Cardiac Robert S. Balaban, PhD Frederick W. Heineman,

#{149} Scott

PhD,

Magnetization in MR Imaging

Chesnick MD

Kyle

Hedges,

#{149}

PhD

Frederick

Samaha,

#{149}

M

resonance

AGNETIC

ing

of the

(MR)

heart

imag-

cots

is beginning

to

have a significant impact on the evaluation of cardiac function (1-4). One requirement of such studies is to accunately differentiate between ventriculan cavity and myocardium to determine ejection fraction (2) or local motion (5,6). Furthermore, long-echotime (TE) (ie, T2-weighted) images have been shown to be potentially useful

in the

detection

regions (3,4,7), tumors, or transplant rejection (8-10). One of the many difficulties encountered

in cardiac

MR

studies

artifacts.

One

solution

is the

to this

problem, with regard to contrast between heart chamber and wall, has been the use of Ti-weighted sequences that rely cause an effective

on blood shortening

flow to of blood

Ti. These sequences are usually applied in a rapid-acquisition mode (echo-planar imaging or fast low-angle shot [ii]) to acquire a series of cardiac images (3,4). Yet Ti-weighted images provide little information

about

macromolecular

on

biochemical changes in the cardiac muscle in various pathologic conditions that is apparently provided with

Radiology

distorting long-TE images, makes the determination

180:671-675

a T2-weighted

and

apparent

ficult

(12).

image.

In addition

this

to

motion of the real

T2 of cardiac Indeed,

tissue

motion

dif-

inter-

ference may be partially responsible for the variability in studies designed to evaluate the effectiveness of T2 weighting ies.

The the Laboratory Heart, Lung,

of Cardiac Energetics, Blood Institute, National Institutes of Health, Bldg 1, Rm B3-07, Bethesda, MD 20892. Received December 19, 1990; revision requested February 14, 1991; revision received March 15; accepted April 18. Address reprint requests to R.S.B. RSNA, 1991 and

in cardiac

purpose

evaluate transfer

contrast

cavity

myocardium

determine lan and transfer

the

study

stud-

was

magnetization (MTC) could

to increase and

pathology

of this

whether contrast

relative

water-proton rates in MR

between

effectively

to

with

short

saturates

macromolecules an effective tween

the

TEs.

The

the

protons

in the

of the tissue. Where coupling mechanism bemacromolecules

and

water

a selective decrease in intensity of water will Thus,

MTC

is caused

by

a decrease in MR signal intensity in regions where an effective coupling mechanism macromolecules stant of the

between

water and exists. The rate proton magnetization

transfer between the and water can also be a pixel-by-pixel basis cases, MTC is similar tensity likely dence

lecular chemistry, and concentration. ever, be generated

tissue con-

macromolecules quantitated on (16). In many to T2 signal in-

(13-i6). This similarity occurs because of the of both processes on

significant

structure

chest

be observed.

degradation of spatial resolution in long-TE sequences due to heart-motion

of the

generation of MTC in MR images has been described previously (13-16). To be brief, it involves the application of a low-power radio-frequency (RF) irradiation off-resonance from the bulk water resonance. This irradiation

protons exists, the MR signal

of ischemic

Index terms: Heart, experimental studies, 51.1214 #{149}Heart,MRstudies,51.1214,52.1214 Magnetic resonance (MR), contrast enhancement #{149}Magnetic resonance (MR), experimental #{149} Magnetic resonance (MR), pulse sequences

I From National

MD

Transfer Contrast ofthe Heart’

The use of magnetization transfer contrast (MTC) in short-echo-time (TE) cardiac magnetic resonance (MR) imaging was evaluated. For most cardiac MR imaging protocols, either long TE and short repetition time or exogenous intravascular agents are used for generating contrast between the ventricular wall and cavity as well as detecting pathologic conditions of the ventricular wall. The major problem with longTE images is that the motion of the heart degrades the spatial resolution of the image during the TE period. However, MTC is generated by an off-resonance irradiation during the interpulse delay period that is relalively insensitive to motion artifacts. Short-TE (5-15 msec) gradient-recalled echo sequences were used for imaging the heart with and without MTC. These studies revealed that MTC can be used to greatly improve the contrast between the myocardial wall and blood chamber in short-TE images and may provide useful parameters for tissue characterization in pathologic cardiac muscle.

