Contrast-enhanced MR Imaging of Coronary Arteries: Comparison of ...

Debiao Li, PhD Jie Zheng, PhD Hanns-Joachim Weinmann, PhD

Index terms: Animals Coronary vessels, MR, 54.121412, 54.121413, 54.12143, 54.12149 Magnetic resonance (MR), contrast enhancement, 54.12143, 54.12149 Magnetic resonance (MR), vascular studies, 54.12143, 54.12149 Radiology 2001; 218:670 – 678 Abbreviations: CNR ⫽ contrast-to-noise ratio ECG ⫽ electrocardiographic RCA ⫽ right coronary artery SNR ⫽ signal-to-noise ratio 3D ⫽ three dimensional 1

Contrast-enhanced MR Imaging of Coronary Arteries: Comparison of Intra- and Extravascular Contrast Agents in Swine1 PURPOSE: To compare the efficacy of an intravascular contrast agent, gadomer-17, in improving magnetic resonance (MR) imaging of coronary arteries with that of an extravascular agent, gadopentetate dimeglumine, in pigs. MATERIALS AND METHODS: Eight pigs underwent imaging after three injections: 0.20 mmol of gadopentetate dimeglumine per kilogram of body weight and 0.05 and 0.10 mmol/kg gadomer-17. Coronary images were acquired repeatedly after each injection by using an inversion-recovery–prepared segmented three-dimensional sequence with either breath holding (n ⫽ 4) or respiratory gating (n ⫽ 4). Coronary artery–to–myocardium contrast-to-noise ratios (CNRs) were compared between injections.

From the Department of Radiology, Northwestern University Medical School, 448 E Ontario St, Ste 700, Chicago, IL 60611 (D.L., J.Z.); and Schering AG, Berlin, Germany (H.J.W.). Received April 8, 1999; revision requested June 15; final revision received July 19, 2000; accepted July 25. Supported in part by National Institutes of Health grant HL 38698 and research grants from Schering AG, Berlin, Germany; Berlex Laboratories, Wayne, NJ; and Siemens Medical Systems, Erlangen, Germany. Address correspondence to D.L. (email: [email protected]).

RESULTS: At breath-hold imaging, substantial CNR improvement over precontrast images was observed in images acquired during the first pass of gadopentetate dimeglumine in coronary arteries and up to 6 and 10 minutes after 0.05 and 0.10 mmol/kg of gadomer-17 injections, respectively. The CNR with 0.10 mmol/kg of gadomer-17 was 20% (P ⬍ .05) higher than that with gadopentetate dimeglumine at first-pass imaging. At respiratory-gated imaging, significant CNR improvement (P ⬍ .05) over precontrast images was observed in images acquired up to 10, 30, and 50 minutes after gadopentetate dimeglumine and both gadomer-17 injections, respectively. The CNR on the first images obtained after 0.10 mmol/kg gadomer-17 injection was 168% (P ⬍ .05) higher than that on the images obtained after gadopentetate dimeglumine injection.

As a member of Schering Aktiengesellschaft Berlin, H.J.L. has a financial interest in developing gadomer17.

CONCLUSION: Gadomer-17 provided greater and more persistent CNR improvements than did gadopentetate dimeglumine; further evaluation of its utility for coronary imaging in humans is warranted.

©

RSNA, 2001

Author contributions: Guarantor of integrity of entire study, D.L.; study concepts and design, D.L., J.Z., H.J.W.; definition of intellectual content, D.L., J.Z., H.J.W.; literature research, D.L., J.Z.; experimental studies, D.L., J.Z.; data acquisition, D.L., J.Z.; data analysis, D.L., J.Z., H.J.W.; statistical analysis, D.L., J.Z.; manuscript preparation, D.L.; manuscript editing and review, D.L., J.Z., H.J.W.

670

Various T1-shortening contrast agents have been used extensively in magnetic resonance (MR) angiography of the entire body. At the time this article was written, all clinically approved agents in the United States were extravascular and freely diffusible to interstitial space. These agents have dramatically increased the blood signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) between blood vessels and background tissues at MR angiography of the carotid, pulmonary, abdominal, and peripheral arteries and the aorta. Improved coronary artery delineation has also been demonstrated during the first pass of the agents by using a fast three-dimensional (3D) breath-hold imaging technique (1– 4). Several intravascular contrast agents have been developed in recent years by various pharmaceutical companies (5–9). These agents have a larger T1 relaxivity and remain in the blood pool longer than do extravascular agents. Results of animal studies and preliminary clinical trials (10 –13) have shown these agents to be promising in improving coronary MR angiography. The purpose of this study was to compare the efficacy of a newly developed intravascular agent with that of a clinically used extravascular agent,

Figure 1. Schematic diagram of the contrast-enhanced coronary artery imaging sequence. After ECG triggering and a delay time (TD), a nonselective inversion pulse was applied, followed by the inversion time and a frequency-selective fat saturation pulse. Twenty-one in-plane phase-encoding lines were acquired during each heartbeat. A navigator echo (NAV) was acquired after data collection for respiratory-gated imaging. After each x-y plane of k space was covered in consecutive heartbeats in an interleaved pattern, the partition-encoding gradient was increased by one step. At precontrast imaging, the inversion pulse (in dotted box) was not applied. TI ⫽ inversion time.

gadopentetate dimeglumine, in improving coronary artery delineation in pigs. We hypothesized that this comparison would directly demonstrate the advantages of intravascular agents over extravascular agents to justify the need for intravascular agents in coronary artery imaging. Breath-hold and respiratory-gated imaging sequences were performed, and technical issues related to imaging the coronary arteries with contrast agents were addressed.

