Assessment of Myocardial Viability in Rats: Evaluation of a New Method Using Superparamagnetic Iron Oxide
Nanoparticles and Gd-DOTA at High Magnetic Field
Catherine Chapon,
1Laurent Lemaire,
1* Florence Franconi,
2Laurent Marescaux,
3Pierre Legras,
4Benoit Denizot,
1and Jean-Jacques Le Jeune
1The aim of this study was to detect salvageable peri- infarction myocardium by MRI in rats after infarction, using with a double contrast agent (CA) protocol at 7 Tesla. Intra- vascular superparamagnetic iron oxide (SPIO) nanoparticles and an extracellular paramagnetic CA (Gd-DOTA) were used to characterize the peri-infarction zone, which may recover function after reperfusion occurs. Infarcted areas measured fromT1-weighted (T1-w) images post Gd-DOTA administra- tion were overestimated compared to histological TTC stain- ing (52%ⴞ3% of LV surface area vs. 40%ⴞ3%,Pⴝ0.03) or toT2-w images post SPIO administration (41% ⴞ4%, P ⴝ 0.04), whereas areas measured fromT2-w images post SPIO administration were not significantly different from those measured histologically (Pⴝ0.7). Viable and nonviable myo- cardium portions of ischemically injured myocardium were enhanced after diffusive Gd-DOTA injection. The subsequent injection of vascular SPIO nanoparticles enables the discrim- ination of viable peri-infarction regions by specifically alter- ing the signal of the still-vascularized myocardium. Magn Reson Med 52:932–936, 2004.©2004 Wiley-Liss, Inc.
Key words: magnetic resonance imaging; myocardial infarction;
contrast media; myocardial viability
The ability to discriminate between infarcted and normal tissue after myocardial infarction is an important factor in monitoring prognosis and therapeutic options. The detec- tion of ischemic but still viable myocardium is of major interest because this stunned and/or hibernating zone may be able to recover its function and vascularization if reper- fusion occurs (1,2).
Several studies have investigated the ability of contrast- enhanced MRI (CE-MRI) to characterize myocardial inju- ries in humans (3) and animal models (4). Some contro- versies regarding the most appropriate contrast medium for imaging of myocardial infarction persist. Indeed, sev- eral experimental studies have indicated that extracellular
MR contrast media, such as gadolinium-chelates, overes- timate infarct size, which suggests that the enhanced re- gion encompasses both viable and nonviable portions of myocardium (4,5). Therefore, other contrast agents (CAs) have been tested to better define those zones. Porphyrin- based molecules that specifically target necrosis have been investigated (4), and intravascularT1(e.g., macromolecu- lar CAs (6)) and T2 (e.g., iron oxide nanoparticles (7,8)) CAs have been assessed to target the still-perfused myo- cardium (9). However, the peri-infarction zone, which may represent a salvageable area with reduced but residual contractile function, has not yet been characterized by a single technique. A double-contrast protocol using gado- linium-chelates and necrosis-specific CAs has been used to characterize the peri-infarction zone (10,11). However, this protocol does not allow an immediate diagnosis, since MRI has to be performed several hours after the necrosis- specific agent is injected (10 –12). The aim of this study was to determine whether a double-contrast protocol using the different distribution and relaxivity properties of in- travascular SPIO nanoparticles and an extracellular CA (Gadolinium-DOTA) could be used to detect the peri-in- farction zone at an early stage post-infarction.
MATERIALS AND METHODS MR Contrast Media
We prepared and purified SPIO nanoparticles following Molday’s method, with modifications (9). The hydrody- namic diameter of the particles was 50 nm. The magnetic relaxivities at 7 Tesla were: r1⫽1.2 mM⫺1L⫺1s⫺1and r2⫽ 247 mM⫺1L⫺1s⫺1 (13). The relaxivities of commercial Gadolinium-DOTA (Dotarem®, Guerbet, France) used in this study were r1 ⫽ 2.6 mmol⫺1L⫺1s⫺1 and r2 ⫽ 4.9 mmol⫺1L⫺1s⫺1at 7 T (13).
