2016 Volume 80 Issue 12 Pages 2513-2519
Background: Coronary revascularization has been shown to induce left ventricular (LV) reverse remodeling (RR). The serial morphologic changes in enhanced necrotic tissue during RR on cardiac magnetic resonance imaging (CMR) have not been investigated.
Methods and Results: This retrospective study included 26 patients with severe LV systolic dysfunction (ejection fraction [EF], <35% on echocardiography) who underwent CMR before and >6 months after surgical revascularization. Of 26 patients, 20 had a reduction of ≥10% in end-diastolic and end-systolic volumes (classified as RR group). The RR group had improvement in EF after revascularization (28.8±6.6% vs. 40.6±7.8%, P<0.0001), and no change in absolute infarct mass (17.3±10.9 g vs. 17.5±10.4 g, P=0.8), but an increase in relative infarct mass (21.0±13.7% vs. 26.5±19.4%, P=0.01) due to reduction of myocardial mass after revascularization. Significant increase in regional transmural extent (30.3±21.6 vs. 42.6±22.8, P<0.0001) and in thickness of enhanced tissue (4.2±1.5 mm vs. 5.9±1.8 mm, P<0.0001) was found in the RR group. No significant differences were observed in any of the variables in the non-RR group.
Conclusions: In patients with chronic myocardial ischemic dysfunction, significant volume reduction after revascularization led to significant increase in regional transmural extent of the enhanced area without a change in absolute infarct mass, on CMR. (Circ J 2016; 80: 2513–2519)
Inversion-recovery cardiac magnetic resonance imaging (CMR) after administration of extracellular contrast agents allows the direct visualization of the transmural extent of scars, owing to its high spatial resolution and the high contrast between infarcted and non-infarcted myocardium.1–4 Thus, by preoperatively assessing the transmural extent of enhanced necrotic myocardium on late gadolinium enhancement (LGE) CMR, the functional recovery of dysfunctional segments after surgical revascularization in patients with chronic ischemic left ventricular (LV) dysfunction can be predicted.5–10 CMR in these particular patients frequently shows LV chamber dilatation with significant thinning of infarcted regions. These imaging features are attributed to LV remodeling after acute myocardial infarction (MI), a well-known process involving the expansion of infarcted segments in the early phase, and time-dependent dilatation, distortion of ventricular shape, and volume-overload hypertrophy of non-infarcted segments in the late phase.11–17 LV remodeling has been associated with a worse prognosis in patients with heart failure.18 Therefore, many investigators have focused on the sequential changes of infarct tissue with myocardial remodeling, and have identified several predictive factors for LV remodeling after MI.11–14,19
Editorial p 2427
In contrast, reverse remodeling (RR) refers to the regression of pathologic myocardial hypertrophy, chamber shape distortions, and LV dysfunction.20 Coronary revascularization has been shown to induce LV RR, especially in patients with a substantial amount of viable myocardium.21 Indeed, sequential CMR after surgical revascularization has indicated significant improvements in global LV systolic function and reduction of LV volume in a subset of patients with viable myocardium, indicative of myocardial RR. Until now, however, there have been no reports on the changes in infarcted and non-infarcted myocardium during the process of RR using CMR.
In this study, using gadolinium-enhanced CMR, we analyzed the morphologic changes in infarcted and non-infarcted myocardium during the course of RR. Specifically, we focused on the serial changes in enhanced necrotic tissue and non-enhanced non-infarcted tissue after surgical revascularization in patients with ischemic LV dysfunction, using CMR as a quantification tool.
The institutional review board approved this retrospective study and waived the requirement for informed consent.
