Document NG4ddp8g6GReV5M6bGrnvQgq8
Musculoskeletal
Lia A. Moulopoulos,
MD #{149D} atla G. K Varma,
Norman E. Leeds, MD #{149E} . Edmund Kim, MD
Raymond Alexanian,
MD #{149H}erman
I. Libshitz,
MD #{149M} eletios A. Dimopoulos, #{149D} ennis A. Johnston, PhD
MD
MD
Radiology
Multiple Myeloma:
Spinal
In Patients
with Untreated
Newly Diagnosed
Disease'
MR Imaging
Spinal magnetic resonance (MR) im-
aging was performed in 29 patients
with newly diagnosed, untreated
multiple myeloma. Nineteen (66%)
patients were asymptomatic.
Sagittal
pre- and postcontrast
Ti-weighted
spin-echo images and gradient-re-
called-echo
images of the thoracic
and lumbosacral
spine were ob-
tamed. Marrow involvement
was
identified in 20 (69%) patients. There
were three MR patterns: focal lesions
in nine patients (31%), diffuse in-
volvement
in seven (24%), and an
inhomogeneous
pattern of tiny le-
sions on a background
of normal
marrow in four (14%). A statistically
significant correlation between MR
imaging patterns of marrow involve-
ment and serum hemoglobin
values
(one-way, P = .0899; Kruskal-Wallis, P = .0620) and between MR imaging
patterns and percentage of marrow
plasmacytosis
(Kruskal-Wallis,
P = .0314) was noted, with patterns of
diffuse and focal marrow involve-
ment associated with more abnormal
values. Spinal MR imaging in pa-
tients with early myeloma may reveal
marrow involvement
in both symp-
tomatic and asymptomatic
patients.
Some correlation was found between
MR imaging patterns and laboratory
indexes of disease.
Index terms: Bone marrow, MR. 30.1214
Myeloma, 30.3452 #{149S}pine, MR. 30.1214
Spine, primary neoplasms,
30.3452
Radiology 1992; 185:833-840
I From the Departments
of Diagnostic Radiol-
ogy (LAM., D.G.K.V., N.E.L., E.E.K., H.I.L.),
Medical Oncology (MAD., R.A.), and Biomath-
ematics (D.A.J.), University of Texas M. D.
Anderson Cancer Center, 1515 Holcombe Blvd.
Houston, TX 77030. Received February 21, 1992;
revision requested May 1; final revision received
July 16; accepted July 29. Address reprint re-
quests to LAM. C RSNA, 1992
M ULTIPLE myeloma
is a mono-
clonal proliferation
of malig-
nant plasma cells that usually affects
the bone marrow (1). The peak mci-
dence of multiple myeloma is during
the 6th decade of life. It is slightly
more common in men and affects
three in 100,000 persons per year (1).
The ability of magnetic resonance
(MR) imaging to depict bone marrow
abnormalities
has been documented
(2). The purpose of our study was to
evaluate spinal MR imaging in pa-
tients with newly diagnosed,
un-
treated multiple myeloma to identify
and describe patterns of marrow in-
volvement
in this largely asymptom-
atic population.
We also attempted
to
correlate MR imaging patterns of spi-
nal involvement
and laboratory pa-
rameters to further assess the role of
MR imaging in this group of patients.
MATERIALS
AND METHODS
Twenty-nine
consecutive
patients with
newly diagnosed, untreated multiple my-
eloma were studied with MR imaging of
the thoracic and lumbosacral
spine. There
were 13 men and 16 women (age range,
30-76 years; mean, 59 years). Nineteen patients (66%) had no bone pain. Ten
patients (34%) reported back pain. Con-
ventional radiographs
of the spine were
available for evaluation in all patients.
Laboratory
evaluation
for multiple
my-
eloma and posterior iliac crest aspirations
was performed in each patient. The extent
of tumor mass was staged as high in six
patients (hemoglobin
level, < 8.5 g/dL
[ <85 g/L]; serum calcium mg/dL I > 2.87 mmol/L]),
level, > 11.5 low in 14 pa-
tients (hemoglobin
level, > 10.5 g/dL
[ > 105 g/L]; serum calcium level, < 11.0
mg/dL [ < 2.74 mmol/L];
serum mono-
clonal component level, < 4.5 g/dL), and
intermediate
in nine patients
(all other
cases) (3,4). Three patients (10%) had non-
secretory multiple myeloma with bone
marrow
plasmacytosis
and no evidence of
abnormal protein in the blood or urine (5).
MR imaging was performed with a 1.5-T
unit (Signa; GE Medical Systems, Milwau-
kee) with use of a rectangular
surface coil.