Experimental Protocol
Animal care and use was in strict accordance with the regulations of the French Ministry of Agriculture. Female Wistar rats (N⫽ 10) weighing 250 –280 g were anesthe- tized by spontaneous inhalation of a mixture of isoflurane (5–1.5%) and oxygen (3 L/min). Permanent ligature with surgical suture of the left coronary was performed accord- ing to the method of Selye et al. (14). Four rats died immediately after coronary occlusion occurred.
Five hours after the coronary ligature was performed, six rats received an IV injection of Gd-DOTA (0.2 mmol/kg), followed about 20 min later by an IV injection of SPIO
1INSERM U 646 Inge´nierie de la Vectorisation, Universite´ d’Angers, Angers, France.
2Service Commun d’Analyses Spectroscopiques, Universite´ d’Angers, An- gers, France.
3E´cole Nationale Ve´te´rinaire, Unite´ de Pathologie Chirurgicale, Nantes, France.
4Service Commun Animale´rie Hospitalo-Universitaire, Universite´ d’Angers, Angers, France.
Grant sponsor: CNRS-INSERM; Grant number: 8BG02H; Grant sponsor:
Angers Agglome´ration De´veloppement.
*Correspondence to: Dr. Laurent Lemaire, INSERM U 646, Inge´nierie de la Vectorisation Particulaire, 10, rue Andre´ Boquel, 49100 Angers, France.
E-mail: laurent.lemaire@univ-angers.fr
Received 14 October 2003; revised 13 May 2004; accepted 13 May 2004.
DOI 10.1002/mrm.20210
Published online in Wiley InterScience (www.interscience.wiley.com).
©2004 Wiley-Liss, Inc. 932
nanoparticles (5 mg Fe/kg). Imaging was performed before and after each injection.
MRI
In vivo MRI experiments were performed on a Bruker Avance DRX 300 equipped with a vertical superwide-bore magnet operating at 7 Tesla. The resonant circuit of the NMR probe was a 64-mm-diameter1H birdcage. The rats were anesthetized by the same procedure used for surgery.
ECG was recorded with a PHYSIOGARD SM 785 NMR (Bruker, Wisembourg, France) to provide a trigger pulse at the R-wave signal. We acquired scout gradient-echo im- ages to obtain a short-axis plan of the heart.
Cine ECG-gated T1-weighted (T1-w) imaging was per- formed with 16 phases in order to cover the entire cardiac cycle, with slice thickness ⫽ 2 mm, minimum TR ⫽ 11.8 ms, TE⫽2.5 ms, flip angle⫽13.5°, eight excitations, matrix⫽128⫻128, and FOV⫽40⫻40 mm, leading to an in-plane resolution of 310 m. Cine imaging was per- formed to identify the akinetic zone of the left ventricle (LV) wall corresponding to the infarcted area, and the image corresponding to the diastolic phase was compared to the same phase aquired using a fast ECG-gated T2-w sequence with the following parameters: minimum TR⫽ 1000 ms, effective TE ⫽ 31.7 ms with a train of eight echoes, matrix⫽ 128⫻128, FOV ⫽ 40⫻ 40 mm, slice thickness⫽2 mm, and eight averages. The acquisition of theT2-w images was timed to concur with end-diastole in order to minimize flow artifacts in the ventricular lumen.
A 3– 4-mm saturation band perpendicular to the imaging plan was placed over vessels and arteries close to the myocardium to limit phase-encoding artifacts.
Two reference T1-w and T2-w images were collected prior to any CA injection.T1-w images were then acquired for 20 min after Gd-DOTA injection. Afterwards, sets of T2-w images were acquired before SPIO injection, and for 40 min after the injection.
