Nov 1, 1985 - Malignant bone tumors can be resected and function can be restored by ..... MRI demonstrates a large tumor (ar- .... with giant cell tumor of.
F-.
Magnetic resonance imaging of primary malignant bone tumors
index terms: Bones
magnetic Magnetic
resonance resonance
- -
=
Imaging imaging
staging
Johan
L. Bloem,
M.D.*
Allan
Theo
H.M. Falke,
M.D.*
Robert
M. Steiner,
Everett
E.H. Overbosch,
Anthony Joost
H.M. Taminiau, Doornbos,
From
the
Departments
Center,
Medical
The
tuiden
Center,
(s);Department Radiology,
Gasthuis,
Radiology,
of
Deventer,
Thomas
Jefferson
of
r_’r
I
Uni-
Hospital, Philadelphia (1) Study supported by the Netherlands Cancer FoundaGrant
1KW
Address J.L. Bloem, Diagnostic Hospital,
requests
Department
M Leiden,
to of
Radiology, Universily Leiden, Rijnsburger-
weg 10,2333
L1I
.!
TABLEI
!,
Important Criteria in StagiflgofPrifflaryBon#{235}T...... Us Extension
I
Extraosseous
I
8589.
reprint M.D.,
Jr., M.D.*II
The
versify
tion
des Plantes,
St. Geer-
(IB; Department
Netherlands
M.D.*
During the last decade, there has been a growing awareness of the value of wide, but local en bloc resection of malignant soft tissue and skeletal tumors followed by reconstructive surgery to restore function (5, 10, 16, 18, 20). Resection and reconstructive procedures in patients with primary malignant bone tumors can, however, be performed only if accurate information can be obtained concerning the extent ofthe tumor with respect to anatomic planes and concerning the involvement of vessels and nerves flable I, Figure 1) (1 1).
Presently atthe Department of Radiology and Radiological Sciences, Vanderbilt UniNashville Diagnostic
M.D.* #{182}
Introduction
Netherlands.
versily
Ziedses
of
Leiden,
M.D4
The authors report a refrospective study that suggests that MRI may be superior to CT for the preoperative evaluation of bone tumors
Diagnostic Radiology (#{176}), Orthopedic Surgery (Ii, and Oncology (fl, of the Universily Medical
Oosterom,
M.D.*
B. George
THIS EXHIBIT. A SELECTION OF THE RADIATION THERAPY PANEL, WAS DISPLAYED AT THE 70TH SCIENTIFIC ASSEMBLY AND ANNUAL MEETING OF THE RADIOLOGICAL SOCIETY OF NORTH AMERICA. NOVEMBER 25-30. 1984, WASHINGTON. D.C.
M.D.t
1. Van
I
The
Netherlands.
Volume
5, Number
6
November,
#{149}
1985
‘
RadioGraphics
853
MRI of malignant
tumors
Bloem
C
0 U
I
Figure 1 Malignant bone tumors can be resected and function can be restored by reconstructive procedures, as in this patient with Ewing’s sarcoma of the tibia, ifaccurate information concerning the intra- and extraosseous extensian oftumor can be obtained.
854
RadioGraphics
November,
#{149}
1 985
‘
Volume
5, Number
6
et al.
Bioem
MRI ofmalignant
et al
tumors
0
a.
This change in operative approach has been facilitated by improved surgical technique, by computed tomography (CT) and other imaging technologies (1, 3, 4, 6, 9, 13, 15, 17)which provide abundant information about the extent of tumor involvement, and by an increase in the life expectancy of the patient owing to effective chemotherapy (1, 7, 8, 14, 16, 19, 21, 22). The combined modalily approach with preoperative chemotherapy requires that the response of the tumor to the chemotherapeutic agents be determined promptly in order to avoid unproductive delay in surgical treatment (21).
In this study, we shall compare the abilily of CT, arteriography, and magnetic resonance rnaging (MRI) to provide precise and useful anatomic information as part of the preoperative examination of patients with primary malignant bone tumors. Magnetic resonance images of different stages of intra- and extracompartmental tumor extension are displayed for two different pulse sequences, and in addition, the effect of chemotherapy ofthe lesions on their MR images is illustrated.
Materials
and
Twenly-one patients with malignant bone tumors were studied in retrospect. The pathology of the tumors and their locations are summarized in Table II. The abilily of MRI to define the anatomy of the tumors was compared with that of plain films, CT, and arteriography. All of the patients, with the exception ofthe one with non-Hodgkin’s lymphoma, were operated upon, and in each case the gross
Methods and microscopic pathology were correlated with the diagnostic images. The resected specimens were sliced in axial, sagittal, or coronal planes corresponding to the planes used in the MRI studies. In addition, three patients with osteosarcomas and one with a Ewing’s sarcoma were examined by MRI before and after chemotherapy to establish whether chemotherapy altered the MR images.
