Author + information
- Received January 29, 2008
- Revision received March 20, 2008
- Accepted April 7, 2008
- Published online July 15, 2008.
- Robin Nijveldt, MD⁎,§,⁎ (, )
- Aernout M. Beek, MD⁎,
- Alexander Hirsch, MD§∥,
- Martin G. Stoel, MD¶,
- Mark B.M. Hofman, PhD†,
- Victor A.W.M. Umans, MD, PhD#,
- Paul R. Algra, MD, PhD⁎⁎,
- Jos W.R. Twisk, PhD‡ and
- Albert C. van Rossum, MD, PhD⁎,§
- ↵⁎Reprint requests and correspondence:
Dr. Robin Nijveldt, Department of Cardiology, VU University Medical Center, De Boelelaan 1117, P.O. Box 7057, 1007 MB, Amsterdam, the Netherlands.
Objectives We examined the relation between angiographic, electrocardiographic, and gadolinium-enhanced cardiovascular magnetic resonance (CMR) characteristics of microvascular obstruction (MVO), and their predictive value on functional recovery after acute myocardial infarction (AMI).
Background Microvascular obstruction on CMR has been shown to predict left ventricular (LV) remodeling, but it is not well known how it compares with commonly used criteria of microvascular injury, and earlier reports have produced conflicting results on the significance and extent of MVO.
Methods Thrombolysis In Myocardial Infarction (TIMI) flow grade, myocardial blush grade (MBG), and ST-segment resolution were assessed in 60 patients with AMI treated with primary stenting. Cardiovascular magnetic resonance was performed between 2 and 9 days after revascularization to determine early MVO on first-pass perfusion imaging, late MVO on late gadolinium-enhanced imaging, and infarct size and transmural extent. Cine imaging was used to determine LV volumes and global and regional function at baseline and 4-month follow-up.
Results Early and late MVO were both related to incomplete ST-segment resolution (p = 0.002 and p = 0.01, respectively), but not to TIMI flow grade and MBG. Of all angiographic, electrocardiographic, and CMR variables, late MVO was the strongest parameter to predict changes in end-diastolic volume (β = 0.53; p = 0.001), end-systolic volume (β = 8.67; p = 0.001), and ejection fraction (β = 3.94; p = 0.006) at follow-up. Regional analysis showed that late MVO had incremental diagnostic value to transmural extent of infarction (odds ratio: 0.18; p < 0.0001).
Conclusions In patients after revascularized AMI, late MVO proved a more powerful predictor of global and regional functional recovery than all of the other characteristics, including transmural extent of infarction.
Infarct size is a strong predictor of prognosis after acute myocardial infarction (AMI), and reperfusion therapy has contributed to an important decrease in mortality by limiting myocardial necrosis (1). However, despite successful recanalization of the infarct-related artery, perfusion of the ischemic myocardium is not or is incompletely restored in up to 30% of patients due to microvascular obstruction (MVO), angiographically referred to as the no-reflow phenomenon (2). The presence of no-reflow in these patients has been found to be a predictor of adverse events, with higher incidence of left ventricular (LV) remodeling, congestive heart failure, and death (3,4). The diagnosis of no-reflow can be made using angiography (5,6), electrocardiography (7), nuclear scintigraphy (8), myocardial contrast echocardiography (MCE) (3), or cardiovascular magnetic resonance (CMR) (4). Both MCE and CMR allow direct visualization of the no-reflow zone (3,9). Myocardial contrast echocardiography using intracoronary contrast agents was the first technique to show that angiographic reflow does not always imply restoration of myocardial flow (10). Although MCE can be used to predict functional changes and outcome after reperfused AMI, it is still limited by attenuation artifacts and poor visibility of (postero-)lateral segments (11), and is not capable of quantifying total infarct size. CMR allows accurate assessment of function, transmural extent and total size of infarction, and MVO in all segments of the left ventricle (12). Two methods have been described for the detection of MVO using gadolinium-enhanced CMR: first-pass perfusion (9,13) and late gadolinium enhancement (LGE) (14,15). Both techniques have been shown to predict LV remodeling and outcome, but it is not known how they compare with the commonly used angiographic and electrocardiographic criteria of microvascular injury. Also, evidence is still limited in optimally treated patients, and earlier reports have produced conflicting results on the significance and extent of MVO (15–18).