Images Analysis
Images were analyzed with the use of Paravision 2.1 soft- ware (Bruker, Wissembourg, France). Signal intensity (SI) values were obtained from operator-adjustable dimensions and positioned regions of interest (ROIs). The SIs in each ROI were normalized to an external reference tube of nickel chloride. Four ROIs were defined on the images acquired after Gd-DOTA injection and then transposed onto the other images. The ROIs were positioned over the ischemic myocardium (antero-lateral wall of the LV), the normal myocardium (posterior interventricular septum (IS)), the peri-infarction zone (outer rims of necrosis), and the nickel chloride reference tube.
SI variations (⌬S) between control preinjection images (Pii) and each postcontrast media injection image (Pci) were measured for each ROI and defined as follows:⌬S⫽ ((SIPci – SIPii)/(SIPii)) ⫻ 100. SIPii corresponds to the SI mean between the two control images.
The contrast-to-noise ratio (CNR) between infarcted and peri-infarcted myocardia was calculated prior to and after CA injection according the formula CNR ⫽ (SIregion A –
SIregion B)/(SDnoise), where region A⫽infarcted myocardium,
B ⫽ peri-infarction zone, and SD ⫽ standard deviation.
The CNR between normal and peri-infarction zones was calculated the same way, with A⫽ normal myocardium and B⫽peri-infarction zone. Infarct sizes were measured manually on post-Gd-DOTAT1-w images, and they corre- sponded to the enhanced areas of myocardium. On the post-double-contrastT2-w images, an infarct was defined as myocardium whose signal was not modified by the SPIO injection.
Postmortem Measurements
The rat hearts were sliced into 2-mm-thick sections corre- sponding to MR images for TTC staining. The size of the infarcted region was measured on numbered sections by means of NIH image-analysis software, and was compared with MRI measurements in a blinded manner.
Statistics
All results are expressed as mean⫾standard error of the mean (SEM). The SI variations in the different myocardial regions and infarct sizes measured with MRI and TTC were compared by means of a two-factor ANOVA. Multi- ple comparisons were determined by the least-squares method. Values ofP⬍0.05 were considered statistically significant.
RESULTS
Characterization of the Gd-DOTA-Enhanced Region on T1-w Images
Cine MRI enabled detection of the akinetic LV wall, and evidence of myocardial infarction was found in all six rats.
However, on the pre-CA images we were unable to dis- criminate between reversibly injured and irreversibly in- jured myocardia (Fig. 1a). Six to seven minutes after a bolus injection of Gd-DOTA was administered, the edge of the akinetic LV wall was characterized by a signal increase of 27.3%⫾8.2% (P⫽0.02), while in the central infarcted area and the normal myocardium the signals remained unaffected (–5.08%⫾4.85% (P⫽0.6) and 7.9%⫾11.4%, (P ⫽ 0.5), respectively) (Fig. 1c). The CNR significantly increased after Gd-DOTA injection between the infarcted and peri-infarcted areas (– 0.53⫾ 0.34 and –2.43⫾ 0.92 (P⫽0.04) prior to and after Gd-DOTA injection, respec- tively), but not between the normal and infarcted zones (0.68⫾0.47 and 0.02⫾0.81 (P ⫽0.13) before and after Gd-DOTA injection, respectively).
When the images were collected at a later time (18 – 40 min after injection (Fig. 1b)), no significant variation in the SI of the normal myocardium was observed (20.25⫾ 3.65, P ⫽ 0.52), whereas a general enhancement in the edge of the akinetic LV and the central infarct area was observed (respectively 65%⫾9.5% (P⫽0.0001) and 49%
⫾8% (P⫽0.0001; 18 min after Gd-DOTA injection), with no significant difference between the two zones (P⫽0.16;
Fig. 1c). In terms of CNR, a value of –2.07⫾1.17 between the edge of the central infarct area and the akinetic LV was calculated (P⫽0.2 vs. preinjection CNR), and a value of –1.53⫾0.64 between the normal and peri-infarcted zones (P⫽0.03 vs. preinjection CNR) was obtained.
