Cardiovascular magnetic resonance characterization of peri-infarct zone remodeling following myocardial infarction

All animal studies were approved by the Johns Hopkins University Institutional Animal Care and Use Committee and comply with the "Guide for the Care and Use of Laboratory Animals" (NIH Publication no. 80-23, revised 1985). Fourteen female Göttingen minipigs, at 15 ± 1 months of age, were purchased from Marshall BioResources (North Rose, NY). Myocardial infarcts were created in a closed chest approach by 120 minutes balloon occlusion of the left anterior descending artery, followed by restoration of flow, as described in detail previously [9]. One animal did not survive the infarct procedure and 4 animals died during the follow-up period at day 3, day 4, day 30, and day 90. One Göttingen minipig could not be imaged at all time points because of poor venous accessibility.

CMR images were acquired sequentially -- prior to the infarction procedure -- at baseline (BSN) and at 3, 10, 30, 60, and 90 days post MI, using a 1.5 T MR scanner (CV/i, GE Medical Systems, Waukesha, WI).

Global LV function was assessed using a steady-state free precession pulse sequence with breath-holding acquisitions [10]. A total of six to eight contiguous short-axis slices were prescribed to cover the entire LV from base to apex. Image parameters were as follows: TR/TE = 4.2 ms /1.9 ms; flip angle = 45°; 256 × 160 matrix; 8 mm slice thickness/no gap; 125 kHZ; 28 cm FOV and 1 NSA.

To assess circumferential strain (Ecc), CMR tagged images were acquired with an electrocardiography-gated, segmented K-space, fast gradient- recalled- echo pulse sequence with spatial modulation of magnetization, to generate a grid tagged pattern [11, 12]. Images were obtained at the same location as the cine-CMR images with image parameters as follows: TR/TE = 6.7 ms /3.2 ms; flip angle =12; 256 × 160 matrix, views/s: 4; 8-mm slice thickness/no gap; 31.25 kHz; 28 cm FOV; 1 NSA; and 6 pixels tagging space.

After an intravenous injection of Gd-DTPA (0.2 mmol/kg body weight, Magnevist, Berlex, Wayne, NJ), LGE CMR images were acquired 15 minutes later using an ECG-gated, breathhold, interleaved, inversion recovery, FGRE pulse sequence. LGE- CMR images were acquired in the same location as the short axis cine-images. Imaging parameters were TR/TE/TI = 7.3 ms, 3.3 ms and 180 to 240 ms; Flip angle = 25°; 256 × 196 matrix; 8 mm slice thickness/no gap; 31.2 kHZ; 28 cm field of view (FOV) and 2 NSA. Inversion recovery time was adjusted as needed to null the normal myocardium [13].

CMR images, covering the entire ventricle, were analyzed using a custom research software package (Cine Tool, GE Medical Systems, Waukesha, WI). Infarct size and infarct mass in the LGE- CMR images were defined based on 3 standard deviations of the signal intensity from the remote mean (non-infarct) myocardium for the core infarct, and 2 standard deviations of the signal intensity from the remote mean for the PIZ; the full-width half-max (FWHM) method was also applied, as described in detail previously [6, 7, 14]. Infarct remodeling parameters were evaluated on delayed contrast enhanced CMR images using commercially available imaging analysis software (Sigma Scan® Pro5, Systat Software Inc., San Jose, CA). Infarct expansion and thinning were defined as follows:

Cine CMR images were analyzed with QMass MR® (Medis medical imaging systems, Leiden, The Netherlands). To evaluate global left ventricular (LV) function and volumes, and LV wall motion and wall thickness, endo- and epicardial borders of the LV were defined in the end-diastolic and end-systolic frame in contiguous slices, and then LV parameters and volumes were calculated. We assessed contraction in the mid-wall layer of 24 circumferential locations of the LV myocardium. Tagged images were quantitatively analyzed with a custom software package (Diagnosoft HARP, Diagnosoft Inc., Palo Alto, CA), as described in detail previously [15]. The peak systolic circumferential strain (Ecc) was determined from the strain map of each point. Negative Ecc values represent myocardial contraction, whereas values of increasing strain (toward positive values) reflect worsening of cardiac contractile function in that region. Less negative Ecc values represent hypokinetic myocardium. A value of 0 represents akinetic non-contractile myocardium, whereas a positive value represents dyskinetic myocardial segments.

