A reperfused acute MI was experimentally induced in 12 three-month-old castrated large male white pigs by using an angioplasty-based ischemia/reperfusion model. Acute MI was created as previously described.16–18 Continuous electrocardiography, blood pressure monitoring, and pulse oximetry were available for all animals. Anesthesia was induced by intramuscular injection of ketamine (20mg/kg), xylazine (2mg/kg), and midazolam (0.5mg/kg) and was maintained by intravenous administration of midazolam (0.2mg/kg/h). Buprenorphine (0.3mg/kg) was used for analgesia. All animals were intubated and mechanically ventilated with oxygen (FiO2, 28%). Continuous infusion of amiodarone (300mg, 150mg/h) was maintained during the procedure as prophylaxis for malignant ventricular arrhythmias. The animals were anticoagulated by 300 IU/kg heparin administration (an activated clotting time was above 250seconds in all cases).

The study design is summarized in Figure 1. The percutaneous procedure included a femoral arterial 6-F line and an over-the-wire coronary angioplasty balloon (Voyager, Abbott) that was placed in the midleft anterior descending artery (LAD) (n = 4), in the proximal left circumflex (LCx) (n = 4) or in the proximal right coronary artery (RCA) (n = 4). As in previous ischemia/reperfusion models in our group,17,18 the LAD was occluded after the first diagonal branch to reduce mortality secondary to arrhythmias or cardiogenic shock compared with proximal LAD occlusions while still resulting in large infarctions. After a coronary angiogram confirmed correct catheter position, the pigs were carefully taken to the CMR suite located wall-to-wall with the experimental catheterization laboratory. An in vivo perfusion CMR-based criterion standard was performed during selective intracoronary gadolinium injection at the subsequent coronary occlusion site to delineate the AAR (Video 1 of the supplementary material) and the balloon was then inflated to induce an acute MI for 40minutes. After 7 days post-MI, a follow-up CMR was performed and subsequently all animals were sacrificed to validate the CMR-based criterion standard against classic Evans blue staining. All animals received experimental care in accordance with the Guide for the Care and Use of Laboratory Animals. The study protocol was approved by the Institutional Animal Research Committee at the Spanish National Center for Cardiovascular Research (CNIC).

Figure 1.

Study design. Cathlab, catheterization laboratory, CMR, cardiac magnetic resonance; LAD, left anterior descending artery; LCx, left circumflex artery; RCA, right coronary artery; T2W-STIR, T2-weighted short tau triple-inversion recovery.

(0.09MB).

All studies were performed using a 3 Tesla (3T) Philips Achieva-TX whole body scanner (Philips Healthcare, Medical Systems, Best, The Netherlands) equipped with a 32-element cardiac phased-array surface coil using electrocardiographic gating. Anesthesia was induced and maintained as described above in each study. In the 7-day post-MI CMR, the following procedures were performed: SSFP sequences for function (11-13 contiguous left ventricle (LV) short-axis slices; field of view 280 x 280mm; slice thickness 8mm without gap; echo time/repetition time 1.4/2.8ms, flip angle 45°; cardiac phases 25; voxel size 1.8 x 1.8mm; SENSE factor 1.5; NEX 3) and black blood T2W-short tau triple-inversion recovery (T2W-STIR) sequences for edema (11-13 contiguous LV short-axis slices; field of view 280 x 280; slice thickness 8mm without gap; TE 80ms; voxel size 1.4 x 1.4mm; echo-train length 16; NEX 2). A coil sensitivity correction algorithm for all T2W images was implemented in the scan acquisition. The T2W-STIR sequence was tuned for the 3T scanner by adjusting the inversion time to 200ms to correct from longest T1 at 3T and the acquisition was triggered for every 2 beats to avoid signal saturation due to very short repetition time. Homogenous excitation was ensured by multichannel excitation technology. Thus, the results reported here are applicable to 1.5 T scanners.

