Cardiac imaging techniques play a critical role in the evaluation of novel therapies providing reliable surrogate endpoints in the ongoing effort to explore new approaches to treat heart diseases. Myocardial transplantation of different cell types and preparations has been widely investigated as a potential therapy for myocardial infarction (MI) and heart failure (HF) in recent years (1). Cardiac magnetic resonance (CMR) is one of the preferred imaging tools, in addition to the frequently used echocardiography and nuclear techniques, to provide reliable assessment of left ventricular ejection fraction (LVEF) and infarct size, which are generally accepted surrogate endpoints (2- 3). Previous animal studies using CMR showed that mesenchymal stem cells (MSC) reduced infarct size and improved left ventricular (LV) function in acute and chronic MI (4- 6). However, the use of CMR in the clinical setting is limited in patients with metallic implants and pacemakers and is relatively complex and time consuming.

Over the last decade, multidetector computed tomography (MDCT) has emerged as a novel tomographic cardiac imaging modality. Improvement in spatial and temporal resolution has allowed the implementation of myocardial applications, which are necessary to evaluate regenerative therapies. Although these MDCT applications are still investigative, current MDCT technology is able to provide reliable and reproducible assessment of cardiac function (7- 8), myocardial viability with delayed contrast enhancement (DCE) (9- 11), and myocardial blood flow (MBF) (12- 14). However, whether MDCT can be used for follow-up studies and is able to show therapeutic effects of regenerative cell therapies has not been investigated. Accordingly, the purpose of this study was to 1) use MDCT LV function and myocardial viability as surrogate endpoints in a randomized animal study to evaluate outcomes of intramyocardial-delivered cell therapy and 2) evaluate the robustness and comparability of MDCT in relation to CMR.

All animal studies were approved by the Johns Hopkins University Institutional Animal Care and Use Committee and complied with the “Guide for the Care and Use of Laboratory Animals” (National Institutes of Health publication 80-23, revised 1985). Thirty-three female Göttingen minipigs were purchased from Marshall BioResources (North Rose, New York). MIs were induced by occlusion of the mid-left anterior descending artery with an inflated coronary angioplasty balloon as previously described (15). Five animals did not survive the infarct procedure. Two animals died during the early follow-up in the first 4 days after MI induction, and 4 animals died between weeks 8 and 12 post-MI. Therefore, a total of 22 animals were used in the study.

The 13 animals randomized to cell therapy received bone marrow–derived porcine MSC. MSC were obtained, isolated, and expanded as previously described (6). In brief, MSC were harvested when they reached 80% to 90% confluence. Cells were placed in cryo bags at a concentration of 10 to 15 million MSC/ml and then frozen in a control-rate freezer to −180°C until the day of implantation. Trypan blue staining was performed to attest viability of thawed MSC lots before injection. Only MSC lots containing 85% or more of viable cells were used in the study.

Twelve weeks after MI, animals were randomized to receive either intramyocardial injections of porcine MSC or phosphate-buffered saline to serve as a treatment (n = 13) or placebo (n = 9) group, respectively. Myocardial injections were performed as previously described (5- 6).

MDCT images were acquired at 2 time points, before randomization at week 12 post-MI and 12 weeks after the intramyocardial injection procedure at week 24.

Each animal was scanned with electrocardiographic (ECG) monitoring using a 0.5 mm × 64-slice MDCT scanner (Aquilion 64, Toshiba Medical Systems Corporation, Otawara, Japan). Data acquisition for LV function parameters and LV volumes were initiated manually at a threshold value of 180 Hounsfield units in the descending aorta. A 60-ml bolus of iodixanol (Visipaque 320 mg iodine/ml, Amersham Health, Buckinghamshire, United Kingdom) was injected intravenously at rate of 5.0 ml/s, opacifiying the LV chamber during first pass. An additional 90 ml of iodixanol was given intravenously after first-pass acquisitions, and DCE images were acquired 10 min after initial contrast delivery; 150 ml of iodixanol (0.91 ± 0.04 × 103 mg iodine/kg body weight) is 1.5 times the human equivalent dose of contrast currently used for MDCT angiography.