The purpose of the present study was, therefore, to explore the relation between angiographic, electrocardiographic, and CMR characteristics of microvascular injury, and to investigate their predictive value on recovery of global and regional LV function after optimal treatment for AMI.
We screened consecutive patients presenting with a first ST-segment elevation AMI, according to standard electrocardiographic and enzymatic criteria. All patients had undergone primary percutaneous coronary intervention (PCI) with (bare-metal) stent implantation within 12 h of symptom onset. Exclusion criteria were: unsuccessful angiographic reperfusion (Thrombolysis In Myocardial Infarction [TIMI] flow grade <2), hemodynamic instability, left bundle branch block, or (relative) contraindications for CMR. Sixty-seven patients were prospectively enrolled in the study. Patients were treated with aspirin, heparin, abciximab, clopidogrel, statins, beta-blocking agents and angiotensin-converting enzyme inhibitors, according to American College of Cardiology/American Heart Association practice guidelines (19). All of the patients gave informed consent to the study protocol, which was approved by the local ethics committee.
Coronary angiography was performed at the end of the PCI procedure for off-line analysis of TIMI flow grade (5) and myocardial blush grade (MBG) (6). Images were assessed by 2 blinded observers, using the following definitions: TIMI flow grade 2: complete filling of the entire vessel, but slower than nonaffected vessels; TIMI flow grade 3: normal flow; MBG 0: no myocardial blush; MBG 1: minimal myocardial blush or contrast density; MBG 2: moderate blush or contrast density, but less than a contralateral or ipsilateral noninfarct-related artery; and MBG 3: normal myocardial blush or contrast density similar to a contralateral or ipsilateral noninfarct-related artery. Angiographic incomplete reperfusion was defined as TIMI flow grade 2 or MBG <2.
The ST-segment resolution was evaluated on a 12-lead electrocardiogram acquired on admission and 1 h after PCI. The sum of ST-segment elevation was measured 60 ms after the J point in leads I, aVL, and V1 to V6 for anterior and leads II, III, aVF, V5, and V6 for nonanterior AMI. The percentage resolution of ST-segment elevation from before to after PCI was calculated and categorized as complete (≥70%), partial (30% to <70%), or no (<30%) ST-segment resolution (7). Incomplete reperfusion was defined as <70% ST-segment resolution on electrocardiography.
Cardiovascular magnetic resonance examination was performed on a 1.5-T clinical scanner (Sonata/Symphony, Siemens, Erlangen, Germany) using a phased-array cardiac receiver coil. Baseline scan was scheduled between 2 to 9 days after reperfusion and follow-up at 4 months. Electrocardiogram-gated breath-hold cine imaging was performed to determine LV function, using a segmented steady-state free-precession pulse sequence in multiple short-axis views every 10 mm covering the entire left ventricle. Typical in-plane resolution was 1.6 × 1.9 mm2, with slice thickness 5.0 to 6.0 mm (repetition time/echo time 3.2/1.6 ms, flip angle 60°, matrix 256 × 156, temporal resolution 35 to 50 ms). First-pass perfusion was performed during administration of a gadolinium-based contrast agent (Magnevist, 0.1 mmol/kg; Schering, Berlin, Germany) at a rate of 3.0 ml/s, using a single-shot saturation recovery gradient-echo pulse sequence. Three short-axis slices were obtained per heartbeat, every 10 mm, covering the infarct area as seen during cine imaging (90° pre-pulse, repetition time/echo time 2.1/1.0 ms, saturation time 120 ms, flip angle 12°, matrix 128 × 93, in-plane resolution 3.0 × 3.3 mm2, slice thickness 8.0 mm, total scan duration ∼1.5 min). Immediately after first-pass perfusion, an additional 0.1 mmol/kg gadolinium-based contrast agent was administered (cumulative dose 0.2 mmol/kg). Late gadolinium enhancement images were obtained 12 to 15 min after the second contrast administration (20), using a 2-dimensional segmented inversion recovery gradient-echo pulse sequence, with slice position identical to the cine images. Typical in-plane resolution was 1.4 × 1.7 mm2, with slice thickness 5.0 to 6.0 mm (repetition time/echo time 9.6/4.4 ms, flip angle 25°, triggering to every other heartbeat). The inversion time was set to null the signal of viable myocardium and typically ranged from 250 to 300 ms.