MATERIALS AND METHODS Domestic swine (n ⫽ 9; weight, 18 –25 kg) (Oakhill Genetics, Ewing, Ill) were used. All animal preparation and imaging procedures were approved by the Animal Care and Use Committee at Northwestern University School of Medicine. The intravascular agent, gadomer-17 (SHL 643A; Schering, Berlin, Germany), is a 24-gadolinium cascade polymer with a molecular weight of 35 kDa. It has the same concentration of gadolinium (0.50 mmol/mL) as do conventional extravascular agents and has a T1 relaxivity of 13 L/(mmol 䡠 sec) at 1.5 T, which is approximately three times that of gadopentetate dimeglumine (Magnevist; Berlex, Wayne, NJ) (4 L/(mmol 䡠 sec). It is eliminated rapidly at glomerular filtration but does not extravasate much into interstitial space.

Experimental Design First, blood T1 values were measured in a single pig after each of the three intravenous contrast agent injections (0.20 mmol of gadopentetate dimeglumine per kilogram of body weight, 0.05-mmol/kg gadomer-17, and 0.10-mmol/kg gadomer17). Coronary artery images were then acquired repeatedly by using a fast 3D sequence in a group of four pigs after each of the three contrast agent injections. Each examination was completed within Volume 218

䡠

Number 3

a single breath hold by switching off the ventilator at the end of expiration. Finally, coronary arteries were imaged repeatedly in another group of four pigs by using retrospective respiratory gating after each of the three contrast agent injections. For T1 measurement and coronary artery imaging, there was a time delay of at least 80 minutes between consecutive injections. Our T1 measurements and simulations have indicated that because of this time delay, the effects of residual contrast agent from previous injections on later studies are negligible. In addition, the order of the three injections was varied among pigs to randomize these effects.

Animal Preparations Each pig received an intramuscular injection of a preanesthetic that consisted of atropine sulfate (0.10 mg/kg), acepromazine (0.12 mg/kg), and ketamine (20.0 mg/kg) by using a 19-gauge butterfly injection set. A needle was secured in the ear or leg vein for intravenous contrast agent administration. Anesthesia was maintained during imaging by administering inhaled isoflurane (30% isoflurane and 70% oxygen) through an anesthesia ventilator (Ohio Medical Products, Madison, Wis). Ventilation was maintained at a rate of 15 breaths per minute and an inspiration ratio of 50%. After anesthesia was induced, the animal was placed in the supine position on the MR imaging patient table. Electrocardiographic (ECG) leads and a pulse oximeter were then attached to the pig for monitoring with an MR imaging– compatible hemodynamic monitoring system (Omni-Trak 3100 MRI Vital Signs Monitoring System; Invivo Research, Orlando, Fla).

Blood T1 Measurements Blood T1 values were measured dynamically after each of the three contrast

material injections by performing a twodimensional fast gradient-echo sequence with a 90° preparatory pulse (14). Imaging parameters included a repetition time msec/echo time msec of 2.4/1.2, a data acquisition matrix of 64 ⫻ 128 (phase encoding by readout), a section thickness of 8 mm, an inversion time of 90 msec, and a flip angle of 8°. The shape of the section excitation profile was taken into consideration to improve the accuracy of T1 estimation (15). The contrast material injection time was 20 seconds. The blood signal intensities were measured at the descending aorta. Blood inflow effects were eliminated by acquiring data at diastole.

Contrast-enhanced Coronary Artery Imaging An inversion recovery–prepared 3D segmented gradient-echo sequence was performed for contrast-enhanced coronary artery imaging, as shown in Figure 1. Sequence parameters included 4.0/1.5; a field of view of 115 ⫻ 230 mm2; a data acquisition matrix of 105 ⫻ 256 (phase encoding by readout), an in-plane resolution of 1.1 ⫻ 0.9 mm2, and a section thickness of 2 mm. Twenty-one in-plane phaseencoding steps were acquired during each heartbeat by using a centric reordering scheme; this resulted in a data acquisition window of 84 msec. After data were acquired, the images were sinc-interpolated in the section direction, such that the number of partitions was doubled. The inversion-recovery preparation was used to suppress the myocardial signal, as illustrated in Figure 2. By using a centric encoding scheme, low-frequency components were acquired first in each cardiac cycle when blood and myocardium had the greatest contrast. To determine the appropriate inversion time for maximal blood-myocardium contrast, computer simulations were performed by using actual imaging parameters. As shown in Figure 3, maximal blood-myocardium contrast occurs at an inversion time of approximately 200 msec, which was used for all postcontrast imaging.

MR Imaging: Intravascular and Extravascular Contrast Agents in Swine

䡠

671

At a typical imaging session, transverse scout images were first obtained to find the origin of the right coronary artery (RCA). A sagittal-oblique section was acquired along the proximal portion of the RCA. A double-oblique imaging plane was then prescribed along the atrioventricular groove to cover the proximal portions of the left and right coronary arteries. At coronary artery imaging, a precontrast examination was first performed without the inversion pulse and with a constant flip angle of 15°, followed by postcontrast imaging with inversion-recovery preparation and a constant flip angle of 25°.

Breath Holding The imaging sequence shown in Figure 1 was performed. The number of partitions was 10, and the imaging time was approximately 25–35 seconds (pig heart rates were 90 –110 beats per minute). During data acquisition, the ventilator was turned off at the end of expiration to freeze respiratory motion. Contrast media were injected over 20 seconds manually and uniformly through the established intravenous needle. Each injected dose was diluted by mixing it with normal saline to 20 mL. At first-pass imaging, it is important to match the acquisition of central k-space lines with the maximal arterial signal. First, a test bolus (1 mL of gadopentetate dimeglumine) was injected and its circulation time estimated from blood signal intensities measured in the descending aorta. The time delay between the start of full-dose injection and that of data acquisition was determined with an empiric formula (2). After the first-pass images were acquired, images were acquired repeatedly after each contrast material injection. The order and timing of the three injections in the four pigs in this group are shown in Table 1.