Long-Lasting Detection of the Peri-Infarction Zone by a Combination of Gd-DOTA and Nanoparticle Injection UsingT2-w Imaging
For improved detection of the peri-infarction zone, the injection of paramagnetic MR contrast media (Gd-DOTA) was followed by an injection of superparamagnetic (SPIO) CA, andT2-w images were acquired.
T2-w images acquired 20 min after Gd-DOTA injection showed that the SI increased compared to the reference image (Fig. 2) in the infarcted area (64% ⫾ 5.5%, P ⫽ 0.0001) and the peri-infarction zone (51% ⫾ 5.8%, P ⫽ 0.0001), but remained unchanged in the normal myocar- dium (14.1%⫾11.7%,P⫽0.2) (Fig. 2c). The Gd-DOTA induced a significant variation in CNR on theseT2-w im- ages between normal and peri-infarcted areas (– 4.32 ⫾ 0.71 preinjection compared to –9.23⫾1.09 post Gd-DOTA injection,P⫽0.002), but not between infarcted and peri- infarcted areas (–1⫾0.86 preinjection compared to 0.59⫾ 1.24 post Gd-DOTA injection,P⫽0.41).
After SPIO injection, a signal decrease was observed in the normal myocardium (– 60.4%⫾7% (P⫽0.0001)) and the peri-infarction zone (–30.7%⫾ 10% (P⫽ 0.01)), but the infarcted area was not affected (2.0%⫾3.4%,P⫽0.8) (Fig. 2a and c). The CNR then significantly increased be- tween the infarcted and peri-infarcted areas from –1.00⫾ 0.86 (P⫽0.001) prior to contrast media injection to 0.59⫾ 1.24 post-injection of Gd-DOTA (P⫽0.017), and then to 5.2⫾1.71 post SPIO (P⫽ 0.004). The CNR between the normal and peri-infarction zones was unaffected after FIG. 1. T1-w MR images prior to CA injection (a) and 20 min after
administration of 0.2 mmol/kg Gd-DOTA (b), with the corresponding SI time curves (c). RV, right ventricle wall; LV, left ventricle wall; IS, interventricular septum (■); ILV, infarcted LV (‚); PIZ, peri-infarction zone ( ).
FIG. 2. T2-w MR images acquired prior to CA injec- tion (a) and after double CA administration (b), with the corresponding SI time curves (c). RV, right ven- tricle wall; LV, left ventricle wall; IS, interventricular septum (■); ILV, infarcted LV (‚); PIZ, peri-infarction zone ( ).
SPIO injection (–10.91⫾0.75) compared to that measured after Gd-DOTA injection (–9.23⫾1.09,P⫽0.41).
Histopathologic Examination—Comparison of Infarct Size Measured Using Gd-DOTA, SPIO, and TTC
TTC staining outlined a transmural myocardial infarction in all six animals. We assessed the infarct size usingT1-w imaging after Gd-DOTA, andT2-w imaging after Gd-DOTA and iron oxide administration. Values were compared to TTC-stained histological slice measurements (Fig. 3). The true infarct size demarcated by TTC (40% ⫾ 3% of LV surface area) was significantly smaller than that defined by Gd-DOTA (52%⫾3% of LV surface area;P⫽0.03) on the T1-w images. However, the infarct size measured on the T2-w images after Gd-DOTA and SPIO administration (41% ⫾ 4% of LV surface area) correlated with that de- marcated by TTC (P⫽0.7). Furthermore, the infarct area detected onT2-w images was significantly different from that defined on theT1-w images (P⫽0.04).
DISCUSSION
The present study describes a strategy whereby the poten- tially salvageable cardiac peri-infarction zone can be mapped early by a combination of a diffusive positiveT1 MR CA and a vascular negativeT2MR contrast media. The double-contrast MR technique was performed in one im- aging session 5 hr post-infarction.
