key: cord-0010088-all97xwr
authors: Sharma, Puneet; Socolow, Josh; Patel, Salil; Pettigrew, Roderic I.; Oshinski, John N.
title: Effect of Gd‐DTPA‐BMA on blood and myocardial T(1) at 1.5T and 3T in humans
date: 2006-02-02
journal: J Magn Reson Imaging
DOI: 10.1002/jmri.20504
sha: a6a9ceae02cb976b7a781e64fc6e1a1211c716fa
doc_id: 10088
cord_uid: all97xwr

PURPOSE: To compare T(1) values of blood and myocardium at 1.5T and 3T before and after administration of Gd‐DTPA‐BMA in normal volunteers, and to evaluate the distribution of contrast media between myocardium and blood during steady state. MATERIALS AND METHODS: Ten normal subjects were imaged with either 0.1 mmol/kg (N = 5) or 0.2 mmol/kg (N = 5) of Gd‐DTPA‐BMA contrast agent at 1.5T and 3T. T(1) measurements of blood and myocardium were performed prior to contrast injection and every five minutes for 35 minutes following contrast injection at both field strengths. Measurements of biodistribution were calculated from the ratio of ΔR(1) (ΔR (1myo)/ΔR (1blood)). RESULTS: Precontrast blood T(1) values (mean ± SD, N = 10) did not significantly differ between 1.5T and 3T (1.58 ± .13 sec, and 1.66 ± .06 sec, respectively; P > 0.05), but myocardium T(1) values were significantly different (1.07 ± .03 sec and 1.22 ± .07 sec, respectively; P < 0.05). The field‐dependent difference in myocardium T(1) postinjection (T(1)@3T – T(1)@1.5T) decreased by approximately 72% relative to precontrast T(1) values, while the field‐dependent difference of blood T(1) decreased only 30% postcontrast. Measurements of ΔR (1myo)/ΔR (1blood) were constant for 35 minutes postcontrast, but changed between 1.5T and 3T (0.46 ± .06 vs. 0.54 ± .06, P < 0.10). CONCLUSION: T(1) is significantly longer for myocardium (but not blood) at 3T compared to 1.5T. The differences in T(1) due to field strength are reduced following contrast administration, which may be attributed to changes in ΔR (1myo)/ΔR (1blood) with field strength. J. Magn. Reson. Imaging 2006. © 2006 Wiley‐Liss, Inc.

This includes a number of applications that use contrast agents for cardiac studies, such as delayed enhancement imaging, first-pass myocardial perfusion, and MR angiography (MRA). The pharmacokinetics of gadolinium-based contrast agents (e.g., Gd-DTPA-BMA) have been well described in both animals and humans (5) (6) (7) (8) (9) (10) , and it has been shown that determining the regional contrast agent concentration is an important predictor for signal enhancement in pathologic regions (9, 11) . However, signal enhancement depends not only on pharmacokinetics and imaging parameters, but also on magnetic field strength (B 0 ), which causes the spin-lattice relaxation times (T 1 ) in most biological tissues to change (2, (12) (13) (14) . However, the magnitude (and variation among individuals) of T 1 differences between 1.5T and 3T for blood and myocardium pre-and postcontrast are unknown.

One of the indirect consequences of circulating contrast media is that the modulation in T 1 is not constant over time and among all patients. As a result, it is challenging to obtain consistent image contrast postinjection. An impression of the change in T 1 over time post-injection can be realized by studying the contrast kinetics of blood and myocardium in healthy human volunteers. From the time course of T 1 change, conclusions can be made about the role of the paramagnetic agent in each of these tissue compartments (blood and myocardium). By extending the study to 3T, one can both compare relaxation times between fields and assess the dependency of the contrast media on the field strength and tissue compartment. Postcontrast T 1 values in blood and myocardium are necessary to optimize pulse sequences at different field strengths.

The measurement of postcontrast T 1 values also enables evaluation of the partition coefficient ( myo ), which has significance in describing the pathologic state of injured myocardium (5, 15, 16) . Since the partition coefficient is an inherent physiological property, its value should not depend on MR properties, such as field strength. Although measurements of myo have been performed at 1.5T (16) , they have not been evaluated at 3T in the same subset of people. It is important to evaluate partition coefficient values if perfusion and biodistribution studies are to be extended to high field strengths.