CMR data analysis and definitions
All CMR data were analyzed on a separate workstation using dedicated software (Mass version 2006beta, Medis, Leiden, the Netherlands). Cine, first-pass perfusion, and LGE images acquired during the same imaging session were matched by using slice position. Registration of follow-up to baseline cine and LGE images was achieved by consensus of 2 observers using anatomic landmarks, such as papillary muscles and right ventricular insertion sites. On all short-axis cine slices, the endocardial and epicardial borders were outlined manually on end-diastolic and -systolic images. Left ventricular volumes and left ventricular ejection fraction (LVEF) were calculated. Each short axis was divided in 12 equiangular segments, starting at the posterior septal insertion of the right ventricle. Segmental wall thickening (SWT) was calculated by subtracting end-diastolic from end-systolic wall thickness. Myocardial segments were considered to be dysfunctional if SWT was <3 mm, based on the mean SWT of 4.4 ± 0.7 mm (mean ± 2 SD) in a group of 10 healthy volunteers (age 50 to 75 years). Complete recovery of dysfunctional segments was defined as SWT ≥3.0 mm at follow-up.
First-pass perfusion was evaluated qualitatively. Microvascular obstruction was considered to be present if a region of hypoperfusion persisted for >1 min after contrast bolus arrival in the left ventricle and was located in the subendocardial layer of the infarct core (9,13) in at least 1 of the slices. To verify that a true perfusion deficit persisted after passage of the contrast agent, all acquired phases were evaluated. Microvascular obstruction on first-pass perfusion imaging was termed early MVO.
The assessment of LGE images and infarct size was done as previously described (15). Total infarct size was expressed as percentage of LV mass. Transmural extent of infarction was calculated by dividing the hyperenhanced area by the total area of the pre-defined segment (%). A transmurality score was calculated in each patient, expressed as the sum of segments with >75% infarcted myocardium, as a percentage of the total number of segments scored. On LGE images, MVO was defined as any region of hypoenhancement within the hyperenhanced area (14,15), and was termed late MVO. Microvascular obstruction was included in the calculation of total infarct size. The extent of late MVO was calculated in each patient, expressed as the sum of the segments with MVO, as a percentage of the number of segments scored.
For analysis of segmental function and transmural extent of infarction, the 2 most basal and 2 most distal slices were excluded, because segmental evaluation at these levels is not considered to be reliable due to the LV outflow tract and partial volume effect respectively. All CMR studies were supervised by 1 operator, and all images were analyzed by 2 experienced observers who were blinded to patient data. There was no disagreement between observers regarding the presence of early or late MVO in each patient.
Data are expressed as mean ± SD for continuous variables and as frequency with percentage for categorical variables. Comparison of categorized angiographic and electrocardiographic variables and microvascular injury was done by the chi-square test, or by the Fisher exact test if an expected cell count was <5. The paired-samples t test was used to compare differences in global LV parameters between baseline and follow-up, and the independent-samples t test to compare means between subgroups. To identify independent predictors of global LV indexes at baseline and the change of these parameters between baseline and follow-up, multivariable linear regression analyses with a forward selection procedure were used. Variables entered the model if p < 0.10.
