| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the Departments of Cardiology (G.K.L., K.M., A.A.B., M.K., U.S., M.M., T.M.) and Diagnostic and Interventional Radiology (A.S., M.P.B., G.A.), University Hospital Eppendorf, Hamburg, Germany; Coordination Center for Clinical Trials, University of Düsseldorf, Düsseldorf, Germany (P.E.V.); and Department of Radiology, University of California–San Francisco, San Francisco, Calif (M.S.). Received July 14, 2006; revision requested September 18; revision received November 21; accepted January 8, 2007; final version accepted February 12. Address correspondence to G.K.L., Roentgeninstitut, Kaiserswerther Str 89, 40476 Düsseldorf, Germany (e-mail: lund{at}roentgeninstitut.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
Materials and Methods: The study was approved by the institutional ethics committee and all patients gave written informed consent. In 55 patients (48 men, seven women; mean age ± standard deviation, 56 years ± 13), contrast material–enhanced and cine MR imaging were performed 5 days ± 3 and 8 months ± 3 after myocardial infarction (MI). Microvascular obstruction (MO) and infarct size were estimated at first-pass enhancement (FPE) and delayed enhancement (DE) MR, respectively. Remodeling was defined as an increase in LV end-diastolic volume index of 20% or higher at follow-up. Differences in continuous and categorical data were analyzed by using Student t test and Fischer exact test as appropriate.
Results: Patients with remodeling (n = 13, 24%) had higher creatine kinase MB (P < .05), more anterior infarcts (P < .05), more often a reduced Thrombolysis in Myocardial Infarction flow (P < .05), larger infarct size at DE MR (P < .001), a greater extent of MO at FPE MR (P < .01), lower ejection fraction (P < .001) and higher LV end-systolic volume index (P < .01). Infarct size at DE MR was a powerful predictor for remodeling (odds ratio: 1.18, P < .001), demonstrating that the risk for remodeling increased 2.8-fold with each 10% increase in infarct size. Infarct size of 24% or more of LV area predicted remodeling with high sensitivity (92%), specificity (93%), and accuracy (93%). Infarct resorption was larger in patients with remodeling (P < .01).
Conclusion: Infarct size 24% or more of the LV area constitutes an important threshold to predict remodeling. Patients with remodeling develop disproportionate infarct resorption.
© RSNA, 2007
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
Contrast material–enhanced magnetic resonance (MR) imaging has the ability to generate several parameters predictive of remodeling including LV volume, LVEF, MO, and infarct size (9–11). Recent studies (11,12) showed that delayed enhancement (DE) MR correctly estimates infarct size in acute and chronic MI. Experimental data showed the accuracy of first-pass enhancement (FPE) MR in determining the extent of MO (13). The accuracy of MR imaging renders this technique as an attractive tool to identify patients at risk for remodeling. Furthermore, serially performed contrast-enhanced MR offers the possibility to study infarct resorption after AMI. Previous studies have uniformly shown that the infarct size decreases after MI (14–17). However, factors that have an impact on infarct resorption and the relative changes in infarct size in patients with and those without remodeling are not well understood.
The purpose of the study was to prospectively evaluate the accuracy of clinical and cardiac MR imaging parameters for predicting LV remodeling by using follow-up imaging as reference standard, and to prospectively evaluate infarct resorption in patients with reperfused first MIs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
Patients were treated by using either direct angioplasty (n = 58) or thrombolytic therapy (n = 10). Coronary angiography was performed in all patients to identify the infarct-related artery. Perfusion of the infarct-related artery was determined at consensus reading by two observers (G.K.L. and M.P.B., 12 and 4 years experience reading coronary angiograms, respectively) by using the Thrombolysis in Myocardial Infarction trial criteria (18). Attending physicians performed medical treatment without knowledge of MR data because the initial MR study was completed after medical treatment was assigned.
Follow-up MR imaging was not performed in 13 patients: one died, two were lost to follow-up, three had an ischemic event in the follow-up period, and seven refused repetition of MR (Fig 1). The final study group comprised 55 patients (48 men, seven women; mean age ± standard deviation, 56 years ± 13). This number of patients guaranteed a study power of 90% or greater, owing to a clinical relevant probability to develop remodeling of 25% ± 10 (8,19) and a significance level of 5%.