Relaxation rates (R 1 ϭ 1/T 1 ) are often used to describe the effect of contrast media on tissue relaxation. Many mechanisms contribute to the relaxation rate, and these mechanisms can be linearly combined to express an "observed" R 1 value (R 1obs ). As such, the contribution from contrast media can be isolated from the precontrast R 1 (R 1pre ) to express meaningful information about tissue enhancement (17) and regional distribution of the contrast agent (18) from the following expression (15, 18) : (1) where fECV is the extracellular volume fraction, r 1 is the longitudinal relaxivity of the contrast agent (19) , and [CA] and [CA] EC are the contrast agent concentration in tissue and extracellular compartment, respectively. Note that R 1contrast is equivalent to the change in R 1 over time (⌬R 1 (t)), i.e., R 1contrast ϵ ⌬R 1 (t) ϭ R 1obs -R 1pre . The primary term in R 1contrast that may be subject to field dependency is r 1 , since [CA] and fECV are constant for a given dose and individual, respectively. Although it has been shown that r 1 is less sensitive to field changes above 1.5T (20) , other reports suggest a tissue-specific change in r 1 (9, 19, 21, 22) . Additional data are needed to determine whether contrast media affects relaxation uniquely at different field strengths.

An informative index of the distribution behavior of Gd-DTPA-BMA in myocardium and blood can be assessed with MRI by considering the ratio of ⌬R 1 (t) between myocardium and blood:

This expression has been used to determine the tissue distribution volume and assess the cellular integrity in ischemic and necrotic myocardium (5, 15, 16) , under the assumption that there is fast exchange and a steadystate concentration between extracellular compartments ([CA] EC-myo ϭ [CA] EC-plasma ). The factor fECV/(1.0 -Hct) relates the extracellular volume of distribution between myocardium and blood, and is equivalent to the partition coefficient ( myo ). If the ratio of relaxivities (r 1myo /r 1blood ) is unity, then myo ϭ ⌬R 1myo /⌬R 1blood (t), which should be the same between 1.5T and 3T. A measurable change in ⌬R 1myo /⌬R 1blood (t) between 1.5T and 3T may indicate a field-dependent change in r 1myo / r 1blood . The focus of this investigation was to measure T 1 of blood and myocardium in human subjects at two imaging field strengths (1.5T and 3T) before and after contrast agent injection. T 1 values were determined precontrast injection and at every five minutes postcontrast for 35 minutes. In addition, the distribution of contrast media between myocardium and blood at 1.5T and 3T was quantitatively investigated.

All experiments were performed using commercially available Gd-DTPA-BMA (Omniscan, Amersham, Oslo, Norway) from 20-mL prefilled syringes. Precontrast T 1 values of blood and myocardium were calculated from a set of four ECG-gated, inversion recovery (IR), singleshot, balanced steady-state free precession (b-SSFP) sequences (FOV ϭ 300 ϫ 285 mm, 90 lines acquired (using 80% scan percentage) reconstructed to 256 matrix, thickness ϭ 8 mm, TR/TE/␣ ϭ 2.5 msec/1.2 msec/35°, readout duration ϭ 225 msec). This sequence is similar to one presented previously (23) , and hereafter is referred to as IRss. Magnetization preparation was performed with a nonselective adiabatic inversion pulse. Single-shot imaging as a T 1 measurement technique was used before contrast injection to eliminate the long repetition (and breath-hold) times necessary to calculate longer blood and myocardial T 1 s at 3T. The inversion times (TIs) were slightly different at both field strengths to ensure points on either side of the zero-crossing (1.5T: 400, 600, 1000, 1400 msec; 3T: 500, 800, 1100, 1500 msec). Because of the long TIs, a trigger delay was applied to provide imaging in diastole of the second heartbeat.

Postcontrast T 1 values were calculated from two ECG-gated, segmented IR b-SSFP images (FOV ϭ 300 ϫ 285 mm, matrix ϭ 256, 42 lines/segment, thickness ϭ 8 mm, TR/TE/␣ ϭ 3.1 msec/1.05 msec/40°, readout duration ϭ 125 msec, three R-R intervals, acquisition time ϭ 12 heartbeats), with trigger delays set to ensure imaging of the same phase of the cardiac cycle (diastole). This sequence will be referred to as 2pt-IR. The TI of the first image was set Յ 150 msec, while the second TI was set maximally depending on the subject's heart rate (usually Ն 650 msec). The temporal resolution for each postcontrast T 1 measurement (two images) was less than one minute.

