Cardiac Markus Jungehulsing, Hans Hermann Hilger,
MD MD
#{149} Udo
Sechtem, MD #{149} Peter Schicha, MD
Thrombi: Gradient-Echo
Gradient-echo (GRE) and spin-echo (SE) magnetic resonance (MR) imaging was performed in 3i patients with chronic left ventricular (LV) thrombi. Thrombi were confirmed or excluded at surgery or by means of other corroborative diagnostic techniques. MR images were evaluated by three reviewers without knowledge of results of corroborative studies. Diagnoses were graded unequivocal if agreed on by three observers and probable if agreed on by two observers. With SE imaging, i2 of iS confirmed thrombi were detected unequivocally, five were considered probable, and one was not detected. With GRE imaging, i6 of the 18 thrombi were visualized unequivocally; two were considered probable. With SE technique, thrombus was unequivocally excluded in nine of 13 cases and exclusion was considered probable in four. One finding was false-negative. Exclusion of thrombus with GRE imaging was unequivocal in 10 of 13 cases and probable in two, and one finding of thrombus was false-positive. GRE imaging resulted in improved differentiation of thrombi from the surrounding blood pool and myocardium and thus was diagnostically superior to SE imaging in detection of LV thrombi. Index terms: Aneurysm, cardiac, 524.773, 524.775 #{149} Heart, MR. 524.1214 #{149} Heart, thrombosis, 524.731, 524.775 #{149} Magnetic resonance, comparative studies #{149} Thrombosis, MR. 524.1214 1992;
MD
#{149} Harald
Left Ventricular Spin-Echo and
Radiology
Theissen,
182:225-229
I
Evaluation with MR Imaging’
patients with left ventricular (LV) aneurysms, embolism is a complication that causes considerable morbidity (1,2). Detection of thrombi in these patients has important consequences; immediate institution of anticoagulant therapy may reduce the possibility of systemic embolism (3,4). Methods such as LV angiography, scintigraphic techniques, computed tomography (CT), and two-dimensionab echocardiography (7) arc used clinically to detect thrombi (4-7). The advantages that two-dimensional echocandiognaphy has over the other mentioned techniques are portability, unlimited imaging planes, bow cost, and noninvasiveness, combined with reasonable sensitivity and specificity. Another noninvasive technique for detecting LV thrombi is magnetic nesonance (MR) imaging. The value of spin-echo (SE) MR imaging in the dctection of LV thrombi has recently been shown (8-10). Gradient-echo (GRE) MR imaging of the heart has added important dynamic information to the morphologic detail demonstrated by SE MR imaging (11-13). GRE MR imaging also is capable of enhancing the signal intensity of flowing blood and thus enhancing contrast between blood and myocandium, especially in patients with slow intraventricular blood flow (14,15). Therefore, the purposes of this study were to evaluate the appearance of LV thrombi on GRE images and to compare the diagnostic value of SE and GRE MR imaging for the diagnosis or exclusion of LV thrombi in pa-
PATIENTS
N
tients
with
LV
aneurysms.
AND
Thirty-one
women,
patients (22 men and nine years old) with chronic underwent SE and GRE
36-77
LV aneurysms
MR imaging.
Thrombi
18 patients
and
cepted M.J. (:
the Clinic and Outpatient Medicine Received
July 15. Supported
RSNA,
1992
Clinic
for Nuclear
Medicine
(U.S., H.H.H.), Universit#{228}t zu KOln, January 29, 1991; revision requested
by the Deutsche
(M.J., PT., H.S.) and the Third
Address
reprint
in
requests
confirmed
in
13 patients
at
on by techniques.
use of Sungi-
cal confirmation was available in 13 patients. Confirmation or exclusion in all other patients was accomplished by con-
roborative diagnosis with the following techniques: sional
at least two of two-dimen-
echocardiography
(n = 5), and (Table).
MR
Imaging
Proton
(ti
(n
CT 15)
=
Technique
MR
imaging
was
with a superconducting at 1.5 T (Gyroscan; erbands).
18),
=
LV angiography
SE MR
performed
magnet Philips, Best,
imaging
was
operating The Neth-
performed
with electrocardiographic gating and two echoes with echo times (TEs) of 30 and 60 msec. Repetition time (TR) ranged from 470 rate
to 1,060 msec, of the patient.
depending The heart
on the heart was imaged
from the LV apex to the aortic root in transverse sections with 8-mm thickness and 0.8-mm gap. The image matrix was
128 x
256
pixels.
