| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, British Heart Foundation Cardiac MRI Unit, General
Infirmary at Leeds, Great George St., Rm. 170, D Fl., Jubilee Wing, Leeds LS1
3EX, UK.
2 Cardiac MRI Unit, Franz-Volhard-Klinik, Charité/Humboldt University,
Berlin, Germany.
3 Department of Medical Physics, University of Leeds, Leeds, UK.
Received July 21, 2005;
accepted after revision November 3, 2005.
Supported by a Royal College of Radiologists Priming Grant. D. R.
Messroghli was supported by a Marie Curie fellowship grant by the European
Commission.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Eight patients with chronic aortic regurgitation and normal coronary arteries awaiting surgical valve replacement underwent a comprehensive MRI examination that included assessment of left ventricular function, severity of valvular regurgitation, and presence of overt myocardial scar evidenced by delayed enhancement. For each patient, myocardial T1 relaxation times determined with a modified Look-Locker technique before and after contrast administration were compared with values previously established for 15 healthy volunteers.
RESULTS. There was no statistical difference (p > 0.05) in slice-averaged myocardial T1 relaxation times either before or after gadolinium administration in the patient group compared with the normal range of times. Segmental averaged T1 relaxation times in segments with abnormal wall motion did, however, show statistically significant differences from healthy controls 10, 15, and 20 minutes after administration of gadolinium (510 vs 476 milliseconds, p = 0.001; 532 vs 501 milliseconds, p = 0.002; 560 vs 516 milliseconds, p = 0.001, respectively). Two of the aortic regurgitation patients also had focal areas of myocardial delayed enhancement.
CONCLUSION. Segment-based myocardial T1 mapping has the potential for showing differences between relaxation times in aortic regurgitation and in normal hearts, suggesting the existence of a diffuse myocardial fibrotic process.
Keywords: aortic regurgitation • cardiac imaging • left ventricle • MRI • myocardial fibrosis
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It is accepted that the loading conditions associated with chronic aortic regurgitation lead to insidious myocardial fibrosis and that this condition precedes and may contribute to the development of heart failure [3, 4]. Current recommendations for surgical replacement of the valve depend on the degree of dilatation of the left ventricle and reduction in ejection fraction, because symptoms occur relatively late [5]. However, evidence of dilatation or dysfunction seen with standard imaging techniques such as echocardiography and angiography may imply the presence of fibrosis, and the presence of considerable preoperative fibrosis leads to suboptimal operative results [3].
Cardiac MRI can depict areas of overt scar or fibrosis after administration of gadopentetate dimeglumine, particularly after myocardial infarction [6, 7]. These delayed enhancement-based techniques, however, are limited by issues of spatial resolution. This limitation is particularly true in disease processes that produce diffuse patchy rather than overt fibrosis. T1 mapping techniques are based on images produced with a standardized scale, that is, milliseconds rather than the arbitrary signal intensities used in standard MRI studies. The implication is that signal intensity within a pixel depends on the absolute longitudinal relaxation properties of the tissue voxel [8]. Calculation of the relaxation time of each pixel within a parametric image can help detect subtle differences in regional tissue characteristics, allowing for contrast resolution not related to arbitrary differences in signal intensity. The purpose of this study was to use high-resolution T1 mapping before and after contrast administration to investigate the feasibility of using MRI to detect areas of fibrosis in myocardium exposed to chronic aortic regurgitation and to compare the results with a previously established normal range.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
The control group comprised 15 healthy volunteers (nine men; 33.1 ± 6.5 years old) as previously described [9]. All volunteers gave written informed consent; separate ethical approval had been given for this study. None of the volunteers had a history of cardiovascular or systemic disease, and none had symptoms at scanning. Tympanic temperature was recorded for patients and control subjects before scanning to exclude unsuspected pyrexia that could affect T1 relaxation time.