We evaluated 4 outcome variables of regional myocardial function (change in end-diastolic wall thickness, change in end-systolic wall thickness, change in SWT, and complete recovery) in relation to the presence of late MVO. Only dysfunctional segments at baseline were included, and outcomes were stratified by the transmural extent of infarction. Because regional function in different segments within 1 patient is strongly related, outcomes were analyzed using multilevel analyses (linear and logistic regression) with 3 levels: segments within slices and slices within patients (MLwiN, version 1.02.0002, Centre for Multilevel Modeling, London, United Kingdom) (21). In each analysis, a correction was made for the baseline variable of regional myocardial function in question (i.e., if the dependent variable was change in end-diastolic wall thickness, presence of microvascular injury and baseline end-diastolic wall thickness were included as covariates).
All statistical tests were 2 tailed, and a p value of <0.05 was considered to be statistically significant.
Baseline patient characteristics and medication are listed in Table 1. Seven patients did not undergo the follow-up study and were therefore excluded from CMR analysis (refusal of follow-up CMR in 3, claustrophobia in 1, cardiac death in 2, and noncardiac death in 1 patient). There was no reinfarction, revascularization, or hospitalization for heart failure between baseline and follow-up study in the remaining 60 patients.
Angiography and electrocardiography
All angiographic and electrocardiographic characteristics are listed in Table 1. Of the 11 patients with TIMI flow grade 2, 5 had MBG 2 (46%), 4 had MBG 1 (36%), and 2 had MBG 0 (18%). Of the 49 patients with TIMI flow grade 3, 30 had MBG 3 (61%), 16 had MBG 2 (33%), and 3 had MBG 1 (6%). There was no statistical association between TIMI flow grade or MBG and the time to reperfusion.
The mean sum of total ST-segment elevation was 17.9 ± 10.8 mV before PCI and 7.9 ± 7.3 mV after the procedure, resulting in 10.0 ± 8.6 mV absolute ST-segment resolution. The mean relative ST-segment resolution was 55.7 ± 32.0%. There was a significant relationship for ST-segment resolution (categorized as no, partial, or complete) with TIMI flow grade 2 to 3 (p value for trend = 0.005) and with MBG 0 to 3 (p value for trend = 0.004). Of the 24 patients with complete ST-segment resolution, 23 had TIMI flow grade 3 (96%) and 21 had MBG 2 to 3 (88%). Of the 36 patients with incomplete ST resolution, 10 had TIMI flow grade 2 (28%), and 6 had MBG 0 to 1 (17%). There was no statistical relationship between ST-segment resolution and the time to reperfusion.
The CMR examinations were performed 5 ± 2 days and 116 ± 22 days after primary PCI. The LVEF at baseline was 42.6 ± 9.0%, which significantly improved to 45.0 ± 9.5% at follow-up (p = 0.001). There was no statistically significant change in mean global LV volumes between baseline and follow-up in the total group (data not shown). Mean total infarct size was 16.9 ± 9.7% at baseline.
Early MVO was present in 41 patients (68%), and 34 patients (57%) also demonstrated late MVO. Seven patients (12%) had early MVO without late MVO. Both early MVO and late MVO presence were related to incomplete ST-segment resolution (p = 0.002 and p = 0.01, respectively; chi-square test) (Table 2). There was no statistically significant relation between early or late MVO and TIMI flow grade or MBG (Table 2). The time to reperfusion was not different between patients with or without early MVO (3.4 ± 2.5 h vs. 3.7 ± 4.0 h, respectively; p = 0.73) and between patients with or without late MVO (3.4 ± 2.3 h vs. 3.8 ± 4.6 h, respectively; p = 0.70).