All images were obtained prior to initia-
tion of treatment for multiple myeloma.
Sagittal TI-weighted spin-echo (600/20
[repetition
time msec/echo
time msec])
images were obtained in all patients. Gra-
dient-recalled
(GRE) acquisition in the
steady state (GRASS; GE Medical Systems)
(200-600/13-20;
13#{176}-25#{fl1ip76} angle) images
were obtained in 26 patients. GRE images
(75/15; 12#{17if6i}p angle) were obtained
in
one case. Unenhanced
and gadolinium-
enhanced Ti-weighted images with simi-
lar window widths and levels were ob-
tamed in 26 patients. Ti-weighted
MR
images of the lumbar and thoracic spine
obtained before and after administration
of contrast material, viewed with the same
window widths and levels in 10 patients
with solid primary malignancies,
were re-
viewed to study the presence or lack of
enhancement
in the uninvolved
marrow.
All images were acquired
in the sagittal
plane by using a 256 x 192 or 256 x i28
data acquisition matrix with a 28-cm field
of view. Section thickness was 4 mm with
a i-mm intersection
gap.
MR images were analyzed for patterns
of myelomatous
involvement
(Figs i-7).
Marrow involvement was characterized as
focal when focal areas of fatty marrow re-
placement with a signal intensity similar
to or lower than that of muscle were noted
on Ti-weighted
images in one or more
vertebrae (Figs 1, 2, 3, 6). A variegated pat-
tern of involvement
was considered
present when multiple innumerable
to be tiny
foci of marrow replacement
on a back-
ground of uninvolved
marrow were
observed
throughout
the spine on Ti-
weighted images, with subsequent
enhancement of these foci on gadolinium-
enhanced images (Fig 4). Diffuse involve-
ment was defined as total replacement
of
fatty marrow throughout
the spine on Ti-
weighted images, with evidence of mar-
row enhancement
on gadolinium-en-
hanced Ti-weighted
images (Fig 5). When
one or more vertebral bodies were dif-
fusely involved, the pattern of involve-
ment was classified as focal, provided that
uninvolved
marrow was present in other
vertebral bodies. We did not observe more
than one pattern of marrow involvement
in any one patient in this study.
The number and conspicuity
of the le-
Abbreviation:
GRE = gradient-recalled.
833
a. b. C. d.
Figure 1. Symptomatic myeloma appearing as a focal pattern of marrow involvement in a 30-year-old
age (600/20) of the thoracic spine reveals an unusual hyperintense
appearance of marrow involvement.
man. (a) Ti-weighted
sagittal MR im-
The seventh dorsal vertebral body is
compressed
(arrow in a, b). There is slight impingement
on the thecal sac. (b) Ti-weighted
postcontrast
MR image (600/20) reveals enhance-
ment of the lesions. (c) Radiograph of the thoracic spine reveals compression
and myelomatous
lesion in the seventh thoracic vertebral body
(arrow). Involvement
of the fifth thoracic vertebral body is also noted. (d) Ti-weighted
sagittal MR image (600/20) of the thoracic spine 13
months after treatment
reveals focal lesions (arrows) appearing
hypointense
relative to muscle.
sions with individual
sequences
were
noted. Compression
fractures
were listed
and classified as malignant in pathogene-
sis when either diffuse or focal and
rounded signal intensity abnormality
was
seen within the marrow.
They were con-
sidered of benign pathogenesis
when ei-
ther no signal intensity
abnormality
or a
band of low signal intensity
parallel to the
vertebral end plate was identified on Ti-
weighted
images.
The MR imaging
findings
were corre-
lated with tumor mass and values of he-
moglobin,
serum ft,-microglobulin,
per-
centage of bone marrow plasmacytosis,
serum creatinine,
serum calcium, and se-
rum monoclonal
protein. In addition, the
type of abnormal protein (immunoglobu-
lin C, 22 patients;
immunoglobulin
A, four
patients;
nonsecretory
myeloma,
three
patients) was correlated with MR imaging
patterns.
Statistical
analysis was performed
with
the one-way analysis of variance, Kruskal-
Wallis one-way analysis of variance
by
ranks, Student
t test, 2, and Mann-Whit-
ney test (6). With the small numbers
of
patients
in each MR imaging
category
the
assumption
of normality
of the means and
approximate
equality of variances re-
quired for the one-way
analysis of vari-
ance may not hold and are not testable
with any degree of power. For this reason,
the parametric one-way analysis of vari-
ance and the Kruskal-Wallis
tests were
both performed.