All signal values used for T 1 fitting were normalized using the mean background noise to offset the differences in receiver gain and display scaling factors between the individual TI images. Precontrast T 1 values (using IRss) were calculated from a two-parameter fit assuming monoexponential relaxation and ideal spin inversion:

The fit enabled estimation of T 1 and the scaling factor, F, from the measured (scaled) signal intensity, S, and TIs. The goodness of fit was determined from the r 2 value of the fit. Although it has been shown that the "observed" T 1 (T 1 *) in IR b-SSFP is less than the "true" T 1 due to the transient approach to steady state (24) , the acquisition duration (T acq ) of the current method is short (ϳ225 msec), which makes the magnetization decay rate during readout, E 1 * (E 1 * ϭ exp(-T acq /T 1 *)), negligible relative to "true" E 1 . As a result, the difference between the observed and true T 1 s is minor. Postcontrast T 1 was determined via a two-point ratio method (using 2pt-IR), as described elsewhere (25) . Briefly, the two IR images (TI-high and TI-low) were identically scaled and divided (TI-high/TI-low). Since the imaging parameters were held constant (except for TI) and images scaled equivalently, the ratio of intensity values (intensity-ratio) in each pixel was related to T 1 by intensity-ratio ϭ (1 Ϫ 2exp(ϪTI high /T 1 ))/ (1 Ϫ 2exp(ϪTI low /T 1 )) (4) where TI high and TI low represent the TIs used in each image. Equation [4] was solved numerically for T 1 using a modified Newton-Rapshon method performed in Matlab 6.5 (MathWorks, Natick, MA, USA). Note that the intensity-ratio can have a positive or negative value, depending on the sign of the magnetization at the chosen TIs. Since only magnitude images were acquired in this analysis, the intensity-ratio was always positive, which led to uncertainty about the true sign of the magnetization. This uncertainty resulted in two unique solutions in the 2pt-IR analysis. However, by choosing a very low TI (TI low Յ 150 msec) coupled with a high TI, we assumed that the intensity-ratio was always negative, even for the low blood T 1 's seen early after contrast injection (15, 16) .

All images were taken offline for T 1 calculation. In vivo T 1 maps were generated from the postcontrast 2pt-IR in Matlab by performing T 1 calculations on a pixel-bypixel basis. From the T 1 maps, region-of-interest (ROI) measurements were taken from blood (one ROI: center of left ventricle) and myocardium (two ROIs: septum and posterior wall), and the mean and standard deviation (SD) of the measurement were recorded.

The T 1 measurement techniques were validated in MR phantoms. Thirteen 50-mL plastic vials containing distilled water were made with varying concentrations of Gd-DTPA-BMA contrast media (0 mM-2 mM Gd). The tubes were arranged in a head coil at 3T (Magnetom Trio; Siemens Medical Systems, Erlanger, Germany) and reference T 1 relaxation times were measured using a segmented IR b-SSFP sequence with 20 TI values spanning 90 -6500 msec. The parameters for the sequence were FOV ϭ 300 mm 2 , matrix ϭ 256 ϫ 256, TR/TE/␣ ϭ 3.4 msec/1.26 msec/45°, 21 lines/segment, slice thickness ϭ 8 mm, and segment interval ϭ 7000 msec.

The IRss (used for precontrast T 1 measurements) and 2pt-IR (used for postcontrast T 1 measurements) methods were used to measure T 1 in the same 13 tubes, using TIs equivalent to those proposed for the in vivo experiments. A physiology simulator was used to supply a constant heart rate of 75 beats per minute (bpm). T 1 maps were generated for both techniques, as described in the previous section, since sample motion was absent. The similarity with the reference T 1 measurement technique was determined from the percent difference between the two values.

Ten healthy human subjects (six males and four females, age ϭ 29.7 Ϯ 4.7 years) were recruited to participate in the study. The protocol in this study was approved by the university's institutional review board, and informed consent was provided by each volunteer. Each subject underwent two MRI studies: one at 1.5T (Philips Intera, Best, The Netherlands; five-element phased-array receive coil) and one at 3T (Siemens Magnetom Trio; eight-element phased-array receive coil). Both studies involved contrast administration of either 0.1 mmol/kg (N ϭ 5, three males and two females) or 0.2 mmol/kg (N ϭ 5, three males and two females) intravenously. The appropriate dose was determined according to each subject's weight. The subjects underwent each study at least 3 days apart and no more than 3 weeks apart, and 1.5T and 3T imaging studies were performed in random order.