Predictors of global function and recovery
Tables 3 and 4⇓⇓ demonstrate univariable and multivariable linear regression analysis for the prediction of baseline LVEF and LV end-systolic volume and the change in LVEF and LV end-systolic volume between baseline and follow-up CMR. Infarct size, transmural extent of infarction, and the extent of late MVO were all highly significant predictors of LVEF and LV end-systolic volume at baseline. Multivariable analysis revealed LGE infarct size as the strongest and single independent predictor of baseline LVEF (β = −0.58; p < 0.0001) and LV end-systolic volume (β = 1.20; p < 0.0001). The results of univariable and multivariable regression analysis of LV end-diastolic volume at baseline were similar to results for LVEF and LV end-systolic volume.
The presence of late MVO was the strongest predictor of change in LVEF and LV end-systolic volume at follow-up (β = −3.94; p = 0.006 and β = 8.67; p = 0.001, respectively). Univariable analysis revealed the presence of late MVO, the extent of late MVO, and the transmural extent of infarction as highly significant predictors of change in end-diastolic volume (p < 0.005). The extent of late MVO was the strongest significant predictor of change in end-diastolic volume after multivariable analysis (β = 0.53; p = 0.001).
The changes in LV volumes and LVEF according to late MVO status are shown in Figure 1. Patients with late MVO showed a significant increase in LV end-diastolic volumes (p = 0.01), with a trend toward an increase in LV end-systolic volumes (p = 0.10), and absence of improvement in LVEF (p = 0.14). Patients without late MVO showed a significant decrease in volumes and a significant improvement in LVEF. Within the group of patients with late MVO, no significant correlation was found between the extent of MVO and the change in LVEF (Pearson r = 0.05; p = 0.78), end-systolic volume (Pearson r = 0.19; p = 0.29), or end-diastolic volume (Pearson r = 0.21; p = 0.24).
Regional function and recovery
A total of 3.924 segments were analyzed, of which 2.158 segments (55%) were dysfunctional at baseline. On the LGE images, 1.709 segments (43.6%) demonstrated hyperenhancement and 380 segments (9.7%) demonstrated late MVO at baseline. The presence of late MVO significantly increased with infarct transmurality: 0.6%, 17%, and 49% in segments with 0 to 25%, 26% to 75%, and 76% to 100% extent of infarction, respectively (p < 0.001).
Figure 2 shows the observed changes in regional wall thickness and function in dysfunctional segments, according to infarct transmurality and presence of late MVO. Segments with significant hyperenhancement (>25%) and late MVO showed a larger decrease in end-diastolic and end-systolic wall thickness during follow-up than segments without MVO (Figs. 2A and 2B). Also, in segments with >75% hyperenhancement and late MVO, improvement in wall thickening was significantly less than in segments without MVO (p = 0.007) (Fig. 2C). Out of all the dysfunctional segments, 654 segments showed complete recovery at follow-up. The likelihood of complete recovery was highest in segments with no or minimal hyperenhancement without MVO and lowest in segments with >75% hyperenhancement with MVO (Fig. 2D). Only 6% (23 of 372) of the dysfunctional segments with late MVO showed complete recovery during follow-up, compared with 35% (631/1,786) of the dysfunctional segments without MVO (odds ratio: 0.18; 95% confidence interval: 0.08 to 0.38; p < 0.0001).
The present study is the first to directly compare angiographic, electrocardiographic, and gadolinium-enhanced CMR characteristics of microvascular injury and their predictive value on functional outcome in a homogeneous group of patients after successful primary stenting for AMI. The main findings can be summarized as follows: 1) of all characteristics of microvascular injury, late MVO was the strongest predictor of change in global LV function and volumes at follow-up; 2) ST-segment resolution, but not TIMI flow grade and MBG, correlated with the presence of MVO on first-pass perfusion and LGE CMR; and 3) late MVO was a stronger predictor of regional functional outcome than infarct transmurality.