MRI
All examinations were performed on a 1.5-T scanner (Intera CV, Philips
Medical Systems) with a Master gradient system (30 mT/m peak gradient
amplitude, 150 m/T/s slew rate) and a five-element cardiac synergy surface
coil. A four-lead vectorcardiogram was used for ECG gating.
A comprehensive cardiac MRI protocol was used in the assessment of each patient. LV function was assessed with contiguous multiple-slice short-axis cine images in a balanced steady-state free precession (SSFP) pulse sequence. Severity of aortic regurgitation was quantified with a standard free-breathing T1-weighted gradient-echo phase velocity-encoded sequence. Three T2-weighted spectral presaturation with inversion recovery fat-saturation fast spin-echo slices (apical, middle, and basal) were acquired before the T1 mapping sequences to exclude the presence of regional active inflammation as a cause of abnormal findings. This sequence involves application of a slice-selective fat-saturation pulse between the preparatory inversion pulse pair and the readout used in a standard double inversion recovery black blood fast spin-echo sequence. This method results in suppression of signal from both fat and inflowing blood. Presence of delayed enhancement was visualized by use of a standard inversion recovery T1-weighted gradient-echo technique.
A modified Look-Locker inversion recovery sequence was used for generation of T1 relaxation time pixel maps of a single midventricular slice, as previously described [10]. This sequence essentially consisted of three inversion recovery-prepared ECG-synchronized Look-Locker trains executed within one breath-hold (16-20 seconds) over consecutive heartbeats. Each train was preceded by a specific inversion pulse (inversion times, 100, 200, and 350 milliseconds, respectively) followed by multiple single-shot readouts over several heartbeats with a balanced SSFP sequence combined with sensitivity-encoding parallel imaging [11]. A balanced SSFP sequence was used for readout because it offers superior signal-to-noise ratios and less modulation of the relaxation curve than conventional gradient-echo imaging [12]. Sensitivity encoding is a parallel imaging technique performed with a multiple-element array of parallel receiver coils. With use of a rectangular field of view, a series of back-folded images are acquired that are reconstructed as a single unfolded image by means of predetermined sensitivity maps for each coil element. These sensitivity maps are determined on a preceding low-resolution reference scan. Use of this technique leads to undersampling of k-space with subsequent increased speed of acquisition. This reduced acquisition time allows restriction of the acquisition window to less than 200 milliseconds during diastole, reducing cardiac motion artifact. All single-shot data acquisitions were performed at the same post-R wave trigger delay time in end diastole (determined by heart rate), so that each readout image for the differing inversion times was identical in anatomic location and phase of cardiac cycle. A total of 11 source data sets (three, three, and five from the respective inversion pulses) per midventricular slice were acquired. Each set of Look-Locker experiments was performed before and 2, 4, 6, 8, 10, 15, and 20 minutes after administration of 0.15 mmol/kg gadopentetate dimeglumine (Magnevist, Schering) by automatic injector (Spectris, Medrad) through a 20-gauge cannula in an antecubital vein at an injection rate of 5 mL/s with a 10-mL saline chaser. The exact scanning parameters were as follows: balanced SSFP sequence combined with sensitivity-encoding factor of 2; partial Fourier acquisition; TR/TE, 3.9/1.95; 50° flip angle; 380 x 342 mm field of view; 240 x 151 matrix; 1.58 x 2.26 mm pixel size; 8-mm slice thickness. The acquisition window for readout was 191 milliseconds.
Three patients with LV dysfunction (ejection fraction < 55% or resting regional wall motion abnormalities) underwent follow-up assessment of LV function 9 months postoperatively.
Image Analysis
Parametric T1 relaxation time pixel maps were generated from the modified
Look-Locker sequence images off-line on a PC running customized software
developed in-house on an IDL 6.0 platform (RSI). The program performed a
nonlinear three-parameter curve fit on the source data sets, automatically
calculated T1 time for each pixel, and constructed a parametric T1 map.