Automatic shimming was performed at 1.5T, while at 3T local shim volumes were manually placed over the heart to reduce artifacts due to field inhomogeneities. T 1 measurements were performed on a midventricular short-axis slice. Following the precontrast T 1 imaging protocols, contrast media was administered through a bolus injection in the antecubital vein, and subsequent postcontrast T 1 measurements were made every five minutes for 35 minutes.

To characterize the effect of field dependence on relaxation times, the T 1 difference between 1.5T and 3T was determined before and after contrast injection for blood and myocardium in each subject (T 1 @3T -T 1 @1.5T). This T 1 difference was averaged over all subjects at a given dose, and represented a time course of T 1 differences between 1.5T and 3T, before and after contrast injection.

The partition coefficient of a tissue can be approximated by measuring the ratio of ⌬R 1 in each tissue compartment as outlined in Eq. [2] . ⌬R 1myo /⌬R 1blood (t) measurements were performed at each time point postcontrast by converting the T 1 information to R 1 (R 1 ϭ 1/T 1 ). Since ⌬R 1myo /⌬R 1blood (t) was quantified temporally and between field strengths, the steady-state distribution assumption was directly assessed along with the field dependence of Eq. [2] . Since direct measurements of hematocrit and myocardial extracellular volume fraction were not made, the precise value of r 1myo and r 1blood could not be quantified.

The data are presented as the mean Ϯ SD. Comparisons of measurement results between 1.5T and 3T were made by analysis of variance (ANOVA) and deemed significant if P Ͻ 0.05.

Measurements from the T 1 maps of both T 1 measurement techniques are shown in Table 1 , and compared with the reference T 1 measurements in the phantoms. Two-parameter T 1 fitting to the IRss images was determined with low error (relative to reference measurements), particularly for long T 1 s (Ͼ0.50 sec; Ͻ5% dif-ference), which span the T 1 values seen precontrast in vivo. The 2pt-IR method yielded low error for T 1 values in the range of 0.12-0.50 sec (Ͻ3% difference), which are similar to T 1 values seen postcontrast in vivo. Accuracy for measuring long T 1 s using the 2pt-IR method was compromised by the use of only three RR intervals between inversions.

For measuring very low T 1 s (Ͻ0.50 sec), IRss was not as accurate as the 2pt-IR method. Even so, errors were low with the single-shot technique for measuring T 1 values typically seen postcontrast (Ͻ10% difference for 0.12 sec Ͻ T 1 Ͻ 0.50 sec), suggesting its use for cases in which subject breath-holding is an issue.

The T 1 measurement techniques were performed in all subjects without any complications. Since the TIs used precontrast ranged from 400 to 1500 msec spanning two heartbeats, some misregistration between images occurred despite the use of trigger delays. Therefore, fitting was performed using signal intensities from manually drawn ROIs. From the twoparameter fit of the precontrast T 1 measurement data, low fit errors were observed (r 2 Ͼ 0.95), while the 2pt-IR produced T 1 maps with no registration errors (Fig. 1) . A histogram analysis of ROI measurements from the T 1 maps (by measuring the SD within each ROI) resulted in 95% confidence intervals of Ϯ7.3 msec and Ϯ8.0 msec, respectively, for T 1 s measured using this technique (26) .

Blood and myocardium T 1 values pre-and postcontrast at 1.5T and 3T are summarized in Table 2 , respec- 

Top: Precontrast T 1 measurement of one subject using four separate IRss images at selected TIs. In vivo T 1 was estimated using ROIs and a least-squares approximation to Eq. [3] . Bottom: T 1 calculation from a subject 15 minutes postinjection (0.2 mmol/kg) at 1.5T. The ratio method for calculating T 1 involves two IR images using two different TIs: TI high ϭ 650 msec and TI low ϭ 150 msec. The images were rescaled and then divided (TI high /TI low ) to produce an intensity-ratio map. A T 1 map was numerically calculated on a pixelwise basis using Eq. [4] , assuming an appropriate T 1 value.

tively. Precontrast T 1 values for blood (N ϭ 10) did not differ significantly between 1.5T and 3T despite a mean increase (1.5T: 1.58 Ϯ 0.13 sec; 3T: 1.66 Ϯ 0.06 sec; P Ͼ 0.05). Significant differences were observed between precontrast T 1 values for myocardium (1.5T: 1.07 Ϯ 0.03 sec; 3T: 1.22 Ϯ 0.07 sec; N ϭ 10, P Ͻ 0.05).