When the one-way anal-
ysis of variance
and the Kruskal-Wallis
tests are performed
and the data satisfy
the assumptions
of the one-way
analysis
of variance,
the Kruskal-Wallis
test is not
as significant
as the one-way
analysis of
variance by 5%-15%, whereas when the
one-way
analysis of variance
assumptions
are violated, the Kruskal-Wallis more significant.
test is
834 #{14R9}adiology
a b.
c.
Figure 2. Symptomatic myeloma appearing as a focal pattern of marrow involvement in a
71-year-old man. (a) Ti-weighted
sagittal MR image (600/20) of the thoracic spine shows in-
volvement of the entire compressed third dorsal vertebral body (white arrow in a-c) and sev-
eral focal lesions in the posterior
spinal elements
(black arrows in a-c). (b) GRE MR image
(217/20; 20#{17fl6ip} angle) better demonstrates
the lesions in the posterior spinal elements (black
arrows). (c) Ti-weighted
postcontrast image (600/20) reveals enhancement
in all areas of mar-
row involvement.
The lesion in the third dorsal vertebra is only faintly seen (white arrow).
RESULTS
Spinal MR imaging revealed mar-
row involvement
in 20 of 29 (69%)
patients. Three MR imaging patterns
of marrow involvement
were noted:
focal lesions in nine (31%) patients,
diffuse involvement
in seven (24%),
and a variegated
pattern of multiple
small lesions on a background
of nor-
mal marrow in four (14%). In nine
(31%) patients,
there was no MR im-
aging evidence of myelomatous
in-
volvement.
Seven patients with focal
lesions, two patients with diffuse in-
volvement, and one patient with a
variegated pattern reported back
pain. All patients with a normal MR
imaging study were free of bone pain.
There were 66 focal lesions (lumbo-
sacral spine [n = 30]; thoracic spine
[n = 36]). Only 12 (18%) lesions were
demonstrated
as lytic foci on conven-
tional radiographs.
On Ti-weighted
images, 52 lesions (79%) were hypoin-
tense relative to muscle, whereas 14
(21%) lesions, all in one patient, were
hyperintense
relative to muscle (Fig
1). Conventional
radiographs
of the
spine in this patient revealed lytic le-
sions consistent with multiple my-
eloma. On follow-up MR images of
the spine 13 months after initiation
of
treatment,
the focal lesions were hy-
pointense
relative to muscle, suggest-
ing the presence of hemorrhage
on
December 1992
a. b.
c
Figure 3. Symptomatic myeloma appearing as a focal pattern of marrow involvement in a
48-year-old
man. (a) TI-weighted
sagittal MR image (600/20) of the thoracic spine reveals focal
areas of decreased
signal intensity (arrows). (b) GRE image (75/15; 12#{1f7l6ip} angle) better dem-
onstrates
the lesions and reveals two additional
lesions in the spinous process of the ninth
vertebral body (black arrow) and in the body of the eighth dorsal vertebra (white arrow).
(c) Ti-weighted
postcontrast image (600/20) clearly demonstrates
the lesions and reveals an
additional
small lesion in the 11th dorsal vertebral body (arrow).
the initial study. On GRE images, all
visualized lesions were hyperintense
relative to muscle. All 37 lesions stud-
ied after contrast enhancement
re-
vealed various degrees of uptake of
contrast material.
Nineteen (39%) of the 49 lesions
studied with both Ti-weighted
and
GRE sequences were equally well
seen on both series. Eleven
(22%)
lesions were better seen on Ti-
weighted images. GRE images better
depicted ii (22%) lesions, six of
which were located in the posterior
spinal elements (Fig 2). Eight (i6%)
lesions identified on Ti-weighted
im-
ages were not seen on GRE images.
Eleven (30%) of 37 lesions studied
with pre- and postcontrast
Ti-
weighted images could be identified
only on the unenhanced
images,
whereas five (14%) lesions in a patient
with a paucity of fatty marrow in the
spine were better seen on the en-
hanced images (Fig 3).
In five of eight patients studied
with both Ti-weighted
and GRE se-
quences, Ti-weighted
images better
demonstrated
focal lesions. In two of
eight patients, focal lesions were
equally well seen with both se-
quences, while in one patient with a
relative lack of spinal fatty marrow,
focal lesions were better seen on GRE
images. Focal lesions were better de-
picted on unenhanced
Ti-weighted
images compared with gadolinium-
enhanced images in five of seven pa-
tients studied with both sequences,
equally well seen in one patient, and
less well visualized
in one patient
with a paucity of spinal fatty marrow.
