| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Institut für Klinische Radiologie, Klinikum Großhadern,
Marchioninistrasse 15, Munich 81377, Germany.
2 Department of Cardiology, Klinikum Innenstadt, Munich, Germany.
Received May 10, 2004;
accepted after revision December 22, 2004.
Address correspondence to A. Huber.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Forty-three patients with myocardial infarction were examined on a 1.5-T MR system 10 min after administration of contrast material (gadodiamide, 0.2 mmol/kg) with a single-shot 2D multislice technique (single-shot inversion recovery trueFISP) that allows one to image the entire short axis during one breath-hold (18 heartbeats) and with a segmented 2D single-slice technique (inversion recovery turboFLASH) that requires one breath-hold per slice (12 heartbeats). Signal intensity was determined in normal myocardium, in infarcted myocardium, and in the left ventricle. The contrast-to-noise ratio (CNR) of normal and infarcted myocardium was determined. The areas of hyperintense infarctions on selected slices and the entire volumes were compared for both sequence techniques.
RESULTS. The inversion recovery trueFISP sequence has a lower CNR than the inversion recovery turboFLASH sequence (mean values, 10.0 vs 12.9, respectively; p = 0.005) for differentiation of viable from nonviable myocardium. The CNR of injured myocardium and blood in the left ventricular cavity also has a lower value for the multislice technique compared with the single-slice technique (0.6 vs 1.2, respectively; p = 0.045). Assessment of the area of infarction within one slice (r = 0.97, p < 0.002) and of the volume of the entire infarction (r = 0.96, p < 0.003) is possible with excellent correlation of both techniques.
CONCLUSION. Despite having a lower CNR, the inversion recovery 2D single-shot trueFISP sequence allows fast and accurate identification of the area and volume of infarction with high spatial resolution within a single breath-hold.
Keywords: ischemia • heart • MR contrast agents • MRI • myocardial infarction
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
For cardiac MRI, high-end MR systems with strong gradient systems, dedicated surface coil techniques, and advanced pulse sequence techniques are necessary. The reason is that both high spatial and temporal resolution and a high signal-to-noise ratio have to be achieved, and artifacts caused by respiratory motion and cardiac contraction have to be suppressed [5]. Shorter scan times are advantageous to achieve better patient compliance; to avoid contrast change in the myocardium over time; and to shorten the examination time for a comprehensive examination of function, perfusion, and viability during one examination.
For imaging viability, single-slice MR techniques are widely used. The most established well-documented single-slice technique in the literature is an inversion recovery turbo fast low-angle shot (turboFLASH) technique for the detection of myocardial infarction and accurate assessment of viability [2, 4, 6]. This pulse sequence technique served as the gold standard in our study for the evaluation of a new fast multislice technique: an inversion recovery 2D single-shot true fast imaging with steady-state precession (trueFISP) sequence. This sequence allows imaging of nine slices during one breath-hold. The purpose of the study was to investigate the contrast-to-noise ratio (CNR) of infarction and normal myocardium for both pulse sequence techniques, to determine diagnostic accuracy of the multislice technique in detecting myocardial infarction (MI), and to compare the area and volume of infarction with both pulse sequence techniques. The single-shot inversion recovery trueFISP sequence acquires one slice during one R-R cycle every second R-R cycle. In addition to the shorter scanning time, the single-shot technique may help to avoid respiratory motion artifacts in uncooperative patients compared with a segmented pulse sequence technique. A recent study using a similar MR technique showed promising initial results in a small patient population [7]. In that study, a clear decrease of CNR was found. The current study with a large patient population can show if the change of CNR has an impact on diagnostic accuracy or if CNR is still high enough to use the technique in the future with high-spatial-resolution and parallel imaging techniques.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MRI
All MR images were obtained with a 1.5-T unit (Magnetom Sonata, Maestro
Class, Siemens Medical Solutions) with 8 channels and a dedicated 12-element
phased-array body coil. ECGs and heart rate were monitored with the
physiologic monitor of the imager (MR-compatible monitoring system). All
patients were examined in the supine position. A 20-gauge cannula was placed
in an antecubital vein for injection of contrast material. Scout images were
obtained to determine the exact position and standard orientations: the
long-axis view, the four-chamber view, and the short-axis views of the left
ventricle. Gadodiamide (Omniscan, Amersham Health) was injected at a dosage of
0.2 mmol/kg of body weight, followed by 20 mL of saline. The examination of
viability started after a 10-min delay.