Following contrast injection there was a significant decrease in blood and myocardium T 1 values at both field strengths and doses (P Ͻ 0.001). The mean postcontrast T 1 values for blood and myocardium tended to be higher at 3T compared to 1.5T by 5-10% at each time point; however, the increase was not significant (P Ͼ 0.05). It was observed that the change in T 1 from 1.5T to 3T was greater precontrast than postcontrast. Furthermore, the magnitude of the change between fields was relatively insensitive to the administered contrast dose, as determined from Table 2 . T 1 values were significantly lower at double dose (0.2 mmol/kg) than single dose at both field strengths (P Ͻ 0.05), but the change from 1.5T to 3T was the same for each dose. Figure 2 depicts these results by plotting the difference in T 1 between fields (T 1 @3T -T 1 @1.5T) for blood and myocardium preand postcontrast. As shown, the difference in myocardium T 1 between 1.5T and 3T seen prior to contrast injection (0.16 Ϯ 0.06 sec, N ϭ 10) was reduced by 72% (0.04 Ϯ 0.06 sec, N ϭ 10, P Ͻ 0.05) after 10 minutes. The amount of the reduction was almost constant over all time points and was insensitive to dose. A similar trend was observed in blood, but the decrease was 30% after 10 minutes. Figure 3 shows the cumulative ⌬R 1myo /⌬R 1blood (t) values for 0.1 mmol/kg and 0.2 mmol/kg. A constant value of ⌬R 1myo /⌬R 1blood (t) over time existed at both 1.5T and 3T, which confirms the fast-exchange assumption by showing a steady-state distribution of the contrast agent between blood and myocardium compartments. ⌬R 1myo /⌬R 1blood (t) did not differ significantly between single and double doses (1.5T (single and double): 0.49 Ϯ .05 and 0.44 Ϯ .06; 3T (single and double): 0.56 Ϯ .05 and 0.53 Ϯ .07, P Ͼ 0.05). Despite the constant value over time, a large difference in ⌬R 1myo / ⌬R 1blood (t) was observed between 1.5T and 3T (0.46 Ϯ .06 and 0.54 Ϯ .06, P Ͻ 0.10). This implies that there may be a difference in compartmental contrast agent relaxivities (r 1myo and r 1blood ) between field strengths (see Eq. [2] ), assuming that the ratio of compartmental extracellular volumes ( myo ) did not change between 1.5T and 3T.