GRE images were superior in demon-
strating focal lesions in five of six pa-
tients studied with both GRE and
gadolinium-enhanced
images. In one
patient, both series failed to reveal
any of the lesions noted on the unen-
hanced Ti-weighted
images.
Follow-up MR imaging studies
were performed in three patients
with focal lesions 6-i3 months after
initiation of treatment; marked im-
provement
was noted in two, and in
another patient, who had nonsecre-
tory myeloma, no change was noted.
The variegated on Ti-weighted
pattern was noted
images as multiple
tiny foci hypointense
relative to unin-
volved marrow that enhanced on
gadolinium-enhanced
Ti-weighted
images (Fig 4). Multiple hyperintense
foci of marrow involvement
were
noted in only one of three patients
studied with GRE imaging. Follow-up
MR imaging 38 months later revealed
marked progression
of the variegated
pattern in one patient with low tumor
mass myeloma
who never received
treatment and remained asymptom-
atic. A 6-month
follow-up
MR imag-
ing study after treatment in a patient
with intermediate
tumor mass re-
vealed marked interval improvement
in the appearance
of the variegated
pattern (Fig 4).
In seven patients with diffuse involve-
ment of the marrow, Ti-weighted
im-
ages revealed lack of fatty marrow in
the spine, with the intervertebral
disks being iso- or hyperintense
rela-
tive to the abnormal vertebrae (Fig 5).
The signal intensity
of the marrow
was slightly lower than that of adja-
cent muscle groups. A diffuse increase
in signal intensity was noted in only
two of six cases studied with GRE im-
aging (Fig 5). Gadolinium-enhanced
images revealed diffuse enhancement
of the marrow in all cases, with the
disks becoming hypointense
relative
to the vertebral bodies (Fig 6). Fol-
low-up MR imaging i month after
bone marrow transplantation
in one
patient with diffuse marrow involve-
ment revealed an almost normal ap-
pearance
of the marrow with no defi-
nite evidence
of enhancement.
A retrospective
review of pre- and
postcontrast
Ti-weighted
images in
10 patients with solid primary malignancies, viewed with similar window
widths and levels, did not reveal any
change in the appearance of unin-
volved vertebral bodies and relative
signal intensities of the intervertebral
disks after administration
of contrast
material.
Twenty-eight
compression
frac-
tures were observed
in 13 patients
(ii [39%] of the lumbosacral spine; 17
[61%] of the thoracic spine). Twenty-
two (79%) were considered
of malig-
nant pathogenesis
on MR images
(Figs 1, 2, 4, 6). Lytic lesions were
identified in four (14%) compressed
vertebral bodies on conventional
ra-
diographs of the spine. Six (21%) of
the 28 compression
fractures
showed
no evidence of myelomatous
involve-
ment on MR images (Fig 7).
Statistical correlation of the MR im-
aging patterns with hemoglobin
val-
ues revealed
a marginally
significant
relationship by means of one-way
(P = .0899) and Kruskal-Wallis
(P =
.0620) analysis, with the diffuse and
focal lesion patterns
associated
with
lower hemoglobin
values than the
variegated and normal patterns (Ta-
ble 1). A graphic illustration of these
observations
is presented
in Figures 8
and 9. The relationship
between MR
imaging patterns and bone marrow
plasmacytosis
was not significant at
one-way analysis (P = .14) but was
Volume 185 #{14N9}umber 3
Radiology
#{14893} 5
Figure 4. Symptomatic
myeloma
appearing
as a variegated
pattern of marrow involve-
ment in a 45-year-old man. (a) Ti-weighted
sagittal image (600/20) of the lumbar spine
reveals a variegated appearance of the mar-
row. The hyperintense
focus in the second
lumbar vertebral body is an area of focal fatty
marrow. The first lumbar vertebra is corn-
pressed, with retropulsion of turnorous bone.
(b) GRE sagittal MR image (650/20; 20#{17fl6ip}
angle) reveals multiple hyperintense
foci of
disease throughout the lumbar spine. Note
involvement of the posterior spinal elements.
(c) Ti-weighted
postcontrast
image (600/20)
reveals inhomogeneous
enhancement
throughout the lumbar spine. (d) Ti-weight-
ed sagittal image (600/20) of the lumbar
spine 6 months after treatment
reveals an
increase in the fatty marrow content of the
spine. Several vertebral compression
frac-
tures have developed.
significant at Kruskal-Wallis
analysis
(P = .0314); MR imaging evidence of
marrow involvement
was associated
with higher values of bone marrow
plasmacytosis.
A significant
relation-
ship between MR patterns and serum
132-microglobulin, serum creatinine,
monoclonal
protein,
and serum calcium
was not observed
with either test. No
correlation
was found between types
of abnormal protein and MR imaging
patterns.