1991;

Radiology

most depenmacromo-

correlation time, MTC can, howin the absence of a

TE in comparison

with

T2

weighting. In addition, in comparison with T2 measurements (12), MTC can be quantitated with little concern fon motion (16). In previous studies, we found that off-resonance irradiation had a strong effect on muscle tissue but little or no effect on blood (13-16). The lack of irradiation effect on blood is due to the low concentration of macnomolecules

and

flow.

Blood

flow

MTC effect, since it is difficult rate the few macromolecules in blood as it moves through gion

that

is irradiated

(13).

inhibits

the

to satupresent the reIn addi-

be used

the

as well

heart as to

macromolecumagnetization imaging proto-

Abbreviations: Bi = surface coil magnetic field, MTC = magnetization transfer contrast, Q = quality factor of electric charge, RF = radio frequency, SAR = specific absorption rate, TE echo time, TR = repetition time.

=

671

tion, MTC magnetization plane,

is generated is not

making

to motion

while the in the transverse

it relatively

artifacts.

On

insensitive the

basis

was

these observations, MTC candidate for increasing

limes,

and

formation structure available different

100

ventricular-

11

surface

chemistry

(17,18),

useful

about by using pathologic

is especially emia which

the

C 0

in

true

and transplant T2 changes

0

wall

may

this approach conditions.

with

regard

Figure

inalso be in

This to isch-

rejection, have already

AND

at 1.5 T.

The

RF coil used

in been

custom-designed

surface

as a rectangle

coil

a

Hoult

was

a

away wave

The region the induced a rectangular The power

the resonant frequency cmrange of 63.8 MHz. Final tun-

capacitors

was performed and

fed

with

a

with capaci-

through a Bazooka Balun (21) made with a triaxial Beldon 9222 cable (Beldon, Richmond, Idaho). The entire coil was placed in a sealed Delrin container (FPI Industries, Pittsburgh). The coil had an unloaded quality factor of electric charge (Q) of 150 with a loaded Q of 30. There was no appreciable decrease in nesonant frequency ( < 50 KHz) with loading on a human chest, indicating that the dielectric coupling to the sample was minimat (19). This is important for the penformance of the coil as well as minimization of dielectric heating in the MTC generation by the off-resonance irradiation. Power deposition from this coil was evaluated to ensure that human subjects tive

coupling

were not exposed

to excessive

RF power.

For surface coils, the risk of excessive power deposition is greatest in the tissue dosest to the coil elements. The risk is both from small “hot spots,” due to con672

#{149} Radiology

with

ited the was

detected

with

I cm

of subject)

Figure surface

2.

Effect

distance from the effect. A phantom of 6% agar and 75 mmoVL saline solution was used in this study. Mo refers to control data, Ms to data with off-resonance irradiation.

of radial

coil on the MTC

Data are presented as the contrast-to-noise ratio for the agar and saline as a function distance

from

protocol pulse

the

power

deposited

by

was approximated as just above the coil. in a 1-cm-thick anuvolume, Vane as a fraction, fn

deposited

determined

separation)

in a 1-cm-thick from

from

with

Channel

coil.

The

and

2 cm

specific

Q measure-

coil from

a human

absorption

echo

with

chest.

the expression Deposition of depends on not just the local assumes current

used in this approach is not exact. However, we believe that it is adequate for the purpose at hand. With an SAR limit of 8 W/kg (guideline limit of the U.S. Food and Drug Administration) in any I cm of tis-

the coil, our measurements

set

a limit on total power to the coil for human studies at approximately 32 W. An RF

fuse (23) in the irradiation sured that this limit could

rraciaon

Transmitter

4 s&eo

Grerbent

Phase

Gractont

Encode

Readout

Gradient

3.

sequence

Schematic diagram used in these studies.

of the pulse

The radiation was applied during the delays when no imaging gradients or data acquisition was occurling ( = 43 msec). In single-section experiments, this resulted in a duty cycle in excess of 90%. With multisection sequences, depending on the heart rate and number of sections, the duty cycle was 70%-85%. The diagram is not drawn to scale.

channel

en-

not be ex-

effective MTC phantom because vides a reasonable model of most

it pro-

tissues

that exhibit a strong MTC effect (i3,15). Gradient-recalled-echo images were then collected with repetition-time (TR) and TE conditions nearly identical to those used

with the human subjects, with varying levels of off-resonance irradiation power. A specific decrease in signal sistent with the magnetization

effect plot

is observed

in the agar.

of the signal

intensity

intensity transfer

con-

Figure

1 is a

of agar

versus

that of saline (both measured 1 cm from the coil) as a function of maximal power deposition

at the surface

calculated

as dis-

ceeded.

