Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS.

For personal use only; all rights reserved

317

MR Imaging
and Ewings

of Osteogenic
Sarcoma

Twenty
patients
with biopsy-proven
osteogenic
(1 1 cases)
or Ewings (nine cases)
sarcoma were evaluated
by MR imaging on a O.15-T resistive unit to determine
the value
of MR in the diagnosis
and treatment
of these two neoplasms
and to develop the best
protocol for MR imaging. In all 20 cases,
MR identified
tumor spread into bone marrow,
and it was superior to CT in five cases. Extension
of tumor into the soft tissues adjacent
to bone was shown better
by MR than CT in six cases. Improved
anatomic
information
from MR is the result of the ability
to image in the axial, coronal, and sagittal planes.
Compared
with CT, MR identifies
cortical disease but has inferior spatial resolution
and
defines calcium poorly.
MR can be used to monitor tumor
response
to chemotherapy,
and the relationship
of tumor to adjacent vasculature
can be determined
without the use
of contrast agents. Two pulse sequences
are necessary
for maximum display of disease,
since, in general,
tumor involvement
of the bone marrow
is best assessed
on Tiweighted
sequences,
and tumor involvement
of the soft tissue is best seen on T2-

Orest B. Boyko1
David A. Gory1
Mervyn
D. Cohen1
Arthur Provisor
David Mirkin3
G. Paul DeRosa4

weighted

sequences.

Additional

information

about bone-marrow

tumor extent, and the relationship

of tumor to blood
adjunct to CT in the evaluation
of these neoplasms.

Clinical experience
with
1 1 1. MR is also becoming

MR imaging
increasingly

vessels

involvement,
makes

MR

soft-tissue
a valuable

of primary
important

bone tumors is accumulating

[i in the imaging
of bone-marrow
images, identifies bone-marrow
and soft-

disease [1 1 -1 6]. MR provides multiplanar

tissue extension of tumor, and holds promise as a means of monitoring
tumor
response to therapy [5-8]. We reviewed the MR findings in 1 1 cases of osteogenic
sarcoma and nine cases of Ewings sarcoma to evaluate the clinical usefulness of
the technique and to determine the best MR imaging strategy for these two tumors.
Received
February 27, 1986; accepted
vision September
1 0. 1986

after re-

Materials

Presented at the annual meeting of the American

Roentgen
1

Ray Society.

Department

Boston.

of Radiology.

School of Medicine,
anapolis,
IN 46223.

April

1985.

Indiana University

Indito 0.
B. Boyko, Department of Radiology, Indiana University School of Medicine, 926 W. Michigan St.,

Indianapolis,
2

diana
strief

and Regenstnef
Address
reprint

Institute,
requests

IN 46223.

Department

of Hematology

University
Institute,

School

and Oncology,

of Medicine,

Indianapolis,

and

In-

Regen-

IN 46223.

3Department
of Pathology,
Indiana
University
School of Medicine,
and Regenstrief
Institute,
Indianapolis,
IN 46223.
4Department
of Orthopaedic
University
School of Medicine,
stitute, Indianapolis,
IN 46223.
AJR 148:317-322,

0361 -803x/87/1
C American

February

Surgery,
Indiana
and Regenstrief
In1987

482-0317

Roentgen

Ray Society

and Methods

Twenty patients aged 9-21 years with either osteogenic sarcoma (1 1 cases) or Ewings
sarcoma (nine cases) had MR scans before chemotherapy
or radiation therapy. A 0. 15-T
resistive
scanner
(Technicare
Teslacon)
was used. All patients were examined with at least
one Ti
and one T2-weighted
spin-echo
(SE) pulse sequence,
and three patients were
examined with inversion-recovery
(IR) pulse
sequences.
The echo-delay
time (TE) was 30,
32, 60, or 120 msec; the repetition time (TA) was 250, 500, 550, 1000, or 2000 msec. (SE
250/32, 550/32, and 500/30 are Ti weighted; SE 1000/30, 1000/60, 2000/60, and 2000/
120 are more T2 weighted.) For the IR sequence, a 90#{176}-i
80#{176}
signal-reading
pulse pair
followed the initial 180#{176}
inverting pulse. The inversion time (TI) was 450 msec. The 90#{176}
and
180#{176}
refocusing
pulses were separated
by 1 5 msec. The TR was 1500 msec. IA images are
heavily Ti weighted. In one case, multiecho images were obtained with a TA of 2000 msec
and a TE ranging from 30 to 240 msec.
Contiguous
slices
of either
10 or 15 mm were obtained in the coronal,
axial,
or sagittal
plane. Images were reconstructed
into a 256 x 128 matrix and interpolated to 256 x 256 for
-

display.