Gd-DOTA Enhancement
T1-w MRI after Gd-DOTA administration showed a non- homogeneous enhancement in a permanent myocardial occlusion (Fig. 1b and c). Six minutes after administration, an enhancement occurred in the border zone of the aki- netic LV wall, but not in the normal myocardium or the core of the occlusion, where an enhancement was detected only after 15– 40 min. This inhomogeneous enhancement after Gd-DOTA injection may be attributed to the absence of a direct path into the tissue for the contrast media, since the artery was occluded. The delayed enhancement of the infarct core associated with the stable enhancement of the
peri-infarction zone led to a loss of contrast between those two zones, and therefore we were unable to obtain an accurate measurement of the peri-infarction area. Even though an early contrast enhancement at the outer rim of infarction was observed after Gd-DOTA injection, we were unable to detect the potentially viable area with precision.
Indeed, since the size and level of enhancement evolved with time, the accuracy of the measurement was highly dependent on the time period between the Gd-chelate injection and the imaging process (15,16). It may also have been affected by the physiological status of the rats with respect to the invasive procedure they had undergone
⬍5 hr previously. Consequently, as illustrated by data from the literature, estimation of the infarct size after ex- tracellular CA administration is controversial. Our results appeared to either match (10) or overestimate (4,5) the infarct area determined histologically by TTC examina- tion. However, the discrepancy in the results may also reflect differences in the infarction model (reperfused or not) used and/or partial volume effects related to the im- aging parameters (17) and animals used (dogs, rabbits, or rats) in the different studies. Although high-resolution MR images were acquired in this study (spatial in-plane reso- lution⫽310m), the latter point may also partly account for the overestimation of the infarct size measured after Gd-DOTA injection (Fig. 3), since the slices were 2 mm thick. Thus, under the experimental conditions of this study, a diffusive CA (Gd-DOTA) did not allow an accurate assessment of myocardial infarction.
Double-Contrast Protocol
The double-contrast protocol appears to be one way to document both the infarcted zone, which will not recover function, and the peri-infarction zone, which may recover function. The principle behind this method is to use two CAs that differ in terms of distribution, as described in recent studies (10,11) in which a combination of necrosis- specific porphyrin-based and nonspecific (Gd-DTPA) CAs were used to measure the size of the peri-infarction zone.
Since both CAs used in this study are signal enhancers, we used the specific CA before the nonspecific CA. The mech- anism by which the porphyrin-based molecule targeted the necrosis is unknown, but it required several hours to produce a significant effect on the images (10 –12).
In the present study, the two CAs used differed not only in terms of distribution, but also in terms of relaxivity properties. SPIO distributed in the vascular compartment and had a negative effect on T2 contrast, whereas Gd- DOTA distributed in infarcted/peri-infarcted areas and had a positive effect onT1contrast. In this case, the order in which the CAs were injected was essential. Indeed, if a superparamagnetic CA is injected prior to Gd-DOTA, the important drop in SI in the vascularized myocardium will result in a loss of quality inT1-w images.
The accuracy of measurement was determined by a sim- ple procedure involving SPIO nanoparticle injection, as previously shown (8), and the injection of Gd-DOTA did not improve the results. However, we hypothesized that partial volume effects were at least partly responsible for the overestimation of the infarct size measured with Gd- DOTA. The use of the double procedure enabled us to rebut this hypothesis in our experiment. Indeed, if such FIG. 3. Assessment of infarct size using Gd-DOTA-enhancedT1-w
MRI (black bar),T2-w MRI after Gd-DOTA and iron oxide nanopar- ticle injection (gray bar), and TTC-stained histologic slices (white bar).
partial volume effects existed, they should have been ob- served equally on the post-SPIO images, leading to an overestimation of the measurements. This was not the case, however, since the values obtained by histology cor- related with those obtained from the T2 set of images.