Optimization of Inversion Time (TI)
Ten minutes after contrast material administration, a segmented inversion
recovery cine trueFISP pulse sequence (TI scout) acquisition was performed at
a midventricular short-axis location. This acquisition was used to determine
the TI at which the signal of normal myocardium is null for the subsequent
delayed enhancement acquisition. The details of the sequence, based on an
approach proposed by Scheffler and Henning
[8], are shown schematically in
Figure 1. Thus, after a
nonselective hyperbolic secant pulse applied after the R wave trigger,
multiple cardiac phases were acquired using a segmented trueFISP readout. The
resulting 23 cine images show the recovery of the longitudinal magnetization
at multiple time intervals—in increments given by the temporal
resolution—after inversion. In our implementation, 7 k-space lines were
acquired per segment per cardiac phase, resulting in a temporal resolution of
15 msec (TR/TE, 2.2/1.1; flip angle, 50°) and a breath-hold duration of 14
cardiac cycles. The spatial resolution was typically 2.6 x 1.8 x
10 mm3, the field of view was 340 x 265 cm, and the matrix
was 192 x 102. From the resulting images with TI values from 185 to 515
msec, the optimum TI value was determined by visual assessment in agreement by
two operators. The optimum TI was determined as the image in which the signal
intensity of normal myocardium was near null and the signal intensity of the
left ventricular cavity was lower compared with that of the infarcted region.
There was positive contrast between infarction and normal myocardium
[3]. The TI scout sequence is
available as a standard product.
|
Imaging Delayed Enhancement
A segmented inversion recovery turboFLASH sequence was used to cover the
entire left ventricle in the short-axis view. One slice was acquired per
breath-hold. This sequence technique was used as standard of reference,
because it is well established and was investigated in the current literature
[2,
4,
6]. For the inversion recovery
turboFLASH sequence, the TR was 11 msec; TE, 4 msec; flip angle, 25°; and
bandwidth, 140 Hz/pixel. Thirty-two k-space lines were acquired during one
cardiac cycle. With data acquisition every second heart beat, imaging of one
slice was possible during nine cardiac cycles. Data acquisition was performed
at mid diastole at a time window of minimal cardiac motion during 352
msec.
The inversion recovery turboFLASH sequence was compared with a new pulse sequence technique: a single-shot 2D inversion recovery trueFISP technique. That new technique allows imaging of nine slices during one breath-hold with data acquisition every second heart beat. The single-shot 2D inversion recovery trueFISP sequence was performed first, just after the TI scout, because the acquisition takes only a single breath-hold. Thus, no relevant change of optimal TI can be expected before the inversion recovery turboFLASH images are acquired. For the single-shot inversion recovery trueFISP sequence, the TE was 1.1 msec; TR, 2.2 msec; flip angle, 60°; and the bandwidth, 1,220 Hz/pixel. The acquisition window was 352 msec for 160 k-space lines. With data acquisition every second heart beat, imaging of one slice was possible during one cardiac cycle. Seventeen cardiac cycles were necessary to image nine slices during one single breath-hold. The single-shot inversion recovery trueFISP sequence [6] involves constant radiofrequency pulsing and gradient refocusing of transverse magnetization. The nonselective inversion recovery pulse is used to improve the T1 contrast compared with the T2 contrast.