For follow-up analysis, short axis LGE CMR images were compared to regional strain values. Images were divided in to 24 segments for detailed analysis. Regional strain values were matched with the infarct, PIZ, and remote areas, defined by LGE CMR images. Mechanical segmental dysfunction was then assessed by the circumferential strain uniformity ratio estimate (CURE) measurement, as previously described [16, 17]. Briefly, circumferential strain (y-axis) was plotted versus sector position for the 24 evenly distributed segments in each slice (x-axis), and subjected to Fourier analysis. A perfectly synchronous ventricle provides a CURE value of 1, whereas for a perfectly dyssynchronous heart the CURE index is 0. Although the CURE index has been used previously to quantify dyssynchrony in the whole ventricle, we applied this method to index segmental dysfunction.

Three days post MI the infarct territory was large, accompanied by the largest peri-infarct zone detected by CMR (Figure 1). At 10 days, the infarct territory and the PIZ decreased by 28 ± 5% percent and 46 ± 6% percent, respectively -- using the 2 and 3 standard deviations threshold. We also applied the FWHM criterion and found a decrease from 5.9 ± 0.5 gram to 3.5 ± 0.4 gram at day 3 to day 10, respectively, (p < 0.001) -- which represent a 39.6% in PIZ. There were no differences in decrease with either method (p = 0.41). The myocardial scar expanded and thinned until day 90, consistent with ongoing remodeling processes (Table 2). The infarct scar mass decreased until week 4, and then modestly increased thereafter in accordance with infarct expansion. In contrast, the PIZ decreased continuously until day 90. The results of the absolute and derived values for the infarct size and the peri-infarct zone -- using the 2 and 3 standard deviation threshold --- are shown in Figures 2 and 3. The PIZ analyzed with the FWHM criterion followed the same pattern.

We examined wall motion and end-diastolic wall thickness as regional remodeling parameters by using the American Heart Association 17-segment model. As shown in Figure 4, the end-diastolic wall thickness initially increased after acute MI, in the infarcted territories, and then decreased over time in parallel with a loss of regional function. We analyzed the course of myocardial function in the PIZ in detail by following single slices over time (Figure 4). Since the 17-segment model averages data from large sectors of the left ventricle, we focused on regional strain analysis in single slices divided by the 24 segments (Figure 5). At day 30, Ecc showed increased strain values in the PIZ, compared to the infarct scar (p < 0.05), which is consistent with a tethering effect (of the infarct scar on the viable myocytes) in the PIZ (Figure 5). The time to peak strain also showed a delay between the infarct scar and the PIZ (203.0 ± 11.9 ms vs.171.2 ± 7.4 ms, respectively; p < 0.05). In addition, segmental contractive dysfunction, quantified by CURE, was increased acutely after MI (BSN vs day3; p < 0.05) and showed the most dysfunction 30 days post MI, represented by the lowest CURE value (0.74 ± 0.06).

Cardiac wound repair, after MI, involves temporarily overlapping phases which include an inflammatory phase and tissue remodeling phase. The first phase starts shortly after coronary artery occlusion-- with or without reperfusion-- and involves degradation of normal extracellular matrix, invasion of inflammatory cells at the site of initial injury, and the induction of bioactive peptides and cytokines [19]. Therefore, the early imaged MI substrate is characterized by a high cellularity, which is replaced by dense collagen fibers. These synthetic and degradative events, within the myocardial extra-cellular matrix, occur in a time and region-dependent manner following MI - which is reflected by histological and ultrastructural morphological changes. Within the MI region, the newly formed scar provides a means to tether viable myocyte fascicles, and thereby forms a substrate to resist deformation from the intra-cavitary stresses, generated during the cardiac cycle [20, 21].