The major findings of this study were as follows: 1) the T 1 of myocardium was 1.07 Ϯ 0.03 sec at 1.5T and 1.22 Ϯ 5T and 3T decreases following contrast injection, and the decrease is more significant for myocardium than blood. Following contrast injection, the T 1 difference remains relatively constant over time and is insensitive to dose and tissue type. Ten minutes postinjection the T 1 difference decreased 72% for myocardium and 30% for blood. Although there was significant variability in T 1 among subjects, the relaxation times at 3T were still generally greater than 1.5T. 0.07 sec at 3T; 2) the T 1 difference due to field strength (between 1.5T and 3T) was significantly reduced for myocardium, but not for blood, following contrast administration; 3) ⌬R 1myo /⌬R 1blood (t) differed between 1.5T and 3T, suggesting field and tissue dependence of the contrast agent relaxivity; and 4) there was significant inter-subject variability in T 1 postcontrast. In all 10 subjects, the T 1 of myocardium was 1.07 Ϯ 0.03 sec at 1.5T and 1.22 Ϯ 0.07 sec at 3 T, while blood T 1 was 1.58 Ϯ 0.13 sec at 1.5T and 1.66 Ϯ 0.06 sec at 3T. The 3T values found in our study were roughly 8% higher than those reported by Noeske et al (2) (1.55 sec for blood and 1.12 sec for myocardium at 3T), which may reflect the different measurement techniques used. Noeske et al (2) utilized a partially refocused gradient-echo in the steady-state technique (GRASS) to measure T 1 of myocardium and blood; however, they did not specify the segment TR and quantification method used. The T 1 of blood at 1.5T was approximately 15-30% higher in our study compared to previously reported values (1.20 sec (3,27), 1.23 sec (16), 1.38 sec (28), and 1.34 sec (29)), but lower than that observed in a recent study in pigs (1.80 sec (30) ). However, these values may also dependon the measurement technique used. Klein et al (16) and Flacke et al (28) utilized a modified Look-Locker technique that is known to measure T 1 * (31). It is also known that blood T 1 at 1.5T depends strongly on hematocrit (32) , with variations between 1.1-2.0 seconds for Hct variations of 0.6 and 0.2. The b-SSFP readout module used with our method has been shown to be less sensitive to saturation effects than spoiled gradient-echo techniques (e.g., fast lowangle shot (FLASH)) due to the refocusing and reuse of transverse magnetization, which makes the sampling of free M z recovery more accurate in IR experiments (33) . This is because the decay rate of magnetization during b-SSFP readout (E 1 *) is slower than that for FLASH (24, 28, 34) . Because of this transient decay rate, the continuous sampling of the T 1 relaxation curve using either b-SSFP or FLASH requires appropriate correction of the measured T 1 relaxation time, with more significant corrections needed when FLASH is used. Since our technique sampled T 1 with discrete TIs using (relatively) short readout times, the majority of the signal intensity at k y ϭ 0 evolved from the free IR preceding data acquisition, and thus the measured T 1 was an accurate estimation of the true T 1 . Even so, the accuracy is a function of the applied flip angle and the number of excitations used during data acquisition, especially if a simplistic recovery expression like Eq. [3] is used to map signal intensities to T 1 . It can be shown (by using the full b-SSFP expression (35) ) that the error in estimation of T 1 using IRss is relatively minor (Ͻ7%) over flip angles of 10 -90°(assuming T 1myo ϭ 1100 msec, T 2myo ϭ 50 msec, T 1blood ϭ 1500 msec, T 2blood ϭ 250 msec). Furthermore, this measurement technique was assessed on phantoms of varying T 1 values and resulted in accurate measurements compared to reference T 1 values.

The postcontrast T 1 values at 1.5T reported here are comparable to those obtained previously with the Look-Locker method (16) . The calculation of T 1 maps using the 2pt-IR method is limited by possible misregistration of the source images, which may cause erroneous measurements near tissue interfaces. This can be attributed to blurring from insufficient breath-holding, or acquisitions during different phases of the cardiac cycle. Erroneous T 1 measurements due to misregistration become significant if the technique is extended to individuals with small subendocardial infarcts. In this case, manually placed ROIs on the source images may be ideal.

The average difference in precontrast T 1 between 1.5T and 3T was larger for myocardium (0.16 Ϯ 0.06 sec) than blood (0.08 Ϯ 0.13 sec). The fact that myocardium T 1 increased more than blood from 1.5T to 3T may be attributed to the greater free water content and shorter molecular correlation times in blood (36) , which would cause less field dependence on T 1 , in much the same way that water and CSF T 1 appear almost insensitive to field change. The trend toward similar blood and myocardium T 1 values at high field strength may lower image contrast between blood and myocardium on T 1weighted images.

A dose of 0.2 mmol/kg caused T 1 values to be significantly lower compared to 0.1 mmol/kg at both field strengths (P Ͻ 0.05), as shown in Table 2 . However, the measured ⌬R 1myo /⌬R 1blood (t) did not differ significantly between single and double doses. As a result, the ⌬R 1myo /⌬R 1blood (t) data are shown cumulatively. The partition coefficient is not known to be dose dependent, since it is an inherent physiologic property. Relaxivity also should not be dose dependent. Indeed, different doses of Gd-DTPA-BMA in solution yield specific R 1 values (Table 1) , and the slope of this relationship is equal to the relaxivity and is assumed to be constant.

All postcontrast T 1 values were generally higher at 3T compared to 1.5T, but the change was neither significant nor constant over all subjects. Some subjects revealed a marked T 1 change between fields (ϩ0.15 sec), whereas others experienced only a subtle change at the same time point (ϳ0.06 sec). As a result, when data at each time point were analyzed cumulatively, there was no significant difference in blood and myocardium postcontrast T 1 between 1.5T and 3T (Table 2 ). This observation reveals that the change in T 1 seen prior to contrast administration is obscured following injection, Figure 3 . The ratio of ⌬R 1 (⌬R 1myo /⌬R 1blood ) exhibits constancy over time. This verifies that a steady state exists between compartments ([CA] myo ϭ [CA] plasma ). However, a measurable difference was observed between fields, which suggests that the compartmental influence of the contrast agent favors myocardium at 3T.