Table 2 shows the relationship
be-
tween tumor mass and MR imaging
patterns.
There were no high-tumor-
mass variegated
or normal MR imag-
ing patterns. If the Table is collapsed
by combining intermediate
and low
tumor mass and compared
with com-
bined diffuse and focal patterns ver-
sus normal and variegated
patterns,
the resultant
x2 is 6.15 (P = .013), mdi-
cating a potential relationship
be-
tween tumor mass and MR imaging
patterns.
As the number of patients in mdi-
vidual groups was small and because
the statistical analysis revealed
a ten-
dency for diffuse and focal lesion pat-
terns to be associated nounced laboratory
with more pro-
abnormalities
and
higher tumor mass, diffuse and focal
lesion patterns were pooled into a
single group and variegated
and nor-
mal patterns,
into a second group.
Statistical analysis with the Student
test revealed significantly
lower val-
ues of hemoglobin
(P = .024) and sig-
nificantly
higher values of bone mar-
row plasmacytosis
(P = .029) for the
diffuse and focal lesion group in com-
parison with the variegated
and
normal group. Mann-Whitney
non-
parametric
tests corroborated
the as-
sociation
of MR imaging
patterns
and
bone marrow plasmacytosis
(P < .01)
but not the association
of MR imaging
patterns
and hemoglobin
(P =
indicating the need for further
with larger numbers of patients
individual groups.
.21), studies in
DISCUSSION
In multiple myeloma,
infiltration
of the marrow
plasma occurs
cell in
836 #{14R9}adiology
b.
d.
both a nodular and interstitial
fashion
(7). Therefore,
various patterns
of in-
volvement
are expected
on MR im-
ages of the spine. In our study, spinal
MR imaging revealed three patterns
of myelomatous
involvement
of the
marrow in 20 of 29 (69%) patients.
In
agreement
with other authors,
we
saw that focal areas of marrow
in-
December 1992
a. b.
C.
Figure 5. Symptomatic
myeloma appearing
as a diffuse pattern of marrow involvement
in a
56-year-old
woman. (a) TI-weighted
sagittal MR image (600/20) of the thoracic spine reveals
complete
replacement
of the fatty marrow. Note the isointense
appearance
of vertebral bodies
relative to intervertebral
disks. (b) GRE image (250/20, 20#{17fl6ip} angle) reveals diffuse inhomo-
geneous signal hyperintensity
throughout
the thoracic spine. (c) Ti-weighted
postcontrast
image (600/20) reveals diffuse enhancement
of the marrow. The intervertebral
disks are now
hypointense
relative to the enhanced
vertebral bodies.
volvement
were the most frequently
observed abnormality
(8-10) (Fig i).
In addition,
total marrow replacement
(diffuse pattern) and inhomogeneous
marrow replacement
(variegated pat-
tern) were noted in seven (24%) and
four (14%) patients, respectively.
All three patterns of marrow in-
volvement
were seen on Ti-weighted
MR images. Both variegated and dif-
fuse patterns, however, when noted
on Ti-weighted
images, had to be
differentiated
from normal patterns
of
inhomogeneous
distribution
of fatty
marrow often observed
in older per-
sons or from the persistent red mar-
row pattern occasionally
seen in
younger patients with myeloma (ii).
Gadolinium-enhanced
Ti-weighted
images showed diffuse or inhomoge-
neous enhancement,
respectively,
in
all patients with diffuse or variegated
involvement
of the marrow.
To our knowledge,
gadopentetate
dimeglumine
has not been previously
employed in the evaluation of myelo-
matous involvement
of the spine. Al-
though normal marrow may mark-
edly enhance
in children
younger
than 7 years, enhancement
of the
marrow is subtle in adults and can be
appreciated
only when intensity
mea-
surements
are calculated
(12,13). In-
tervertebral
disks are isointense rela-
tive to spinal marrow in patients with
diffuse marrow involvement.
After
administration
of contrast material,
the disks become hypointense
relative
to marrow as the abnormal vertebral
bodies enhance.
GRE images were not helpful in
depicting
diffuse or variegated
mar-
row involvement.
On GRE images,
bone is markedly
hypointense
to
muscle, and in the majority
of cases,
neoplastic aggregates
will appear as
hyperintense
foci; it is possible that a
diffuse or variegated
pattern of
plasma cell infiltration of the marrow
may not produce
sufficient
increase
in
cellularity
to overcome
the magnetic
susceptibility
effect of bone with a
GRE sequence (14). Spin-echo T2-
weighted or inversion recovery im-
ages, though more time-consuming,
may be more sensitive
in depicting
these patterns
(10).
