Enhancement patterns detected by multidetector computed tomography are associated with microvascular obstruction and left ventricular remodelling in patients with acute myocardial infarction

Abstract

Aims

This study evaluated the clinical value of myocardial contrast-delayed enhancement (DE) with multidetector computed tomography (MDCT) for detecting microvascular obstruction (MVO) and left ventricular (LV) remodelling revealed by DE magnetic resonance imaging after acute myocardial infarction (AMI).

Methods and results

In 92 patients with first AMI, MDCT without iodine reinjection was performed immediately following successful percutaneous coronary intervention (PCI). Delayed-enhancement magnetic resonance imaging performed in the acute and chronic phases was used to detect MVO and LV remodelling (any increase in LV end-systolic volume at 6 months after infarction compared with baseline). Patients were divided into two groups according to the presence (n = 33) or absence (n = 59) of heterogeneous enhancement (HE). Heterogeneous enhancement was defined as concomitant presence of hyper- and hypoenhancement within the infarcted myocardium on MDCT. Microvascular obstruction and LV remodelling were detected in 49 (53%) and 29 (32%) patients, respectively. In a multivariable analysis, HE and a relative CT density >2.20 were significant independent predictors for MVO [odds ratio (OR) 13.5; 95% confidence interval (CI), 2.15–84.9; P = 0.005 and OR 12.0; 95% CI, 2.94–49.2; P < 0.001, respectively). The presence of HE and relative CT density >2.20 showed a high positive predictive value of 93%, and the absence of these two findings yielded a high negative predictive value of 90% for the predictive value of MVO. Heterogeneous enhancement was significantly associated with LV remodelling (OR 6.75; 95% CI, 1.56–29.29; P = 0.011).

Conclusion

Heterogeneous enhancement detected by MDCT immediately after primary PCI may provide promising information for predicting MVO and LV remodelling in patients with AMI.

Acute myocardial infarction, Microvascular obstruction, Multidetector computed tomography, Delayed-enhancement magnetic resonance imaging, Percutaneous coronary intervention

Introduction

Primary percutaneous coronary intervention (PCI) has improved the prognosis of patients with acute myocardial infarction (AMI).^1,2 However, despite successful recanalization of the culprit vessel, left ventricular (LV) remodelling occurs in an appreciable proportion of patients with AMI.³ Left ventricular remodelling following AMI is a major predictor of morbidity and mortality for overt congestive heart failure and life-threatening arrhythmias.⁴ Left ventricular remodelling is predicted by the extent of transmurality and infarct size after AMI.⁵ Late gadolinium-enhanced cardiac magnetic resonance imaging (MRI) is the current clinical standard for the assessment of LV infarct size,⁶ microvascular obstruction (MVO),⁷ and the prediction of functional recovery.⁸

Myocardial delayed enhancement (DE) on multidetector computed tomography (MDCT) enables the accurate evaluation of infarct size⁹ and has comparable results as an alternative technique to MRI in AMI.^10,11 Myocardial DE on MDCT allows the early evaluation of myocardial viability immediately after PCI without additional injection of contrast agent.¹² In our previous study, we showed that contrast DE patterns might be reliably useful in predicting functional recovery of the post-ischaemic myocardium and LV remodelling,¹³ and myocardial contrast DE size might be a very early predictive marker for future clinical cardiac events in patients with AMI.¹⁴ Recently, Amanieu et al.¹⁵ reported that the presence of a hypoenhanced area had high specificity for the prediction of MVO. Microvascular obstruction, which is an area of hypoenhancement surrounding hyper-enhancement detected by DE-MRI, corresponds to the production of no-reflow regions and reflects myocardial microvascular dysfunction.¹⁶ Furthermore, the extent of MVO has been shown to determine the magnitude of myocardial damage, and it provides an incremental prognostic marker of functional recovery and prognosis beyond that of infarct size.^5,7,17 Therefore, we hypothesized that the presence of hypoenhanced areas detected on DE-MDCT in AMI patients may have a similar potential to predict LV remodelling. The aim of this study was to evaluate the clinical value of myocardial contrast DE with MDCT for detecting MVO and LV remodelling revealed by DE-MRI after AMI.

