neighboring voxels in space and intensity. BF was implemented using MATLAB (MathWorks, Natick, MA), and generally was completed within 15 minutes per volume using the computer system previously listed. Details of the application of (“classic”) BF to murine CT data can be found in other works[20]. A quantitative evaluation of the application of BF to DE micro-CT data has also been previously described[19].

A DE decomposition of calcium and iodine was performed after registration and filtration using corresponding 80 and 40 kVp data by solving the following linear system at each voxel:

$$x=A^{-1}b$$

Expanding the linear system:

$$\left[\begin{array}{c}\begin{array}{c}{C}_I\\ C_{Ca}\end{array}\end{array}\right]={\left[\begin{array}{cc}C{T}_{I,40}& C{T}_{Ca,40}\\ C{T}_{I,80}& C{T}_{Ca,80}\end{array}\right]}^{-1}\left[\begin{array}{c}\begin{array}{c}C{T}_{40}\\ C{T}_{80}\end{array}\end{array}\right]$$

In this formulation, x was the least squares solution for the concentration of the iodine (C_I) and calcium (C_Ca) in mg/mL in the voxel under consideration. A was a constant sensitivity matrix measured in HU/mg/mL for iodine (CT_I,40, CT_I,80) and calcium (CT_Ca,40, CT_Ca,80) at 40 and 80 kVp, respectively. Finally, b was the intensity of the voxel under consideration at 40 kVp (CT₄₀) and 80 kVp (CT₈₀) in Hounsfield Units (HU).

Values for CT_I,40, CT_I,80, CT_Ca,40, and CT_Ca,80 were determined empirically using a calibration phantom and were 36.62, 53.60, 20.74, and 14.94 HU/mg/mL, respectively. Voxels with negative concentrations of both materials were set to zero.

Micro-SPECT imaging

Micro-SPECT imaging was used to provide an in vivo validation of the infarct size for all mice in the MI group. Cardiac micro-SPECT images were obtained using the U-SPECT-II/CT system (Milabs, Utrecht, The Netherlands) fitted with an ultra-high resolution 0.35 mm multi-pinhole collimator. Anesthetized mice were injected with 185–370 MBq of Tc99m-tetrofosmin (GE Healthcare, Arlington Heights, IL) via tail vein catheter. Following injection, animals were placed prone on a heated animal bed with integrated ECG and respiratory monitoring. Field of view was adjusted to the margins of the heart using orthogonal radiographs generated by the attached micro-CT unit. SPECT images were acquired over 30 minutes (3 frames, 10 minutes per frame). Animals were returned to cages and recovered after imaging.

SPECT data was acquired in list-mode and reconstructed using the Pixel-based Ordered Subset Expectation Maximization (POSEM) iterative reconstruction algorithm (6 iterations, 16 subsets, 0.125 mm voxel size). Reconstructed images were viewed and optimized using PMOD v.3.3 biomedical image quantification software (PMOD Technologies Ltd., Zurich, Switzerland).

Measurement of infarct size

The total size of each infarct (% of total left ventricular wall infarcted) was determined for both SPECT and CT by manually segmenting the highly-enhancing regions of the ventricle in every slice in ImageJ. For the CT slices, both the standard CT image (80 kVp) and the iodine map from dual energy decomposition were used to locate the infarcts. The number of voxels in each region of interest was calculated, and the number of voxels from all of the slices was summed to determine the total volume of healthy (enhancing) myocardium and total left ventricular wall volume. Infarcted volume was calculated by subtracting the volume of healthy myocardium from the total volume of the ventricular wall.

Histopathological analysis

On completion of the imaging studies, all MI mice were euthanized for histopathological analysis. The hearts were excised and rinsed in PBS and cut into myocardial rings of 1-mm thickness. Thereafter, midventricular sections were stained with TTC for 20 min, and digital pictures were acquired for quantification of infarct size. Infarct size for each mouse was determined by manually segmenting the images into healthy myocardium and LV wall(including both healthy and infarcted myocardium). The area of each region was calculated and infarct size was determined by subtracting the area of healthy myocardium from the total area of LV myocardium. This process was repeated for corresponding micro-CT images, which were identified within the 3D data set through anatomic landmarks such as papillary muscles.

Radiation dose estimation

The radiation doses associated with the 4D micro-CT and dual energy micro-CT scans were 360 and 260 mGy, respectively [21]. Therefore the cumulative dose for a micro-CT study focused on imaging myocardial infarction has been 620 mGy.

Statistical analysis

Linear regression and paired t-tests of the data were performed using R (R Foundation for Statistical Computing, http://www.r-project.org). The paired t-test tested the hypothesis that the difference between the CT measurements and gold standard measurements (SPECT or TTC) was equal to zero. All statistical analyses were considered to be significant at a p<0.05 level.