To
improve
visualization
of the intracardiac anatomy and to simplify differentiation of flow effects that may occur only in one imaging plane, obbique sagittal sections with 8-mm section thickness were interventricular
performed
obtained septum.
parallel Angulation
by electronic
to the was
axis rotation.
Spa-
tial presaturation was not available at the time the study was conducted and could therefore not be used. GRE images were acquired by using prospective electrocardiographic and TRs of 26-30 msec with angle
compensation
was of blood for
gating TEs of 13
40#{176}. To signal
constant
improve intensity, velocities
Clinic
Joseph Stelzmann Strasse 9, 5000 Cologne March Ii; revision received July 10; ad-
Fonschungsgemeinschaft.
were
excluded
surgery (aneurysmectomy) corroborative imaging
flow
From
METHODS
Patients
msec. The flip the homogeneity
for Internal 41, Germany.
Radiology
to
Abbreviations: left ventricular, spin echo, TE time.
=
GRE = gradient echo, LV = ROI = region of interest, SE = echo time, TR = repetition
was employed as described previously (16). Section thickness of transverse 5cclions
was
8 mm
with
a 1.6-mm
gap.
The
heart was imaged from the LV apex to the base. Sagittal sections were obtained paralbcb to the intervcntricular septum as with the SE technique.
Image
Interpretation
SE and
GRE
images
were
interpreted
by
three independent observers who had no knowledge of the results of corroborative studies.
GRE
images
were
reviewed
in a
movie loop that could be stopped to allow closer observation and measurement of details on still frames. MR imaging findings were graded unequivocal if the three independent observers made the same diagnosis.
If only
two
observers
agreed,
the diagnosis was regarded as being only probable. Thus, there were four diagnostic categories: unequivocal thrombus, probablc thrombus, probably no thrombus, and unequivocally no thrombus. To characterize the contrast among blood, thrombus, and myocardium, signal intensity of thrombus, blood, myocardium, and cpicardial fat was measured in each patient by using regions of interest (ROIs) on transverse sections. ROI area was at beast 10 pixels, and ROIs were positioned in close proximity to thrombus. Signal intensity
was
measured
on
ob-
SE images
tamed with either TE and on GRE images obtained at systobc and diastobe in all patients
in whom
thrombus
cabby diagnosed aging
(n
=
with
12).
Signal
related to fat because dependence of signal terval.
Statistical
was
unequivo-
SE and GRE MR imintensities
were
of the well-known intensity
on
TR
in-
of parts of the thrombus relative to the myocandial wall (17). On second-echo images, slowly flowing blood within the LV ancunysm had higher signal intensity than did thrombus and myocardium in 10 of ii patients with recognizable thrombus. In one patient, flow signal intensity prevented visualization of a thrombus (Fig 3). The presence of sunrounding high-intensity slow blood flow usually facilitated the diffcnentiation between blood and thrombus. Signal intensity of thrombus was darker than that of flowing blood in eight patients. Mean signal intensity of thrombus was significantly lower than that of blood (P = .01) (Fig 4). Three thrombi, however, were brighter than surrounding blood, which was probably because of a nebatively high flow velocity and low blood flow signal intensity as a result of the smaller myocandiab lesions in these patients. Differentiation between thrombus and myocardium was more difficult. Seven thrombi were brighter than myocardium and four were slightly darker, and there was no constant signal intensity relation (P = .29, paired t test). In three patients, diag-
techniques
nosis
in five
patients.
With
both
techniques, diagnosis was false in one of these five. One patient had a false diagnosis of probable thrombus excbusion with SE images and a correct unequivocal diagnosis of thrombus with GRE images.
Analysis
Relative signal intensities of thrombus, blood, and myocardium were compared by using the paired t test and the Bonferroni correction for multiple comparisons. The distribution of unequivocal and uncertain diagnoses, as well as of the correct and false diagnoses with SE and GRE images, was compared by using the x2 test.