No contrast

agents

were

used.

The distribution of the 1 1 cases of osteogenic sarcoma were femur (seven), pelvis (two),
humerus (one), and tibia (one). The nine cases of Ewings sarcoma were located in the femur

Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS. For personal use only; all rights reserved

318

BOYKO

Fig. 1.-Ewings
sarcoma
of the left fibula.
A, Soft-tissue-window
CT shows no soft-tissue
mass or marrow abnormality.
B, IR 1500/450
image shows marrow replacement
by tumor (arrow).

(two), pelvis (five), spine (one), and fibula (one). All patients
film examinations,

and i 9 patients

had CT scans

before

had plain-

MR imaging.

CT was performed on either a Philips Tomoscan 31 0 or Technicare

20600 scanner. Follow-up MR studies were performed in nine patients to monitor

response

to therapy.

Diagnoses

in all 20 cases

were

proven by biopsy. Two resected osteogenic sarcoma specimens were

studied for pathologic correlation of MR findings.
MA studies were reviewed retrospectively
by two of the authors;
their

observations

were

simultaneously,

and

correlated

a consensus

with

plain

impression

films

was

of tumor in the soft tissues and bone marrow

time of open biopsy and/or amputation.

and

reached.

CT scans
Presence

was confirmed

at the

ET

AL.

AJR:148,

C, SE 1000/60
image shows soft-tissue
mass (arrows)
and periosteal
elevation
(arrowhead),
but fails to show difference
between
normal tibia
and abnormal
fibula marrow as was seen in B.

The soft-tissue

sarcoma

MR

identified

soft-tissue

masses

in all 20 cases.

In the

19

but the extent

was

better

shown

by MR

in five cases.

CT

case.
Both Ti - and T2-weighted
images were required in the evaluation of tumor in soft tissue and bone marrow, since on Ti weighted images tumor that infiltrates muscle can have a
showed

signal

marrow

isointense

abnormalities

with

muscle

better

(Fig.

than

MR

in one

1 B). The signal

becomes

images
(Fig. 1 C). Conversely,
sequence
chosen can cause tumor that is
infiltrating
bone marrow to have a bright signal similar to
normal marrow
(Fig. 1 C).
In one case of a pelvic osteogenic sarcoma, the extensive
marrow involvement
shown by MR changed the surgical
approach
to a hemipelvectomy.
MR can also show metastatic
sites of osteogenic
sarcoma even when there is calcified

of both osteogenic

and Ewings

signal intensity,

and these

two tumors
could not be distinguished
based on Ti
signal intensity.
MR showed cortical thickening
and tumor infiltration

or T2
of the

cortex by both types of tumors. In four cases where cortical

disease was seen on plain film and/or CT, no cortical disease
was evident on MR. Cortical involvement
by Ewings or
osteogenic sarcoma is identified on MR by the replacement
of the

dark

signal

of the

signal intensity

Both

cases in which CT was perforrned, MR showed better the

soft-tissue mass in six, including one case of Ewings where
the mass was isodense with muscle on CT (Fig. 1A). Marrow
involvement was evident in all cases imaged by MR and CT,

component

had an inhomogeneous

creased

Resufts

February 1987

CT and

plain

cortex

with

the

of the invading
radiographs

and localization

of calcifications

of calcification

produced

comparatively

in-

tumor.

provided

better

definition

in all cases. On MR large foci

discrete

areas

of signal

void

(Fig.