For both pulse sequences, the previously determined optimal TI was used. Typical values for the field of view were 330 x 286 mm; for the spatial resolution, 1.3 x 1.8 x 8 mm3; and for the matrix size, 256 x 160. Identical values were used for both pulse sequences.
Data Evaluation
Regions of interests were defined for the hyperenhanced and infarcted areas
in the normal myocardium and in the left ventricular cavity with Argus
software at a Leonardo workstation (Siemens Medical Solutions). The regions of
interest provide information about signal intensity, contrast, and area of
infarction. For calculation of CNR, additional regions of interest were
positioned in the background that provided information about the SD of
background noise. The maximum transmural extent of the infarction was
determined in relation to the thickness of the myocardium.
The signal intensity (SI) values and the CNR values were
determined for both sequence types. The CNR value for the infarcted myocardium
in comparison with the normal myocardium was determined as follows:
 | (1) |
 | (2) |
The area of hyperenhanced myocardium was determined for each short-axis slice on inversion recovery turboFLASH images and on the single-shot inversion recovery trueFISP images. First, one representative slice was selected for comparison of both sequence types. Second, the single values of the area of infarction were summarized for each patient and each sequence type to determine the entire infarct volume. The results of the two pulse sequence types were compared.
Paired Student's t tests were calculated for the values of the hyperenhanced area on a representative slice (in centimeters squared) and of the volume (in milliliters) of the entire infarction for a significance level of p = 0.01 (H0: µd = 0), where H0 means null hypothesis, µd = µ1 - µ2, µ1 is the mean value of population 1, and µ2 is the mean value of population 2. The correlation coefficient and the equation of regression were determined for the infarct volume and the area of the infarction on a representative slice for the two pulse sequences. A Bland-Altman's plot was calculated for the volume and area of infarction.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The inversion recovery trueFISP sequence revealed a lower CNR than the inversion recovery turboFLASH sequence (mean values, 10.0 vs 12.9, respectively; p = 0.005) for viable and nonviable myocardium, calculated using equation 1. The CNR of injured myocardium and blood in the left ventricular cavity, calculated using equation 2, was also lower for the multislice technique compared with the single-slice technique (0.6 vs 1.2, respectively; p = 0.045). The mean values of the signal intensity of normal myocardium and infarcted myocardium and the SD of background noise were 12.2, 54.5, and 4.3 for the single-shot inversion recovery trueFISP sequence and 20.0, 83.6, and 4.9 for the inversion recovery turboFLASH, respectively. The mean values for the signal-to-noise ratio (SNR) of infarcted myocardium were 12.6 and 17.1, respectively. The mean imaging time for the inversion recovery single-shot trueFISP sequence was 16 ± 3.2 (SD) sec. The mean imaging time for the segmented inversion recovery turboFLASH sequence was 10 min 13 sec ± 45 sec.
The mean volume of the injured myocardium per patient was 26.3 ±
13.4 mL with the inversion recovery turboFLASH sequence and 26.1 ± 13.5
mL with the single-shot inversion recovery trueFISP sequence. The Student's
t test showed that there is no difference for values of infarct
volume for a significance level of p = 0.01. The assessment of the
volume of the infarction was possible with excellent correlation of both
techniques (r = 0.96, p < 0.003). The regression equation
was
 |
 |
Figures 2A and 2B reveals a comparison of MR images of a 63-year-old patient with occlusion of the left anterior descending artery. Figures 3A and 3B shows an anteroseptal myocardial infarction visualized with the inversion recovery trueFISP sequence (Fig. 3A) and with the inversion recovery turboFLASH sequence (Fig. 3B).