During MDCT acquisition, respiration was suspended and imaging was performed using a retrospective ECG-gated MDCT protocol without dose modulation. Imaging parameters: gantry rotation time = 400 ms; detector collimation = 0.5 mm × 64 (isotropic voxels = 0.5 × 0.5 × 0.5 mm³ − 13 linepairs/cm); helical pitch = variable depending on heart rate (range 6.4 to 6.8); tube voltage = 120 kV; and tube current = 400 mA. Animals with heart rates >100 beats/min during first-pass acquisition received intravenous metoprolol (2 to 5 mg) and /or amiodarone (50 to 150 mg) to achieve lower heart rate for DCE acquisitions.

All raw data were reconstructed at contiguous 0.5-mm slice thickness by an adaptive multisegment reconstruction algorithm (16). First-pass acquisitions were reconstructed in 10% steps from 0% to 90% throughout the entire R-R interval using a standard kernel (FC43); images were reformatted at 4-mm slice thickness in short axis and evaluated in QMass CT 7.1 (Medis Medical Imaging Systems, Leiden, the Netherlands). Endocardial and epicardial borders of the LV were defined in all 10 contiguous slices, and each end-diastolic and end-systolic frame was determined to calculate LVEF, left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), LV stroke volume, and LV remodeling parameters. The temporal resolution based on gantry rotation of all MDCT acquisitions was 211.5 ± 5.4 ms.

DCE-MDCT data (for myocardial scar assessment) were reconstructed at 80% of the R to R′ interval using the FC43 kernel and multisegment reconstruction. ECG editing to account for arrhythmias was performed when necessary. Multiplanar reformation at 4-mm slice thickness in the short axis of the heart was implemented, and MDCT images were analyzed using a custom research software package (Cine Tool, GE Medical Systems, Waukesha, Wisconsin). Infarct mass/volumes in DCE-MDCT images were defined by 3 SD of the signal intensity above the viable myocardium for the core infarct and 2 SD for the peri-infarct zone as described (16). Infarct size was defined as infarct mass as a percentage of LV mass. Both datasets, the cine MDCT acquisitions, and DCE-MDCT images, were analyzed by 2 blinded observers.

CMR was performed at the same time points as MDCT imaging in random order on a 1.5-T MR scanner (CV/i, GE Medical Systems).

LVEF function and LV volumes were assessed using a steady-state free precession pulse sequence (17). Six to 8 contiguous short-axis slices were prescribed to cover the entire LV, from base to apex. Image parameters were as follows: repetition time/echo time = 4.2 ms/1.9 ms; flip angle = 45°; 256 × 160 matrix; 8-mm slice thickness/no gap; 125 kHz; 28-cm field of view, and 1 number of signals averages. After an intravenous injection of gadolinium–diethylene triamine pentaacetic acid (0.2 mmol/kg body weight, Magnevist, Berlex, Wayne, New Jersey), DCE-CMR images were acquired 15 min later using an ECG-gated, breath-hold, interleaved, inversion recovery, and fast gradient echo pulse sequence. DCE-CMR images were acquired in the same location as the short-axis cine images. Imaging parameters were repetition time/echo time/inversion time = 7.3 ms/3.3 ms/180 to 240 ms; flip angle = 25°; 256 × 196 matrix; 8-mm slice thickness/no gap; 31.2 kHz; 28-cm field of view, and 2 number of signal averages. Inversion recovery time was adjusted as needed to null the normal myocardium (16).