Between November 2005 and March 2012, we retrospectively reviewed 35 patients who satisfied the following 3 inclusion criteria: (1) >70% epicardial coronary stenosis in at least 2 vessels on coronary angiography; (2) LV ejection fraction (LVEF) ≤35% on echocardiography; (3) availability of both preoperative and postoperative contrast-enhanced CMR; and (4) the presence of delayed enhancement on preoperative CMR. Among them, 9 patients were excluded due to either no delayed enhancement on CMR (n=4), or poor image quality (n=5). Finally, 26 patients (M/F=23/3; mean age, 64.1±9.1 years; range, 49–78 years) were included in this study. Baseline patient characteristics are listed in Table 1. Patients with acute MI were not included in this study. Complete revascularization was performed in all patients. All patients were divided into either the RR group or non-RR group. RR was defined as reduction of ≥10% in end-diastolic volumes (EDV) and end-systolic volumes (ESV) as measured on CMR.22
Data given as mean±SD or n (%). MI, myocardial infarction.
All CMR was performed with a 1.5-T unit (Sonata Magnetom; Siemens, Erlangen, Germany) using a phased-array body surface coil. CMR was acquired during repeated end-expiratory breath-holds and was electrocardiographically gated. Postoperative CMR was obtained at least 6 months after surgery to avoid the effect of early postoperative myocardial dysfunction related to the operation itself. Mean time interval between the surgery and postoperative CMR was 21±14 months. There were no events between the 2 examinations in all patients.
After localizer imaging, cine true fast imaging with steady-state precession (TrueFISP, repetition time/echo time, 2.2 ms/1.1 ms; flip angle, 80°; typical pixel size, 1.6×1.3 mm; slice thickness, 6 mm; slice gap, 4 mm; and temporal resolution, 42 ms) was performed in 3 long-axis planes (2-, 3-, and 4-chamber views) of the heart. A total of 10–11 parallel short-axis sections per patient were obtained starting from the mitral annulus to the apex, covering the entire LV for volumetric analysis. The most basal section was precisely adjusted to be perpendicular to the mitral annulus plane at 4-chamber and 2-chamber end-diastolic (ED) phases in each patient. Contrast-enhanced imaging was acquired in the same orientation as the cine images 10 min after i.v. administration of 0.2 mmol/kg gadopentetate dimeglumine (Magnevist, Schering AG, Berlin, Germany), using the inversion recovery segmented spoiled-gradient echo and phase-sensitive inversion recovery methods (repetition time/echo time, 9.1 ms/4.2 ms; flip angle, 25°; typical pixel size, 1.6×1.3 mm; slice thickness, 6 mm; slice gap, 4 mm).
A single experienced radiologist (E.-A.P with 9 years of cardiovascular imaging experience), unaware of patient information, performed the following LV measurements: LV function, LV mass, infarct quantification, and myocardial thickness. LV function and mass were analyzed on a computer using dedicated software (QMASS MR; Medis, Leiden, the Netherlands). Cine loops were reviewed, and ED and end-systolic (ES) frames were identified. The most basal section was required to show ≥50% visible myocardial circumference at the mitral valve level in order to be included. Epicardial and endocardial contours were manually traced from the most basal to the most apical slice. Papillary muscles and trabeculae were included in the cavity volume and excluded from the myocardial mass. LV EDV and ESV, stroke volume, cardiac output, and EF were measured and indexed to body surface area. LV mass at ED and ES phases was averaged to obtain the myocardial mass.
Infarct mass was quantified using the 6-SD threshold method on phase-sensitive inversion recovery gradient echo imaging using a dedicated semiautomatic analysis program (cmr42; Circle Cardiovascular, Calgary, Canada). Determination of the most basal section and endocardial contour was performed in the same fashion as that for ventricular volumetrics. The absolute infarct mass was calculated by summing the volume of the hyperenhanced regions in all slices multiplied by 1.05 g/ml. Subsequently, relative infarct mass was expressed as a proportion of the total LV myocardium (% infarct mass). In terms of the analysis of regional transmurality, we used a model in which the LV was divided into 12 circumferential segments in 2–4 representative short axis views in every patient (Figure 1). According to transmural extent of hyperenhanced area on preoperative CMR, segments were classified into 4 groups: 1–25%, 26–50%, 51–75% and 76–100%.
