| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Imaging |
1 From the Department of Radiology, New York University Medical Center, 530 First Ave–MRI, New York, NY 10016 (V.S.L., D.R., S.S.T., J.J.S., C.A.N., G.M.I.); and Siemens Medical Solutions, Chicago, Ill (O.P.S.). Received January 15, 2003; revision requested April 8; revision received April 24; accepted June 16. Supported by a grant from the Society of Thoracic Radiology. Address correspondence to V.S.L. (e-mail: vivian.lee@med.nyu.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Twenty patients underwent stress-rest SPECT examinations with sestamibi. Results were considered equivocal for assessment of myocardial infarct on the basis of fixed perfusion defects that either had normal wall motion or exceeded any wall motion abnormalities. Patients then underwent (a) contrast-enhanced MR imaging for assessment of myocardial infarct and (b) cine MR imaging for assessment of wall motion. For image analyses, the left ventricle was divided into 14 segments. Wall motion and extent of infarct were assessed independently and compared.
RESULTS: Forty-one segments were equivocal for infarct at SPECT, and most (21 of 41 [51%]) involved the posterior or inferior wall. Infarct was confirmed with MR imaging in 10 of 41 (24%) equivocal segments in eight patients (40%). An additional 29 segments in eight patients had infarct at MR imaging that was not suspected at SPECT, including segments in three patients with no clinical history of myocardial infarct prior to imaging. All cases of infarct except one that were equivocal or undetected with sestamibi at SPECT were nontransmural at MR imaging, and most of the unsuspected subendocardial infarcts (15 of 28 [54%]) had no associated wall motion abnormalities.
CONCLUSION: Patients with radionuclide examination findings that are equivocal for infarct may benefit from contrast-enhanced MR imaging, particularly in the setting of nontransmural infarct.
© RSNA, 2003
Index terms: Heart, MR, 524.121412, 524.12143 • Heart, SPECT, 524.12162 • Magnetic resonance (MR), vascular studies, 524.121412, 524.12143 • Myocardium, infarction, 511.771
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Stress-rest single photon emission computed tomography (SPECT) with technetium 99m (99mTc) sestamibi (Cardiolite; Bristol-Meyers Squibb, North Billerica, Mass) is widely used for the diagnosis of coronary artery disease. Estimates of infarct size obtained with sestamibi at SPECT in patients with known coronary artery disease correlate with histopathologic findings and patient outcomes (4–9). Because of limitations in spatial resolution, however, subendocardial infarct remains a diagnostic challenge. In addition, photon attenuation from overlying soft tissue can cause perfusion defects that mimic infarct and has been reported in as many as 20%–50% of cases (10,11). The favorable properties of 99mTc sestamibi enables electrocardiographically (ECG) gated SPECT to provide simultaneous evaluation of perfusion and wall motion. By categorizing fixed defects associated with normal wall motion as artifactual, the false-positive rate for diagnosis of infarct can be lowered (12,13). This may occur at the expense of decreased sensitivity, however, since infarcts that result in preserved function, namely subendocardial infarcts, may be missed.
Magnetic resonance (MR) imaging provides high-spatial-resolution images of myocardial infarct because of the delayed contrast enhancement of scar tissue at rest (14). Studies involving the use of animal models have yielded excellent correlation between the global and transmural extent of infarct depicted at MR imaging and at histologic examination (15,16). In humans, the transmural extent of viable myocardium seen at resting contrast material–enhanced MR imaging accurately predicts the reversibility of associated myocardial dysfunction (3,17). In addition to its higher spatial resolution, MR imaging is less susceptible than radionuclide imaging to artifacts related to patient body habitus, and assessments made with MR imaging are independent of the extent of wall motion abnormality.
We performed this study to evaluate whether contrast-enhanced MR imaging is useful for assessment of myocardial viability in patients with equivocal stress-rest results from SPECT examination with sestamibi.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MR imaging was performed within a mean of 1 month after SPECT with sestamibi (range, 7–95 days). No patients had intervening revascularization or a substantial change in symptoms between the two examinations. Of the 20 patients, eight had a clinical history of myocardial infarct, and five had undergone revascularization prior to SPECT (coronary angioplasty and/or stent placement [n = 2], coronary bypass graft placement [n = 1], or both [n = 2]).