RESULTS Diagnostic GRE Images
Accuracy
of SE and
With SE MR imaging, unequivocal diagnosis of thrombus or absence of thrombus was possible in 21 of 31 cases (Table). In five cases, thrombus was considered probable; in another five cases, thrombus was not probable. Both of these categories included an incorrect diagnosis: one was falsepositive and one was false-negative. For the total of 93 diagnoses made by the three reviewers, SE image interpretation was correct in 81 (87%) and incorrect in 12 (13%). The GRE movie 226
allowed better visual distinction of thrombi from moving blood than did comparison of still SE images. GRE images were preferred for making the diagnosis in 87 of 93 paired observations. In 26 of 31 patients, GRE MR imaging results were unequivocal, and only five of 31 were uncertain (vs SE technique, x2 = 3.7 and P = .05). Among these five, one false-positive diagnosis was made in a patient with a large anterior aneurysm. SE and GRE image quality in this case was decreased because of prominent breathing artifacts. Eighty-seven (94%) of the individual diagnoses made by means of GRE image interpretation were correct, and six (6%) were not. The number of false diagnoses per observer was significantly higher with SE images than with GRE images (x2 = 6.4, P = .01). No diagnostic errors occurred in the 19 patients in whom all three observers formed the same diagnosis from SE or GRE MR images (Table). GRE diagnosis was unequivocal in six patients that had a probable diagnosis with use of SE images (three with and three without thrombus). Diagnosis was considered probable with both
#{149} Radiology
Appearance Images
of Thrombi
of thrombus
was
possible
only
when information from oblique sagittab sections was added. In generab, differences in signal intensities between the measured structures on SE images were slight and varied in direction; therefore, no constant signal intensity relation existed on SE images.
on SE
On first-echo SE images, differentiation between thrombus and myocardium was difficult because slightly higher signal intensities (n = 7) of thrombi were encountered as often as were slightly bower ones (n = 5) (Fig 1). Although the differences in signal intensity between thrombi and blood were also slight, mean signal intensity of thrombus was slightly higher (P = 0.05, paired t test) (Fig 2). In three patients, thrombi had lower signal intensity than did blood. All of these patients showed large aneurysms of the left ventricle, and signal intensity features were probably due to prominent slow blood flow and back of complete signal extinction of the blood. Differentiation from blood in these cases was possible only because a small line of flow signal intensity mdicated the presence of an additional structure; this signal intensity was probably due to slight movement
Appearance Images
of Thrombi
on GRE
On GRE images, blood had higher signal intensity than did all other intracardiac structures on systolic and diastolic images, and thrombus abways showed a bower signal intensity than did blood (P < .01) (Fig 5). The contrast between thrombus and blood was therefore consistent in all cases, which facilitated the identification of the 16 thrombi detected on GRE images. Of the 16 thrombi, ii had signal intensity bower than that of myocardium, three had nearby the same signab intensity, and two had signal intensity intermediate between those of blood and myocardium. In general, thrombus was the structure with the lowest signal intensity (P < .02, Bonfcrroni t test). LV aneurysms were more clearly defined on GRE images, due to the absence of wall thickening of the January
1992
Findings
and GRE MR Imaging
of SE
and Corroborative
Studies
in 31 Patients
with
LV Infarction
Chronic
MR Imaging Diagnosis of Thrombus Patient
No.
GRE
I 2 3 4 5 6
SE
Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal
7
8 9 10 11
Location Thrombus Confirmed
Exclusion Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal Unequivocal
probable*
of Thrombus in Myocardium
x x x x x x x x x x
Echo, Surgery Surgery Surgery Surgery Surgery Surgery Echo, Echo, Echo,
x
Echo, angio
angio
angio, CT anglo
Unequivocal
Unequivocal
Unequivocal
Unequivocal
x x
Echo, Echo,
14
Unequivocal
Probable
x
15
Unequivocal
Probable
x
16 17 18 19 20 21 22 23 24 25 26
Unequivocal Exclusion unequivocal
Unequivocal Exclusion unequivocal
Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Probable Probable Probable Exclusion Exclusion
Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Exclusion Probable Probable Probable Exclusion Exclusion
x
Surgery Echo, Echo, Echo, Echo, Surgery Echo, Echo, Echo, Echo, Surgery Surgery Surgery Echo, Surgery
x
Echo, angio
27 29
30 31 Note.-Angio * The thrombus t False diagnosis
thinned, dium.