Cine MR images were analyzed with QMass MR 7.1 (Medis Medical Imaging Systems). To evaluate LVEF, LV volumes and myocardial mass endocardial and epicardial borders were defined each in the end-diastolic and end-systolic frame in 25 to 30 contiguous slices, and LV parameters and volumes were calculated. The temporal resolution of the cine MR images was 21.6 ± 0.9 ms. DCE-CMR images were analyzed on the same custom research software package (Cine Tool, GE Medical Systems) as DCE-MDCT images. Infarct size, determined by infarct mass as a percentage of LV mass and infarct volume in DCE-CMR images, was defined with the same threshold method described for DCE-MDCT analysis.

All data are presented as mean ± SEM unless otherwise stated. For infarct size and infarct volume comparison, Pearson correlation and linear regression analysis were used to compare MDCT and CMR. Results were confirmed by Bland-Altman analysis and agreement expressed as mean ± SD difference between methods at 95% confidence intervals. The MDCT and CMR data were evaluated with a paired Student t test. All analyses were performed in MedCalc (MedCalc Software, Mariakerke, Belgium). A value of p < 0.05 was considered significant.

Twenty-two animals were used in this study 12 weeks after MI induction (91.9 ± 2.6 days and 87.9 ± 2.0 days after MI induction for placebo and MSC groups, respectively; p = 0.19). After randomization, animals in both groups were of similar age (16.8 ± 0.7 months and 15.5 ± 1.1 months for placebo and MSC, respectively; p = 0.35) and body weight (39.3 ± 1.6 kg and 35.7 ± 1.9 kg for placebo and MSC, respectively; p = 0.11).

One placebo- and 2 MSC-treated animals were excluded from analysis due to poor imaging quality caused by motion artifacts at week 24. Infarct size, defined as infarct mass as percentage of LV mass, was reduced in all MSC-treated animals at week 24, whereas infarct size increased in all animals in the placebo group (Figure 1). We observed a mean expansion of 29.1 ± 5.7% in infarct scar volume in the placebo group (p = 0.001). MSC therapy had limited effect on total scar volume (p = 0.1) ((Figures 1, 2)C, and Figure 2D). At week 24, infarct size and infarct volume were reduced in MSC-treated animals compared with those in the placebo group (p = 0.003 and p = 0.01, respectively), ((Table 1), (Figure 2)A to Figure 2D).

Cardiac MDCT demonstrated that the decrease in infarct size resulted in 14.0 ± 6.7% improvement in LVEF in animals randomized to MSC therapy, whereas LVEF decreased by 12.3 ± 3.1% in the placebo group ((Table 1), (Figure 2)E and Figure 2F). Of note, 2 animals that received MSC therapy did not show improved LVEF (Figure 2F). All MDCT data for global LV function and LV volumes are summarized in (Table 1).

Interobserver correlation for infarct size, total infarct volume, LVEF, and LVEDV were excellent (r = 0.96, r = 0.98, r = 0.93, and r = 0.93, respectively) in MDCT acquisitions. The complete data are presented in (Table 2).

MDCT and CMR were performed on the same day in random order within 2 h following the first imaging study. Three animals had to be excluded from the analysis because CMR image data were not obtained on the day of the MDCT imaging. In 32 CMR studies, available the same day as MDCT acquisitions, we found fair agreement for infarct size, infarct volume, LVEF, and LVEDV ((Figure 3), Table 3). In addition, LVESV showed a good correlation with both imaging modalities (r = 0.81; p < 0.001). Bland-Altman analysis revealed a slight overestimation by 1.4 ml (95% confidence intervals: 16.3 to −13.6) with MDCT. (Table 4) shows a direct comparison of all MDCT and MRI values at the 2 time points.

Finally, we showed that MDCT and CMR demonstrate similar effectiveness of MSC therapy from week 12 to week 24 (Figure 4). Infarct size decreased by mean values of 4.7 ± 1.1 ml and 6.2 ± 2.0 ml when evaluated with DCE-MDCT and DCE-CMR in MSC-treated animals, respectively (p = 0.43). The changes in LVEF with MSC treatment were also apparent with both imaging modalities, with comparable mean values (4.5 ± 1.0 ml and 5.6 ± 1.3 ml, MDCT vs. MRI, respectively; p = 0.4). The changes between the placebo- and MSC-treated animals were also detected with both imaging modalities. Although the detected differences between MDCT and MRI were not significant for all infarct and functional parameters, the changes over time correlated with the assessment (Figure 4).