Scintigraphic Technique and Image Evaluation
Resting images were acquired after a 2–4-hour fast, 30 minutes after administration of 10 mCi (370 MBq) of 99mTc sestamibi. SPECT images were acquired with a commercially available gamma camera and computer (Cardio and Vertex, respectively; Adac, Milpitas, Calif) by using a 180° imaging arc, an all-purpose parallel-hole collimator, and 64 frames with 20 seconds per stop for a total imaging time of 20–25 minutes. Subsequently, an additional 30 mCi of 99mTc sestamibi was given either during peak treadmill exercise (n = 15) (patients underwent imaging after 15 minutes) or following infusion of dobutamine (Baxter Healthcare, Deerfield, Ill) (n = 2) or dipyridamole (Elkins-Sinn, Cherry Hill, NJ) (n = 3) (patients underwent imaging after 1 hour). A gated SPECT study was performed with 16 frames acquired per cardiac cycle by using similar acquisition parameters for a similar time period. Summed gated frames were processed as 6.47-mm tomographic sections by using a Butterworth filter with a cutoff frequency of 0.52 cycles per centimeter and power of 5. The gated images were processed by using a Hanning filter and a cutoff frequency of 0.7 cycles per centimeter.
For all patients, gated SPECT images were viewed prospectively in cinematic format by one reader (S.S.T., with 12 years of experience) by using Autospect and Autoquant software (Adac, Milpitas, Calif). From the stress-rest short- and long-axis images with polar maps, relative perfusion was categorized quantitatively in each of the 14 segments on the basis of the output of the software (Fig 1) and was graded subjectively by using a five-point scale: 0 indicated normal perfusion; 1, mildly decreased perfusion; 2, moderately decreased perfusion; 3, severely decreased perfusion; and 4, no perfusion. Defects were determined to be fixed or reversible (improving by one grade or more). Regional wall motion was assessed by means of subjective evaluation of endocardial excursion and regional wall thickening and was graded on a five-point scale: 0 indicated normal findings; 1, mild to moderate hypokinesis; 2, severe hypokinesis; 3, akinesis; and 4, dyskinesis.
|
MR Imaging Technique
Each patient underwent MR imaging in the supine position with a 1.5-T clinical MR imager (Quantum; Siemens, Erlangen, Germany) and a torso phased-array coil. All acquisitions were ECG gated. Breath-hold ECG-gated cine gradient-echo MR images were acquired with eight to 10 short-axis views (from mitral valve through apex), a horizontal long-axis view, and a vertical long-axis view. For all MR imaging, section thickness was 8 mm with a 2-mm intersection gap.
The first nine subjects were examined by using a segmented spoiled gradient-echo MR sequence for wall motion with the following sequence parameters: 9/6.1 (repetition time msec/echo time msec), 20° flip angle, 126 x 256 matrix, 244–281 x 325–375-mm field of view, 45-msec effective temporal resolution with view sharing, and acquisition time of 14 heartbeats for each section. For the subsequent 11 subjects, a true fast imaging with steady-state precession, or FISP, sequence was used for cine imaging with the following typical parameters: 3.6/1.8, 60° flip angle, 120 x 256 matrix, 244–281 x 325–375-mm field of view, 15 k-space lines per segment per heartbeat, and 54-msec temporal resolution (18).
Patients then received an intravenous bolus of 20–30 mL of gadopentetate dimeglumine (0.15 mmol per kilogram of body weight of Magnevist; Berlex Laboratories, Wayne, NJ). For evaluation of myocardial viability, breath-hold ECG-gated two-dimensional segmented inversion-recovery spoiled gradient-echo MR images (14) were acquired 10–20 minutes after injection in the same imaging planes as those used for cine gradient-echo MR imaging. Inversion times were selected to ensure nulling of uninfarcted myocardium and typically ranged from 225 to 325 msec. Other parameters for the segmented inversion-recovery spoiled gradient-echo sequence were typically 700/5.4, 30° flip angle, 25 k-space lines per segment, and data acquisition triggered every other heartbeat, with matrix and field of view matching those of the cine MR images.
MR Image Analysis
The cine and contrast-enhanced viability images were acquired in the same views and therefore were aligned for interpretation. For image analysis, the left ventricle was divided into a series of three concentric rings (base, middle, and apex) with a total of 14 segments that were identical to those used in scintigraphic analysis (Fig 1). Typically, the three basilar sections acquired with MR imaging were used to define the basilar ring of the bull’s eye diagram (Fig 1), the middle three sections were used to define the middle ring, and the two to three apical sections were used to define the apex.
Short-axis images acquired at end diastole and end systole were used to determine end-diastolic volume, or EDV, and end-systolic volume, or ESV, by using regions of interest drawn manually by one author (D.R.). Ejection fraction was computed as (EDV - ESV)/EDV.