Infarct mass quantification using the 6-SD threshold method. (A) Phase-sensitive inversion recovery image showing delayed enhancement in the lateral wall, indicating necrotic infarct tissue. (B) Endocardial and epicardial borders are manually traced. The delayed enhancement area is overlaid in yellow for pixels with signal intensity higher than the threshold value, 6 SD above the mean. Based on the overlaid pixels, the infarct area is automatically calculated. Regional transmural extent of enhanced regions was calculated using a model in which the left ventricle was divided into 12 circumferential segments.
Myocardial thickness and circumference of the LV were determined on 3 contiguous mid-ventricular sections of inversion recovery gradient echo imaging using electronic calipers on a PACS workstation (Radmax; Marotech, Seoul, Korea) with three 2,048×1,536-pixel 20.8-inch monochrome liquid crystal display monitors (ME315L; Totoku Electric, Tokyo, Japan). Measurements of wall thickness were obtained from the enhanced region of the infarcted segment, non-enhanced rim of the infarcted segment, and remote normal myocardium.15,16,23 Regional spoke lengths were obtained from regions that contained no papillary muscles. LV circumferential length was measured as a curved length of the endocardial border in the enhanced and non-enhanced regions separately.
Descriptive statistics are presented as mean±SD. Wilcoxon matched-pairs signed rank test or paired t-test was used to compare preoperative and postoperative parameters. Mann-Whitney U-test was used to compare the RR and non-RR groups. Fisher’s exact test or Mann-Whitney U-test was used to compare the RR and non-RR groups. All statistical analysis was performed using MedCalc for Windows, version 22.214.171.124 (MedCalc Software, Mariakerke, Belgium). Differences were considered significant for P<0.05.
There were no significant differences in age (RR group, 64.7±9.0 years vs. non-RR group, 62.3±9.9 years, P=0.573), sex (M/F: RR group, 18/2 vs. non-RR group, 5/1, P=1.0), or follow-up duration (RR group, 17.7±12.4 months vs. non-RR group, 14.5±4.3 months, P=0.573) between the 2 groups.
Table 2 lists the serial changes in LV function and mass after revascularization in the 2 groups. The RR group had significant improvement in EF and reduction of EDV and ESV after revascularization (all, P<0.0001): average increase in EF was 11.8% and average reduction of EDV and ESV was 33.4% and 44.1%, respectively. In contrast, the non-RR group had no significant improvement in EDV (P=0.173), with a reduction of 3.4% on average. In the non-RR group, although small, ESV decreased by 10.4% on average (P=0.028), resulting in significant improvement in EF by 5.1% (P=0.028), but the magnitude of LVESV reduction (44.1±11.8% vs. 10.5±5.3%, P<0.001) and EF improvement (11.8±6.8 vs. 5.1±3.4, P=0.019) were significantly greater in the RR group than in the non-RR group. There were no significant differences in LV volume or function on preoperative MRI between the 2 groups (all, P>0.05).
Data given as mean±SD. †Wilcoxon matched-pairs signed rank test. CMR, cardiac magnetic resonance imaging; EDV, end-diastolic volume; ESV, end-systolic volume; RR, reverse remodeling.
Myocardial mass was significantly reduced in the RR group after revascularization (P=0.003) and the average reduction in mass was 7.3 g/m2. No changes, however, were found in the non-RR group (P=0.249). The myocardial mass on preoperative imaging was not significantly different between the 2 groups (P=0.108).
The RR group had no changes in absolute infarct mass but a significant increase in relative infarct mass due to reduction in the myocardial mass after revascularization (Table 3). No changes were observed, however, in absolute or relative infarct mass in the non-RR group.
Data given as mean±SD. Abbreviations as in Table 2.