Methods

Study population

Between June 2012 and March 2014, 98 consecutive patients with first AMI were initially assessed for inclusion into the study. The inclusion criteria for AMI patients were as follows: (i) chest pain >30 min in duration with presentation within 24 h after onset of symptoms; (ii) ST-segment elevation >0.1 mV within two contiguous electrocardiograph leads; and (iii) elevated creatine kinase-myocardial band (CK-MB) isoenzymes within 24 h of chest pain. Exclusion criteria were a history of previous MI, cardiogenic shock, or renal insufficiency (creatinine >1.5 mg/dL). Cardiac rupture caused the death of one patient, re-infarction occurred in one patient due to acute stent thrombosis, cerebrovascular disorder occurred during follow-up in two patients, and two patients refused to give their informed consent. Thus, the remaining 92 patients (77 men and 15 women; mean age, 63 ± 12 years) represent the study population. All patients underwent 64-slice MDCT without administration of additional contrast media immediately following successful PCI. The study protocol was approved by our institutional review board, and all the patients enrolled gave their informed consent.

Sixty-four-slice multidetector computed tomography scanning procedure

Scanning was performed using prospective electrocardiogram (ECG) triggering with a 64 × 0.625 mm slice collimation CT scanner (LightSpeed VCT, GE Healthcare, Milwaukee, WI) with a gantry rotation speed of 350 ms/rotation. Scanning was performed using a tube energy of 120 kV and effective tube current of 200 mA at a table feed of 20 mm/gantry rotation with beam pitch of 0.2 and a radiation dose of 2.0 ± 1.0 mSv. Acquisition of CT data and an ECG trace were automatically started during a 9- to 12-s breath-hold. Effective radiation doses were estimated by multiplying the dose–length product reported by the scanner by a conversion factor of 0.014 mSv/mGy cm according to standard methodology outlined in the recent guidelines.¹⁸

Reconstruction and analysis of 64-slice multidetector computed tomography images

Analysis of the scans was performed on a ZIOSTATION workstation (ZIOSOFT Inc., Tokyo, Japan). The original axial CT images and multiplanar reformations were used for analysis of the myocardium. The images used for interpretation and for quantitative analysis were reconstructed at a slice thickness of 2 mm. Extent of hyper-enhancement was judged as transmural if ≥75% of the LV wall thickness in each segment of the axial, short-axis, and long-axis multiplanar images was visually involved and as subendocardial if <75%. Infarct area was defined as a hyper-enhancement with a signal intensity (SI) >2 SD above the mean SI of remote normal myocardium¹⁹ and was adjusted by manual tracing of the suspected area in three dimensions. Infarct size was calculated on the basis of graphics displayed by SI histogram thresholding using a step-based algorithm to detect voxels within the defined SI range in the three dimensions and to define clusters of voxels that meet the hyper-enhancement definition. Using this method, the dedicated software determined the mean SI of the hyper-enhancement in Hounsfield units (HUs) and calculated total myocardial contrast DE size in grams. We defined heterogeneous enhancement (HE) as the concomitant presence of hyper- and hypoenhancement areas within the infarcted myocardium. We included both the hypoenhanced area surrounded by the hyper-enhanced layer, which is a typical finding of MVO on DE-MRI (Figure 1A) and patchy enhancement areas with hyper- and hypoenhancement segments (Figure 1B) in HE. Figure 1C showed no HE. The CT density values of the contrast-enhanced myocardium and non-enhanced myocardium in the remote area were measured from at least three regions of interest (ROIs) with the dimension of myocardial thickness (Figure 2). Then, to normalize the CT values, the relative CT density was calculated as the rate of average density of contrast-enhanced myocardium to that of non-enhanced myocardium.¹⁵

Figure 1

Short-axis DE-CT in three patients with acute myocardial infarction. (A and B) The presence of HE, and (C) no heterogeneous enhancement. (A) Transmural contrast-delayed enhancement in the anterior wall and a hypoenhanced area surrounded by a hyper-enhanced layer (arrows), which is a typical finding of microvascular obstruction on DE-MRI. (B) Transmural contrast-delayed enhancement in the lateral wall and a patchy enhancement area with hyper- and hypoenhancement segments (arrowheads). (C) Subendocardial contrast-delayed enhancement in the inferior wall and no HE. DE-CT, delayed-enhancement computed tomography; DE-MRI, delayed-enhancement magnetic resonance imaging; HE, heterogenous enhancement.