|
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In contrast to 18F-FDG PET, contrast-enhanced MRI is able to image the myocardium with superior spatial resolution [2]. Thus, complete transmural infarctions can be distinguished from nontransmural infarctions in at least 25% steps of the thickness of the left ventricular myocardium. Viable myocardium has the potential to improve its contractility after revascularization even if it has a transmural extent of 50%, whereas complete transmural infarctions have a low probability to improve their contractility. Especially in patients with a reduced left ventricular ejection fraction, the prediction of possible improvement of regional wall thickening is important for their outcome after surgical revascularization. The spatial resolution we used in our study had a maximum in-plane voxel length of 1.8 mm. Thus, a subdivision of the thickness of the myocardium in four 25% steps for assessment of the transmurality of the infarction was possible for both pulse sequence techniques, because the identical spatial resolution was used for the segmented single-slice technique and the multislice technique.
It has been known for several years that regions of myocardial infarction exhibit higher signal intensity than normal myocardium on T2-weighted images [9, 10] and on T1-weighted images after administration of extravascular contrast agents [2, 4, 6]. Since the initial experience, many studies have been performed using a variety of pulse sequences to differentiate injured from normal myocardium. Kim et al. [2] and Simonetti et al. [6] implemented a breath-hold inversion recovery segmented turboFLASH sequence for T1-weighted postcontrast imaging of infarction that produces strongly T1-weighted images because of the inversion pulse. Simonetti et al. compared 10 pulse sequences to acquire contrast material-enhanced images after myocardial infarction and found the greatest differences in regional myocardial signal intensity for the breath-hold inversion recovery turboFLASH sequence. In the current literature, this sequence is the most investigated and established one [1-4, 6, 7], which is the reason we chose to use it as a standard of reference in our study.
The new pulse sequence technique (inversion recovery trueFISP) allows imaging nine slices during one breath-hold (Fig. 6). Thus, the entire left ventricular myocardium can be covered. Our results show a close agreement of the area and volume of infarction for the reference sequence (inversion recovery turboFLASH) and the inversion recovery trueFISP sequence. A mixed T1 to T2 contrast is known for trueFISP sequences [6]. Therefore, the area of infarction could have been overestimated in patients with subacute and acute myocardial infarctions because the edema of the infarction was larger than the delayed enhancing nonviable myocardium [9, 10]. The results of our study show that the area of infarction is not overestimated on the inversion recovery trueFISP sequence when compared with the inversion recovery turboFLASH sequence. The first reason may be the use of the inversion pulse. The second reason may be the use of the individual sequence parameters, a short TR and TE, and a flip angle of 60°. The inversion pulse and the sequence parameters support the T1 contrast compared with the T2 contrast.
However, in spite of the optimization of the T1 contrast by the pulse sequence, strong signal was observed in the fluid in patients with pericardial effusion or pleural effusion, indicating that a relevant T2 contrast is still present. Therefore, the main reason for accurate detection of the volume of infarction may be the weak T2 signal and contrast caused by the edema in the infarcted muscle and the surrounding area compared with fluid or with contrast enhancement of the infarcted myocardium. Determination of the SNR values showed a higher mean value for the trueFISP sequence than for the turboFLASH sequence. Thus, reduction of CNR is caused by higher signal intensity of the normal myocardium. There may be two possible reasons: either the TI value determined by the TI scout by visual evaluation needs to be corrected by a certain value compared with the inversion recovery turboFLASH sequence or very low contrast uptake in normal myocardium after 10 min causes a higher signal on the trueFISP sequence than on the turboFLASH sequence.
Usually, the TI is adapted individually to null the signal of the normal myocardium and improve contrast between infarcted and viable myocardium. The performance of inversion recovery delayed hyperenhancement is highly sensitive to the TI selected using magnitude images. A suboptimal TI can cause a reduction in contrast or even destroy the contrast between infarcted and normal myocardium completely [6]. Different methods with which to optimize the TI value are available. Different TI values can be used, and the best one can be chosen by trial and error. Another method is to use a TI scout sequence, which acquires data sets with different TI values and reduced spatial resolution during one breath-hold [8, 11, 12]. From the complete image set at one slice position, the TI value of the image with the optimal contrast can be selected for imaging the entire myocardium.