Computation of each T1 map took approximately 7 minutes, and the resultant
images were stored in DICOM format. The LV myocardium was later defined by
standard manual tracing of the epicardial and endocardial contours with
commercially available software (Mass 5.0, Medis) on a workstation (Figs.
1A and
1B). The mean T1 relaxation
time of the myocardial slice was subsequently calculated by summation and
averaging of the values for each myocardial pixel within the slice. Each slice
was also divided into a standard six segments for a midventricular short-axis
image, and T1 relaxation times for each segment were calculated
[13]. These steps were
followed for each time-scale acquisition before and after contrast
administration. Mass 5.0 software was used for calculation of LV functional
parameters (ejection fraction and mass), overt scar quantification on delayed
enhancement, and calculation of aortic regurgitant fraction for each patient
by drawing of manual contours around relevant regions of interest (LV
epicardium and endocardium on SSFP cine images, areas of delayed enhancement
on T1-weighted gradient-echo images, and aortic valve on phase velocity
sequences). Delayed enhancement was defined as a region of increased signal
intensity compared with normal nulled myocardium 20 minutes after injection of
gadopentetate dimeglumine. Visual assessment of regional wall motion in a
corresponding midventricular slice was performed to correlate significant
differences in segmental relaxation times with localized areas of dysfunction.
Wall motion was graded as follows: 1, normal; 2, mild to moderate hypokinesia;
3, severe hypokinesia; 4, akinesia; and 5, dyskinesia.
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
T1 values are not specific for individual tissues, but each tissue has a normal range dependent on field strength. Deviations outside the normal range may indicate an ongoing pathologic state [17]. Specific in vivo measurement of T1 times with parametric maps has already found clinical application in neurologic imaging in the diagnosis of multiple sclerosis [18, 19]. Measurement of T1 relaxation times in the heart is technically demanding because of a number of factors. Cardiac motion limits image acquisition to time of least motion, that is, diastole. This restriction is compounded by the increased T1 relaxation times associated with modern high-field-strength systems. The ability of the high-resolution modified Look-Locker technique we used to overcome these limitations by providing high-resolution T1 measurements acquired within a single breath-hold has been confirmed [9, 10]. Less accurate T1 mapping techniques have been shown feasible in detection of changes in relaxation times associated with acute myocardial infarction, and excellent correlation with conventional delayed enhancement-based infarct imaging has been achieved [20]. It was reasonable to hypothesize that a chronic condition, such as aortic regurgitation, known to cause diffuse myocardial fibrosis would lead to a change in myocardial T1 relaxation properties, particularly after contrast administration.
Overall, this feasibility study did not show a significant difference in myocardial T1 relaxation times between the two groups. There was also no correlation between ejection or regurgitant fraction and myocardial T1 relaxation time. When segmental analysis was performed, however, differences did emerge in T1 times after contrast injection in eight segments with resting wall motion abnormalities. None of these segments had discernible delayed enhancement, and six had persistent wall motion abnormalities after successful valve replacement. Given the persistent dysfunction without overt delayed enhancement exhibited in these segments, it is presumed that myocardial fibrosis was present in a patchy rather than a diffuse global pattern (i.e., only in the segments with abnormal wall motion) and that the differences in T1 relaxation times are attributable to the presence of fibrosis. Accordingly, it is encouraging that a discernible difference in T1 relaxation times existed between these segments and segments exhibiting normal resting motion, although the values did lie within the normal range. An interesting feature of this difference was that the T1 relaxation times in these abnormal segments were longer than in normal control segments, which is the converse of what we expected—that is, shortened T1 times in segments with fibrosis due to increased extracellular distribution volumes of gadopentetate dimeglumine.