cussed

To evaluate the power dependence of MTC by using a surface coil, a 5% agar and 75-mmoVL NaC1 (saline) phantom

power dependence of the off-resonance effect saturates near the 8 W/kg limit

was

mal effects can be obtained level with a surface coil.

used

that

a

of 1,000/13.

rate (SAR)

paths that are only approximations of the true ones in the sample, the SAR analysis

actenistics

of

imaging

-24 misc

Figure

the coil

the coil 1 cm (normal

was then estimated from Peak SAR = (P/F,,)/v,, . power by induced currents the actual current paths, Bi field. Thus, because it

sue around

surface

was a gradient-recalled sequence

Second

=

Q values obtained annular and planar phantoms filled 75 mmol/l saline. The power deposwithin the 1-cm-thick plane closest to coil as a fraction, f,,,, of total power, P1.

The peak

10

50 W of constant-

of peak

made

8

(cm)

RF.

was determined and unloaded

ments

6

from

currents anulus deposited

lus of known of the power

with

were

position

63.8-MHz

on alternate sides of the circuit board to generate a distributed series capacitance in the coil. This was supplemented by eight equally spaced fixed capacitors (88

Cazenovia,

no hot spots

the coil driven

plane loaded

variable

coil magnetic field (Bi). The liquidand resistive-paper method of

(ie, normal

ries capacitance was extensively distributed (20). The copper plating was etched

ing and matching

4

Position

and Chen (22) was used to detect capacitor hot spots. With this ap-

any

itself was made of double-sided, copperplated Teflon circuit board (Rofgens Duroid RT, Chandler, Ariz). To minimize dielectric coupling to the body, the circuit board was placed on edge to minimize contact with the surface (i9), and the se-

NY) to bring cint into the

is for

electric fields that are greatest near the capacitors in the coil, and from the currents induced in the sample by the

surface crystal

constructed

Laboratories,

The curve

servative

(20 x 30 x 1.5 cm). The coil

pF) (Dielectric

irradiation

msec).

proach, studies

of off-resonance

of the radiation.

msec(TE

METHODS

in these

Effect

agar and saline I cm from the coil. The imaging protocol was a gradient-recalled echo with a pulse sequence of 1,000/13 (TR

All experiments were conducted with healthy group composed of three male and four female volunteers. MR imaging protocols were performed with a Signa system (GE Medical Systems, Milwaukee) operating

1.

power on proton signal intensity from a phantom of agar (6%) and 75 mmoVL saline solution. Ms agar/Ms saline = the ratio of the agar and saline water proton intensity in the presence

reported.

MATERIALS

2

20 Pows,(W/kg)

biochemical

of the heart

Ms(saline)Ms(agar)

z

tion has been shown to be specifically dependent on the concentration, conof macromolecules

Mo(saiinc)-Mo(agar)

-....-. 0

of

a good

cavity versus myocardium contrast short-TE MR images. Furthermore, since the mechanism of MTC genera-

relation

___

at

mimicked

of a human

the loading

chest.

char-

Agar is an

cussed

above.

above,

As seen

in this

indicating

that

figure,

near

the dismaxi-

at this power

September

1991

RESULTS Representative single-section, gradient-necalled-echo images are in Figure 4. Figure 4b is a 3-mm-section control image collected with a 24-cm field of view. Eight acquisitions were used per phase-encode step. The TR

was

approximately

TE was

750 msec,

15 msec.

The

was approximately image in Figure

second

channel

and

the

4.

Heart of a healthy man aged 30 years in single-section, gradient-recalled echo MR images. (a) MTC image with 8 W/kg (32 W total power) applied. (b) Control image with no off-resonance irradiation. Imaging parameters were = 750/15 pulse sequence, 256 x 128 acquisition matrix, eight excitations, 24-cm field of view, and a 3-mm section thickness.

as examples

normal-receive

of short-TE

diac gating was standard optical

The off-resonance interpulse separate

the

Car-

by using gating

irradiation

field-gradient

the device.

during

delay was generated RF channel that was

during

stangraused

sequences.

obtained peripheral

the

with a gated off

switches

and

data-acquisition periods (Fig 3). This gating was accomplished by using the Z-gra-

dient

current

monitor

43 msec

for

ing,

data

and

(Signa

3.0; GE Medi-

or the oscilloscope trigger 4.0) as a trigger to turn off the radiation for approximately

cal Systems) pulse (Signa off-resonance

section

selection,

phase

acquisition.

encod-

In a typical

sin-

gle-section duty cycle

gated study, this resulted in a of approximately 90%, and multisection images varied from 75% to 85%. A 4-kHz off-resonance irradiation was

used

effective with

in these

studies

saturation

minimal

the bulk

direct

water

resonance

irradiation

observed

phantoms

or in tissue

with

transfer diagram

in this by

the

coil.