2A). Small calcifications

resulted in an inhornogeneous
appearance in some cases (Figs. 3 and 4). The areas of lowest
signal in tumors on MR did not always correspond
to areas
ofcalcification
(Fig. 4). Fourteen cases had periosteal
reaction
seen on plain films or CT, and this could be identified
by MR
in eight cases (Fig. 1 C).
Pathologic-radiologic

sarcoma

extent

showed

of tumor

a lateral

correlation

the

reliability

in bone marrow

plain film of the distal

in two

of MR

mass

to display

and soft tissue.

femur

lesion with a wide zone of transition,

soft-tissue

cases of osteogenic

(Fig. 5). Comparison

showed

the

true

In one case,

an osteoblastic

periosteal

reaction,

of a sagittal

and

MR image

the T2 pulse

with the gross specimen

proximal extent of marrow

tumor matrix

amount of tumor necrosis was present, with the greatest

amount of viable tumor adjacent to the epiphyseal line. Chondroid as well as osteoid matrix were present, but their distribution did not correlate directly with the inhomogeneous
signal of infiltrated marrow. Soft-tissue extension and periosteal reaction were seen on MR. In this case CT provided

more

intense

on T2-weighted

(Fig. 2).

Histologically

no tumor

was

showed that MR predicted

the
disease within several millimeters.
present

in the

epiphysis.

A large

Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS. For personal use only; all rights reserved

AJR:148,

February

MR

1987

OF

OSTEOGENIC

AND

EWINGS

319

SARCOMA

Fig. 2.-Osteogenic
sarcoma of the right humews.
A, coronal
SE 2000/60
image shows primary
tumor (arrows).
cakifiei
axillary node metastases (arrowheads)
appear
as areas
of decreased signal.
B, Chest radiograph confirms densely calcified axillary node metastases
(arrowheads).

Fig. 3.-Osteogenic
sarcoma
of the femur.
A, Bone-window
CT scan shows several small
areas of calcification
(arrows),
and tumor extension into marrow
and surrounding soft tissue.
B, Axial SE 1000/60
image shows inhomogeneous signal from tumor (arrows) that cannot be
correlated
with CT as being areas of calcification, which shows limitation
of MR in the evaluation of calcium.
Tumor replacement
of normal
marrow signal is seen.

better definition

than MR of calcium

deposition

in the tumor

matrix.

Response to therapy
(Fig. 8) and osteogenic

could be monitored for both Ewings

sarcomas.
In three cases of osteo-

Examination
of osteogenic

of a coronal specimen section in another case

sarcoma showed tumor extension
into the

genic sarcoma,
tumor enlargement
documented
by MR allowed for the early cessation of methotrexate
chemotherapy,

epiphysis

6). The

followed

(Fig.

extent

of tumor

involvement

marrow and soft tissue in the gross specimen

lated with the abnormal signal shown by MR.
Flowing

blood

emitted

no signal

and appeared

directly
black

of the

on all

pulse sequences.
In 12 cases, MR showed the relationship
of tumor to blood vessels more easily than CT did. Scanning
in the sagittal plane was often useful for demonstration
superficial
femoral
and popliteal
arteries.
The coronal

provided excellent demonstration

displacement
of pelvic structures
Improved anatomic information

by amputation.

corre-

of the
plane

of major pelvic vessels and

(Fig. 7).
can be derived from MR by

the opportunity
to image in several
planes.
In Figure 8 a
coronal
MR image showed
extension
of a spinal Ewings
sarcoma
behind the diaphragmatic
crus. The CT scan was
initially interpreted
as an adrenal
mass that was invading
bone.

Discussion

We found infiltration

of the bone marrow

by osteogenic

and

Ewings
sarcoma
to be clearly shown by MR; in five cases it
was superior
to CT. Tumor has a long Ti relaxation
time and

thus has a signal intensity

lower than that of marrow

Ti-weighted

Pathologic

images

[16].

correlation

fat on

of marrow

involvement
by two osteogenic
sarcomas
(Figs. 5 and 6)
showed the marrow involvement
demonstrated
by MR to be
within several millimeters of that documented
on the pathologic specimen.
This supports
the findings
gators of the reliability
of MR in showing
extent

of neoplasm

[10,

16].

by other investithe true marrow

320

BOYKO

ET

AJR:148, February 1987

AL.

Fig. 4.-Ewings
sarcoma of the left ilium.
A, CT shows areas of calcification
(arrows)
within tumor with periosteal reaction and soft-

Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS. For personal use only; all rights reserved

tissue mass.
B, Axial SE 500/30

image shows
inhomogeneous signal in tumor, but areas of lowest signal
(arrowheads)
do not correspond
to calcifications
seen on CT. There is marrow
replacement
by
tumor (arrow).