Angiography and electrocardiography
In AMI, restoration of microvascular flow is considered to be a key factor for post-ischemic repair, functional outcome and prognosis. The presence of microvascular injury can be assessed with a multitude of invasive and noninvasive techniques. Despite their proven clinical relevance, there have been few studies on the relative value of these techniques, and to our knowledge there are no studies that have compared CMR characteristics of MVO to TIMI flow grade, MBG, or ST-segment resolution. A number of reports have addressed angiographic and electrocardiographic techniques and found that MBG and ST-segment resolution provided prognostic information beyond the standard TIMI flow grading and the (semiquantitative) TIMI frame count (22). Myocardial blush grade and ST-segment resolution have each been used as (surrogate) end points in the evaluation of MVO-reducing therapies (23). However, a recent report showed that MBG and ST-segment resolution were discordant in almost 40% of patients (22), which theoretically limits their use both in daily practice and in research. In the present study, ST-segment resolution was discordant with TIMI flow grade in 38% of the patients and with MBG in 48% of the patients. Neither TIMI flow grade nor MBG was predictive of functional outcome. Although incomplete ST-segment resolution was related to baseline function, it did not predict changes at follow-up. This finding is probably attributable to the relatively small number of patients studied.
By allowing the direct visualization of the nonperfused zone, gadolinium-enhanced CMR has provided new insights in the pathophysiology and prevalence of MVO (4,9,13,24). Current first-pass CMR techniques have high sensitivity for the diagnosis of MVO, detecting it in 65% to 87% of patients with successfully reperfused AMI and TIMI flow grade 3 (16,25,26). The reported prevalence of MVO with LGE imaging is lower: 28% to 58% (14,15,17,18,26,27). So far, only 2 studies have directly compared early and late MVO, and both showed a higher prevalence of early MVO but did not assess their relative values for the prediction of outcome (26,27). The difference in prevalence is attributed to ongoing slow diffusion of contrast into regions with less-severe microvascular damage and subsequently smaller areas of hypoenhancement at postponed (late) imaging (Fig. 3). The short-axis LGE images were acquired between 12 and 15 min after contrast administration (20). Consequently, we may have missed areas with MVO which had already disappeared due to wash-in of contrast. Late MVO therefore reflects infarcts with a more severely injured microvasculature. The present data extend previous studies by showing that, although first-pass imaging was more sensitive in detecting MVO (68% vs. 57%), MVO on LGE imaging had greater clinical relevance by identifying patients with worse functional outcome, and was in fact the only predictor of LV volumes and function at follow-up in a multivariable analysis that included all angiographic, electrocardiographic, and CMR measures of MVO.
Furthermore, the present study documents no statistical relationship between the extent of MVO and LV remodelling over time in patients with late MVO. This finding is in line with our previous results, where we reported no difference in LV indexes between patients with small or large areas of MVO (15), suggesting that the size and extent of MVO may be clinically less important than its mere presence.
MVO versus infarct size and transmurality
The present study also demonstrates that late MVO has diagnostic value beyond infarct transmurality, which is an established predictor of functional recovery of stunned or hibernating myocardium (28,29). Several studies have compared MVO (either by first-pass or by LGE imaging) to LGE infarct size or transmural extent, but results have not been conclusive, with some reports favoring MVO and others infarct size or transmurality as the best predictor of outcome (4,14,16,18,25). Although infarct size was the only independent predictor of baseline volumes and function in the present study, late MVO was a better predictor of the changes at follow-up. Regional analysis showed that late MVO was associated with increased wall thinning and less improvement of wall thickening, regardless of the degree of infarct transmurality. These results strongly suggest that, in the acute setting, MVO might be more relevant than infarct size or transmural extent.
Cardiovascular magnetic resonance assessment of MVO and infarction is typically performed between 2 and 9 days after reperfusion, because the extent of both MVO and infarction is stable within that period and has been shown to increase in the first 48 h (24,30). Thus, we assured a fixed infarct and MVO size, although we may have underestimated the full potential of CMR (and other parameters) to predict functional recovery. Additionally, because we performed LGE imaging between 12 and 15 min in order to optimally analyze infarct size and extent, we cannot exclude that earlier or later acquisition would have influenced the predictive value of late MVO.