=
unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal
probable probable
the
probable probable unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal unequivocal
diagnostic
myocandid not
yield.
sionab flowing
The
sometimes
comparison of signal intensity quotients of thrombus versus fat, blood versus fat, and myocardium versus
fat
be confused
(21). In this improved
from the surrounding myocardium, because tween thrombus and
creased
tunes
significantly
on
GRE
images
was
all thrombi than that
DISCUSSION
for
nique,
another
noninvasive
because
studies
method for cardiac echocanthere is a
large
revealed
quality
MR images of thrombi
and bestnuc-
enhanced.
Sig-
was constant,
and
lower blood
to certain bus. For thrombi from
Volume
by
patients
relation. isointense
signal
even
characteristics
anatomic (18,19)
of the throm-
instance, laminated mural may be difficult to distinguish
myocardium
182
with
#{149} Number
two-dimen-
1
or
protruding laminar
Lange, Large, Lange, Small,
protruding laminar laminar protruding
Large,
laminar
Small, protruding
was
degraded
by prominent
breathing
Three with
intensity
greater
than
thrombi
with GRE images, because of the for the cine display of cardiac
motion
and
blood
flow.
In one
of that
two the
tion
between
flow
diac structures.
artifacts
Signal
was
five
patients
high signal-intensity bus. An alternate that these thrombi origin (22). The
certainty
led
to
spuriously
values explanation were of diagnosis
also
papillary
study,
we
have
shortening
used flip
of blood
or throm-
constant diastolic
signal im-
a GRE
flip
angles
of 20#{176} or
accentuated effects
intracanthe carsuch as
muscle
of 40#{176}. Shallower
the (15)
of the
T2*
anand
para-
magnetic compounds contained in the thrombus and further increased contnast between blood and thrombus. SE
for thromwould be
of more
inferior
30#{176} might 12
and
intensity
changed slightly during cycle, whereas structures
In our
diamag-
a
observers felt more comfortable making a diagnosis with GRE images, because the cine display simplified differentia-
gle
thrombi
of
patient,
freely movable thrombus was clearly depicted on GRE MR images, but was not diagnosed on SE images. In general,
were
myocardium,
artifacts.
proved capability
the
these
caused
Large, Large,
wall
flow diac
echocandiograms
the
sep-
CT
netic myocardium. It cannot be cxcluded that volume averaging of signal intensity of blood and thrombus in
of
Small, protruding
Apex
false-positive, or equivocal findings are not uncommon. Those findings may be attributable to technically inadequate abnormalities
sep-
bus maintained a nearly intensity on systolic and ages.
and
false-negative,
intenventricular
Comparison between the signal intensity of myocardium and that of thrombus on GRE images showed no
nearly
echocardiognaphic
that
Ant
thrombus
signal intensity surrounding
and lat wall wall wall
pool.
constant
tech-
had of the
and lat wall; apex and ant wall
protruding protruding laminar laminar protruding protruding laminar laminar laminar laminar
angio
blood pool the contrast surrounding
relation
Small, Small, Large, Large, Small, Small, Large, Large, Large, Lange,
angio angio angio CT
and slowmay
with
considerably
nab intensity
The primary screening patients with a suspected thrombus is two-dimensional diognaphy. Nevertheless,
CT angio angio angio
Image
study, use of GRE the differentiation
on SE and GRE MR images revealed that the difference between the signal intensity of measured structures in(Fig 6).
angio angio
echocardiography (20), blood in an aneurysm
wall
Apex;
Echo, angio, Surgery
probable probable
angiography, ant = anterior, echo = echocardiography, lat = lateral. was moving and therefore not detected with SE MR imaging. of large aneurysm of the anterior wall was made with both techniques.
infarcted anterior Sagittab GRE images
improve
x
CT
Ant Apex Ant Apex Apex Apex Ant Ant Ant Apex
Features of Thrombus
tum Ant wall Apex; interventnicular tum Apex; ant wall Apex Apex
12 13
28
need
Corroboration Techniques
MR
imaging
has
good
and specificity in detection thnombi (8). Nevertheless,
recent
image interpretation this method include
im-
largely
differing
values
sensitivity
of LV difficulties
in
encountered with motion artifacts, for
blood
Radiology
signal
#{149} 227
I
I
1-
I
P 0,05
-j--------
P
S
I
a. Figure
b.