Two observers (G.M.I. and V.S.L. with 3 and 5 years of cardiac MR imaging experience, respectively), who were unaware of patient identities and scintigraphic results, evaluated the cine and contrast-enhanced MR viability images independently and resolved differences by means of consensus. The extent of segmental wall thickening and motion at MR imaging was graded on the same five-point scale used for SPECT with sestamibi. Delayed hyperintensity 10–20 minutes following contrast material administration indicated myocardial infarct. Each segment was graded visually according to the extent of tissue in each segment (percentage of segmental cross-sectional area) that was hyperintense, for which a score of 0 indicated no hyperintensity or infarct; 1, hyperintensity in 1%–25% of tissue; 2, 26%–50%; 3, 51%–75%; 4, 76%–99%; and 5, complete hyperintensity. Hyperintensity on a single section was considered sufficient for the diagnosis of infarct. Disagreements that required consensus occurred in seven subjects for wall motion abnormalities and in eight subjects for infarct, usually resulting from differences in one grade of severity between readers.
Clinical Correlation
After a mean follow-up interval of 1 year (range, 8–826 days), referring physicians were contacted, and charts were reviewed for all patients. Clinical history of myocardial infarct, coronary catheterization, and surgical or interventional procedures performed prior to MR imaging and SPECT were recorded. The results of diagnostic tests performed after MR imaging (including coronary catheterization and ECG) and the performance of subsequent revascularization procedures were noted.
Data Analysis
All continuous data are presented as means ± SDs. To evaluate whether the detection of infarct with sestamibi at SPECT was independent of the presence of wall motion abnormalities, 
2 analysis was performed, with P < .05 used to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Scintigraphic Results: Segments Indeterminate for Infarct
Of the 280 segments in 20 patients evaluated with stress-rest scintigraphy with sestamibi, 41 (15%) were indeterminate for viability. These indeterminate segments were primarily in the posterior or inferior (21 segments in 17 patients) (Figs 2, 3), septal (10 segments in nine patients), and anterior (eight segments in five patients) (Fig 4) locations. Two indeterminate segments were in the lateral wall in two patients. Five patients had one indeterminate segment, 10 had two, four had three, and one had four.
|
|
|
|
|
|
|
|
MR Imaging Results: Equivocal Segments
Of the 41 segments considered equivocal for infarct at SPECT with sestamibi in 20 patients, 10 (24%) segments in eight (40%) patients had abnormal enhancement on MR viability images (Figs 2, 4). Of the 21 posterior and inferior wall segments equivocal for infarct at scintigraphy, five (24%) showed infarct at MR imaging; similarly, of the eight segments in the anterior wall, two (25%) were abnormal at MR imaging. Only one of the 10 equivocal septal segments showed infarct at MR imaging, while both equivocal segments in the lateral wall demonstrated infarct.
The distribution of infarct was subendocardial in all 10 equivocal segments confirmed to be infarcted at MR imaging (Figs 2, 4). The transmural extent of infarct measured less than 75% in all cases (1%–25% in one, 26%–50% in five, and 51%–75% in four). While all equivocal segments that were shown to have infarct at MR imaging had decreased perfusion with relatively preserved wall motion at scintigraphy, only one segment showed completely normal wall motion (Fig 4).
We also examined stress-induced ischemia in the segments considered equivocal at scintigraphy in relation to MR imaging findings. Four of 15 (27%) equivocal segments had evidence of ischemia at both scintigraphy and MR imaging, whereas six of 26 (23%) equivocal segments without evidence of ischemia at scintigraphy were proved to have infarct at MR imaging.
MR Imaging Results: Unsuspected Infarct
At MR imaging, eight patients had unsuspected infarct in 29 segments that were considered viable at scintigraphy. Most nonviable segments were normal at SPECT (n = 24) (Fig 2), while five segments were interpreted as ischemic without infarct. In seven of these eight patients, the diagnosis of myocardial infarct was assigned at scintigraphy on the basis of abnormalities in other segments of the left ventricle. In one patient, however, scintigraphy showed only ischemia without infarct, and thus, the diagnosis of infarct was made only with MR imaging in this case.