Retrospective Respiratory Gating The sequence structure was the same as that of a breath-hold imaging sequence, except that retrospective respiratory gating (16) was performed. The first postcontrast imaging sequence started 20 seconds after the initiation of each injection. Contrast material injection lasted 1 minute. Navigator echoes were collected at the dome of the diaphragm by using a pair of 90° and 180° pulses. The number of data acquisitions was four and the number of 3D partitions was 32. The imaging time was 640 cardiac cycles (approximately 672

䡠

Radiology

䡠

March 2001

Figure 2. Graph illustrates the improved contrast between blood and myocardium with inversion-recovery preparation. In each cardiac cycle, the longitudinal magnetizations of blood and myocardium recover toward their respective equilibrium values during the delay time (TD). They are inverted by a 180° pulse and again recover toward equilibrium values. When an appropriate inversion time (TI) is chosen, the myocardial signal is completely nulled at the start of data acquisition, whereas the magnetization of blood is nearly fully recovered because of its short T1.

6 – 8 minutes). The imaging sequence was repeated after each contrast material injection. The order and timing of the three injections in the four pigs in this group are shown in Table 1. All imaging examinations were performed with a 1.5-T whole-body imaging system (Magnetom Vision; Siemens) with a maximum gradient strength of 25 mT/m and a gradient rise time of 600 ␮sec. The conventional body coil was used as the radio-frequency transmitter, and a four-element body phased-array coil was used for signal reception. To avoid image wraparound from a small field of view, only the two anterior elements of the coil were used.

Image Reformatting and Data Analysis Image reformatting.—Multiplanar reconstruction and maximum intensity projection were performed by using the standard software provided with the commercial Siemens MR imaging system.

3D volume-rendered images were generated at a commercial imaging processing station (3DVirtuoso; Siemens) to visualize the coronary arteries. SNR and CNR measurements.—Signal intensities were measured in the first 3 cm of both the left and right coronary arteries and in the adjacent myocardium. The SD of noise was estimated as the mean signal intensity of the background air divided by 1.25 (17). Statistical analysis was performed with a two-tailed Student t test. A P value of less than .05 was considered to indicate a significant difference.

SNR Simulations Computer simulations were performed to confirm the blood signal improvements with the use of contrast agents. The sequence parameters for the simulations included a repetition time of 4.0 msec, a flip angle of 25°, and 21 lines acquired per cardiac cycle. A precontrast T1 of 1,300 msec and an R-R interval of 600 Li et al

Figure 3. Graph shows simulated longitudinal magnetizations of blood and myocardium at the start of data acquisition in each heartbeat as a function of inversion time (TI) in an ECG-triggered inversion-recovery– prepared sequence. According to our measurements, the T1 of blood was 50 –100 msec in early phases after contrast material injections. By assuming a 10% myocardial blood volume, myocardial T1 was estimated as 300 –500 msec. Imaging parameters used for simulations included a TR of 4.0 msec, a flip angle of 25°, and 21 lines acquired per cardiac cycle. An R-R interval of 600 msec was assumed. An inversion time of approximately 200 msec (dotted vertical line) offers maximal blood-myocardium contrast for the T1 values considered.

TABLE 1 Order and Timing of Contrast Agent Injections at Breath-hold and Respiratorygated Imaging Type of Imaging and Pig No. Breath-hold imaging 1 2 3 4 Respiratory-gated imaging 1 2 3 4

0.20 mmol/kg Gadopentetate Dimeglumine

0.05 mmol/kg Gadomer-17


1 (0) 2 (120) 2 (90) 3 (87)

3 (88) 1 (0) 3 (87) 2 (85)

2 (83) 3 (85) 1 (0) 1 (0)

1 (0) 2 (87) 2 (89) 3 (83)

2 (82) 3 (82) 1 (0) 2 (87)

3 (85) 1 (0) 3 (98) 1 (0)

Note.—Numbers represent the order of contrast agent injection. Numbers in parentheses represent time delay (in minutes) from the previous injection.

msec were used. Because constant flip angles were used in data acquisition, blood signal decreased with each phase-encoding line during the cardiac cycle. The blood signal of the central k-space line (the first line acquired in each cardiac cycle) was used to represent the blood signal intensity. The time-of-flight effect caused by blood flow was not considered, so the simulations represented the worstcase scenario for the precontrast blood signal. T2* effects were not taken into account in the simulations. For gadolinium-based contrast agents, T2* effects were negligible because of the the 1.5msec echo time used in our study.

RESULTS Blood T1 Measurements The postcontrast arterial blood T1 values measured in a single pig are shown in Volume 218

䡠

Number 3

Figure 4. The blood R1 was greatest with 0.10-mmol/kg gadomer-17 throughout the measurements. The blood R1 with 0.05-mmol/kg gadomer-17 was similar to that with 0.20-mmol/kg gadopentetate dimeglumine during first-pass imaging but became greater afterward. Although these measurements had no significance, they provided a useful reference for interpreting our experimental results.