1. 30) obtained
(a) Diagram in patient
thrombus
is lower
blood,
which
blood
appears
tnicular ventricle,
than
has
that
higher
intensity of cardiac features on (b) transverse SE section had LV aneurysm and laminar thrombus. Signal intensity
of myocandium
signal
=
echo
VCI
time,
and
intensity.
slightly darken than LV = left ventricle,
septum,
TE
of signal i5, who
=
The
does RA inferior
can be distinguished
remaining
thrombus. =
right
vena
from
intraventnculan,
AD
atrium,
descending ROIS = region =
slowly
flowing
normally
aorta,
IVS
of interest,
(740/ of
flowing intenven-
=
RV
=
Figure
3. (a) Diagram of signal intensity of cardiac features on (b) transverse section (740/60) obtained in the same patient as in Figure 1. The LV thrombus identified because of prominent flow artifacts. Signal intensity of thrombus not be measured. See Figure i for key to abbreviations.
right
I Ft
second-echo SE could not be could therefore
and during
ratio
more, there
resulting from a diminished the second
as demonstrated is no constant
the thrombus ond echoes.
There
blood flow yesignal-to-noise echo. Further-
in this study, signal intensity of
during
the
first
and
5cc-
T2-weightcd plexes such gbobin,
are several
potential
explana-
sections,
and
it was
therefore
always possible to measure dbc section, accomplishment would have avoided partial
ron. On the other
hand,
not
in the midof which volume en-
partial
volume
and sometimes calcifiboss of protons leads to intensity on Ti- and
MR images. as hemoglobin,
dcoxyhcmogbobin,
or fenritin
tions for the varying signal intensities of thrombus. First, partial volume effects from averaging thrombus and blood in the same section cannot be excluded. Thrombi often did not extend oven
three
iron complexes, cation (23). The decreased signal
show
Iron commethemohemosidenin,
different
magnetic
intensity
on T2-wcightcd A T2-wcighted
intensity
of thrombi
images
and signal images (24-27). sequence can be
organization.
un-
thus
as loss
tween blood.
of water, 228
An
aging
condensation
#{149} Radiology
changes
thrombus
such
of panamagnetic
with
in increased
achieved by gating to every second, third, or fourth heartbeat and might have accentuated the T2 shortening fect of the paramagnetic compounds
metabolic
thrombi,
however, therefore quence.
signal
boss
enhanced
thrombus and It is conceivable
contrast
be-
surrounding that the
more
because
used in this intensity of of some
undesin-
the
standard
GRE
and longer not employed Furthermore, would
have
sequence,
TR intervals were for the SE severy long TR inincreased
imaging
time considerably. Some thrombi were isointense myocardium or had even higher than
which was the infarct sult
the thrombotic material and reduced its signal intensity, whereas the echo rephasing effects would have intensifled the signal of slow blood flow and
possibly
tenvals
gbobin
efof
SE sequence to higher signal
able Ti shortening effects that made the distinction of thrombus from blood more difficult. The study was designed to evaluate the diagnostic potential of the standard SE sequence in companison
effects that may have contributed to diagnostic problems with use of SE mmages did not substantially interfere with diagnosis with GRE images, in which contrast between thnombi and the sunrounding blood pool was constant. Second, thrombotic material differs with respect to its age and degree of dergocs
some
characteristics: Oxyhemoglobin is diamagnetic, deoxyhemogbobin and methemoglobin are paramagnetic, and hemosidenin and fennitin are supenparamagnetic (22). Paramagnetic agents cause Ti and T2 shortening for tissues in which they are located, which results on Ti-weighted
I Ft
flI=d
Figure 4. Comparison (pained t test) of nelative signal intensities on SE MR images (TE, 60 msec) obtained in 11 patients. In general, blood showed higher signal intensity than did myocardium, whereas there was no consistent signal intensity relation between thrombus and cardiac structures.
Ti-weighted study led bocity,
/ F.,t
cava.