All cases of infarct undetected with sestamibi except for one were nontransmural at MR imaging. Of the 28 segments with subendocardial infarct at MR imaging that were not suspected at scintigraphy, most (n = 15 [54%]) had normal wall motion at both gated SPECT with sestamibi and cine gradient-echo MR imaging. The transmural extent of infarct in these 15 segments with normal contractility was either less than 25% (n = 9) or between 26% and 50% (n = 6). Segments with undetected infarct at SPECT with sestamibi were significantly more likely to show normal wall motion than segments in which infarct was suspected (P = .02).
Clinical Correlation
Contrast-enhanced MR imaging resulted in a diagnosis of myocardial infarct in nine patients, eight of whom had areas of infarct not suspected at scintigraphy. Six patients had a clinical history of infarct, and the distribution of infarct at MR imaging was compatible with previous catheterization results. Three patients had no history of infarct, and therefore, infarct was newly diagnosed at MR imaging. In one patient, stress-rest SPECT with sestamibi showed a fixed perfusion defect in the anterior wall without wall motion abnormality, which was considered equivocal for infarct; MR imaging showed unsuspected subendocardial infarct (Fig 4). In the other two patients, SPECT images obtained with sestamibi showed evidence of ischemia that was superimposed on a fixed defect, which was thought to represent infarct versus attenuation or bowel artifact. In both patients, MR imaging showed single-territory infarcts in the equivocal segments.
Clinical Follow-up and Angiographic Correlation
Six patients underwent coronary catheterization within 15 days of MR examination. Four had evidence of infarct at contrast-enhanced MR imaging in 14 segments, of which four were considered equivocal for infarct, and 10 were not suspected on the basis of findings at SPECT. At angiography, all four patients had severe coronary artery disease in the territory corresponding to infarct at MR imaging. This group included two patients with no clinical history of myocardial infarct prior to imaging (Fig 4).
The remaining two patients had MR imaging findings that were negative for infarct. They underwent coronary angiography because of findings suggestive of ischemia at SPECT. One patient was found to have severe three-vessel disease; SPECT showed one segment equivocal for infarct, in addition to ischemia in several other segments. Because the study was performed only at rest, MR imaging was not used to assess ischemia. The other patient with no MR imaging evidence of infarct who underwent catheterization had only mild stenosis in the middle left anterior descending artery and did not undergo revascularization.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The favorable properties of 99mTc sestamibi have enabled use of gated SPECT techniques, whereby wall motion and perfusion can be assessed simultaneously (12,13). By assuming that fixed defects caused by infarcts should demonstrate decreased wall motion and thickening, while those caused by attenuation artifacts should demonstrate normal motion and thickening, gated SPECT can reduce the false-positive rate of sestamibi studies. For example, Dogruca et al (13) showed that gated analysis increased the specificity of detection of inferior wall disease from 46% (12 of 26 cases) to 88% (23 of 26 cases).
DePuey and Rozanski (12) published a larger study of 551 patients who underwent gated SPECT with sestamibi, of whom 180 (33%) had fixed defects. By reclassifying the 60 patients with fixed defects and normal or mildly impaired wall motion as having normal segmental function, the percentage of patients with unexplained fixed defects (no clinical history of myocardial infarct) decreased from 14% (78 of 551 patients) to 3% (18 of 551 patients). Moreover, 55 of 60 (92%) patients with fixed defects and relatively normal function were women with anterior defects (n = 29) or men with fixed inferior wall defects (n = 26). The authors cited this distribution as additional evidence that the defects were artifactual.
The assumption that all fixed perfusion defects with preserved function (even those in the inferior or anterior wall) represent artifacts clearly leads to an increase in false-negative findings. In the study by Dogruca et al (13), 16 of the 51 segments with fixed defects had normal segmental function and were considered artifactual. Yet, on the basis of findings of critical stenoses at angiography and ECG changes, six of the 16 (38%) segments were established to have been infarcted previously and therefore represented false-negative findings. Similarly, DePuey and Rozanski (12) found that four of 102 (4%) patients had fixed defects but no wall motion abnormality and yet had a clinically established myocardial infarct in the associated territories.
In our series, five of the 17 (24%) subjects with fixed defects in the inferior wall and relatively normal function had infarct at MR imaging. On the basis of our observation that the distribution of the infarct was subendocardial in all cases, we suspect that the rim of viable myocardium superficial to the infarct is responsible for relatively preserved function. Clearly, the ability of MR imaging to depict and enable quantification of infarct independent of wall motion abnormalities is an important advantage.