Figure 2

Short-axis delayed-enhancement computed tomography of a 49-year-old man. The CT density values of the contrast-enhanced myocardium and non-enhanced myocardium were measured from at least three ROIs with the dimension of myocardial thickness. To normalize the CT values, the relative CT density was calculated as the rate of average density of contrast-enhanced myocardium to that of non-enhanced myocardium. In this image, the relative CT density was calculated to be 3.18. The abbreviations were defined in the legend for Figure 1.

Magnetic resonance imaging protocol

Delayed-enhancement magnetic resonance imaging was performed in the acute phase (median 6 days after AMI onset) and chronic phase [median 198 days (median 6 months) after onset]. All MRI studies were performed with the patient in the supine position in a 1.5-T clinical scanner (MAGNETOM Avanto, Siemens, Erlangen, Germany) using a dedicated six-element phase-array cardiac coil. The ECG-gated images were acquired during repeated breath-holds of varying duration depending on the heart rate. Cine images were obtained using steady-state free-precession cine MRI in the two-chamber, four-chamber, and short-axis views. In the cardiac short-axis direction, the LV was completely encompassed by contiguous 8-mm-thick slices with a 2-mm interslice gap. Next, T2-weighted MRI was performed in the cardiac short-axis direction using a dark-blood T2-weighted short-tau inversion-recovery (STIR) fast-spin echo sequence. Finally, 5–10 min after injection of 0.2 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer, Berlin, Germany), delayed contrast-enhanced images were acquired in the same orientation as the cine images using a segmented inversion-recovery gradient-echo pulse sequence.

Magnetic resonance imaging analysis

Analysis of the scans was performed using Argus viewing software (Siemens). End-diastolic volume, end-systolic volume, ejection fraction (EF), and myocardial mass were obtained from short-axis cine-MR images. Late gadolinium-enhancement images were assessed both for infarct area and MVO area. T2-weighted STIR images were used to determine the presence of myocardial haemorrhage in the ischaemic myocardium. Infarct area was defined on late contrast images as a hyper-intense area. Areas of hyper-enhancement were defined as myocardium with an SI >2 SD above the mean SI of remote normal myocardium.¹⁹ Microvascular obstruction was defined as a late hypoenhancement signal within the myocardial wall surrounded by a hyper-enhanced region. Myocardial haemorrhage was defined as a hypointense area in the centre of the infarct area on T2-weighted images.²⁰ The MDCT and MRI images were analysed by two independent observers to calculate inter- and intra-observer agreements.

Coronary angiography and percutaneous coronary intervention procedure

Experienced interventional cardiologists performed coronary angiography (CAG) through the radial or femoral approach with 6–7 F catheters. The culprit coronary artery was defined on the basis of ECG and CAG results. All patients underwent PCI with both aspirin and a loading dose of 300 mg clopidogrel. A glycoprotein IIb/IIIa receptor inhibitor is not yet available in Japan, so no patient received this drug.

Before the start of the procedure, 6000–10 000 units of heparin were given intravenously, and the activated clotting time was maintained at >300 s. The contrast agent used was iopamidol 370 mg I/mL (Schering AG, Berlin, Germany). Post-procedural optimal coronary flow was defined as a thrombolysis in myocardial infarction (TIMI) flow grade of ≥3, and incomplete reperfusion or lack of procedural success was defined as a TIMI flow grade of 0–2.²¹ Time to reperfusion was defined as the period from time of onset of symptoms to time of reperfusion. Total contrast volume and time from last coronary injection to initiation of the CT scan were also measured.