In addition, a CMR-based AAR criterion standard was acquired before MI induction using perfusion imaging (field of view 280 x 280mm; slice thickness 8mm without gap; flip angle 15°; TR 2.8ms; TE 1.4ms; saturation time 45ms; voxel size 2.4 x 2.4) and inversion recovery gradient-echo sequences as those typically used for scar assessment (11-13 contiguous short-axis slices; field of view 280 x 280mm; slice thickness 8mm without gap; flip angle 15°; TR 5.6ms; TR 2.8ms; voxel size 1.6 x 1.6mm; inversion time 500ms, shot interval, 2 beats) during selective coronary injection of gadolinium. Inversion recovery images were obtained in addition to intracoronary gadolinium perfusion images to improve the spatial resolution for border delineation. Preliminary experiments were performed to determine the best intracoronary gadolinium concentration and infusion volume/rate. Accordingly, all animals received 0.2 mmoL/kg gadolinium-based contrast media through the coronary catheter at 2mL/s for intracoronary gadolinium perfusion and at 0.15mL/s for inversion recovery imaging. All short-axis slices matched those of the cine imaging within the same study and, similarly, follow-up images were done with the same slice thickness and position as baseline according to anatomical landmarks.

Cardiac magnetic resonance images were analyzed using dedicated software (QMASS 7.5, Medis, Leiden, The Netherlands). Basal LV slices proximal to the site of coronary artery occlusion without AAR or apical LV slices with slow flow and/or partial volume artifacts were excluded. Border delineation on the CMR-based criterion standard for anatomical AAR delineation was mainly performed on inversion recovery images due to better contour visualization. However, special care was taken to confirm concordance between the extent of hyperenhanced areas on intracoronary gadolinium perfusion and inversion recovery images (Figure 2). The epicardial and endocardial contours in each LV short axis were manually traced including papillary trabeculations within the cavity. Areas at risk on T2W-STIR images were identified as hyperenhanced regions using the full-width half-maximum method8 and the extent of AAR was calculated with respect to the total LV slice area (%). Extreme care was taken to avoid including any artificially high signal intensity due to inadequately suppressed slow flow within the cavity space. If present, a central core of hypointense signal within the area of increased signal was included in the T2W-STIR analysis. Additionally, regions of interest sized 0.5mm2 were manually drawn on T2W-STIR images to calculate the average signal intensity for all pixel values within the hyperenhanced area (AAR) and the normal signal area (remote myocardium). Hypointense areas within the AAR in cases of microvascular obstruction or hemorrhage were carefully avoided. Then, the signal intensity change between the AAR and remote myocardium for each study was calculated as (signal intensityAAR – signal intensityremote)/signal intensityremote.9,19

Figure 2.

Cardiac magnetic resonance-based criterion standard images in the left anterior descending artery, left circumflex and right coronary artery myocardial territories using intracoronary gadolinium perfusion and inversion recovery sequences to visualize the anatomical area at risk.

(0.44MB).

All animals were transferred to the catheterization laboratory for AAR assessment by Evans blue staining after the 7-day CMR study. The LAD (n = 4), LCx (n = 4) or RCA (n = 4) was reoccluded by angioplasty balloon inflation at the same level as the initial occlusion, following the same angiographic landmarks. Then, 100mL of Evans blue dye (5%) were infused into the LV cavity through a pigtail catheter while the balloon was inflated. With this technique, nonischemic areas appear blue, whereas the AAR remains unstained.20 The animals were then euthanized with a lethal injection of pentobarbital sodium after heparin administration and their hearts were excised. After 3hours of freezing, the hearts were sliced into 8-mm section slices using a commercial rotator blade, and midventricular consecutive slices were selected as they matched to the 8-mm short axis CMR images of interest. Finally, high-resolution photographs were taken using a digital camera (Pentax K-m) and ImageJ software v.1.45s (National Institutes of Health, Bethesda, Maryland, United States) was used for AAR quantification. The AAR was expressed as a percentage of the total slice LV area (%) after manual delineation of cardiac contours.

Continuous variables are expressed as mean ± standard deviation. The correlations between the measures were assessed by Spearman rank order coefficients and Bland-Altman plots were used to study the agreement. To take into account repeated measures and clustering (several slices by animal), analyses were performed using the mixed-model procedure. A P value < .05 was considered statistically significant. All statistical tests were performed with IBM SPSS Statistics software, v.20.0 (SPSS, Inc, Chicago, Illinois).