Results with animal-based models of chronic aortic regurgitation suggest that the extracellular matrix in areas of fibrosis does not contain excess collagen compared with normal tissue but contains increased levels of glucosamine and fibronectin [21]. Collagen is a polypeptide molecule, whereas fibronectin is a glycoprotein, and glucosamine is a polysaccharide-containing proteoglycan. Changes in T1 relaxation after administration of paramagnetic agents such as gadolinium are not a direct manifestation of the agent but reflect its effect on the behavior of hydrogen protons within the tissue. Glucosamine is hydrophilic. It is possible that the slightly increased T1 times in these segments may reflect subtle variations in the water exchange properties or spin-lattice energy transfer efficiency within the extracellular matrix.
A number of limiting factors of the study may explain the failure to show differences in T1 times in more than three patients. These limitations include both patient and technical considerations. The most obvious is the small number of patients, but the study was an initial feasibility study. Because we did not seek ethical approval within our study protocol for histologic confirmation of the presence of LV fibrosis by biopsy at surgery, we cannot state with certainty that any patient had incurred clinically significant myocardial fibrosis as a consequence of valvular disease. Only three patients in the cohort had LV dysfunction, and the presence of fibrosis in two of these can only be inferred. LV dysfunction has been shown to be reversible postoperatively as a result of correction of the high afterload associated with aortic regurgitation. The implication is that LV dysfunction may not invariably mean presence of myocardial fibrosis [22]. This situation was elegantly shown in our subject whose LVEF went from 47.8% before to 71.4% after valve replacement. Therefore it is not unreasonable to suggest that according to the available evidence, only two of the eight patients may have had clinically significant fibrosis.
A technical limitation was that only a single midventricular slice was sampled in each patient because of the limited acquisition times available for each Look-Locker train. Accordingly, as with endomyocardial biopsy techniques, pathologic abnormalities may have been present but in an anatomic location different from the area interrogated. This limitation is perhaps borne out by the finding of focal overt fibrosis on delayed enhancement imaging, on which total LV coverage was obtained. To our knowledge, delayed enhancement has not been described in the context of aortic regurgitation. It is reasonable to conclude, however, that the examples identified in our study represent areas of focal fibrosis as a consequence of chronic volume overload. Both patients had normal coronary arteries and no history of myocardial infarction, angina, or chest pain suggestive of an ischemic cause of the symptoms. Neither patient had a medical history of sarcoidosis, myocarditis, Löffler's syndrome, arrhythmogenic right ventricular dysplasia, or any other condition that would account for such focal patterns of enhancement [23]. Neither patient had significant LV dysfunction, but patient 4 had severe (70%) regurgitation, and patient 5 had the longest known duration of regurgitation. Delayed enhancement is not specific to previous myocardial infarction and is believed to occur as the result of delayed washout of gadopentetate dimeglumine in areas of increased distribution volume [24, 25]. Focal areas of delayed enhancement have been described in relation to chronic pressure overload in aortic stenosis and in hypertrophic cardiomyopathy. In both instances the delayed enhancement was presumed to be secondary to focal myocardial fibrosis [26, 27]. However, given the small numbers involved in this study, other than the observation that the phenomenon occurs, no inferences can be drawn about pathophysiologic risk factors for development of delayed enhancement in chronic aortic regurgitation.
In summary, in a pilot cohort of patients with chronic aortic regurgitation, myocardial T1 mapping did not show a difference in T1 relaxation times suggestive of a diffuse fibrotic process. A segment-based analytic strategy, however, may yield regional differences within a slice that suggest fibrosis in areas of regional wall motion abnormality identified on functional cine images. Furthermore, focal areas of delayed enhancement suggestive of areas of overt fibrosis secondary to chronic volume and pressure overload may be seen in a minority of patients with chronic aortic insufficiency.
Acknowledgments
We wish to thank the Department of CardioThoracic Surgery, General
Infirmary at Leeds, for assistance in completing this study.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. H. Marwick and M. Schwaiger The Future of Cardiovascular Imaging in the Diagnosis and Management of Heart Failure, Part 1: Tasks and Tools Circ Cardiovasc Imaging, July 1, 2008; 1(1): 58 - 69. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