Study endpoints

Study endpoints were defined as the incidence of MVO detected by MRI and the occurrence of LV remodelling. Left ventricular remodelling was defined as any increase in LV end-systolic volume (LVESV) at chronic phase after infarction compared with baseline.²⁰ Study endpoints were blinded to the findings of CT imaging.

Cardiac events were defined as cardiac death, hospitalization for worsening heart failure, or sustained ventricular tachycardia/ventricular fibrillation. After hospital discharge, all patients were followed-up every month at our outpatient clinic of our hospital up to the 6 months. Cardiac events were documented by clinical visits, standardized follow-up phone calls to the patients or their families, and followed by a review of medical records, electrocardiogram, and cardiac enzyme data. To establish the diagnosis of cardiac events, an outcome panel of two experienced cardiologists reviewed patient data forms containing prospectively collected information as well as medical records pertaining to the hospital admission.

Statistical analysis

All data are expressed as the mean ± SD or median and interquartile range for non-normally distributed data. Comparisons of continuous variables were analysed by unpaired t-test or Mann–Whitney U test according to the data distribution. Comparisons of categorical variables between groups were performed by the χ² test without correction for multiplicity. Inter- and intra-observer agreements were performed by intraclass correlation coefficients for continuous variables and kappa (κ) test for categorical variables. Receiver-operating characteristic (ROC) analysis was used to determine optimal cut-off values of relative CT density for prediction of MVO. The best cut-off value was defined as the point with the highest sum of sensitivity and specificity. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), and diagnostic accuracy of the CT characteristics alone or in combination were calculated for prediction of MVO. We performed comparison between two groups about various parameters in MVO and LV remodelling, respectively, and we used the significantly different parameters for the univariable analysis. Multivariable logistic regression analysis was then developed to calculate odds ratios and 95% confidence intervals (CIs). We assessed collinearity for multivariable models by Pearson's correlation and excluded the parameters which had collinearity. We performed stepwise forward selection considering any valuables with values of P < 0.05 to identify potential risk factors for MVO and LV remodelling. A P-value of <0.05 was considered to indicate statistical significance.

Results

Patient, lesion, and procedural characteristics

Microvascular obstruction was detected in 49 (53%) patients and LV remodelling was detected in 29 (32%) patients. Patients were divided into two groups according to the presence (n = 33) or absence (n = 59) of HE detected by MDCT. The median time delay between last injection of contrast medium and MDCT scanning was 17 ± 8 min (range 5–34 min). During the 6-month follow-up periods, three cardiac events (one rehospitalization for worsening heart failure and two sustained ventricular tachycardia/ventricular fibrillation) were observed in HE group, but none in non-HE group (P = 0.042). Baseline patient characteristics, procedural characteristics, and medications are summarized in Table 1. There were no statistically significant differences in sex, coronary risk factors, onset to reperfusion time, stent size, stent length, contrast volume, culprit lesion, and TIMI flow between the two groups, whereas age was higher in the non-HE group. Peak level of CK and its MB fraction and brain natriuretic peptide level at 7 days were higher in the group with HE. Killip Class II or greater heart failure and beta-blocker therapy were more commonly observed in the HE group.