Table 2 also lists the serial changes in transmural extent of hyperenhanced area after revascularization. The RR group had significant increase in the transmural extent of hyperenhanced area after revascularization in segments with ≤75% transmural extent, but no significant change in segments >75% transmural extent. In contrast, the non-RR group had no significant change in transmural extent after revascularization in any group of segments.
A significant increase in the thickness of the enhanced region was found in the RR group after revascularization: 4.2±1.5 mm vs. 5.9±1.8 mm (P<0.0001; Figure 2; Table 4). There were no significant changes, however, in the thicknesses of the non-enhanced rim of infarcted segments and remote normal myocardium after revascularization (all, P>0.05). In the non-RR group, no differences were found in myocardial thickness between preoperative and postoperative scans (all, P>0.05; Figure 3).
Representative magnetic resonance imaging in a 49-year-old man showing reverse remodeling after revascularization. After revascularization, left ventricular (LV) end-diastolic (ED) volume markedly decreased from 243.3 to 124.4 ml/m2 with a reduction of 48.9%. LV end-systolic volume also dramatically decreased from 194.0 to 81.4 ml/m2 with a reduction of 58.0%. Although there were no significant changes in the thickness of non-enhanced regions, enhanced regions (arrowheads) were much thicker on postoperative than on preoperative imaging. CH, chamber; DE, delayed enhancement.
Data given as mean±SD. LV, left ventricular. Other abbreviations as in Table 2.
Representative delayed enhancement (DE) magnetic resonance imaging in a 69-year-old man without reverse remodeling after revascularization. Preoperatively, left ventricular (LV) volume was markedly dilated with end-diastolic (ED) volume of 156.5 ml/m2 and end-systolic (ES) volume of 109.8 ml/m2. Reduction of ED and ES volumes was minimal, 1.6% and 11.9% respectively, after revascularization. There were no significant changes in the thickness of enhanced regions (arrowheads). CH, chamber.
After revascularization, mean subendocardial circumferential length of both enhanced and non-enenhanced regions was dramatically decreased in the RR group (Table 4). Magnitude of reduction of circumferential length in the enhanced (23.7±18.1% vs. 1.9±3.3%, P<0.0001) and non-enhanced regions (12.5±14.4% vs. 0.6±5.0%, P=0.003) was significantly greater in the RR group than the non-RR group.
The principal findings of the present study can be summarized as follows: (1) although there were no significant changes in absolute infarct mass after revascularization, a reduction in myocardial mass due to RR led to an increase in relative infarct mass; (2) in the RR group, significant volume reduction after revascularization led to a significant increase in the transmural extent of hyperenhanced area in segments ≤75% of transmural extent but no significant change in segments >75% of transmural extent; and (3) the RR group had significant increase in the thickness of enhanced necrotic tissue without changes in the thickness of the non-enhanced rim of infarcted segments and remote non-enhanced myocardium after revascularization. Overall, in the RR group, wall thickness of the non-enhanced region remained unchanged, while its circumferential length decreased. This contributed to decrease in mass. In contrast, in the enhanced regions, circumferential length decreased but infarct mass remained unchanged. This could explain the increase in the thickness of enhanced necrotic tissue.