In our study, focal lesions were bet-
ter seen on Ti-weighted
images in the
majority of cases. GRE images were
useful in depicting lesions in the pos-
terior elements
by suppressing
the
subcutaneous
fat and thus increasing
contrast between
lesion and back-
ground. In patients with relatively
little fatty spinal marrow and poor
lesion-to-marrow
contrast on Ti-
weighted
images, both GRE and en-
hanced Ti-weighted
images clearly
depicted
the marrow involvement.
GRE images failed to demonstrate
16% of focal lesions noted on Ti-
weighted
images.
The "masking"
effect of gado-
pentetate dimeglumine
on the MR
imaging appearance
of marrow le-
sions has been previously
stressed
(15). In our study, 30% of focal lesions
could not be detected on the en-
hanced images because they became
isointense
relative to uninvolved
mar-
row. In one patient, a solitary focal
lesion could not be identified
on the
Volume 185 #{14N9}umber 3
Radiology
#{14893} 7
6a. 6b. 7a.
7b.
Figures 6, 7. (6) Malignant
compression
fractures in a 63-year-old
woman with symptomatic
myeloma.
(a) Ti-weighted
sagittal image (600/20)
of the lumbar and lower thoracic spine reveals compression fractures of the 11th and 12th dorsal vertebrae, with focal marrow-replacing
le-
sions. (b) GRE MR image (300/20; 200 flip angle) shows an increase in the signal intensity of the focal marrow involvement
in the 11th and 12th
vertebral bodies. (7) Asymptomatic
myeloma and an acute benign compression fracture in a 55-year-old man. (a) Ti-weighted
sagittal MR im-
age (600/20) of the lumbar spine reveals compression of the fifth lumbar vertebra, with a hypointense band adjacent to the superior end plate
(arrow). Focal areas of fatty marrow are noted in the L-4 and L-i vertebral bodies. (b) GRE MR image (267/20; 20#{1f7li6p} angle) reveals an increase
in the signal intensity
of the band of signal alteration
in the fifth lumbar vertebra noted in a (arrow).
14
12
10
H8 b (g/dl) 6
4
2
0 0#{176}
.
0 A1
0 0
0 A
0 20 40 60 80
BM
8.
Figures 8, 9. (8) Graph shows MR imaging patterns versus hemoglobin
centage. L = diffuse pattern, 0 = focal pattern, D = variegated
bone marrow plasmacytosis
(BM ), expressed as a percentage.
pattern,
1 UI..
0 80
B 60 M
40 (Z)
20
C 100
9.
level (Hb) and bone U = normal pattern.
0 00
0 0
00 r:
0
n
[31)
I_I [1
Diffuse
H
r) 0 DOD
Focal
j
Variegated
MR Patterns
000
0
8
p Normal
marrow plasmacytosis
(BM ), expressed
as a per-
(9) Graph shows MR imaging patterns versus
gadolinium-enhanced
images. Other
authors have reported focal lesions in
myeloma to be better or exclusively
seen on T2-weighted
MR images (10).
An unusual
finding in this study
was the hyperintense
appearance
of
the focal lesions relative to muscle
and disk in a 30-year-old
patient with
myeloma
(Fig 1). Amyloid
deposition
was initially considered
a possible
pathogenesis
of this finding; in a case
of amyloid deposition
in the spine
reported in the literature, however,
the abnormality
was seen as a de-
crease in the signal intensity of the
spinal marrow relative to the interver-
tebral disks on Ti-weighted
images
(16). Resolution
of the hyperintense
signal with a hypointense
appearance
of the focal lesions relative to muscle
on a 13-month follow-up MR imaging
study of the spine suggested
the pres-
ence of hemorrhage
within foci of
myelomatous
involvement
of the
marrow at the time of the initial study
(17). Conventional
radiographs
ob-
838 #{14R9}adiology
tamed in this patient revealed
lytic
lesions consistent
with myeloma. Ra-
diologists
should be aware that le-
sions that are hyperintense
relative to
marrow on Ti-weighted
MR images
of the spine do not always indicate
benign pathogenesis such as those of
hemangiomas
or focal fatty islands
and may rarely be encountered
in
patients with myeloma (18).
Several MR imaging criteria have
been employed
to differentiate
be-
nign from malignant compression
December 1992
fractures
(19,20). Absence of marrow
signal alteration has been reported to
indicate a benign vertebral compres-
sion fracture (20). On spin-echo MR
images, diffuse marrow replacement
involving
the entire vertebral body in
a compressed
vertebra has been de-
scribed as highly suggestive of a ma-
lignant pathogenesis
(19). Diffuse in-
volvement, however, was noted in
only eight of 28 (29%) compression
fractures in this study. In half of the
28 compression fractures, a focal or
variegated
pattern of marrow in-
volvement
was noted.