The intrinsic signal-to-noise ratio in human cardiac imaging at 1.5, 3, and 4T

Human cardiac imaging at 3T using phased array coils

Double inversion black-blood fast spin-echo imaging of the human heart: a comparison between 1.5T and 3.0T

Preliminary report on in vivo coronary MRA at 3 Tesla in humans

Measurements of the distribution volume of gadopentetate dimeglumine at echo-planar MR imaging to quantify myocardial infarction: comparison with Tc-DTPA autoradiography in rats

Pharmacokinetics of Gd-DTPA/dimeglumine after intraveneous injection into healthy volunteers

Comparison of the biodistribution of gadolinium-153 DTPA and technetium-99m DTPA in rats

In vivo tissue extracellular volume fraction measurement by dynamic spin-lattice relaxometry: application to the characterization of muscle fiber types

Tissue distribution and magnetic resonance spin-lattice relaxation effects of gadolinium-DTPA

Biodistribution and toxicity of MR imaging contrast media

Myocardial magnetic resonance imaging contrast agent concentrations after reversible and irreversible ischemic injury

MR relaxation times in human brain: measurement at 4 T

In vivo field dependence of proton relaxation times in human brain, liver and skeletal muscle: a multicenter study

Proton magnetic resonance studies of water in normal and tumor rat tissue

Alterations in T1 of normal and reperfused infarcted myocardium after Gd-BOPTA versus Gd-DTPA on inversion recovery EPI

The influence of myocardial blood flow and volume of distribution on late Gd-DTPA kinetics in ischemic heart failure

How much contrast is enough? Dependence of enhancement on field strength and MR pulse sequence

Dependence of MR signal intensity on Gd tissue concentration over a broad dose range

Measurements of the relaxivity of gadolinium chelates in tissues in vivo

Field strength and dose dependence of contrast enhancement by gadolinium-based MR contrast agent

Gadodiamide T1 relaxivity in brain tissue in vivo is lower than in saline

Gd-DTPA relaxivity depends on macromolecular content

A fast, TI insensitive infarct imaging technique

Inversion recovery True-FISP: quantification of T1, T2 and spin density

Measurements of relaxivity (R1) post contrast in patients with prior myocardial infarction

Magnetic resonance imaging: physical principles and sequence design

Black blood" T2-weighted inversion-recovery MR imaging of the heart

Measurement of the gadopentetate dimeglumine partition coefficient in human myocardium in vivo: normal distribution and elevation in acute and chronic infarction

T1-relaxation kinetics of extracellular, intracellular and intravascular MR contrast agents in normal and acutely reperfused infarcted myocardium using echo-planar MR imaging

Characterization of T1 relaxation time and blood-myocardial contrast enhancement of NC100150 injection in cardiac MRI

A novel method for analysis of TOMROP data

MR imaging of various oxidation states of intracellular and extracellular hemoglobin

T1 quantification with inversion recovery TrueFISP

On the transient phase of balanced SSFP sequences

Characterization and reduction of the transient response in steady-state MR imaging

Relaxation effects in nuclear magnetic resonance absorption

Day-to-day variability in glomerular filtration rate in normal dogs by scintigraphic technique

Measurement of single-kidney glomerular filtration rate using a contrast-enhanced dynamic gradient-echo sequence and the Rutland-Patlak plot technique

Studies of Gd-DTPA relaxivity and proton exchange rates in tissue

Highresolution intracranial and cervical MRA at 3.0T: technical considerations and initial experience

MR imaging at high magnetic fields

Quantification of myocardial blood flow and extracellular volumes using a bolus injection of Gd-DTPA: kinetic modeling in canine ischemic disease