The CNR of infarcted myocardium versus viable myocardium was significantly lower (22%) for the multislice single-shot inversion recovery trueFISP sequence than for the segmented inversion recovery turboFLASH sequence. However, in our study the difference of CNR between the inversion recovery trueFISP sequence and the inversion recovery turboFLASH sequence for infarction and normal myocardium was lower compared with that reported in a recent study with a small number of patients [7]. Thus, it may be possible to increase spatial resolution by factor of 2 or 3 by combining the single-shot inversion recovery trueFISP pulse sequence with parallel imaging. It may also be possible to use turboFLASH techniques as single-shot sequences when they are combined with parallel imaging—for example, a sensitivity-encoding (SENSE) factor of 4 or 8 with an 8-channel system. However, further studies are necessary to show whether CNR is still high enough with the use of higher acceleration factors.
The contrast between injured myocardium and the left ventricular cavity is relatively low for both pulse sequence types and lower for the multislice technique. However, a positive contrast that allows accurate determination of the area and volume of infarction is still present. To increase contrast between the left ventricular cavity and injured myocardium, using a longer time interval between contrast material application and imaging viability could help. An advantage of the single-shot technique compared with the inversion recovery turboFLASH technique is that one slice is acquired during one R-R cycle, not one slice in a segmented way over several R-R cycles. Therefore, the single-shot technique may help to reduce respiratory artifacts in uncooperative patients and cardiac motion artifacts in patients with arrhythmias. With a short time interval for imaging the entire left ventricle, contrast change in the myocardium over time can be avoided [13].
Various methods of nuclear medicine are available to distinguish viable from nonviable myocardium. 18F-FDG PET is known as the most reliable method [14, 15]. The thallium SPECT examination has a lower diagnostic accuracy for differentiation of viable from nonviable myocardium than 18F-FDG PET. The reason is that infarcted and viable myocardium can be hypoperfused at rest and under pharmacologic stress or exercise [16-19]. An important disadvantage of the 18F-FDG PET examination is a lower spatial resolution compared with delayed enhancement MRI. Furthermore, the examination of glucose metabolism can be difficult, especially in diabetic patients who are among the group of patients with elevated risk of coronary heart disease [18, 19]. In some cases, it is possible that 18F-FDG PET shows false-positive results in normokinetic segments of the myocardium [15].
In contrast to the methods of nuclear medicine, delayed enhancement MRI with a multislice approach can be performed during a shorter scanning time with less extensive preparation of the patient and contrast material. With short examination times, MRI has the potential to become a more available and more economic examination than 18F-FDG PET [20]. The MR examination of myocardial viability can be combined with an examination of function and perfusion at rest and under pharmacologic stress [21].
In conclusion, single-shot inversion recovery trueFISP allows reliable detection of infarction for an accurate assessment of the area and volume of infarction and an examination with high spatial resolution. This new multislice MR technique seems to be able to replace the established single-slice technique without loss of relevant diagnostic information.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. U. Bauner, O. Muehling, D. Theisen, C. Hayes, B. J. Wintersperger, M. F. Reiser, and A. M. Huber Assessment of Myocardial Viability with 3D MRI at 3 T Am. J. Roentgenol., June 1, 2009; 192(6): 1645 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. C. M. van den Bosch, J. J. M. Westenberg, J. C. Post, G. Yo, J. Verwoerd, L. J. M. Kroft, and A. de Roos Free-Breathing MRI for the Assessment of Myocardial Infarction: Clinical Validation Am. J. Roentgenol., June 1, 2009; 192(6): W277 - W281. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Saremi, J. D. Grizzard, and R. J. Kim Optimizing Cardiac MR Imaging: Practical Remedies for Artifacts1 RadioGraphics, July 1, 2008; 28(4): 1161 - 1187. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