Table 1

Baseline patient characteristics, procedural characteristics, and medications

	HE (−) n = 59	HE (+) n = 33	P-Value
Age (years)	66 ± 12	58 ± 12	0.002
Male (%)	81	88	0.416
Diabetes mellitus (%)	25	33	0.419
Hypertension (%)	59	67	0.486
Dyslipidaemia (%)	61	73	0.258
Smoker (%)	51	61	0.367
BMI	24.6 ± 3.0	25 ± 2.9	0.374
Laboratory findings
CRP (mg/dL)	0.29 ± 0.38	0.25 ± 0.34	0.596
LDL (mg/dL)	126 ± 39	131 ± 37	0.409
Peak CK (IU/mL)	1931 ± 1228	4856 ± 2241	<0.001
Peak CK-MB (IU/mL)	170 ± 101	398 ± 230	<0.001
BNP on day 7 (pg/mL)	145 ± 146	227 ± 188	0.042
Onset-reperfusion time (h)	6.4 ± 6.4	7.4 ± 6.5	0.467
Stent size (mm)	3.3 ± 0.4	3.3 ± 0.3	0.502
Stent length (mm)	26.3 ± 7.6	25.0 ± 7.7	0.439
Contrast volume (mL)	200 ± 58	216 ± 64	0.210
Culprit lesion:
LAD (%)	49	64	0.181
LCx (%)	14	18	0.554
RCA (%)	37	18	0.056
First TIMI flow 0 (%)	53	64	0.303
Final TIMI flow (0–2) (%)	8	18	0.169
Killip class ≥II (%)	3	15	0.041
Medication after on admission
Aspirin (%)	98	100	0.452
Clopidogrel (%)	98	100	0.452
ACE-inhibitor/ARBs (%)	78	88	0.240
Beta-blockers (%)	73	97	0.004
Statins (%)	98	100	0.452

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; BNP, brain natriuretic peptide; CK, creatine kinase; CRP, C-reactive protein; HE, heterogeneous enhancement; LAD, left anterior descending artery; TIMI, thrombolysis in myocardial infarction.

Multidetector computed tomography and magnetic resonance imaging measurements

Representative images from the groups of patients with and without HE are presented in Figures 3 and 4. The intra- and inter-observer intraclass correlation coefficients for relative CT density were 0.90 (95% CI 0.81–0.95) and 0.89 (95% CI 0.78–0.94), infarct size on MDCT were 0.93 (95% CI 0.85–0.96) and 0.94 (95% CI 0.88–0.97), ESV in the acute phase were 0.99 (95% CI 0.98–0.99, P < 0.001) and 0.97 (95% CI 0.85–0.99), and ESV in the chronic phase on MRI were 0.99 (95% CI 0.94–0.99) and 0.99 (95% CI 0.96–0.99), respectively, and inter- and intra-observer agreement (κ) for the interpretation of the presence of HE was 0.91 and 0.81, respectively, and the with all values indicating good intra- and inter-observer agreement. Computed tomography and MRI findings are summarized in Table 2. The CT value of the DE area, relative CT density, transmural enhancement, and myocardial contrast DE size was significantly higher in the patients with HE pattern than in those without this pattern. On MRI, LVEF was lower, and end-diastolic and end-systolic volumes were significantly larger, in the heterogeneous group. These differences were consistently significant in both the acute and chronic phases. The incidences of MVO, myocardial haemorrhage, and LV remodelling were significantly more frequent in the heterogeneous group.

Table 2

Computed tomography and magnetic resonance imaging findings

	HE (−) n = 59	HE (+) n = 33	P-Value
CT
CT value: DE area (HU)	94 ± 34	127 ± 28	<0.001
CT value: remote area (HU)	49 ± 8	48 ± 7	0.287
Relative CT density	1.9 ± 0.8	2.7 ± 0.6	<0.001
Transmural enhancement (%)	20	73	<0.001
Myocardial contrast DE size (g)	9.0 ± 9.2	36.1 ± 17.2	<0.001
MRI
LVEF baseline (%)	45 ± 9	38 ± 8	0.001
Chronic phase (%)	49 ± 10	39 ± 8	<0.001
LVEDV baseline (mL)	119 ± 31	143 ± 38	0.002
Chronic phase (mL)	110 ± 32	149 ± 40	<0.001
LVESV baseline (mL)	67 ± 25	90 ± 33	<0.001
Chronic phase (mL)	56 ± 25	93 ± 34	<0.001
MVO (%)	31	94	<0.001
Haemorrhage (%)	19	67	<0.001
LV remodelling (%)	19	55	<0.001

CT, computed tomography; DE, delayed enhancement; EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; HE, heterogeneous enhancement; HU, Hounsfield unit; LV, left ventricular; MRI, magnetic resonance imaging; MVO, microvascular obstruction.