possibly as a result of different contrast kinetic behavior among subjects, or a substantial T 2 * dephasing effect at 3T during T 1 measurement. The former may be attributed to differences in the glomerular filtration rate (GFR), left ventricular ejection fraction, or extracellular volumes among the subjects, while the latter may be due to field inhomogeneities and susceptibility effects. Although imaging procedures were implemented to reduce T 2 * dephasing (short TE), and the dose was identical at both field strengths, it is possible that the concentration of the contrast agent in circulation at any given time point was not the same during both study sessions. A difference in Gd-DTPA-BMA concentration among studies may be related to day-to-day changes in GFR that are largely controlled by food or fluid intake. Large differences in GFR among humans and dogs (37, 38) and over day-to-day periods (38) were recently observed. This degree of variation in postcontrast T 1 measurements in vivo was also observed in previous studies (15, 16) .The ratio ⌬R 1myo /⌬R 1blood (t) at 1.5T was similar to values measured previously by MRI at 1.5T (16) and 2T (5, 15) . These previous reports assumed that myo ϭ ⌬R 1myo /⌬R 1blood (t), which implies r 1myo /r 1blood ϭ 1 (Eq. [2] ). However, in the present study, there was an observable difference in ⌬R 1myo /⌬R 1blood (t) between 1.5T and 3T, which (from Eq. [2] ) suggests there may be some tissue and field dependency of Gd-DTPA-BMA relaxivity (r 1 ). Previous investigations that directly quantified Gd-DTPA relaxivity reported marginal decreases in r 1 in vivo and in vitro at high field strengths (39 -41) . Since explicit contrast agent concentrations were not determined in the present study, the relaxivity of Gd-DTPA-BMA in blood and myocardium (r 1blood and r 1myo , respectively) could not be directly calculated using serial R 1 measurements (Eq. [1] ). The ratio of compartmental relaxivities (r 1myo /r 1blood (t)) is only valid under the conditions of contrast agent steady state, which was confirmed by a constant value for ⌬R 1myo /⌬R 1blood (t) over time. Even though Fig. 3 demonstrates this constancy at both 1.5T and 3T, it is apparent that a difference exists for this measure between the two field strengths (P Ͻ 0.10). Using approximate values of fECV ϭ 0.35 and Hct ϭ 0.40 (15) , myo ϭ fECV/(1 -Hct) Ϸ 0.58 in Eq. [2] , making r 1myo /r 1blood approximately 0.80 at 1.5T and 0.93 at 3T. It can be inferred from these data, therefore, that the relaxivity of Gd-DTPA-BMA may be greater in blood than in myocardium (r 1myo /r 1blood (t) Ͻ 1). Similarly, using Eq. [1] , the ratio of ⌬R 1 (t) between 1.5T and 3T can be determined in the same individual for either blood or myocardium to reveal the field dependence of r 1 : ⌬R 1 ͑t͒ 1.5T /⌬R 1 ͑t͒ 3T ϭ r 1.5T /r 3T . It can be shown with this equation that r 1.5T /r 3T is 1.18 Ϯ 0.15 in blood and 1.01 Ϯ 0.10 in myocardium, implying that the relaxivity in blood may decrease with field strength (r 1.5T /r 3T Ͼ 1) while the relaxivity in myocardium remains constant. However, because of the broad range of measured myocardial extracellular volumes (0.25-0.40 (5, 15, 16, 42) ), there will be uncertainty in generalizing r 1myo /r 1blood and r 1.5T / r 3T without precise measurements of Hct or myo .Longer T 1 s will exist at higher fields in a given individual, even after injection of a contrast agent. Thus, for imaging sequences that rely on preparation pulses, such as inversion and saturation recovery, it is simpler to suppress longer T 1 s because the slope of magnetization recovery becomes shallower as it crosses or originates from zero, allowing some leeway in TI selection. However, this may also decrease image contrast if the other tissue of interest is also suppressed. Our results suggest that the possible decrease in blood r 1 seen at 3T would produce lower image contrast between blood and myocardium postcontrast at 3T relative to 1.5T. This may have significance in delayed enhancement imaging at 3T, since image contrast will likely decrease between blood and normal myocardium after normal myocardium is suppressed using IR. However, this may benefit image contrast between enhanced infarct tissue and blood for delineating subendocardial infarcts. Unfortunately, predictions can not be made at this time about the field-dependent behavior of contrast-enhanced infarct tissue.In conclusion, T 1 increased from 1.5T to 3T, and more significantly for myocardium than blood. Following contrast administration, T 1 differences between 1.5T and 3T were obscured by field and contrast agent effects such that there was no significant difference in T 1 between 1.5T and 3T following contrast injection. The ratio of contrast agent distribution (⌬R 1myo /⌬R 1blood (t)) exhibited some field dependence, which suggests that contrast agent relaxivity may also be field dependent. These factors may play a role in the reduced T 1 difference observed between 1.5T and 3T.