Contrast-enhanced Coronary Artery Imaging Breath-hold images.—Figure 5 shows breath-hold images obtained in a single pig. On the precontrast image, the proximal RCA was barely visible. In first-pass images, vessel delineation was markedly improved, as compared with that on the precontrast image, and the coronary artery SNR and CNR were slightly better

with 0.10 mmol/kg gadomer-17 (Fig 5, part 7) than were those with the two other injections (Fig 5, parts 2 and 4). No apparent improvement in RCA delineation was observed at 5 minutes after gadopentetate dimeglumine injection (Fig 5, part 3). Therefore, images were not acquired with further delays. Clear SNR and CNR increases were seen on images acquired up to 5 minutes after 0.05-mmol/kg gadomer-17 injection and up to 9 minutes after 0.10-mmol/kg gadomer-17 injection. The coronary artery SNRs and coronary artery–myocardium CNRs for breathhold images in the four pigs in this group are shown in Table 2. Significant SNR and CNR improvements over those on precontrast images were observed on images acquired up to 6 minutes after 0.05-mmol/kg gadomer-17 injection and up to 10 minutes after 0.10-mmol/kg gadomer-17 injection (the entire duration of data collection), whereas after gadopentetate dimeglumine injection, only first-pass images had significantly improved SNRs and CNRs. In addition, the CNR with 0.10 mmol/kg gadomer-17 was 20% higher than that with gadopentetate dimeglumine at first-pass imaging (P ⬍ .05). The SNR and CNR after 0.10-mmol/kg gadomer-17 injections were also significantly higher than those after 0.05-mmol/kg gadomer-17 injections with the same delay times (P ⬍ .05). Retrospective respiratory gating.—Figure 6 shows images acquired with retrospective respiratory gating in a single pig. Delineation of the coronary artery was improved on all postcontrast images, as compared with that on precontrast images. The images obtained with 0.10mmol/kg gadomer-17 had the best SNR and CNR. The SNR and CNR obtained with 0.05-mmol/kg gadomer-17 were also greater than those obtained with 0.20-mmol/kg gadopentetate dimeglumine. It should be noted that the RCA delineation on the images obtained 30 minutes after 0.10-mmol/kg gadomer-17 injection was still better than that on the first images obtained after 0.20mmol/kg gadopentetate dimeglumine injection.


䡠

673

Figure 4. Graphs show blood R1 values measured (a) during the first 2 minutes and (b) up to 80 minutes after various contrast material injections by using a two-dimensional gradient-echo sequence with a 90° preparatory pulse. Blood T1 is the inverse of R1. Curves I, II, and III correspond to 0.10 mmol/kg gadomer-17, 0.05 mmol/kg gadomer-17, and 0.20 mmol/kg gadopentetate dimeglumine injections, respectively. The shortest T1 was approximately 30 msec with 0.10 mmol/kg gadomer-17 and 50 msec with 0.05 mmol/kg gadomer-17 and 0.20 mmol/kg gadopentetate dimeglumine. Blood T1 remained at less than 100 msec for approximately 15 minutes with 0.10 mmol/kg gadomer-17 and for 5 minutes with 0.05 mmol/kg gadomer-17 but for only 30 seconds with gadopentetate dimeglumine. It became longer than 500 msec at 80 minutes after each injection.

By using relatively large volume coverage per examination with respiratory gating and 0.10-mmol/kg gadomer-17 injection, it was possible to visualize all major coronary arteries through data reformatting, as shown in Figure 7. Table 3 summarizes the coronary artery SNRs and coronary artery–myocardium CNRs measured on respiratorygated images in the four pigs. In the images obtained with gadopentetate dimeglumine, the SNR was not different from that on the images obtained before contrast enhancement. Nevertheless, the CNR was significantly improved because of myocardial suppression. The SNR and CNR on the first images obtained after 0.05-mmol/kg gadomer-17 injection were 28% (P ⬍ .05) and 52% (P ⬍ .05) greater, respectively, than on those obtained after gadopentetate dimeglumine injection. The SNR and CNR on the first images obtained after 0.10-mmol/kg gadomer-17 injection were 100% (P ⬍ .05) and 168% (P ⬍ .05) greater, respectively, than on the images obtained with gadopentetate dimeglumine. Significant CNR improvements over those on precontrast images were observed on images acquired up to 20 and 40 minutes after 0.05- and 0.10mmol/kg gadomer-17 injections, respectively. The CNRs on images acquired 674

䡠

Radiology

䡠

March 2001

TABLE 2 Coronary Artery SNR and Coronary Artery–Myocardium CNR at 3D Breath-hold Imaging in Four Pigs Delay after Contrast Agent Injection (min) Coronary artery SNR* First pass 2–4 4–6 6–8 8–10 Coronary artery–myocardium CNR‡ First pass 2–4 4–6 6–8 8–10

0.20 mmol/kg Gadopentetate Dimeglumine



6.1 ⫾ 0.1† 3.7 ⫾ 0.8 3.1 ⫾ 0.4 Not applicable Not applicable

6.0 ⫾ 0.1† 5.9 ⫾ 0.2† 5.2 ⫾ 0.3† 4.3 ⫾ 0.7 Not applicable

6.6 ⫾ 0.1† 7.1 ⫾ 0.3† 7.0 ⫾ 0.2† 6.6 ⫾ 0.3† 6.3 ⫾ 0.5†


4.4 ⫾ 0.1† 4.1 ⫾ 0.2† 3.6 ⫾ 0.4† 2.7 ⫾ 0.6 Not applicable

4.9 ⫾ 0.1† 5.2 ⫾ 0.2† 5.1 ⫾ 0.3† 4.6 ⫾ 0.2† 4.5 ⫾ 0.4†

* The precontrast coronary artery SNR was 4.2 ⫾ 0.1. † Significantly higher than at precontrast imaging (P ⬍ .05). ‡ The precontrast coronary artery–myocardium CNR was 1.4 ⫾ 0.1.