Th,=bs
intensity
IId
Ft
Figure 2. Comparison (pained f test) of nelative signal intensities in first-SE MR images (TE, 30 msec) obtained in 12 patients. Signal intensities of thrombus and myocardium were similar. There was no consistent nelationship in signal intensity among the three structures.
that
myocandium,
measured well away region. Paramagnetic
degradation in
of
signal
products loss
within
volume the age and
can
effect of the
be
attributed
of blood thrombi
hemoglobin
from hemo-
would the
which was indeed the case thnombi. The higher signal thnombus
with signal
in most intensity to
of
partial
or to the fact was unknown,
degradation
re-
thrombus,
that
products
could have been present in fresh thrombotic material, but in insufficient amounts to cause severe signal loss of the
thrombus
on
GRE
images.
January
1992
I Ft
Figure
6.
tensities a.
b.
Figure
5.
(a) Diagram
obtained
of signal
in the same
patient
intensity
patients
of cardiac
as in Figures
1 and
features
on
3. Thrombus
(b) transverse
outlines
GRE
are more
due to a marked difference in signal intensity of blood and thrombus. dium have similar signal intensity on this image and can be differentiated blood between them. See Figure 1 for key to abbreviations.
clearly
Thrombus by the
image
angle,
defined
among
MR
and thin
13.
its bow cost and
msec).
GRE sequences
Identification
ages slight sity
of thrombus
may also difference of
tunes
and
lack
relation
and
myocardium.
TE,
among
and
intensity
crabby
on
thrombus, Blood
phase
signal
images.
to vary
With
blood has consistently intensity in the absence which makes differentiation LV
study
indicates
thrombi
with
ated when addition to tients with prominent
GRE nor
to
MR
MR imaging SE
thrombi.
imaging
choice.
SE
omy
may
and
unknown quicker aging
the
left
5.
is used
imaging,
6.
of only
to
site, technique
7.
however,
overview
of cardiac be
anat-
of additional
of the infarcted
ventricle
would
Echocardiognaphy will primary imaging technique patients with LV aneurysms
8.
area
and must be determined than time-consuming GRE of
4.
to be supe-
detection
therefore
use if the extent
of
is facili-
seems
imaging
MR
a fast
1.
3.
are acquired in especially in paflow artifacts and flow.
at a prespecifled seems to be the
provides
imag-
is im9.
allow.
remain the to screen because
10.
of I 1.
Volume
182
#{149} Number
1
be
accurate
its diagnostic
further
enhanced
in patients
intraventricuban
14.
MR imag-
a noninvasive,
capabilby
with thrombus.
F.I
signal
in-
on GRE MR images
obtained
in 11
during
13 msec;
flip
diastole
40#{176}). Differences
structures
(TE,
in signal
intensity
are consistent.
use
Utz JA, Herfkens RJ, Heinsimer JA, et al. Cine MR determination of left ventricular ejection fraction. AJR 1987; 148:839-843. Haase A, Matthaei D, H#{228}nickeW, Merboldt KD. Flash imaging: rapid NMR imaging using low flip-angle pulses. J Magn Reson 1986; 67:258-266.
low-quality
however,
technique; can
for real-time
with
15.
of
susU
References
2.
detection
imaging
for
If MR
detect thrombi GRE imaging of
that
GRE images SE images, prominent slow blood
itics
can cause consid-
GRE
capability
patients
represents
second
higher signal of turbulence, easier.
ing,
Our
ing
pected
im-
velocity,
cycle
of blood
In
echocardiograms,
blood,
flow
of cardiac
SE
SE
be difficult because of the between signal intenand surrounding strucof consistent signal inten-
thrombus
sity
on
imaging.
/
of relative
myocarline of 12.