The agreement between the CMR-based criterion standard and T2W-STIR in the different coronary artery territories is summarized in Figure 3. The correlation between the CMR-based criterion standard AAR delineation and T2W-STIR-based AAR quantification for LAD and RCA infarctions was high, with an r coefficient of 0.73 (P = .001), mean error 0.50% and limits of agreement of −12.68% and 13.68% for LAD infarctions and an r coefficient of 0.87 (P = .001), mean error of −1.5% and agreement limits of −8.0% and 5.8% for RCA infarctions. Conversely, the correlation for LCx territories was poor (r = 0.21; P = .37), showing a higher mean error and wider limits of agreement for this territory. Similar results were obtained on comparison of the classic histopathology criterion standard and T2W-STIR images (Figure of the supplementary material). In terms of intensity values, statistically significant differences were found between LAD and RCA infarctions (224.5 ± 36.5% vs 153.1 ± 90.3%; P = .003), and the differences were even higher between LAD and LCx infarctions (224.5 ± 36.5% vs 41.7 ± 23.1%; P < .001) (Figure 4).

Figure 3.

Agreement between the CMR-based criterion standard and T2W-STIR images according to each coronary artery territory in animals undergoing 40minutes of ischemia/reperfusion. Spearman correlation is shown in upper panels and Bland-Altman plots in lower panels where the blue line represents the mean error and black lines the limits of agreement. AAR, area at risk; CMR, cardiac magnetic resonance; LAD, left anterior descending artery; LCx, left circumflex artery; LV, left ventricle, RCA, right coronary artery; T2W-STIR, T2-weighted short tau triple-inversion recovery.

(0.32MB).

Figure 4.

T2W-STIR signal intensity values in animals with ischemia/reperfusion affecting different ventricular territories. Statistically significant differences were found between the signal intensity values for infarctions affecting the LAD (224.5 ± 36.5%) and RCA (153.1 ± 90.3%); P = .003. Differences between LAD and LCx infarctions were even higher (224.5 ± 36.5% vs 41.7 ± 23.1%; P < .001). Bars represent mean ± standard deviation. LAD, left anterior descending artery; LCx, left circumflex, RCA, right coronary artery; T2W-STIR, T2-weighted short tau triple-inversion recovery. *P < .001. **P = .003 for comparisons with the LAD Group.

(0.06MB).

There was highly significant good correlation (r = 0.84; P < .001) and agreement (mean error = 0.91%; limits −7.55%-9.37%) between the extent of AAR as measured by the CMR-based criterion standard (intracoronary gadolinium perfusion) and by the histological Evans blue evaluation (Figure 5). Similarly, the extent, shape, and contours of the hyperintense area depicted by both standards and by T2W-STIR images were strongly concordant for each of the coronary artery territories (Figure 6).

Figure 5.

Concordance between CMR-based criterion standard and classic histological standard (negative Evans blue staining) for AAR delineation. Left panel shows the scatter plot and right panel illustrates Bland-Altman plots where blue line represents the mean error and black lines the limits of agreement. AAR, area at risk; CMR, cardiac magnetic resonance; LV, left ventricle.

(0.14MB).

Figure 6.

Comparison between the extent of area at risk visualized by the CMR-based criterion standard, T2W-STIR at day 7, and the pathology staining in each coronary territory. CMR, cardiac magnetic resonance; T2W-STIR, T2-weighted short tau triple-inversion recovery.

(0.39MB).

We used a black blood T2W-STIR sequence for edema depiction. Alternative sequences (T2W-bright blood sequences, T2P-SSFP, or T2 and T1 mapping) have recently been proposed for imaging myocardial edema and could potentially overcome the limitations of T2W-STIR for AAR quantification in the lateral wall. However, T2W-STIR imaging is currently the most commonly used method to assess the presence and extent of myocardial edema both in the clinical scenario24 and, more importantly, in clinical trials with the salvage index as endpoint.25–29 Given that this study was designed to explore the accuracy of the classic T2W-STIR imaging for AAR quantification according to the MI territory, T2-mapping sequences were not included in the imaging protocol. Measurements of collateral flow were not available in this study, which could have influenced the AAR size. However, the swine model of acute reperfused MI is known to have very little or no collateral flow.30 This is an experimental model with a small sample (n = 12) and thus human validation studies are needed to confirm our findings.