This is the first study to use cardiac MDCT for the evaluation of therapeutic effects of cell therapy on infarct size, LV volumes, and LV function. Our results suggest that MDCT imaging reliably assessed LVEDV, LVEF, infarct size, and infarct volumes with DCE acquisitions; MDCT imaging was able to assess the effect of intramyocardial-delivered cell therapy compared with placebo. These findings confirmed previous studies that used CMR to evaluate the therapeutic potential of intramyocardial-delivered MSC in animals with HF with chronic infarct scars (6,18). This study introduces MDCT imaging as an alternative cardiac imaging technique to determine efficacy of novel myocardial therapeutics.

MDCT has the capability to distinguish between viable and nonviable myocardium with DCE for the detection of acute and chronic infarct scars, and its use has recently been reported in animal experiments and human studies; however, it is still considered an investigative technique (9- 11,16,19). After intravenous delivery, iodine-based contrast agents accumulate in myocardial tissue damaged by MI. The discrimination between viable and nonviable myocardium results from increased attenuation values caused by the accumulation of iodine molecules in the infarct area detected by MDCT. This is in contrast to enhancement mechanisms by CMR that rely on alteration of the contrast media via interactions with water molecules (9). The infarct sizes detected by DCE-MDCT in this study were significantly smaller than matching CMR values; these results are in accordance with previously published data for acute and chronic MI (16,20). Thus, DCE-MDCT– and DCE-CMR–derived values for infarct size and infarct volume should not be used interchangeably at the current status of MDCT technology.

The extent of contrast enhancement detected by CMR has been shown to predict the response to medical treatment in patients with HF (21). Reduction of infarct size has been noted with DCE-CMR following myocardial cell therapy in animal studies and patients (4- 6,22- 23). Thus, quantification of infarct size and infarct volume by DCE-MDCT could be a valuable surrogate endpoint, allowing the prediction of response to therapy. DCE-MDCT currently offers the highest spatial resolution for transmural characterization of infarct scars, and this can be particularly helpful in guiding intramyocardial delivery of cell-based therapies. An akinetic segment might be deemed nonviable by nuclear techniques using limited spatial resolution but shows some viability in DCE-MDCT, which increases the likelihood that this segment will eventually recover with therapeutic intervention (24).

Evaluation of cardiac function with MDCT is accomplished with intravenous injection of iodine-based contrast agents for opacification of the blood pool to delineate the borders of the ventricle and application of retrospective ECG-gated protocols to record the complete cardiac cycle during the R-R interval. Various investigators have reported a consistent overestimation of LVESV and an underestimation of LVEF by MDCT with 4-slice and early 16-slice technology compared with CMR; however, with the introduction of 64-slice MDCT technology and improvement in temporal resolution, these intermethod differences in global LV function variables are no longer reproduced. Temporal resolution, based on gantry rotation, has been reported to be less than 165 ms with single-source systems and 83 ms with dual-source MDCT technology (7). Our results are in accordance with studies comparing single-source and dual-source 64-slice MDCT—with CMR showing small differences between both modalities and narrow Bland-Altman windows (25- 29). In our study, we observed good correlation coefficients for LVEDV and LVEF, and we were able to confirm the linear relationship between both volumetric imaging techniques for LV function and LV remodeling parameters, making both tomographic imaging methods interchangeable for LV function and LV volume assessment.

Evaluation of the cardiac structure, LV volumes, and LV function—specifically LVEF—is an integral step in determining prognosis and therapy for patients with HF (30). Human trials, using a variety of cell-therapy approaches, have used improvements in LVEF as an endpoint (3). Compared with volumetric changes, using LVEF as an endpoint has the advantage of denoting a survival difference and is an accepted surrogate measure for mortality (31). The magnitude of treatment effect on LVEF seen in these trials has ranged from 5% to 9%, whereas the change in the control groups has ranged from −2% to 4% (32).