Early CMR acquisitions for infarct size quantification in the acute phase are complicated by post-MI edema, which leads to an overestimation of the exact MI-size. Although the assessment of the PIZ may be influenced by the tissue edema, the PIZ (by CMR) still reflects the mixture of viable and infarcted myocardium.

Different techniques for analysis have been previously used to determine the PIZ in CMR. Specifically, the assessment of the PIZ is either defined with threshold techniques using standard deviations to the remote myocardium [8], or based on modified FWHM criterion applying different signal intensities [7, 22]. All of these methods claim to match the visual assessment of the PIZ by LGE-CMR. However, with the lack of quantitative histological reference, standard to index the heterogeneity of the PIZ, it is difficult to determine which method is superior over the other. In our analysis, we chose to use the 2 and 3 standard deviations threshold for this report and we found the best agreement with our visual assessment of the PIZ on LGE CMR images. In addition, we quantitatively assessed the PIZ with the FWHM criterion and found a similar remodeling pattern. We discovered that the quantified PIZ mass was larger with the FWHM method, compared to the threshold technique -- which confirms previous reports [7, 8].

CMR has been used for risk stratification for ventricular arrhythmias following MI. Studies conducted in the 1980s show that a low LVEF predicts the risk of death after MI [23, 24], and this data is currently used as one of the main criteria for ICD implantation [25]. Subsequent studies demonstrate that infarct surface area and mass-- measured by CMR-- can more accurately identify patients at risk and present a substrate for monomorphic VT, when compared to LVEF [26]. The ability to perform more detailed morphological analysis of the infarct scar will allow refined information on risk stratification [7, 8].

Histologically, the early period of remodeling is characterized by the ongoing changes of the extracellular matrix and cell composition of tissue scar substrate, whereas the later period is characterized by a fixed substrate of cardiomyoctes and well-defined collagen scar [ref]. We demonstrate that the PIZ--assessed by LGE-CMR imaging-- undergoes significant changes early after MI as well. Until day 10, the PIZ decreases rapidly in mass and the PIZ mass stabilizes thereafter. Our data suggests that after 4 weeks post- MI, the PIZ can reliably evaluate structural remodeling. Additional information for regional strain analysis was acquired, which provides valuable information about the tensile forces working on the scar tissue and in the PIZ.

Recently, Fernandes et al. showed that enhanced PIZ function, defined as greater Ecc and earlier time to peak Ecc, was associated with inducibility in patients with post MI heart failure [27]. In our research, we did not see enhanced PIZ function at any particular time point during the remodeling process up to 90 days. However, at 30 days post-MI, we did observe a significant difference in Ecc and time to peak Ecc, between the infarct scar and PIZ, suggesting heightened mechanical stress working at the interface of viable and infarct tissue (Figure 6) [28, 29]. Tensile forces working on the viable myocytes -- neighboring the scar tissue -- have been proposed for decades as crucial patho-physiological mechanisms, involved in post-MI remodeling and infarct expansion [18]. Until now, this concept could not be experimentally verified using non-invasive imaging methods. A recent study demonstrated that circumferential strain rates are independent predictors of outcomes after MI and are predictive of cardiac remodeling following MI [30]. However, Sasano et al. reported the highest inducibility in a porcine model, 30 days after MI-induction [31]. This is the same time-point that we observed the highest difference of strain between infracted area and PIZ in our pigs. The study of Sasano et al. and our findings suggest that the underlying pathophysiology (of post-MI VT) can be elucidated with imaging techniques.

The most significant limitation in this animal study was the inability to perform serial VT inducibilty studies with programmed stimulation. Having this data would help to determine the relationship between the imaged tissue substrate and susceptibility to VT. Another limitation was that all of the infarcts were created in the LAD territory, and were transmural. Strain patterns in the PIZ for non-transmural infarcts are expected to be very different from those in transmural infarcts, but our research did not explore this theory. Further, we did not acquire T2-weighted CMR images to quantify the post-MI edema, nor did we use any medications that are currently administered to patients in the post-MI period.