To

evaluate

the

of this effect, the influence the contrast-to-noise ratio

and agar was assessed. study

are presented

than

saturating

used

to magnify

the Bi field. decreased function

at this

The results

RF power the

spatial

As predicted, as the

B1 field

of distance power

level.

the MTC decreased

from The

of this

msec

of this

of

effect as a

the surface increase

coil

in con-

trast-to-noise ratio, however, was observed even 8 cm from the coil even with 6-W/kg irradiation. It is also important that

the second-channel increase the noise ies because of the

Volume

180

irradiation did in these phantom isolation between

Number

#{149}

3

not studthe

is available

effect

triggered from

of

sec-

coil used the

authors

on

of placing

the

the surface coil. Gradiimages were gated to

(diastole)

after

the

peak

blood

applications

fin100 con-

and

over a variety of time points in the multisection studies. Images were collected with and without MTC for comparison.

Data were intensity control

processed

by using

the mean

of a region of interest in both the and MTC images. Identical pixel

values were evaluated for fat, skeletal muscle, left ventricular chamber, and septum. Statistical comparisons were made with paired t tests, with each heart serving as its own control. All data are presented as mean

contrast ventricular wall was radiation

signal an

± standard

error

of the mean.

the

blood by the

between

The by

was essenradiation,

the blood

chamber markedly also does

intensity

in the

myocardium signal intensity

and

the

in the

and the increased. not affect

(13-15),

increase

heart The the fat

resulting

contrast

in

between

fat. The drop-off and contrast at the

in

posterior wall of the heart is due to the limited penetration of the surface coil used for transmission and reception in this study. None of the subjects reported any heating during the procedure, as predicted with our esti-

mates

of power

deposition

discussed

above. A summary of the six subjects studied with the parameters used in Figure 4 is presented in Tables 1 and 2. No significant difference was ob-

between

the

male

and

female

volunteers; thus, the results were pooled. Table 1 presents the signal intensity in the different regions of interest as a ratio of the signal intensity in the presence of radiation (Ms)

to that

in the

control

This ratio provides titating the magnitude

generated consisted

tent in single-section

was

dependence

little

(ie, fat). A complete

the peripheral blood content (optical ger probe), with the signal collected

2. A lower

(6 W/kg)

of off-

in saline

and the surface

protocol

subject prone ent-recalled-echo

magnitude

effects

request.

The

of position on between saline

in Figure

study

on

(16). This was

resonance

13 minutes. 4a was collected

proach. Since tially unaffected

served

effects

of any

ond RF channel

from

irradiation

by the lack

schematic

an

macromolecules,

confirmed

magnetization

The B1 field and receptivity of the surface coil are also a function of distance

to provide

of the

the time

using identical conditions with an 8 Wfkg irradiation (90% duty cycle) 4 kHz off resonance from the bulk waten resonance. As seen in these images, the signal intensities of heart wall and skeletal muscle were selectively decreased with the MTC ap-

pathway in the imager. For the cardiac imaging protocols, dard single-section and multisection dient-recalled echo sequences were Figure

and

acquisition

tion. heart duced while were Table

by the

situation

(Mo).

a method of the

of quaneffect

off-resonance

to-noise

data

ventricular

for myocardium chamber

and

versus fat in the control ages. MTC significantly contrast

the

radia-

The intensity of skeletal and muscle was significantly rein all cases ( - 60% reduction), the ventricular cavity and fat not affected signfficantly (see 1). Table 2 presents the contrast-

between

subjects

MTC section

images

and

im-

studied.

from are

MTC

increased the tissues in all of

was also evaluated gradient-recalled-echo

Examples study

these

vensus myocardium

this

presented

were

obtained

in multiimages.