Angiographic and electrocardiographic parameters of no-reflow are obtained in the acute setting of myocardial infarction, immediately or very early after reperfusion. The difference in timing may explain the relatively poor correlation between CMR and the other parameters. However, all parameters have been effectively used for the prediction of outcome after MVO, which justifies their comparison despite the difference in timing.
Our findings may become relevant for selecting patients that may benefit from adjunctive (e.g., cell) therapy to promote the repair of infarcted myocardium. In addition, because gadolinium-enhanced CMR accurately visualizes both infarct and MVO, it should be strongly recommended as a principal imaging technique in trials evaluating new therapeutic strategies to limit microvascular injury in the setting of AMI.
We found late MVO to be the most powerful predictor of functional outcome after AMI when directly comparing angiographic, electrocardiographic, and gadolinium-enhanced CMR characteristics of microvascular injury in a homogeneous group of patients after successful PCI. Its predictive value exceeded that of early MVO and assessment of transmural extent of infarction.
The authors thank the technicians, research nurses, and staff of the Cardiology and Radiology departments of the VU University Medical Center and Medical Center Alkmaar for their skilled assistance.
Supported by the Netherlands Heart Foundation, grant 2003B126.
- Abbreviations and Acronyms
- acute myocardial infarction
- cardiovascular magnetic resonance
- late gadolinium enhancement
- left ventricular
- left ventricular ejection fraction
- myocardial blush grade
- myocardial contrast echocardiography
- microvascular obstruction
- percutaneous coronary intervention
- systolic wall thickening
- Thrombolyis In Myocardial Infarction
- Received January 29, 2008.
- Revision received March 20, 2008.
- Accepted April 7, 2008.
- American College of Cardiology Foundation
- Ito H.,
- Maruyama A.,
- Iwakura K.,
- et al.
- Wu K.C.,
- Zerhouni E.A.,
- Judd R.M.,
- et al.
- van't Hof A.W.,
- Liem A.,
- Suryapranata H.,
- Hoorntje J.C.,
- de Boer M.J.,
- Zijlstra F.,
- Zwolle Myocardial Infarction Study Group
- Schroder R.,
- Dissmann R.,
- Bruggemann T.,
- et al.
- Kondo M.,
- Nakano A.,
- Saito D.,
- Shimono Y.
- Judd R.M.,
- Lugo-Olivieri C.H.,
- Arai M.,
- et al.
- Ito H.,
- Tomooka T.,
- Sakai N.,
- et al.
- Choi E.Y.,
- Seo H.S.,
- Park S.,
- et al.
- Kim R.J.,
- Chen E.L.,
- Lima J.A.,
- Judd R.M.
- Lima J.A.,
- Judd R.M.,
- Bazille A.,
- Schulman S.P.,
- Atalar E.,
- Zerhouni E.A.
- Hombach V.,
- Grebe O.,
- Merkle N.,
- et al.
- Baks T.,
- van Geuns R.J.,
- Biagini E.,
- et al.
- Ryan T.J.,
- Antman E.M.,
- Brooks N.H.,
- et al.
- Goldstein H.
- Sorajja P.,
- Gersh B.J.,
- Costantini C.,
- et al.
- Silva-Orrego P.,
- Colombo P.,
- Bigi R.,
- et al.
- Rochitte C.E.,
- Lima J.A.,
- Bluemke D.A.,
- et al.
- Gerber B.L.,
- Garot J.,
- Bluemke D.A.,
- Wu K.C.,
- Lima J.A.
- Beek A.M.,
- Kuhl H.P.,
- Bondarenko O.,
- et al.
- Wu K.C.,
- Kim R.J.,
- Bluemke D.A.,
- et al.