Fig. 5.-Osteogenic
sarcoma
of the femur.
A, Preoperative
radiograph
shows osteoblastic
iosteal reaction
and a soft-tissue
mass (arrows).
B, Sagittal
SE 500/30 image shows proximal

(arrow)
marrow

lesion

with

per-

does
tumor

extent

tissue

component

can extend

of both

beyond

of these

the cortical

neoplasms.

not involve

mass

(arrowheads).

Tumor extends proximally

into marrow
epiphysis
as predicted
by MR (B).

12 cm and

in marrow

The inhomogeneous
MR signal of osteogenic sarcoma did
not correlate
with the histologic
distribution
of chondroid
or
osteoid tumor matrix. We did find on MR that tumor spread
into the bone marrow

and soft-tissue

C, Gross specimen.

or soft-

Decreased

signal in bone marrow

on Ti -weighted
images,
however,
is
not specific for tumor since avascular
necrosis or any marrowreplacing
process can produce the same signal decrease
[14,
16].
In six cases MR was superior
to CT in defining soft-tissue

Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS. For personal use only; all rights reserved

AJR:148,

MR

February 1987

OF

OSTEOGENIC

AND

EWINGS

SARCOMA

32i

Fig. 7.-Ewings
sarcoma
arising from iliac
wings in two different
patients
shows displacement of right (A) and left (B) iliac arteries (arrows),
inferior vena cava (A, arrowhead),
and
bladder (B, arrowhead).

and thus has dark signal

Zimmer et al. [5] believed that cortical disease

shown
on longer SE sequences
(TE =
60 msec, TR = 2,000 msec), in 15 cases with evident cortical
disease on MR this was seen as well on Ti - as on T2weighted MR pulse sequences. Cortical disease was appreciated more consistently
on axial images since problems
arose with partial voluming of signal intensity on coronal and
sagittal images; and corticalouthne
was not as sharply defined
when the long axis of the bone took a more angled or oblique
course, as has been reported
by other workers
[14]. Periosteal reaction could be identified in eight cases by MR.

sequences.
Tumor invasion
of bone cortex can
because
of its prolonged
Ti and T2 relaxation

In both osteogenic
and Ewings
sarcoma,
MR can show
the relationship
of tumor to vasculature
without
the use of

contrast and tumor extension.

soft-tissue
contrast between
Ti -weighted
MR

was

T2-weighted
images enhanced
tumor and muscle better than

images.
found

to be inferior to CT in showing

soft-tissue

calcification
and tumor matrix mineralization
in all i 6 cases in
which calcium
was shown by either CT or plain films. This

was more of a problem

in the interpretation

of osteogenic

sarcomas.
MR can show

cortex

by both

Bone cortex
on all pulse
be identified

thickening,

infiltration,

osteogenic

sarcoma

lacks mobile

protons

and destruction

and

Ewings

of the
sarcoma.

times. Although
was

more

obviously

322

BOYKO

ET AL.

AJR:148, February 1987

Downloaded from www.ajronline.org by 180.253.21.101 on 09/27/13 from IP address 180.253.21.101. Copyright ARRS. For personal use only; all rights reserved

Fig. 8.-Ewings
sarcoma
arising from spine.
A, Coronal
SE 500/30
image shows soft-tissue mass (arrows) and decreased
marrow signal
of involved vertebral body (arrowhead).
B, Coronal
SE 2000/30 image after therapy
shows decrease
in size of soft-tissue mass (arrows).

contrast

agents.

of MR to image

Part of this advantage

in the coronal

rests with the ability

and sagittal

planes.

Pettersson

et al. [1 0] have also commented

on the advantage of MR in
providing information on tumor relationship to major vessels.
Although CT can be used in follow-up
to monitor
the
response

to therapy

of osteogenic

sarcoma

[1 7],

we

found

MR to be a suitable alternative.
Interval growth of tumor
shown by MR allowed cessation of chemotherapy
and early
amputation
in three cases of osteogenic
sarcoma. MR was
useful as a follow-up to monitor therapy response of three
cases of Ewings sarcoma.
In summary, we have found MR to provide adjunct information in the preoperative
planning and posttherapy
followup in cases

of Ewings

and

osteogenic

sarcoma.