Therefore, the overestimation of the infarct size with Gd- DOTA was due to the presence of the peri-infarction zone.
Study Limitations
It should be pointed out that there were two limitations in this study. First, only non-reperfused myocardial infarc- tions were investigated. Even if several experimental stud- ies used the permanent ligature to discriminate infarcted from viable myocardium (10,18) reperfused infarctions may present with different patterns of enhancement.
Second, we used a cine gradient-echo sequence—not an inversion recovery (IR) sequence—to follow Gd-DOTA en- hancement. Several previous studies performed at 1.5 or 2T (19) reported that the CNR between normal and in- farcted myocardium is improved with the use of IR se- quences compared to standard T1-w imaging. We chose not to use IR in order to simplify a long experimental protocol. Indeed, IR delayed enhancement is highly sen- sitive to the inversion time selected, which depends on operator expertise, pulse sequence parameters (19), and the CA clearance rate (20). Moreover, Barkhausen et al.
(10) suggested that the inversion pulse of the IR sequence that is used to minimize the SI of the normal myocardium can diminish the magnitude of the SIs within the myocar- dium. It is possible that we missed the margins of the infarct, which enhanced slightly more than the normal myocardium, because of the extreme contrast of the image.
Finally, the inversion time required to obtain a good-qual- ity image at 7T would be fairly long due to the increase of the longitudinal relaxation time at high magnetic field.
This would increase the acquisition time compared to that required at 1.5T.
CONCLUSIONS
The use of an extracellular CA in combination with vas- cular superparamagnetic nanoparticles appears to be a new method for discriminating the potentially viable peri- infarction zone at an early post-infarction stage. SPIO nanoparticle injection induced an SI decrease in normal myocardium and an early-enhanced Gd-DOTA peri-infarc- tion area, showing that this zone was still vascularized.
Furthermore, the size of the unvascularized myocardium demarcated after the double CA injection matched the true infarct size measured by histological TTC staining, whereas the myocardial infarction was overestimated with the use of Gd-DOTA.
ACKNOWLEDGMENTS
The authors thank M. Moreau for synthesizing the iron oxide nanoparticles, J. Roux and D. Gilbert for housing and caring for the animals used in this study, and Jennifer Godbee for editing the manuscript. This work is part of the Imagerie du Petit Animal project, which is supported by grant 8BG02H from CNRS-INSERM. C. Chapon is sup-
ported in part by a grant from Angers Agglome´ration De´- veloppement.
REFERENCES
1. Scheinman MM, Abbott JA. Clinical significance of transmural vs.
nontransmural electrocardiographic changes in patients with acute myocardial infarct. Am J Med 1973;55:602– 608.
2. Weisman HF, Healy B. Myocardial infarct expansion, infarct extension and reinfarction: pathophysiologic concepts. Prog Cardiovasc Dis 1987;
30:73–110.
3. Dendale P, Franken PR, Block P, Pratikakis Y, De Roos A. Contrast enhanced and functional magnetic resonance imaging for the detection of viable myocardium after infarction. Am. Heart J 1998;135:875– 880.
4. Saeed M, Bremerich J, Wendland MF, Wyttenbach R, Weinmann HJ, Higgins CB. Reperfused myocardial infarction as seen with use of necrosis-specific vs. standard extracellular MR contrast media in rats.
Radiology 1999;213:247–257.
5. Judd RM, Lugo-Olivieri CH, Arai M, Kondo T, Croisille P, Lima JA, Mohan V, Becker LC, Zerhouni EA. Physiological basis of myocardial contrast enhancement in fast magnetic resonance images of 2-day-old reperfused canine infarcts. Circulation 1995;92:1902–1910.
6. Roberts HC, Saeed M, Roberts TP, Muhler A, Brasch RC. MRI of acute myocardial ischemia: comparing a new contrast agent, Gd-DTPA-24- cascade-polymer, with Gd-DTPA. J Magn Reson Imaging 1999;9:204 – 208.