multisection in Figure

with

5. The

a 923/5

Radiology

673

#{149}

pulse sequence, 5-cm section thickness, 256 x 128 acquisition matrix, eight excitations, and a 24-cm field of view. The total imaging time was about 16 minutes, and the secondchannel irradiation had a duty cycle of 82%. As expected, improvements in contrast obtained in these studies were similar to those seen in the singte-section studies. DISCUSSION These

data

can improve ventricular

demonstrate

that

MTC

the contrast between wall and cavity in short-

TE images of the human heart. This contrast is generated during the interpulse delay period and thus does not compromise the high spatial nesolution of these short-TE sequences. Indeed, MTC could benefit any short-TE sequence, including echo-planar-imaging, spin-echo, or projection-reconstruction techniques, in which the magnetization is conditioned before the rapid imaging sequence by the off-resonance radiation generating the MTC effect. Since MTC generally is simitan

to T2 signal

intensity

in

many tissues, the addition of MTC to T2-weighted sequences may also provide a contrast-to-noise advantage. With regard to multisection or cine methods, irradiation periods as short as 50 msec in an imaging sequence have been shown to be effective in saturating the macromolecules and result in significant increases in contrast

in brain

tissue

(14).

These

in-

creases are due to the rather long Ti of the macromolecules, which permits the interpulse delay radiation to accumulate an effect until a steady state between relaxation and saturation is achieved in the macnomolecule. In oun multisection studies, we have found that by maintaining the duty cycle of the irradiation at greater than 50%,

significant

MTC

effects

can

be

realized. These studies were conducted with a surface coil for both transmission and reception. Thus, the images obtamed do not cover the entire heart as do the more common body-coil studies.

This

difference

is due

to the

tim-

ited transmission and reception characteristics of the surface coil used, which roughly limit the field of view to the radius of the coil. A surface coil was

used

vide

images

ratios

in this

of signal

initial

of the

heart

to noise

study

with and

to pro-

high to limit

the RF power deposition to the volume of interest. Adequate MTC was achieved with a nominal 8 W/kg with an approximately 90% duty cycle. 674

Radiology

#{149}

a.

b.

Figure

5.

Multisection, gradient-recalled echo images of the heart of a healthy woman aged 24 years. (a) Control images obtained with no off-resonance irradiation. (b) MTC images obtamed with 8 W/kg applied 4 kHz off-resonance. Imaging parameters were = 923/5, 256 x 128 acquisition matrix, eight excitations with four sections, 5-mm section thickness, and a 24-cm field of view.

This power level coincides with that of the phantom study shown in Figune 1, in which the MTC effect was realized at a surface power level of about 8 W/kg. Due to the exponential nature of the power curve, which fits theoretic predictions (16), even much lower power levels result in significant improvements in contrast. Thus, application of this method to large body coils with which larger volumes of tissue are irradiated may be feasible. The improvement in contrast between the myocardium chambers as well as fat

and heart occurs because

of the selective from the heart

decrease in signal wall. This decrease

curs

of an

by

means

between bulk water romoleculan protons (13). This magnetization cess does not occur blood on fat tissues,

effect

of off-resonance

these

tissues.

The

effective

oc-

coupling

protons in heart

and macmuscle

transfer prosignificantly in the resulting in no

irradiation high

on

contrast

be-

of magnetization water and macro-

molecular

protons

in heart

unknown.

Studies

of various

molecules, especially have demonstrated

muscle

try of the tant

in

of the

chemis-

are importhe

(25).

effectiveness

These

of the macromolecular In addition, exposed on

studies with

struchydnoxyl

the surface of the macrois one of the prominent groups capable of catalyzing

the magnetization with water (i8).

The

exchange process magnetization

transfer effect has also been found to be consistent with a spin diffusion mechanism between the macromolecule and water protons (25). With specific regard to heart muscle, the offresonance irradiation effect on water

of the

protons

cle were similarity filaments cess,

heart

event,

since

interaction

and water, characterization

under

and

skeletal

mus-

similar (see Table 1). This suggests that the rigid myomay be involved in this prothis has yet to be confirmed.

but

In any

this of

it may

various

including thies.

is a

prove useful in the of cardiac muscle

pathologic

ischemia

process macromolecules

conditions,

or myopa-

U

References 1.

is

lipid bilayers, that both the con-

surface

that the effect increases “rigidity” (ie, correlation

groups molecule chemical

2.

macro-

the

determining

revealed increasing

time) tune.

and

macromolecules

coupling

specific

tween cardiac wall and blood should be quite useful in functional imaging of the heart, while the contrast between fat and myocardium should provide a good measure of cardiac fat (24). The most interesting aspect of MTC in the heart is that it may provide a more reproducible and even more accurate measurement of water-relaxation processes. The actual mechanism on chemistry transfer between

time

relation

3.

de Roos A, van Voorthuisen AE. Magnetic resonance imaging of the heart: morphology and function. Curr Opin Radiol 1989; 1:166-173. Higgins CB, Holt W, Pflugfelder P, Sechtem U. Functional evaluation of the heart with magnetic resonance imaging. Magn Reson Med 1988; 6:121-139. Spritzer CE, Herfkens RJ. Magnetic resonance imaging of the heart. Magn Reson Annu 1988; 1:217-244.