In a recently reported study, only
two of 38 (5%) compressions
had
complete
replacement
of the marrow
by myeloma and 17 of 38 (45%)
showed lesser extent of marrow in-
volvement
(10). Therefore,
in multiple
myeloma,
a vertebral
compression
fracture may occur without
the MR
imaging observation
of marrow being
totally replaced by tumor cells, possi-
bly because, in addition to tumorous
marrow infiltration,
osteoclast activat-
ing factors also contribute
to the de-
velopment
of vertebral compression
fractures
(1). In such cases, the config-
uration of the marrow abnormality
(rounded
focal versus parallel to end-
plate band) will aid in differentiating
a malignant from a benign fracture
(20) (Figs 6, 7). As previously
noted, at
spinal MR imaging in patients with
malignancies
other than myeloma,
enhanced
images do not always help
in evaluating
the nature of compres-
sion fractures
because
enhancement
may occur in malignant as well as in
posttraumatic
fractures,
depending
on the age of the fracture (21,22).
An important
finding in our study
was the correlation
between
MR im-
aging patterns of marrow involve-
ment and hemoglobin
and bone mar-
row plasmacytosis
values: The diffuse
and focal lesion patterns were associ-
ated with more pronounced
labora-
tory abnormalities.
In evaluating
these results, one must keep in mind
that hemoglobin
and bone marrow
plasmacytosis
more closely reflect the
status of the marrow than do other
laboratory
parameters
(23). At our
institution,
patients with indolent my-
eloma (asymptomatic,
low tumor
mass) do not receive treatment until
they become symptomatic
or their
disease proceeds to a higher tumor
stage. Preliminary
results of a pro-
spective clinical study show that pa-
tients with indolent
myeloma
and
focal or diffuse involvement
noted on
MR images of the spine needed treat-
ment earlier than did patients with a
variegated
or normal MR imaging
appearance of the spine. Libshitz et al
(10) found no correlation between MR
imaging appearances
and serum 32-
microglobulin
or monoclonal
protein
values in 32 patients with multiple
myeloma; this is in agreement
with
our results. They also, however, noted
no relationship
between MR imaging
appearances
and bone marrow plas-
macytosis, possibly because different
criteria were employed
for MR imag-
ing pattern characterization
(10).
Long-term
prospective
studies are
required to establish the significance
and prognostic value of the different
MR imaging patterns of marrow in-
volvement.
Return of marrow appearance
to
normal on spinal MR images a month
after bone marrow transplantation
in
one patient correlated with lack of
plasmacytosis
on a marrow aspirate,
suggesting a potential role of MR im-
aging in evaluating response to treat-
ment.
Nonsecretory
myeloma
is charac-
terized by involvement
of the bone
marrow but no evidence
of abnormal
protein in the blood or urine (5). In
nonsecretory
myeloma (three of 29
patients in this series), assessment
of
the clinical course depends on re-
peated marrow aspirations
because
there is no detectable monoclonal
peak in the serum (5). An iliac aspirate
may not always be representative
of
the marrow involvement
due to both
nodular and interstitial tumor growth
patterns (7). MR imaging may prove
to be a useful, noninvasive
method of
assessing the course of disease in
these patients.
In summary, spinal MR imaging
may reveal marrow involvement
(focal, diffuse, or variegated)
in both
symptomatic
and asymptomatic
pa-
tients with multiple myeloma. Ti-
weighted
images should be obtained
in all patients. Enhanced Ti-weighted
images can be of great value in cases
of diffuse or variegated
involvement.
GRE images may be employed
in ad-
dition to Ti-weighted
images in se-
lected cases of focal lesions (posterior
elements,
paucity of fatty marrow).
MR imaging findings of marrow in-
volvement
do correlate with certain
laboratory indexes of tumor burden,
with diffuse involvement
and focal
lesions tending
to occur in patients
with more advanced disease. MR im-
aging appears promising for assessing
response
to treatment,
especially
in
patients with nonsecretory
myeloma.
Long-term prospective
studies may be
helpful in defining the prognostic
value of MR imaging patterns tients with multiple myeloma.
in pa-
#{149}
Acknowledgments:
The assistance provided
by Debbie Smith and Mary Carr in the prepara-
tion of this manuscript
is gratefully acknowl-
edged. We thank Dana Holloway, Brenda
Hutsko, Ricardo Estrada, and Joseph Kurian for
technical assistance and Glenda Deal for secre-
tarial assistance.