20 –30 minutes after 0.10-mmol/kg gadomer-17 injection were still greater than those on the first images acquired within 10 minutes of 0.20-mmol/kg gadopentetate dimeglumine injection. In addition, the SNRs and CNRs obtained after 0.10-mmol/kg gadomer-17 injection were significantly greater than those obtained after 0.05-mmol/kg gadomer-17 injections at the same delay times (P ⬍ .05).

SNR Simulations Figure 8 shows the simulated postcontrast blood signal as a function of T1 and as normalized by the precontrast signal. The signal at a T1 of 60 msec (approximately 1/3 of the inversion time) was 191% higher than that at precontrast imaging. Further T1 shortening leads to only small signal increases because the Li et al

Figure 5. MR images demonstrate the improved coronary artery delineation at breath-hold imaging with gadomer-17 over that with gadopentetate dimeglumine. Images were acquired with a 3D breath-hold segmented gradient-echo sequence (4.0/1.5) in a double-oblique (transverse to coronal to sagittal) orientation. Arrows indicate the proximal portion of the RCA. In the precontrast image (1), there is little contrast between the RCA and the background. The RCA signal intensity is increased, whereas background signal intensity is suppressed in the first-pass image obtained after 0.20-mmol/kg gadopentetate dimeglumine injection (2). At 5 minutes after gadopentetate dimeglumine injection (3), no SNR or CNR increases were observed. Second-row images were acquired after 0.05-mmol/kg gadomer-17 injection at various delays (4, first pass; 5, 5 minutes; 6, 9 minutes). Note the improvement in RCA delineation in images 4 and 5, as compared with that in image 1. The RCA delineation started to deteriorate in images acquired at further delays. Third-row images were acquired after 0.10-mmol/kg gadomer-17 injection at various delays (7, first pass; 8, 5 minutes; 9, 9 minutes). Substantial improvements in RCA delineation from 1 are observed in all three third-row images.

longitudinal magnetization of blood is nearly fully recovered during the inversion time for T1 values less than one-third of the inversion time. As T1 increases to greater than 60 msec, the blood signal decreases sharply. It becomes approximately the same as that at precontrast imaging as T1 approaches inversion time. It should be noted that the use of variable flip angles, inclusion of blood inflow enhancement, and changes of the R-R interval and the number of phase encoding lines per cardiac cycle alter the relationship between blood signal and T1 (18).

DISCUSSION

Figure 6. MR images demonstrate the improved coronary artery delineation with gadomer-17 over that with gadopentetate dimeglumine at respiratory-gated imaging. Images were acquired with a 3D respiratory-gated segmented gradient-echo sequence (4.0/1.5) in a transverse to coronal to sagittal orientation. Arrows with solid heads indicate the RCA. The first row contains single partition images. In the precontrast image (1), there is little contrast between the RCA and adjacent myocardium (arrowhead). After 0.20-mmol/kg gadopentetate dimeglumine injection, there is little increase in RCA signal intensity (2). Nevertheless, myocardial signal intensity (arrowhead) is suppressed, which results in improved RCA delineation. Apparent SNR and CNR increases over those in precontrast and gadopentetate dimeglumine– enhanced images are visible in the first images acquired after injections of 0.05mmol/kg (3) and 0.10-mmol/kg gadomer-17 (4). Wraparound artifacts (arrow with open head) are seen on image 1 and were eliminated on postcontrast images by using inversionrecovery preparation. Second- and third-row images were created with multiplanar reconstruction. Again, a gadopentetate-dimeglumine– enhanced image (6) shows certain improvement in RCA delineation over the precontrast image (5). The SNR and CNR are substantially greater in images obtained by using 0.05-mmol/kg gadomer-17 (7, first examination; 8, 10-minute delay at the start of image acquisition) and 0.10-mmol/kg gadomer-17 (9, first examination; 10, 10-minute delay; 11, 20-minute delay; and 12, 30-minute delay at the start of image acquisition). In addition, the RCA delineation with 0.10-mmol/kg gadomer-17 is substantially better than that with 0.05-mmol/kg gadomer-17 at the same delay times. Volume 218

䡠

Number 3

Coronary MR angiography is usually ECG triggered for diastolic data acquisition. Therefore, it has a low imaging efficiency in terms of the available SNR per unit time. In addition, the idle time during each cardiac cycle allows for substantial magnetization recovery before data acquisition. As a result, there is little contrast between blood and myocardium; this lack of contrast severely hinders the delineation of coronary arteries that are surrounded by myocardium. The results of our pig studies demonstrate the marked improvements of coronary artery SNR and coronary artery–myocardium CNR with the administration of contrast agents. It should be noted, however, that because of the reduced T1 weighting secondary to ECG triggering, the blood signal improvement at contrast-enhanced coronary MR angiography was less dramatic than that at non– ECG-triggered imaging. For example, according to our simulations, a reduction of blood T1 from 1,300 to 50 msec could result in a signal intensity increase by a factor of four at non–ECG-triggered im-


䡠

675

aging but by a factor of only two at ECGtriggered imaging. Nevertheless, the coronary artery SNR improvement at contrastenhanced imaging was still substantial. More important, the short blood T1 allows the use of inversion-recovery preparation to suppress the myocardial signal, with little reduction in the blood signal. An additional benefit of the dramatic suppression of the background tissues is a small field of view without the usual wraparound artifacts.