As a novel cardiac imaging option for LV function and viability, MDCT offers some solutions for the difficulties faced in CMR. MDCT can be safely performed in patients with metallic implants, such as pacemakers, defibrillators, and prostheses, who are inappropriate candidates for CMR. Although CMR is considered the reference standard for cardiac function and myocardial viability—based on the accuracy and reproducibility of measurements—the duration of CMR acquisitions with multiple prolonged breath-holds are required to obtain adequate datasets, are time consuming, and are impossible in patients with claustrophobia. Thus, CMR is limited to specialized centers, which is in contrast to the widely accessible cardiac MDCT. MDCT imaging is fast and offers good temporal and excellent spatial resolution; however, MDCT acquisitions require iodinated contrast agents and involve radiation exposure.

To translate cardiac function and viability assessment with MDCT into clinical trials, standardized low-dose protocols are needed, especially to minimize overall radiation exposure because follow-up studies are required for outcome assessment. In the latest generation of MDCT scanners, ECG-gated tube current modulations are commonly applied in retrospective acquisitions for evaluation of cardiac function (7). Low-dose DCE protocols have been reported (9). We recently described a prospectively gated protocol for high-resolution DCE-MDCT imaging that lowered the radiation dose by an order of magnitude (33).

In addition to cardiac function and viability, MDCT is able to assess MBF. Cardiac CT perfusion is based on the same first-pass principle as CMR (14), which has been used to evaluate cell therapy (5). Semiquantitative and quantitative assessments of MBF with MDCT have been reported in experimental animals (9,13). Cardiac MDCT is poised to offer a comprehensive evaluation of novel myocardial therapeutics.

This animal study was conducted without any medical therapy given to patients after MI. Therefore, the chronic MIs expanded and the infarct size increased in the placebo group, consistent with results from a clinical study using noninvasive imaging techniques conducted in the early 1980s (34). With modern medical therapy, decreases in infarct size are anticipated in chronic MI, as shown in a recent CMR study (35). To complicate matters, additive effects of angiotensin-converting enzyme inhibition with beta-blockade in combination with cell therapy have been described (36). Although these limitations are evident in the current study, the goal of this study was to demonstrate the usefulness of MDCT to evaluate the effect of myocardial therapies.

However, there are some limitations that may affect the transfer of the reported MDCT protocols and results to human trials. First, the animals presented had elevated heart rates, which are usually considered a relative contraindication for MDCT. Although the reported MDCT protocols proved reliable even under these unfavorable conditions, patients with elevated heart rates above 60 to 70 beats/min may receive negative chronotropic medication to reduce the heart rate for coronary CT angiography. This will affect the value of functional analysis because ventricular volumes change if beta-blockers are applied; however, beta-blockers only have to be given if coronary CT angiography is mandated. Second, no tube current modulation was applied, as is routinely done now in patient studies to reduce radiation exposure. Increased image noise on systolic images of ECG-gated tube current modulation may affect the accuracy of global and regional LV function assessment. Third, because of the passive contrast kinetics of iodine-based contrast agents and the limited collateral circulation in pigs, DCE-MDCT imaging of chronic infarct scars in this animal model required a relatively large dose of iodine to cause a sufficient change in the volume distribution in the myocardial bed. In this study, we used 1.5 times a human equivalent dose, although successful viability imaging in humans has been reported in several studies with standard contrast volumes used for coronary CT angiography (37).

This animal study suggests that cardiac MDCT can be reliably used to evaluate infarct size, infarct volumes, LV volumes, and LV function after intramyocardial-delivered MSC therapy with sensitivity similar to CMR and showing similar changes at follow-up. These findings support the use of cardiac MDCT in preclinical and clinical studies as a noninvasive imaging technique for detecting therapeutic effects of novel myocardial therapies.