Figure 3

A 64-year-old man from the group of patients with heterogeneous enhancement. Short-axis delayed-enhancement computed tomography imaging (A), delayed-enhancement magnetic resonance imaging at 1 week (B), and cine magnetic resonance imaging (end-systolic time frame) at 1 week (C), and at 6 months (D). Delayed-enhancement computed tomography imaging (A) shows complete transmural contrast-delayed enhancement with heterogeneous enhancement (arrows) and high relative CT density. Delayed-enhancement magnetic resonance imaging at 1 week (B) shows complete transmural enhancement of the left ventricular lateral wall with microvascular obstruction (arrowheads). On cine magnetic resonance imaging (C and D), the left ventricular lateral wall was obviously thinned at 6 months. Left ventricular end-systolic volume was increased from 150 mL at 1 week to 170 mL at 6 months. The abbreviations were defined in the legend for Figure 1.

Figure 4

A 64-year-old man from the group without heterogeneous enhancement. Short-axis delayed-enhancement computed tomography imaging (A), delayed-enhancement magnetic resonance imaging at 1 week (B), and cine magnetic resonance imaging (end-systolic time frame) at 1 week (C), and at 6 months (D). Delayed-enhancement computed tomography imaging (A) shows subendocardial contrast-delayed enhancement without heterogeneous enhancement and with low relative CT density. Delayed-enhancement magnetic resonance imaging at 1 week (B) shows subendocardial enhancement of the left ventricular inferior wall without microvascular obstruction. On cine magnetic resonance imaging (C and D), left ventricular end-systolic volume was reduced from 82 mL at 1 week to 58 mL at 6 months. The abbreviations were defined in the legend for Figure 1.

Predictors of microvascular obstruction

According to the ROC analysis, the area under the curve of relative CT density for predicting the presence of MVO was 0.889 (95% CI 0.825–0.968). When the cut-off value of 2.20 was applied, sensitivity was 86% (95% CI 72.1–94.7%) and specificity was 82% (95% CI 65.7–89.8%). According to the univariable analysis (Table 3), EF < 40%, peak CK > 4000 IU/mL, the presence of HE, relative CT density >2.20, and transmural enhancement were significantly associated with the presence of MVO. In the multivariable logistic regression analysis, the presence of HE, relative CT density >2.20, and transmural enhancement were significant predictors for the presence of MVO after adjustment for multiple confounders. With regard to diagnostic accuracy of CT characteristics for predicting MVO, the presence of HE and relative CT density >2.20 showed a high PPV of 93% (95%CI 75.7–99.1%) and the absence of these two findings yielded a high NPV of 90% (95% CI 76.9–97.3%) for the predictive value of MVO.

Table 3

Univariable and multivariable logistic regression analyses for prediction of microvascular obstruction

Factor	Univariable			Multivariable (P< 0.050)
	OR	95% CI	P-Value	OR	95% CI	P-Value
Age	0.96	0.93–1.00	0.076
Sex (male)	1.37	0.45–4.16	0.577
EF <40%	3.12	1.19–8.14	0.019	1.45	0.33–6.27	0.622
Peak CK >4000 IU/mL	10.66	2.31–49.12	0.002	0.90	0.08–10.10	0.934
HE	25.90	5.61–119.48	<0.001	13.50	2.15–84.90	0.005
Relative CT density >2.2	18.40	5.61–60.00	<0.001	12.00	2.94–49.20	<0.001
Transmural enhancement	29.56	9.34–93.52	<0.001	6.50	1.32–32.00	0.021

CI, confidence interval; CK, creatine kinase; CT, computed tomography; EF, ejection fraction; HE, heterogeneous enhancement; OR, odds ratio; TIMI, thrombolysis in myocardial infarction.

Predictors of left ventricular remodelling

According to the univariable analysis, the presence of HE and peak CK >4000 and relative CT density were significantly associated with the occurrence of LV remodelling. In the multivariable logistic regression analysis, the presence of HE was the only significant independent predictor for the occurrence of LV remodelling after adjustment for multiple confounders (Table 4).