Magnetization Preparation Schemes for Contrast-enhanced Coronary MR Angiography Several methods have been used to suppress the myocardial signal, including magnetization transfer (19,20) and T2 preparation (21–23) for nonenhanced imaging and steady-state preparation (1,10) for contrast-enhanced imaging. However, with these methods, the blood signal is reduced as well. At contrast-enhanced inversion-recovery–prepared imaging, the myocardial signal is better suppressed, with little loss of the blood signal; this results in a better CNR than do other methods. In Figure 9, images acquired with inversion recovery and steady-state preparations are compared; note the greater SNR and CNR of the RCA with inversion recovery than with steady-state preparation. The concept of background suppression with inversion-recovery preparation was previously used at nonenhanced MR angiography of the renal (24), pulmonary (25), and carotid arteries (26). In recent years, it has become the method of choice for corastenhanced coronary artery imaging (13,27). The choice of inversion time for myocardial nulling depends on the myocardial T1 and the residual magnetization of the myocardium before application of the inversion pulse; the residual magnetization is, in turn, affected by the time allowed for magnetization recovery at each heartbeat. Therefore, the inversion time is dependent on the R-R interval to a certain degree. For the estimated myocardial T1 values of 300 –500 msec after contrast material injections and an R-R interval of around 600 msec, an inversion time of 200 msec was found to yield dramatic myocardial suppression in all pig studies. A different inversion time will need to be selected for different myocardial T1 values and R-R intervals.

Comparison of Extravascular and Intravascular Agents Our measurements showed that the blood T1 values during the first pass of 676

䡠

Radiology

䡠

March 2001

Figure 7. Reformations acquired with 0.10-mmol/kg gadomer-17 by using a 3D respiratorygated segmented gradient-echo sequence (4.0/1.5) in a double-oblique (transverse to coronal to sagittal) view shows the major coronary arteries. The first examination performed after gadomer-17 injection (2) shows dramatic improvements in SNR and CNR of the left coronary arteries over those in the precontrast image (1)(in 2, Ao ⫽ aorta, LAD ⫽ left anterior descending artery, LCX ⫽ left circumflex artery). A curved multiplanar reconstruction (4) shows long portions of the left anterior descending (LAD) and left circumflex (LCX) arteries. The route of the curved multiplanar reconstruction is shown in image 3. Another multiplanar reconstruction image (5) clearly delineates the RCA and left circumflex artery. The image created with volume rendering (6) shows major coronary arteries and their spatial relationship with major cardiac structures. Ao ⫽ aorta, LAD ⫽ left anterior descending artery, LCX ⫽ left circumflex artery , RA ⫽ right atrium, RV ⫽ right ventricle.

TABLE 3 Coronary Artery SNR and Coronary Artery–Myocardium CNR at 3D Respiratory-gated Imaging in Four Pigs Delay after Contrast Agent Injection (min) Coronary artery SNR* 0–10 10–20 20–30 30–40 40–50 Coronary artery–myocardium CNR‡ 0–10 10–20 20–30 30–40 40–50

0.20-mmol/kg Gadopentetate Dimeglumine

0.05-mmol/kg Gadomer-17

0.10-mmol/kg Gadomer-17

7.4 ⫾ 0.8 Not applicable Not applicable Not applicable Not applicable


14.8 ⫾ 1.3† 11.6 ⫾ 0.8† 9.9 ⫾ 0.9† 8.7 ⫾ 1.2 5.3 ⫾ 0.9

4.4 ⫾ 0.6† Not applicable Not applicable Not applicable Not applicable

6.7 ⫾ 0.9† 5.1 ⫾ 0.6† 3.6 ⫾ 0.6† Not applicable Not applicable

11.8 ⫾ 1.0† 8.7 ⫾ 0.3† 7.0 ⫾ 0.9† 5.3 ⫾ 0.8† 2.7 ⫾ 0.6†

* The precontrast coronary artery SNR was 7.0 ⫾ 0.8. † Significantly higher than at precontrast imaging (P ⬍ .05). ‡ The precontrast coronary artery–myocardium CNR was 1.4 ⫾ 0.9.

0.05-mmol/kg gadomer-17 and 0.20mmol/kg gadopentetate dimeglumine injections were approximately the same. Therefore, coronary artery images acquired with 0.05-mmol/kg gadomer-17 were compared with those acquired with

0.20-mmol/kg gadopentetate dimeglumine, which has become the standard dose for contrast-enhanced MR angiography of the entire body. Images obtained with gadomer-17 injections of the standard dose (0.10 mmol/kg) also were acLi et al

Figure 8. Graph shows simulated postcontrast blood signal as a function of T1, normalized by the precontrast signal. With the sequence parameters used (4.0/1.5; flip angle, 25°; and number of lines acquired per cardiac cycle, 21) and an R-R interval of 600 msec, there is a sharp decrease of signal with the blood T1 increase when T1 becomes longer than 60 msec (dotted vertical line). Postcontrast signal becomes identical to precontrast signal (dotted horizontal line) when postcontrast blood T1 is approximately 200 msec.