|
FPE MR was used to detect regional MO, whereas DE MR was used to estimate infarct size. For FPE MR, a bolus of 0.1 mmol/kg gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) per kilogram of body weight was injected into an antecubital vein followed by 20 mL of saline solution. Injection was performed at 3 mL/sec by using a power injector (Spectris MR Injector; Medrad, Indianola, Pa). Immediately after injection one image per R-R interval was acquired for the following 60 heartbeats by using an inversion-recovery prepared, T1-weighted turbo fast low-angle shot sequence. FPE MR was performed in the first 36 patients at a single midventricular level, whereas multisection imaging was performed in the subsequent 19 patients at the apical, midventricular, and basal LV levels. The imaging parameters were: 2.4/1.2; inversion time msec, 118; section thickness, 10 mm; field of view, 350 x 306 mm (8:7 rectangular field of view); matrix, 128 x 90; pixel size, 2.7 x 3.4 mm. Patients were asked to hold their breath for as long as possible and then allowed shallow breaths. The sequence lasted 50–60 seconds.
At 10 minutes after injection, multisection DE MR was performed in all patients at the apical, midventricular, and basal levels by using a segmented inversion-recovery prepared T1-weighted turbo fast low-angle shot sequence (20). The imaging parameters were: 7.6/3.4; inversion time msec, 220–300 to null the signal intensity of normal myocardium; delay after trigger, 400 msec; section thickness, 6 mm; field of view, 350 x 262 mm (8:6 rectangular field of view); matrix, 256 x 132; pixel size, 1.37 x 2.0 mm. Images were obtained every other heartbeat to give time for more complete inversion recovery. The sequence lasted approximately 15 seconds.
MR Data Analysis
Initial and follow-up images were transferred to a computer (Macintosh; Apple Computers, Cupertino, Calif) and data were analyzed by using a public domain program (NIH Image, version 1.62; U.S. National Institutes of Health, available at: http://rsb.info.nih.gov/nih-image/). All MR measurements were performed independently by two of three observers blinded to all other information (G.K.L., M.K., U.S., with 7, 3, and 3 years experience interpreting MR, respectively). Data are given as the mean of two observers. Presence of MO was identified on FPE MR images as an area of persistent subendocardial hypoenhancement with signal intensity less than two standard deviations than that of surrounding hyperenhanced myocardium (10).
Delayed enhancement was considered present if the signal intensity was more than two standard deviations than that of remote nonenhanced myocardium on DE MR (10). Signal intensity of remote myocardium was measured in three regions of interest with a size of 0.5 cm2, which were equally distributed in the middle of the nonenhanced myocardium. Mean signal intensity and standard deviation were obtained and the threshold level was calculated. Images were thresholded to these levels, and extent of MO and infarct size was measured and expressed as a percentage of LV area. On DE MR images, a zone of persistent subendocardial hypoenhancement surrounded by a hyperenhanced area was included in the measurement of infarct size (Fig 2) (21). Measurements were performed on each section and values were averaged between sections to obtain mean infarct or MO size per LV area.
|
Statistical Analysis, Reference Standard
Continuous measurements are summarized by using mean ± standard deviation. Differences in continuous data between or within groups were analyzed by using a two-sided Student t test for unpaired or paired data as appropriate. Categorical variables were compared by using the Fischer exact test. Logistic regression was performed by using multiple stepwise regression analysis to identify independent predictors of remodeling at 8 months after MI. In this analysis, all parameters were selected that had a P value of less than .1. LVEF was excluded because of known interdependency between LVEF and infarct size (1,15).
Receiver operating characteristic (ROC) analyses were performed to evaluate the diagnostic accuracy of the model to predict remodeling by using an increase in LV end-diastolic volume index of 20% or higher on follow-up imaging as the reference standard. Areas under the curve were determined and cutoff values were chosen to maximize the Youlden criterion for each parameter. Sensitivities, specificities, and diagnostic accuracies were determined with regard to each cutoff. A P value of less than .05 was required for significance. Statistical analysis was performed with SPSS for Windows (version 13.0; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
|
Clinical and MR Predictors of Remodeling
Stepwise multiple regression analysis was performed to evaluate independent predictors of remodeling. Clinical variables used for analysis were age, peak creatine kinase, peak creatine kinase MB, presence of q-wave infarct, infarct location, initial Thrombolysis in Myocardial Infarction flow, treatment with direct angioplasty, and stent placement in the infarct-related artery. MR variables were infarct size, presence and extent of MO, LVEF, LV end-diastolic volume index, LV end-systolic volume index, and LV mass index. For multiple regression analysis, factors showing a P value of less than .1 in univariate analysis were selected.