Structural and histologic evidence of LV RR achieved via mechanical unloading with the use of ventricular assist devices support the present data.18,24–27 Ventricle dimension rapidly reduced, resulting in total decompression of wall stress, often by 70%, immediately after ventricular assist device placement.24,25 In another study, Bruckner et al confirmed the return of the myocyte to its normal length and mass at the cellular level, showing an average reduction of 26% in myocyte size.26 CMR is ideally suited for the assessment of RR because it is the current gold standard for measuring cardiac volumetry and mass, and offers the ability to track changes in ventricular metrics with low interobserver variability.16,28 Moreover, its non-invasiveness makes it suitable for repetitive analysis. We also found significant decrease in LV mass as well as volume reduction after coronary revascularization in the RR group on CMR. This indirectly suggests that myocyte hypertrophy may have regressed in the RR group. We also noted similar clinical results on RR after aortic valve replacement in patients with chronic aortic regurgitation.29–31 Their study noted normalization of LV volume and a decrease in LV mass to near normal after aortic valve replacement.29,31
Interestingly, we found that the increases in the thickness of infarcted segments in patients with significant LV volume reduction after revascularization resulted from an increase in the thickness of enhanced necrotic tissue, but not in the viable non-enhanced rim. This observation is contrary to the belief that an increase in wall thickness of the infarct segment showing function recovery after revascularization would have resulted from an increase in the thickness of viable non-enhanced myocardium. There are 2 possible explanations for this. The first is the inverse process of infarct expansion. Infarct expansion, occurring within hours of myocyte injury, results in wall thinning and ventricular dilatation. The thinning of the infarct region is a consequence of stretching and slippage between muscle bundles, resulting in reduction of the number of myocytes across the infarct region.11 During the course of healing, connective tissue cells enter the myocyte compartment and connect to disrupted myocyte fibers, providing resistance to further stretching.11 In this scenario, during the inverse process, or “RR”, an increase in the thickness of the infarct region after revascularization might be explained by reverse slippage. It is unlikely, however, because whereas infarct tissue is a dead but active dynamic structure in the acute stage, it becomes stable and relatively unchanged at the chronic stage.19 Therefore, the second explanation, passive thickening of chronic infarct scars due to loss of stretching after volume reduction, is more plausible. This may be described as similar to rubber returning to its original condition after it is stretched and released.
The present result raises an issue regarding the interpretation of viable myocardium via analysis of the transmural extent of necrotic tissue in patients with severe LV remodeling on LGE CMR. Ichikawa et al also raised the same issue on using the transmural extent of enhanced necrotic tissue for predicting preserved regional contractile function, given that the optimal threshold of percent transmural enhancement may be different between acute and chronic MI.32 According to the present data, infarcted segments with smaller transmural extent had a tendency toward greater increase in transmural extent after revascularization. In contrast, infarcted segments with transmural extent >75% showed no significant change after revascularization. Therefore, infarcted segments with transmural extent between 26% and 75% can be troublesome when viable myocardium is defined as transmural extent <50% or <75% in patients with chronic LV dysfunction.
There were some limitations to the present study. First, the quantification of infarct mass and thickness may have been influenced by the signal intensity, depending on the degree of contrast enhancement. Individual hemodynamic differences may have influenced the results, even though an identical protocol for contrast media injection and imaging was used in all patients to minimize this effect. Second, measuring the thickness on LGE imaging is largely dependent on the cardiac cycle during image acquisition. In this regard, all LGE imaging was carried out during the mid-diastole phase. Third, the number of patients in the non-RR group was not sufficient for adequate statistical analysis. Fourth, due to the retrospective nature of this study, selection bias may have been present. This study, however was not designed to investigate the predictive factors for RR but rather to explore the influence of RR on infarct tissue after revascularization. Thus, selection bias owing to the retrospective design may have been less affected in the latter exploration than in the former investigation. The last limitation was the lack of pathology confirmation regarding RR. Future animal studies are warranted to validate this concept on pathology.
In conclusion, CMR was able to show that in patients with chronic myocardial ischemic dysfunction, significant volume reduction after revascularization led to significant increase in transmural extent of enhanced area without a change in absolute infarct mass. Significant increases in transmural extent of enhanced area were found only in segments with ≤75% transmural extent. This suggests that the transmural extent of infarcted tissue can change depending on ventricular status, such as ventricular remodeling after infarction or RR after revascularization, and therefore, increased transmural extent of infarcted tissue after revascularization in patients with LV dysfunction should be carefully interpreted in the context of RR.
We are indebted to Dr Seung-Pyo Lee, a cardiologist for his help, support, and advice. We would also like to acknowledge Chris Woo, BA, for assistance with the manuscript.