Images

obtained in multiple planes and with both Ti- and T2-weighted

pulse sequences are necessary to evaluate the bone-marrow
and soft-tissue extent of these two tumors. The axial plane
allows best assessment of bone cortex without partial-volume
artifacts.

display
provide

Sagittal

and

coronal

planes

provide

an

excellent

of the relationship
of tumor to major vessels and
images of the entire length of a bone for evaluation

of proximal

bone-marrow

or epiphyseal

extent

of tumor

and

skip lesions. In general, Ti -weighted

images provide the
greatest contrast between tumor and marrow, T2-weighted
images show the most contrast between tumor and muscles,
and tumor-infiltrating-bone
cortex
may be seen with either.
One limitation of MR, compared with CT, is the inability of
MR to definitively
identify tumor matrix calcification
and
penosteal

reaction

in all cases.

healthy volunteers

1982;144:549-552
Moon KL Jr. Genant HK, Helms CA, Chafetz NI, crooks LE, Kaufman L.
Musculoskeletal
applications
of nuclear magnetic resonance.
Radiology

1983;147:161-17i
3. Brady TJ, Rosen BR, Pykett IL, McGuire MH, Mankin HJ, Rosenthal Dl.
NMR imaging of leg tumors. Radiology
1983; 149:181-187
4. Berquist TH. Magnetic resonance imaging: preliminary experience
in orthopedic

radiology.

Magnetic

5. Zimrner WD, Berquist

Resonance

Imag

1984;2

:41-52

TH, McLeod RA, et al. Bone tumors: magnetic

versus computed
tomography.
Radiology
1985;

resonance
imaging
155:709-718
6. Bloem JL, Falke THM, Taminiau AHM, et al. Magnetic resonance imaging
of primary malignant
bone tumors. Radiographics
1985;5 :853-886

7. Baker HL Jr, Berquist TH, Kispert DB, et al. Magnetic resonance imaging
in a routine

clinical

setting.

Mayo

Clin Proc

1985;60:75-90

8.

Hudson TM, Hamlin DJ, Enneking WF, Pettersson H. Magnetic resonance

imaging of bone and soft tissue tumors:
early experience
in 31 patients
compared
with computed
tomography.
Skeletal Radiol 1985;13: 134-146
9. Pettersson H, Hudson T, Hamlin D, et al. Magnetic resonance imaging of
sacrococcygeal
tumors. Acta Radiol (Diagn] (Stockh)
1985;26: 161-165
10. Pettersson H, Hamlin DJ, Mancuso A, Scott KN. Magnetic resonance
imaging
of the musculoskeletal
system.
Acta Radiol (Diagn]
(Stockh)
1985;26:225-234

1 1 . McMilIin KI, Pennes DR. Aisen A, Braunstein, Martel W. MRI and CT

evaluation of primary bone and soft tissue tumors. Presented at the annual
meeting of the American
Roentgen
Ray Society, Boston, April 1985
12. Cohen MD, Klatte EC, Baehner R, et al. Magnetic resonance
imaging of
bone marrow

disease

in children.

Radiology

1984;1 51:715-718

13. Fletcher BD, Scoles PV, Nelson AD. Osteomyelitis in children: detection
by magnetic resonance. Radiology
1984;150:57-60
14. Totty WG, Murphy WA, Ganz WY, Kumar B, Daum WJ, Siegel BA.
Magnetic resonance imaging of the normal and ischemic femoral head.
AJR 1984;143:1273-1280

15. Tjon-A-Tham
in Paget

RT, Bloem JL, Falke THM, et al. Magnetic resonance imaging

disease

of the skull. AJNR

1985;6:879-881

16. Daffner RH, Lupetin AR, Dash N, Deeb ZL, Sefczek RJ, Schapiro RL. MRI
in the detection
of malignant
infiltration
of bone marrow. AiR
1986;146:353-358

REFERENCES
1 . Brady TJ, Gebhardt

2.

17. Mail JT, Cohen MD, Mirkin D, Provisor AJ. Response of osteosarcoma
MC, Pykett IL, et aI. NMR imaging of forearms in
and patients with giant-cell tumor of bone. Radiology

preoperative
uation. AiR

intravenous
high-dose
1985;144:89-93

methotrexate

chemotherapy:

to

CT eval-