At initial imaging, a strong correlation was found between infarct size when measured by using DE MR and LVEF (r = 0.68, P < .001), displaying interdependency between the two parameters. LVEF was not included into subsequent multiple regression analyses because LVEF is a dependent variable of infarct size. Stepwise multiple regression analysis identified infarct size when measured by using DE MR as a powerful predictor for remodeling with an odds ratio of 1.18 (95% confidence interval: 1.07, 1.30; P < .001). This odds ratio demonstrated that the risk for remodeling increased 2.8-fold with each 10% increase in infarct size.
ROC Analysis
Infarct size, when measured by using DE MR, had the best test performance to predict remodeling with an area under the curve of 0.891 (Fig 3). Other estimates of myocardial injury such as extent of MO or peak release of creatine kinase MB were inferior to predict remodeling (Table 2). ROC analysis revealed that an infarct size of 24% or more of LV area was the best cutoff to predict remodeling. This cutoff resulted in a sensitivity of 92% and a specificity of 93% to predict remodeling. The correlation between infarct size at baseline and change in LV end-diastolic volume index 8 months after MI is shown in Figure 4.
|
|
|
Also, the relative reduction in infarct size was larger in patients with remodeling (–34% ± 26) compared with patients without remodeling (–12% ± 35, P < .05). A good correlation was found between infarct size at initial imaging and infarct resorption at follow-up (r = 0.65, P < .001, Fig 5), indicating that the initial infarct size determines later infarct resorption. At follow-up, the reduction in infarct size correlated with the increase in LV end-diastolic volume index (r = 0.41, P < .01), indicating that infarct resorption and remodeling are related processes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
Prediction of Remodeling
Animal studies (1,24) showed that remodeling and subsequent changes in LV function are linearly related to infarct size. Clinical studies (2–4,7) suggested that several parameters are predictive of remodeling, including anterior infarct location, patency of the infarct-related artery, LVEF, and infarct size. In agreement with these findings, our study identified a number of parameters predictive of remodeling.
At initial imaging, an intense correlation was found between infarct size when measured by using DE MR and LVEF. This interdependency was of importance for multiple regression analysis. The two parameters had equal impact in the model, so one parameter had to be excluded. For identification of independent predictors, LVEF was removed from the analysis because this parameter is, by nature, the dependent variable of infarct size. Final multivariate analysis showed that infarct size when measured by using DE MR was a powerful independent predictor of remodeling.
The finding that MO was not an independent predictor of remodeling is somewhat in contradiction to previous MR studies (9,25). In an animal study, Gerber et al (25) showed that extent of MO was superior to infarct size to predict remodeling. However, the experimental findings are not directly transferable to our data for several reasons.
First, the animal study analyzed occurrence of remodeling at 10 days after MI. Previous data showed that early remodeling is mainly related to myocardial lengthening and thinning of the infarcted region, whereas the chronic period after MI is characterized by more complex geometric changes in the infarcted and the noninfarcted region (1). Therefore, the ultimate infarct size may represent a more important factor for later remodeling because it has an impact on the entire LV geometry (1).
Second, MI was induced in the animal study by using balloon occlusion of the left anterior descending artery for 90 minutes resulting in presence of MO in all animals. In our study, MO occurred in 38% of patients, which is typical for patients receiving modern reperfusion therapy (4,9). Presence of MO in all animals explains the dominant effect of this parameter in the experimental setting. Conversely, the lower frequency of MO in our study is most likely responsible for the reduced predictive value of this parameter in the clinical setting.
Threshold of Infarct Size for Remodeling
Previous studies suggested that a certain amount of myocardium has to be infarcted before remodeling occurs. On the basis of theoretical assumptions, Klein et al (26) suggested that LV enlargement occurs if 20%–25% of the LV area is noncontractile. A clinical study analyzed the effect of infarct size estimated by the extent of wall motion abnormality on remodeling (2). That study revealed that patients with a mean infarct size of 20% ± 3 LV area showed progressive LV dilatation at 3 years follow-up. Chareonthaitawee et al (7) studied the impact of infarct size obtained by means of sestamibi nuclear imaging on later LV function. That study found that only patients with an infarct size larger than 25% LV area demonstrated a relevant decrease in LVEF in the subsequent year.
Our data confirmed and extended the knowledge about the impact of infarct size on remodeling. ROC analysis revealed that an acute infarct size of 24% or more of LV area was the best cutoff to differentiate between remodeling and preservation of LV geometry. Furthermore, infarct size measured by using DE MR had a higher diagnostic accuracy for prediction of remodeling than enzymatic estimation of infarct size. An infarct size of 24% or higher of LV area constitutes an important threshold for later remodeling, indicating that the myocardium is severely injured and not able to maintain the LV geometry in future. Patients with an infarct size of 24% of LV area or more should be regularly monitored and intensively treated with antiremodeling drugs to prevent the progression of LV enlargement to overt heart failure.