Figure 9. MR images obtained with a respiratory-gated 3D sequence (4.0/1.5) with steadystate (1, 2) and inversion-recovery (3, 4) preparations after separate 0.05-mmol/kg gadomer-17 injections are shown to compare the delineations of the RCA (arrows). The plane of view is transverse to coronal to sagittal. RCA delineation is better in images 3 and 4 than in images 1 and 2. Delay times after injection were 20 seconds (1, 3) and 3 minutes (2, 4). The RCA SNR with inversion recovery was 20% higher than that with steady-state preparation, and the myocardial SNR surrounding the RCA was 50% lower. As a result, CNR with inversion-recovery preparation was 60% higher.

quired to demonstrate the effectiveness of the contrast agent in improving coronary artery depiction. Significant improvements in coronary artery SNR and CNR were observed during the first pass of all gadopentetate dimeglumine and gadomer-17 injections. Although the blood T1 with 0.10mmol/kg gadomer-17 was 40% shorter than those with 0.20-mmol/kg gadopentetate dimeglumine and 0.05-mmol/kg gadomer-17, coronary artery SNR was only Volume 218

䡠

Number 3

10% higher; this finding was consistent with those of simulations and resulted from the fact that the T1 weighting of the imaging sequence was determined mainly by the inversion-recovery preparation. All blood T1 values during the first pass of the three injections were less than one-third of the inversion time. The longitudinal magnetization is expected to be nearly fully recovered at the start of data acquisition and leads to similar blood signals when a centric phase encoding scheme and constant flip angle are used. If the exact blood T1 for each examination is known, variable flip angles can be used to optimize the blood signal; in such a case, further shortening of blood T1 will result in even higher signal (18). Coronary artery SNR on breath-hold images acquired after the first pass of gadopentetate dimeglumine injections was not significantly higher than that on precontrast images because of the rapid elimination of the contrast agent from the blood pool. gadomer-17 shortens blood T1 for a longer period and allows for repeated breath-hold imaging to cover the entire coronary system. If we assume that up to 13 breath-hold examinations are required to cover the entire coronary artery tree (20), 0.1-mmol/kg gadomer-17 is required because it permits an imaging time of at least 10 minutes, according to our study findings, during which coronary artery SNR and CNR are significantly enhanced. Neither 0.20mmol/kg gadopentetate dimeglumine nor 0.05-mmol/kg gadomer-17 is adequate for this purpose. One drawback of breath-hold imaging is that spatial resolution and coverage of each examination are limited by the im-

posed acquisition time. Respiratory gating with free breathing permits a longer imaging time and thus greater resolution and coverage per examination. However, because of the long imaging time and rapid extravasation of the contrast agent from the blood pool, use of gadopentetate dimeglumine did not result in significant SNR enhancement of the blood. Although it is possible to improve the SNR by optimizing the slow injection mode of the extravascular agent (28), 0.10-mmol/kg gadomer-17 produces a much greater SNR and CNR for a longer time and is highly desirable for respiratory-gated imaging. An injection of 0.05-mmol/kg gadomer-17 resulted in improved coronary artery SNR in only the first examination and may not be adequate for the required SNR and acquisition time at respiratorygated coronary artery imaging.

Comparison of 3D Breath-hold and Respiratory-gated Imaging SNR improvements with contrast agents in our study were within the limits predicted from simulations, in which inflow effect was not included. With gadomer-17 injections, coronary artery SNR and CNR on respiratory-gated images were much greater than those on breath-hold images because of the larger numbers of partitions and signal averages at respiratorygated imaging. In addition, the SNR and CNR improvements over precontrast levels with gadomer-17 injections at respiratory-gated imaging were substantially higher than those at breath-hold imaging. This indicates that an intravascular contrast agent is more useful for imaging with larger volumes because of the limited blood inflow enhancement at precontrast imaging. Advantages of imaging with a larger volume include higher SNR, less severe k-space truncation artifacts, and easy setup of slab location and orientation. With a large coverage, the coronary arteries can be visualized as a whole by using image reformatting. However, it should be noted that in our pig studies, respiration was controlled with a mechanical ventilator and its effects were eliminated with either breath


䡠

677

holding or respiratory gating. In studies in humans, the choice of technique is affected by the subject’s breathing patterns and ability to cooperate. In our study, three injections were given to each pig at the same imaging session. According to our measurements, the blood T1 values at 80 minutes after each of the three contrast material injections did not return to baseline values but were all greater than 500 msec. A simple calculation indicates that for blood T1 values less than 100 msec (which are the values of interest for contrast-enhanced coronary artery imaging), the shortening of blood T1 caused by residual effects of the previous injection was less than 10% and led to a blood signal increase of less than 5%. In addition, the order of the three injections was randomized. Therefore, the residual effects of previous injections on later studies should not have had a major effect on our results. Practical application: We have demonstrated in pigs that both a conventional extravascular contrast agent and a newly developed intravascular agent, gadomer-17, can substantially improve coronary artery SNR and coronary artery– myocardium CNR. However, gadomer-17 offers shorter blood T1 values and for a longer period, which allows the performance of multiple examinations of thin slabs to target various parts of the coronary system and is particularly useful for respiratory-gated imaging that covers a large volume and requires a long imaging time. Therefore, an intravascular contrast agent such as gadomer-17 is more effective than a conventional extravascular agent for coronary artery imaging. Although 0.05-mmol/kg gadomer-17 was more effective than 0.20-mmol/kg gadopentetate dimeglumine, we believe that 0.10-mmol/kg gadomer-17 is desirable for better CNR and higher spatial resolution in coronary artery imaging in humans. From a technical point of view, our results confirm that inversion recovery is the method of choice for suppressing the myocardial signal at contrast-enhanced coronary MR angiography. In conclusion, the results of our pig studies show that gadomer-17 shows promise in improving coronary MR imaging; further investigation into its utility in humans is warranted.