Table 4

Logistic regression analysis for prediction of left ventricular remodelling

Factor	Univariable			Multivariable (P< 0.050)
	OR	95% CI	P-Value	OR	95% CI	P-Value
Age	0.96	0.93–1.00	0.067
Sex (male)	8.00	0.99–64.10	0.050
EF <40%	0.57	0.22–1.51	0.264
Peak CK >4000 IU/mL	3.45	1.32–9.06	0.012	1.44	0.42–4.85	0.561
HE	5.24	2.03–13.51	<0.001	3.97	1.26–12.60	0.019
Relative CT density	1.77	1.01–3.12	0.048	1.14	0.57–2.26	0.708
Myocardial contrast DE size	1.02	0.99–1.05	0.056
Transmural enhancement	2.14	0.874–5.26	0.096

CI, confidence interval; CT, computed tomography; DE, delayed enhancement; EF, ejection fraction; HE, heterogeneous enhancement; OR, odds ratio.

Discussion

The major important findings of the study are as follows. First, the presence of HE and high relative CT density were significant predictors of MVO detected by MRI. For prediction of MVO, the presence of both HE and high relative CT density showed a high PPV of 93%, and their absence offered a high NPV of 90% for the detection of MVO. Furthermore, the presence of HE was significantly associated with adverse LV remodelling in the chronic phase. Thus, in the setting of traditional prognostic markers, detection of HE on MDCT images may provide promising information for prediction of MVO and LV remodelling.

Detection of microvascular obstruction by multidetector computed tomography

The presence of MVO after PCI for AMI corresponds to the angiographic no-reflow phenomenon and is assessed as a hypointense infarct core on late gadolinium enhancement by MRI.¹⁶ We showed that the presence of HE and high relative CT density were significant predictors of MVO detected by MRI. A previous report studying DE-CT explained that the hyper-enhancement area on MDCT is caused by the penetration of iodine molecules to the myocardial cells caused by sarcolemmal membrane dysfunction after myocyte necrosis, and the hypoenhanced area is caused by the death and subsequent blockage of intramyocardial capillaries by cellular debris.⁹ Another study investigating occlusion, microembolization, and reperfusion in an animal model of AMI showed patchy microinfarcts and large infarcts with both hyper-enhanced and hypoenhanced areas by MDCT. Histopathologically, these areas showed patchy microinfarcts adjacent to large infarcts and intramyocardial haemorrhage.²² These reports indicate the possibility that HE is associated with distal embolization and myocardial haemorrhage. In fact, the incidence of myocardial haemorrhage was significantly more frequent in the patients with HE than in those without HE in our study. Besides, these findings are almost consistent with the findings of MVO. Robbers et al.²³ reported that the infarct core of MVO contained extensive necrosis and erythrocyte extravasation without intact vasculature, and the surrounding gadolinium-enhanced area of MVO contained granulation tissue, leucocyte infiltration, and necrosis with morphologically intact microvessels containing microthrombi without erythrocyte extravasation.

The relative CT density was calculated by the average CT values of at least three ROIs for infarcted and remote areas. It has components of both intensity and transmurality of the enhancement. As described above, histopathologically, an area of hyper-enhancement represents the penetration of iodine molecules into the myocardial cell.⁹ Therefore, we consider that a high relative CT density area reflects the magnitude of myocardial damage and might also reflect myocardial haemorrhage. In our study, the patients with MVO and myocardial haemorrhage detected by MRI showed significantly higher relative CT density than those without (2.67 ± 0.70 vs. 1.70 ± 0.58, P < 0.001 and 2.81 ± 0.79 vs. 1.92 ± 0.63, P < 0.001, respectively). Histopathologically, haemorrhage is likely to be present in areas of MVO,²⁴ and therefore, these finding suggests the possibility that relative CT density is associated with myocardial haemorrhage, in that the presence of a low-density area and high relative CT density is an important marker for predicting MVO.