The third reason for inconsistent signab intensity may be the result of diffenent heart rates causing slightly different Ti weighting of SE images. In contrast, TRs are short and nearly constant for GRE imaging (variation, 25-27
I3I(1
Comparison
Hamby RI, Wisoff BG, Davison ET, et al. Coronary artery disease and left ventricular mural thrombi: clinical, hemodynamic, and angiocardiographic aspects. Chest 1974; 66:488-494. Asingen RW, Mikell FL, Elsperger KJ, Hodges M. Incidence of left ventnculan thrombosis after acute transmunal myocardial infarction: serial evaluation by two-dimensional echocardiography. N EngIJ Med 1981; 305:297-301. Simpson MT. Obenman A, Kouchoukos NT, Rogers WJ. Prevalence of mural thnombi and systemic embolization with left ventricular aneurysm: effect of anticoagulation therapy. Chest 1988; 77:463-469. Reeder CS, Lengyel M, Tajik AJ, et al. Mural thrombus in left ventricular aneurysms: mcidence, role of angiography, and relation between anticoagulation and embolization. Mayo Clin Proc 1981; 56:77-81. Ezekowmtz MD, Wilson DA, Smith EO, et al. Comparison of indium-ill platelet scintignaphy and two-dimensional echocardiograph in the diagnosis of left ventricular thrombi. EngI J Med 1982; 306:1509-1513. Tomoda H, Hoshiai M, Furuya H, et al. Evaluation of intracardiac thrombus with comuted tomography. Am J Candiol 1983; 51:943-
,
Visser CA, Kan C, Lie IC, et al. Left ventniculan thrombus following acute myocardial infarction: a prospective serial echocardiographic study of 96 patients. Eur Heart J 1983; 4:333-337. Sechtem U, Theissen P. Heindel W, et al. Comparison of magnetic resonance imaging, computed tomography, echocandiography and angiography in the diagnosis of left ventriculan thrombi. Am J Cardiol 1989; 64:1195-1199. Dooms CD, Higgins CB. MR imaging of cardiac thrombi. J Comput Assist Tomogr 1986; 10:375-420. Comes As, Lois JF, Child J5, et al. Cardiac tumors and thrombus: evaluation with MR imaging. AIR 1987; 149:895-899. Nayler CL, Firmin DN, Longmore DB. Blood flow imagin by cine magnetic resonance. Comput Assist Tomogr 1986; 11:751-756.
16.
Buxton RB, Edelmann RR, Rosen BR, Wismen CL, Brady TJ. Contrast in rapid MR imaging: TI and T2 weighted imaging.] Comput Assist Tomogr 1987; 11:7-16. Wehnli F. Fast-scan imaging: principles and contrast phenomenology. In: Higgins CB, Hnicak H, eds. Magnetic resonance imaging of the body. New York: Raven, 1987; 23-38. Clover C, Pelec NJ. A rapid-gated cine MRI technique. In: Knessel FLY, ed. Magnetic resonance annual 1988. New York: Raven, 1988; 299-312.
17.
18.
19.
20.
21.
22.
23.
Lund JT, Ehman RL, Julsrud PR, Sinak U, Tajik AJ. Cardiac masses: assessment by MR imaging. AJR 1989; 152:469-473. Stratton JR. Ligthy CW, Pearlman As, et al. Detection of left ventricular thrombus by twodimensional echocandiography: sensitivity, specificity, and causes of uncertainty. Cinculation 1982; 66:156-164. Asingen RW, Shanma B, Hodges M. Observations on detectin left ventricular thrombus with two-dimensional echocardiography emphasis on avoidance of false-positive diagnosis. Am J Cardiol 1981; 47:145-156. Feigenbaum H. Coronary artery disease. In: Feigenbaum H, ed. Echocardiography. 4th ed. Philadelphia: Lea & Febigen, 1986; 489-494. Mikell FL, Asingen RW, Elsperger KJ, Anderson WR, Hodges M. Regional stasis of blood in the dysfunctional left ventricle: echocandiographic detection and differentiation from early thrombus. Circulation 1982; 66:755-763. Cohen MD, McGuine W, Cony DA, et al. MR appearance of blood and blood products: an in vitro study. AJR 1986; 146:1293-1297. Baumgartnen HR. The role of blood flow in platelet adhesion, fibrin deposition, and formation of mural thrombi. Micnovasc Rev 1973; 5:167-179.
24.
25.
26.
27.
Bradley WC, Schmidt PC. Effect of methemoglobin formation on the MR appearance of subanachnoid hemorrhage. Radiology 1985; 156:99-103. Saini 5, Frankel RB, Stark DD, Fenrucci JT. Magnetism: a primer and review. AJR 1988; 150:735-743. Renshaw PF, Owen CS, McLaughlin AC, Frey TC, Leigh J5. Ferromagnetic contrast agents: a new approach. Magn Res Med 1986; 3:217225. Rapoport 5, 5ostman HD, Pope C, Camputaro CM, Holcomb W, Core JC. Venous clots: evaluation with MR Imaging. Radiology 1987; 162:527-530.
Radiology
#{149} 229