678

䡠

Radiology

䡠

March 2001

References 1. Goldfarb JW, Edelman RR. Coronary arteries: breath-hold, gadolinium-enhanced, three-dimensional MR angiography. Radiology 1998; 206:830 – 834. 2. Zheng J, Li D, Bae KT, Woodard PK, Haacke EM. Three-dimensional gadolinium-enhanced coronary MR angiography: initial experience. J Cardiovasc Magn Reson 1999; 1:33– 41. 3. Wang Y, Lee HM, Winchester PA, Yang Z, Bergman G, Shaknovich A. Motion matched k-space sampling for breath-hold contrast enhanced 3D coronary MR angiography (abstr). Radiology 1998; 209(P):355. 4. Kessler W, Laub G, Achenbach S, Ropers D, Moshage W, Daniel W. Coronary arteries: MR angiography with fast contrastenhanced three-dimensional breath-hold imaging-initial experience. Radiology 1999; 210:566 –572. 5. Stillman AE, Wilke N, Li D, Haacke EM, McLachlan S. Ultrasmall superparamagnetic iron oxide to enhance MRA of the renal and coronary arteries: studies in human patients. J Comput Assist Tomogr 1996; 20:51–55. 6. Engelbrecht MR, Saeed M, Wendland MF, Canet E, Oksendal AN, Higgins CB. Contrast-enhanced 3D-TOF MRA of peripheral vessels: intravascular versus extravascular MR contrast media. J Magn Reson Imaging 1998; 8:616 – 621. 7. Lauffer RB, Parmelee DJ, Dunham SU, et al. MS-325: albumin-targeted contrast agent for MR angiography. Radiology 1998; 207:529 –538. 8. Grist TM, Korosec FR, Peters DC, et al. Steady-state and dynamic MR angiography with MS-325: initial experience in humans. Radiology 1998; 207:539 –544. 9. Taylor AM, Panting JR, Keegan J, et al. Safety and preliminary findings with the intravascular contrast agent NC100150 injection for MR coronary angiography. J Magn Reson Imaging 1999; 9:220 –227. 10. Li D, Dolan B, Walovitch RC, Lauffer RB. Three-dimensional MR imaging of coronary arteries using an intravascular contrast agent. Magn Reson Med 1998; 39: 1014 –1018. 11. Johansson LO, Nolan MM, Taniuchi M, Fischer SE, Wickline SA, Lorenz CH. High resolution magnetic resonance coronary angiography of the entire heart using a new blood-pool agent, NC100150 injection: comparison with invasive x-ray angiography in pigs. J Cardiovasc Magn Reson 1999; 1:139 –143. 12. Woodard PK, Li D, Zheng J, Malcolm PN, Narra VR, Haacke EM. Coronary MR angiography with MS-325, an intravascular contrast agent (abstr). Radiology 1998; 209(P):305. 13. Stuber M, Botnar R, Danias P, et al. Contrast agent-enhanced, free-breathing, threedimensional coronary magnetic resonance angiography. J Magn Reson Imaging 1999; 10:790–799. 14. Larsson BW, Stubgaard M, Sondergaard L, Henriksen O. In vivo quantification of the unidirectional influx constant for Gd-

15.

16.

17. 18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

DTPA diffusion across the myocardial capillaries with MR imaging. J Magn Reson Imaging 1994; 4:433– 440. Zheng J, Li D, Venkatesan R, Haacke EM. Dynamic measurement of blood T1 after contrast injection with slice profile correction. J Magn Reson Imaging 1999; 10: 576 –581. Li D, Kaushikkar S, Haacke EM, et al. Coronary arteries: three-dimensional MR imaging with retrospective respiratory gating. Radiology 1996; 201:857– 863. Henkelman RM. Measurement of signal intensities in the presence of noise in MR images. Med Phys 1985; 12:232–233. Johansson LO, Fischer SE, Lorenz CH. Benefits of T1 reduction for magnetic resonance coronary angiography: a numerical simulation and phantom study. J Magn Reson Imaging 1999; 9:552–556. Li D, Paschal CB, Haacke EM, Adler LP. Coronary arteries: three-dimensional MR imaging with fat saturation and magnetization transfer contrast. Radiology 1993; 187:401– 406. Wielopolski PA, van Geuns RJ, de Feyter PJ, Oudkerk M. Breath-hold coronary MR angiography with volume-targeted imaging. Radiology 1998; 209:209 –219. Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishimura DG. Coronary angiography with magnetization-prepared T2 contrast. Magn Reson Med 1995; 33: 689 – 696. Botnar RM, Stuber M, Danias PG, Kissinger KV, Manning WJ. Improved coronary artery definition with T2-weighted, freebreathing, three-dimensional coronary MRA. Circulation 1999; 99:3139–3148. Stuber M, Botnar RM, Danias PG, et al. Double-oblique free-breathing high resolution three-dimensional coronary magnetic resonance angiography. J Am Coll Cardiol 1999; 34:524 –531. Li D, Haacke EM, Mugler JP, Berr S, Brookeman JR, Hutton MC. Three-dimensional time-of-flight MR angiography using selective inversion recovery RAGE with fat saturation and ECG-triggering: application to renal arteries. Magn Reson Med 1994; 31:414 – 422. Wang Y, Rossman PJ, Grimm RC, Wilman AH, Riederer SJ, Ehman RL. 3D MR angiography of pulmonary arteries using real-time navigator gating and magnetization preparation. Magn Reson Med 1996; 36:579 –587. Wilman AH, Huston J, Riederer SJ. Threedimensional magnetization-prepared timeof-flight MR angiography of the carotid and vertebral arteries. Magn Reson Med 1997; 37:252–259. Hofman MB, Henson RE, Kovacs SJ, et al. Blood pool agent strongly improves 3D magnetic resonance coronary angiography using an inversion pre-pulse. Magn Reson Med 1999; 41:360 –367. Zheng J, Woodard PK, Haacke EM, Li D. Efficacy of slow infusion in contrast-enhanced coronary artery imaging. J Magn Reson Imaging 1999; 10:799 – 804.

Li et al