7. Bjerner T, Ericsson A, Wikstrom G, Johansson L, Nilsson S, Ahlstrom H, Hemmingsson A. Evaluation of nonperfused myocardial ischemia with MRI and an intravascular USPIO contrast agent in an ex-vivo pig model. J Magn Reson Imaging 2000;12:866 – 872.
8. Chapon C, Franconi F, Lemaire L, Marescaux L, Legras P, Saint-Andre JP, Denizot B, Le Jeune JJ. High field MR imaging evaluation of super- paramagnetic iron oxide nanoparticles in a permanent rat myocardial infarction. Invest Radiol 2003;38:141–146.
9. Rousseau V, Pouliquen D, Darcel F, Jallet P, Le Jeune JJ. NMR investi- gation of experimental chemical induced brain tumors in rats, potential of a superparamagnetic contrast agent (MD3) to improve diagnosis.
Magn Reson Mater Physics Biol Med 1998;6:13–21.
10. Barkhausen J, Ebert W, Debatin JF, Weinmann HJ. Imaging of myocar- dial infarction: comparison of Magnevist and gadophrin-3 in rabbits.
J Am Coll Cardiol 2002;39:1392–1398.
11. Saeed M, Lund G, Wendland MF, Bremerich J, Weinmann H, Higgins CB. Magnetic resonance characterization of the peri-infarction zone of reperfused myocardial infarction with necrosis-specific and extracel- lular nonspecific contrast media. Circulation 2001;103:871– 876.
12. Choi SI, Choi SH, Kim ST, Lim KH, Lim CH, Gong GY, Kim HY, Weinmann HJ, Lim TH. Irreversible damaged myocardium at MR im- aging with a necrotic tissue specific contrast agent in a cat model.
Radiology 2000;215:863– 868.
13. Messager T, Franconi F, De Bray JM, Denizot B, Lemaire L, Jallet P, Le Jeune JJ. Contrast agents relaxivities evolution with magnetic field strength (0.47–7 Tesla). In: Proceedings of the 16th ESMRMB, Sevilla, 1999. Abstract 500.
14. Selye H, Bajusz E, Grasso S, Mendell P. Simple techniques for the surgical occlusion of coronary vessels in the rat. Angiology 1960;11:
398 – 407.
15. Oshinski JN, Yang Z, Jones JR, Mata JF, French BA. Imaging time after Gd-DTPA injection is critical in using delayed enhancement to deter- mine infarct size accurately with magnetic resonance imaging. Circu- lation 2001;104:2838 –2842.
16. Moran GR, Thornhill RE, Sykes J, Prato FS. Myocardial viability imag- ing using Gd-DTPA: physiological modeling of infarcted myocardium, and impact on injection strategy and imaging time. Magn Reson Med 2002;48:791– 800.
17. Kim RJ, Chen E-L, Lima JAC, Judd RM. Myocardial Gd-DTPA kinetics determine MRI contrast enhancement and reflect the extent and sever- ity of myocardial injury after acute reperfused infarction. Circulation 1996;94:3318 –3326.
18. Pereira RS, Prato FS, Lekx KS, Sykes J, Wisenberg G. Contrast-en- hanced MRI for the assessment of myocardial viability after permanent coronary artery occlusion. Magn Reson Med 2000;44:309 –316.
19. Simonetti OP, Kim RJ, Fieno DS, Hillenbrand HB, Wu E, Bundy JM, Finn JP, Judd RM. An improved MR imaging technique for the visual- ization of myocardial infarction. Radiology 2001;218:215–223.
20. Kellman P, Arai AE, McVeigh ER, Aletras AH. Phase-sensitive inver- sion recovery for detecting myocardial infarction using gadolinium- delayed hyperenhancement. Magn Reson Med 2002;47:372–383.