References
1. Longo DL. Plasma cell disorders. In: Wil-
son JD, Braunwald
E, Isselbacher
KJ, et al,
eds. Harrison's principles of internal medi-
cine. 12th ed. New York: McGraw-Hill, 1991; 1412-1416. 2. VoglerJB III, Murphy WA. Bone marrow imaging. Radiology 1988; 168:679-693. 3. Alexanian R, Barlcerzak 5, Bonnett JD, et al. Prognostic factors in multiple myeloma.
Cancer 1975; 36:1192-1201. 4. Dune BGM, Salmon SE. A clinical staging
for multiple myeloma. Cancer 1975; 3:842854. 5. Dreicer R, Alexanian R. Nonsecretory myeloma. Am J Hematol 1982; 13:313-316. 6. Norusis Marija J. SPSS/PC + 4.0. Chicago: SPSS, 1991. 7. Barti R, Frisch B, Diem H, et al. Bone marrow histology and serum beta 2 microglobulin in multiple myeloma: a new prognostic strategy. EurJ Haematol 1989; 43(suppl 51):88-98. 8. Daffner RH, Lupetin AR, Dash N, et al. MRI in the detection of malignant infiltration of bone marrow. AJR 1986; 146:353-
358.
9. Fruehwald
FXJ, Tscholakoff
D, Schwaig-
hofer B, et al. Magnetic resonance imag-
ing of the lower vertebral column in pa-
tients with multiple myeloma. Invest
Radiol 1988; 23:193-199.
10. Libshitz HI, Malthouse SR. Cunningham
D,
et al. Magnetic resonance appearance of
multiple myeloma. Radiology 1992; 182:
833-837.
11. Ricci C, Cova M, Kang YS, et al. Normal
age-related patterns of cellular and fatty
bone marrow distribution
in the axial skel-
Volume 185 #{14N9}umber
3
Radiology
#{14893}9
eton: MR imaging study. Radiology 1990;
177:83-88.
12. Breger RK, Williams AL, Daniels DL, et al.
Contrast enhancement
in spinal cord imag-
ing. AJR 1989; 153:387-391.
13. Sze G, Bravo 5, Baierl P. Shimkin PM. De-
veloping spinal column: gadolinium-en-
hanced MR imaging. Radiology 1991; 180:
497-502.
14. Czervionke LF, Daniels DL, Wehrli FW, et
al. Magnetic susceptibility aftifacts in gra-
dient-recalled echo MR imaging. AJNR
1988; 9:i149-1 155.
15. Sze G. Gadolinium-DTPA
in spinal dis-
ease. Radiol Clin North Am 1988; 26:1009-
1024.
16. OlliffJFC, Hardy LR, Williams MP, et al.
Case report: magnetic resonance imaging
of spinal amyloid. Clin Radiol 1989; 40:632-
633.
17. Atlas SW, Grossman RI, Gomon JM, et al.
Hemorrhagic
intracranial
malignant
neo-
plasms: spin-echo MR imaging. Radiology
1987; 164:71-77.
18. RossJS, Masaryk TJ, Modic MT. Carter JR.
Mapstone T, Dengel FH. Vertebral he-
mangiomas: MR imaging. Radiology 1987;
165:165-169.
19. Baker LL, Goodman SB, Perkash I, Lane B,
Enzmann DR. Benign versus pathologic
compression
fractures of vertebral bodies:
assessment
with conventional
spin-echo,
chemical shift, and STIR MR imaging. Ra-
diology 1990; 174:495-502.
20. Yuh WTC, Zachar CK, Barloon TJ, Sato Y,
Sickels WJ, Hawes DR. Vertebral corn-
pression fractures: distinction
between be-
nign and malignant causes with MR imag-
ing. Radiology 1989; 172:215-218.
21. Staebler A, Gaertner C, Seiderer M, et al.
Differential diagnosis ofbenign and malig-
nant vertebral compression fractures with
MR imaging: use of antiphased
long-TR
gradient-echo
sequences and Gd-DTPA
and STIR sequences (abstr). Radiology
1991; 181(P):203.
22. Baker LL, Gean-Marton
AD, Koeller KK, et
al. Compressed vertebral body: MR imag-
ing features in trauma, osteopenia,
infec-
tion and tumor (abstr). Radiology 1991;
18i(P):203.
23. Alexanian R, Barlogie B, Fritsche H. Beta
2 microglobulin
in multiple myeloma. Am
Hematol 1985; 20:345-351.
840 #{14R9}adiology
December 1992