Of the clinical modalities widely used for evaluation of myocardial viability, none can depict subendocardial infarct reliably. Even positron emission tomography (PET), which is considered the reference standard for viability, is limited by its relatively low spatial resolution. MR imaging offers high spatial resolution (2 x 2 x 8-mm voxels) with high image contrast (14) and can be used to detect and quantify the transmural extent of infarct. Wu et al (23) showed that MR imaging depicted nontransmural hyperintensity in all eight of their subjects with clinical evidence of non–Q-wave infarcts. Animal models with subacute and chronic infarct have demonstrated remarkably close agreement between histopathologic depiction of infarct and MR imaging findings (15,16,24). Investigators in two recent studies on the comparison of MR imaging and PET showed high correlation but found that 11% (113 of 1,023) (25) and 18% (316 of 1,753) (26) of all segments were discordant—they were considered viable at PET but showed areas of hyperintensity at MR imaging. This finding was attributed to the higher spatial resolution of MR imaging, which enables a more sensitive diagnosis of infarct.
In our study, contrast-enhanced MR imaging was used not only to evaluate equivocal segments but also to identify an additional 29 segments in eight patients who had predominantly subendocardial infarct not suspected at radionuclide imaging. Our findings are in agreement with those in a recent report by Wagner et al (24), who showed that of 181 myocardial segments with subendocardial (<50%) hyperintensity at MR imaging, 85 (47%) showed defects at resting 201Tl SPECT. In an animal study of 15 dogs that appeared in the same article (24), Wagner et al confirmed that MR imaging was used to correctly identify 100 of 109 (92%) segments with subendocardial (<50%) infarct, while resting 201Tl SPECT was used to correctly identify 31 of 109 (28%) segments.
The clinical importance of otherwise undetectable infarct remains to be assessed more fully, but results of one study (27) suggest that infarct not recognized with contrast-enhanced MR imaging may be an independent predictor of death in patients referred for angiography. The ability to quantify infarct is important in many patients for determination of treatment by helping to predict who might benefit from surgery or angioplasty versus medical therapy (3). Accurate quantification of infarct may also serve as a valuable measure of efficacy in clinical trials of reperfusion therapies (9). Recently, investigators showed contrast-enhanced MR imaging to be useful for the primary diagnosis of acute myocardial infarct, although issues of overestimation of infarct size remain to be resolved (28,29).
There are recognized limitations of the present study. Our sample size was relatively small. The sestamibi imaging method we used reflected our routine clinical protocol; no additional correction algorithms were applied. Given the low spatial resolution of PET, we relied on MR imaging as the reference standard for the diagnosis of myocardial infarct (3,15,16,23,24). Angiographic correlation was limited in our study population. Our protocol was performed for the purpose of evaluating myocardial viability. Stress imaging with MR imaging was not performed. Therefore, myocardial ischemia could not be assessed in our study, although stress MR imaging has proved promising in other laboratories (30–32).
We have shown that findings of fixed perfusion defects in the setting of relatively normal motion at stress-rest SPECT with sestamibi may be attributable either to attenuation artifacts or to subendocardial infarcts. Patients with radionuclide examination findings that are equivocal for infarct may benefit from contrast-enhanced MR imaging for the evaluation of myocardial viability, particularly in the setting of nontransmural infarct.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, V.S.L.; study concepts, V.S.L.; study design, V.S.L., S.S.T., J.J.S.; literature research, V.S.L., D.R.; clinical studies, G.M.I., V.S.L., S.S.T., J.J.S.; data acquisition, G.M.I., V.S.L., S.S.T., J.J.S., C.A.N.; data analysis/interpretation, D.R., V.S.L.; statistical analysis, V.S.L.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P. Finn, K. Nael, V. Deshpande, O. Ratib, and G. Laub Cardiac MR Imaging: State of the Technology. Radiology, November 1, 2006; 241(2): 338 - 354. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Cury, C. A. M. Cattani, L. A. G. Gabure, D. J. Racy, J. M. de Gois, U. Siebert, S. S. Lima, and T. J. Brady Diagnostic Performance of Stress Perfusion and Delayed-Enhancement MR Imaging in Patients with Coronary Artery Disease. Radiology, July 1, 2006; 240(1): 39 - 45. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Gupta, V. S. Lee, Y.-C. Chung, J. S. Babb, and O. P. Simonetti Myocardial Infarction: Optimization of Inversion Times at Delayed Contrast-enhanced MR Imaging Radiology, December 1, 2004; 233(3): 921 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. R. Edelman Contrast-enhanced MR Imaging of the Heart: Overview of the Literature Radiology, September 1, 2004; 232(3): 653 - 668. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