September

1991

4.

5.

6.

7.

8.

9.

10.

Ehman nance

RL,Julsrud P. Magnetic resoimaging of the heart: current status. Mayo Clin Proc 1989; 64:1134-1146. Axel L, Dougherty L. Heart wall motion:

11.

improved

12.

method

of spatial

modulation of for MR imaging. Radiology

magnetization 1989; 172:349-350. Zerhouni EA, Parish DM, Rogers WJ, Yang A, Shapiro EP. Human heart: tagging with MR imaging-a method for noninvasive assessment of myocardial motion. Radiology 1988; 169:59-63. Ahmad M, Johnson RFJr, Fawcett HD, Schreiber MH. Magnetic resonance imaging in patients with unstable angina: comparison with acute myocardial infarction and normals. Magn Reson Imaging 1988; 6:527-534. Tscholakoff D, Aherne T, Yee ES, Derugin N, Higgins CB. Cardiac transplantations in dogs: evaluation with MR. Radiology 1986; 157:697-702. Lund C, Morin RL, Olivari MT. Ring WS. Serial myocardial T2 relaxation time measurements in normal subjects and heart transplant recipients. J Heart Transplant 1988; 7:274-279. Kurland RJ, WestJ, Kelley S, et al. Magnetic resonance imaging to detect heart transplant rejection: sensitivity and specificity. Transplant Proc 1989; 21:2537-2543.

Volume

180

Number

#{149}

3

13.

14.

15.

16.

17.

18.

FrahmJ, Merboldt KD, Brun H, Gyngell ML, Hanicke W, Chien D. 0.3 second FLASH MRI of the human heart. Magn Res Med 1990; 13:150-157. KatzJ, Boxt LM, Sciacca RR, Cannon PJ. Motion dependence of myocardial transverse relaxation time in magnetic resonance imaging. Magn Reson Imaging 1990; 8:449-458. Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989; 10:135-144. Wolff SD, EngJ, Balaban RS. Magnetization transfer contrast: method for improv-

20.

ing contrast

22.

in gradient-recalled-echo

im-

ages. Radiology 1991; 179:133-137. Wolff SD, Chesnick 5, FrankJA, Lim KO, Balaban RS. Magnetization transfer contrast: MR imaging of the knee. Radiology 1991; 179:623-628. EngJ, Ceckler U, Balaban RS. Quantitative 1H magnetization transfer imaging in vivo. Magn Res Med 1991; 17:304-314. Fralix TA, Ceckler U, Wolff SD, Simon SA, Balaban RS. Lipid bilayer and water proton magnetization transfer: effect of cholesterol. Magn Reson Med 1991; 18:214-223. Wolff SD, Fralix TA, Simon SA, Balaban RS. Magnetization transfer spectroscopy of model systems: a probe for the molecular basis of tissue contrast in MRI (abstr). In: Book of abstracts: Society of Magnetic Res-

19.

21.

23.

24.

25.

onance in Medicine 1990. Berkeley, Calif: Society of Magnetic Resonance in Medicine, 1990; 350. Balaban RS, Koretsky AP, Katz LA. Loading characteristics of surface coils constructed from wire and foil. J Magn Reson 1986; 68:556-560. DeCorps M, Blondet P, Reutenaur H, AlbrandJP, Remy C. An inductively coupled series-tuned NMR probe. J Magn Reson 1985; 65:100-109. Wilson MJ, ed. American Radio Relay League handbook Vol 64. Newington, Conn: American Radio Relay League, 1989; 16-7, 16-8. Hoult DI, Chen C-N. The visualization of probe electric fields reveal intense “hotspots” (abstr). In: Book of abstracts: Society of Magnetic Resonance in Medicine, 1988. Berkeley, Calif: Society of Magnetic Resonance in Medicine, 1988; 872. Hedges LK. A fuse for magnetic resonance imaging probes. Magn Reson Med 1989; 9:278-281. KriegahauserJS,Julsrud P, Lund if. MR imaging of fat in and around the heart. Am J Radiol 1990; 155:271-274. Ceckler TL, Balaban RS. Tritium-proton magnetization transfer as a probe of cross relaxation in aqueous lipid bilayer suspensions. J Magn Reson 1991; 93:572-588.

Radiology

675

#{149}