Relation between heterogeneous enhancement and left ventricular remodelling

Heterogeneous enhancement was significantly associated with adverse LV remodelling in the chronic phase, indicating the clinical usefulness of myocardial contrast DE. The major predictors of adverse LV remodelling in AMI patients are reported to be anterior infarct location, peak CK, perfusion status, LVEF, and infarct size.^3,4 Microvascular obstruction and myocardial haemorrhage are predictors of LV remodelling on MRI.^5,17,20 As explained above, HE is a significant predictor of MVO, and histopathologically, it has the potential to be associated with distal embolization and haemorrhage as with MVO.^9,22 Compared with the patients without HE, those with HE were associated with higher peak CK, larger infarct size, lower LVEF, greater numbers of transmural pattern, haemorrhage, and MVO. Therefore, it is natural that HE should be a predictor of LV remodelling. In addition, previous reports have shown that the presence of MVO and/or myocardial haemorrhage is a stronger independent predictor of LV remodelling or infarct zone contractile recovery than transmural extent of infarction or overall infarct volume.^18,24 The present study similarly showed HE to be an independent predictor of LV remodelling than adjusted by transmural extent of infarction or overall infarct volume.

Clinical implications

Contrast-enhanced MRI is well established for the assessment of myocardial viability without ionizing radiation and injection of nephrotoxic contrast material. However, the facts that MDCT can be performed with a short examination times and that it is generally available and easily performed are considered important advantages in comparison with contrast-enhanced MRI.

Multidetector computed tomography immediately after PCI provides information on transmurality of infarction, infarct size, and future cardiac events, as does MRI. Furthermore, in this study we showed that HE detected by MDCT might provide an early predictive marker for predicting MVO and LV remodelling in patients with AMI. Left ventricular remodelling is an important predictor of morbidity and mortality for overt congestive heart failure and life-threatening arrhythmias.⁴ Therefore, if HE can be detected with MDCT, it would be necessary to plan for prevention of adverse patient outcomes through early treatment interventions and careful follow-up.

Study limitations

First, because of our relatively small sample size, we acknowledge that our findings should be confirmed in a larger number of patients. Secondly, as the incidence of MVO was high in this study, an imbalance between cases and samples might occur in analysis of MVO. Thirdly, there is a difference in timing between the DE-MRI and DE-CT examinations. Acute myocardial infarction is associated with myocardial oedema, and therefore, this might also influence the extent of myocardial contrast DE. A previous study similarly showed that MRI in the very early days after infarction overestimated the true extent of irreversibly injured myocardium because of myocardial oedema.²⁵ Therefore, it is possible that there is a small difference in the extent of DE size and transmurality between DE-MRI and DE-CT. Furthermore, the definition of hyper-enhancement was not standardized in DE-CT, and so there were the limitations due to a global classification. The definition of optimal threshold for infarct area in DE-CT was also not standardized, and therefore standardized analysis using this threshold will further increase reproducibility and facilitate (semi) automated quantification. Fourthly, the patients receive additional radiation exposure from the CT scan after primary PCI. However, we make efforts to reduce the radiation exposure by performing low-dose scanning to the extent possible and using prospective ECG triggering methods. In doing so, the radiation exposure of the patients was reduced to 2.0 ± 1.0 mSv. Furthermore, if we can use a high-specification CT system such as a dual-energy scanner or 320-slice MDCT, the better assessment will be provided and the radiation exposure will be reduced even further. Finally, the lack of pathological correlation might leave the mechanism of these phenomena unclear. Therefore, we consider it necessary to investigate the relation between the MDCT images and histopathology and to confirm the correspondence between these imaging characteristics and histopathological findings.

Conclusion

The present study showed that the presence of HE on MDCT was significantly associated with MVO and LV remodelling in patients with AMI. We consider that MDCT images acquired immediately after primary PCI may provide promising information for predicting MVO and LV remodelling in patients with AMI.

Authors’ contributions

H.N. and T.H.: performed statistical analysis. K.A.: handled funding and supervision. A.S., Y.K., Y.T., H.A., and Y.F.: acquired the data. H.W. and A.S.: conceived and designed the research, and drafted the manuscript. Y.N. and K.A.: made critical revision of the manuscript for key intellectual content.

Conflict of interest: none declared.