Infarct Resorption
In our study, the contrast-enhanced region decreased by 26% from initial imaging to follow-up. Similar data were found by Hombach et al (17), who showed that the infarct size decreased by 28%, from 11.4% ± 7.2 to 7.8% ± 5.3 from 6 days to 8 months, respectively, after MI. Our findings are also consistent with those of two recently published studies performing the follow-up study earlier after the acute MI (14,15). Ingkanisorn et al (15) found a 31% decrease in infarct size more than 2 months after acute MI. Similarly, Choi et al (14) found a 27% decrease in infarct size over the same time period.
Our study additionally showed a link between infarct resorption and remodeling by performing acute and follow-up studies with DE MR and cine MR imaging. DE MR imaging demonstrated that patients with remodeling had substantially larger resorption of infarct size compared with patients without remodeling. Furthermore, the initial infarct size was directly related to later infarct resorption. Our data indicate that the acute infarct size has an impact on later infarct resorption. The linear relationship between the decrease in infarct size and the increase in LV end-diastolic volume index indicates that infarct resorption and remodeling are two related processes. Our data suggest that a larger necrosis in AMI results in a smaller scar, which is not able to maintain the shape of the LV in future. Therefore, limitation of infarct size should be the ultimate goal of infarct therapy to preserve acute LV function and to prevent later disproportionate infarct resorption and remodeling.
Limitations
Unlike infarct size, measurement of MO was not obtained from all patients at multisection imaging. Although subgroup analysis did not reveal any differences in presence and extent of MO in patients with single- and multisection FPE MR, this difference in the imaging strategy may have reduced the predictive value of MO for remodeling. Current imaging strategies enable multisection measurement of MO on four to five sections and multisection DE MR allows more accurate quantification of infarct size. These imaging strategies enable full coverage of the LV and should be included in future studies to overcome the current limitations.
Infarct size, when measured by using DE MR, can predict remodeling and can therefore be a tool for risk stratification of patients who recently experienced MI. An infarct size of 24% or more of LV area constitutes an important threshold indicating that the myocardium is not able to maintain the LV geometry in the future. Patients with such large infarcts should be intensively treated with antiremodeling therapy to prevent progressive remodeling to overt heart failure. Our study also showed a link between acute infarct size and later infarct resorption and remodeling. Patients with large infarcts not only have reduced LV function but also develop disproportionate infarct resorption and remodeling.
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
|
IMPLICATION FOR PATIENT CARE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
|
FOOTNOTES |
|---|
Abbreviations: AMI = acute MI • DE = delayed enhancement • FPE = first-pass enhancement • LV = left ventricular • LVEF = LV ejection fraction • MI = myocardial infarction • MO = microvascular obstruction • ROC = receiver operating characteristic
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, G.K.L., A.S., M.P.B.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, G.K.L., M.S.; clinical studies, G.K.L., A.S., K.M., A.A.B., M.P.B., M.K., U.S., M.M., G.A., T.M.; statistical analysis, G.K.L., K.M., A.A.B., P.E.V.; and manuscript editing, G.K.L., A.S., P.E.V., G.A., M.S.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE...References |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Nijveldt, M. B. M. Hofman, A. Hirsch, A. M. Beek, V. A. W. M. Umans, P. R. Algra, J. J. Piek, and A. C. van Rossum Assessment of Microvascular Obstruction and Prediction of Short-term Remodeling after Acute Myocardial Infarction: Cardiac MR Imaging Study Radiology, February 1, 2009; 250(2): 363 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Gibbons, P. A. Araoz, and E. E. Williamson The year in cardiac imaging. J. Am. Coll. Cardiol., January 6, 2009; 53(1): 54 - 70. [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Yang, J. Linden, S. S. Berr, I. L. Kron, G. A. Beller, and B. A. French Timing of adenosine 2A receptor stimulation relative to reperfusion has differential effects on infarct size and cardiac function as assessed in mice by MRI Am J Physiol Heart Circ Physiol, December 1, 2008; 295(6): H2328 - H2335. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Saeed, A. Martin, A. Jacquier, M. Bucknor, D. Saloner, L. Do, P. Ursell, H. Su, Y. W. Kan, and C. B. Higgins Permanent Coronary Artery Occlusion: Cardiovascular MR Imaging Is Platform for Percutaneous Transendocardial Delivery and Assessment of Gene Therapy in Canine Model Radiology, November 1, 2008; 249(2): 560 - 571. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