TABLEIi Material
Histology
Localization
Field strength
36 18 16 16
osteosarcoma osteosarcoma osteosarcoma teleangiectatic
femur femur femur femur
0.15 T 0.15 T 0.15 T 0.15T
F
14
osteosarcoma
femur
0.15 T
F
11
osteosarcoma
femur
0.15 T
F F F M
15 17 13 47
femur fibula fibula tibia
0.5 T 0.151 0.151 0.151
II
M
12
F
23 22
osfeosarcoma osteosarcoma osteosarcoma clearcell chondrosarcoma chondrosarcoma chondrosarcoma
pelvis lumbarspine
0.151 0.151
13
F
22
chondrosarcoma
lumbarspine
0.151
14
M
15
F
30 16
chondrosarcoma Ewing’ssarcoma
scapula tibia
16
M
25
Ewingssarcoma
femur
17
M
18 19
M M
10 57
Ewing’ssarcoma fibrosarcoma
toe tibia
0.5 T 0.151 0.5 T 0.5 1 0.151
16
fibrosarcoma
femur
0.15 T
20
M
21
F
44 24
fibrosarcoma non-Hodgkin’s
pelvis scapula
0.5 1 0.5 1
Case #
Sex
Age
I 2 3 4
F M M F
5 6 7 8 9 10
osteosarcoma
lymphoma
Volume
5, Number
6
November,
1985
‘
RadioGraphics
855
C
0 0
MRI of malignant
tumors
Bioem
All of the patients were examined on either a 0.15 T or 0.5 T Philips MRI system1 in transverse, sagiffal, or coronal planes. Patient access and image qualily were improved for the 0.1 5 T resistive system by the use of a thick tube saddle coil with geometry tailored to the examination of the lower extremities (Figures 2 and 3). The slice thickness was I cm, and data acquisition was performed by single slice technique, using a two dimensional Fourier transformation and a matrix of 256 x 256. To keep the examination time within reasonable limits, the CT findings were used as a guide to select representative slices for MRI. For the same reason, the number of pulse sequences
was limited to inversion recovery (IR) with inversion time (TI) of400, echo time (TE) of 30, and repetition time (TR) of 1400 msec; and spin echo (SE) with TE of 50, and TR of I 000 msec. Computed tomographic studies were performed on a Pfizer 450 AS & E scanner with a scan time of 4.9 sec, slice thickness of 9 or S mm, and a matrix of 256 x 256. When necessary, studies were performed both before and after the intravenous administration of contrast material for the identification of major blood vessels.
I
Philips Medical
Systems.
Eindhoven,
Figure 2 Asaddle coil especially designed for imaging extremities at 0.1 5 1. The use of a thick copper tube as the conducting material made possible an open design which
856
RadioGraphics
afforded
excellent
November,
#{149}
patient
1985
access.
Volume
#{149}
et ai.
5, Number
6
The Netherlands.
Bloem
et ai.
MRI ofmalignant
tumors
Figure 3 In addition to excellent patient access, the especially designed coil provided for significant improvement in signal to noise ratio. (A & B) are comparative images made with the manufacturer’s coil (A) and the thick tube coil (B) at the
same level and with identical noise ratio in (B).
pulse sequences.
Pulse
Note the improved
Sequence
and
As has been stated, image contrast in MRI varies with the pulse sequence used (24). The influence of the pulse sequence on the signal intensities of the primary bone tumors in this study is demonstrated in ten patients who were examined with both IR and SE techniques at 0. 1 S T. The results are summarized in Table III. In the absence of hemorrhage, all tumors had a relatively low signal intensily on IR, suggesting increased TI (Figure 4C). The signal intensify on SE strongly depended on the presence of calcifications. In the presence of diffuse calcifications, the tumor had a relatively low overall signal intensity, suggesting low spin density or a shortT2 or both (Figure 5). In the absence of calcifications, the tumor had a high signal intensily on the SE image in 90% of cases,
Volume
5, Number
signal to
Signal
intensities
suggesting an increased T2 (2, 12) (Figure 4D). As described below, the presence of extensive tumor calcification was an important factor in the abilily of MRI to determine soft tissue extent.
1
TABLEIII
Observation of Tumor Signal Intensity Relative to Bone Marrow Signal intensity (0.15 T) SElechnique Type of Lesion
Higher
Osteolytic
Osteosclerotic
6
‘
November,
IRlechnique
Intensity
1985
Intensity Lower
Higher
Lower
13
2
0
8
0
9
0
5
‘
RadioGraphics
857
MRI of malignant
tumors
Bloem
Bone
Marrow
Involvement the SE image because oftumor calcifications. Definition of affected bone marrow was more accurate with MRI than with CT in all cases because of superb contrast and the absence of beam hardening artifacts in the MR images (Figure 6). Unlike CT and technetium isotope studies, MRI may be helpful in differentiating marrow replacement by tumor from hyperemic osteoporosis (Figure 7).
Bone marrow has a high signal intensily on both TI and T2 weighted images. Therefore, contrast between tumor and normal marrow was higher on our IR images than on our SE images in seven out of the ten cases that were studied by both pulse sequences (Figure 4). In one case, no difference could be observed. In two cases, affected bone marrow was beffer delineated on
Cortical
Invoivement
Although cortical destruction is best demonstrated on plain radiographs and CT scans, it could be depicted on MR images. Cortical destruction was correctly diagnosed in I 8 cases on both CT and MRI. Normal cortical bone on MRI was
Soft
Tissue
On the IR image, contrast between the low signal intensily of tumor and the low signal intensily of normal or slightly edematous muscle was poor in all patients (Figures 4C and aD). On the SE images, however, contrast between the high signal intensity of tumor and the low signal intensify of normal muscle was high in all patients (Figures 4D, 6C and 7C). In six of ten patients examined with
more
858
both
pulse
accurate
sequences,
tumor
delineation
identified as a thin line of low signal intensify. In cases of destruction, interruption of this line could be appreciated owing to an increase in signal intensily on the SE images (Figure 8).
Extension Soft tissue extension ofthe neoplasm was better delineated on MRI than on CT, especially when no calcifications were present (Figure 9). Focal edema had an even higher signal intensify than tumor on the SE image, as is illustrated in one patient with an osteosarcoma accompanied by a compartment syndrome (Figure 10).
was
on the SE image.
RadioGraphics
November,
#{149}
1 985
Volume
#{149}
et ai.
5, Number
6
Bloem
MRI of malignant
et al
Joint The free choice of imaging planes permiffed accurate visualization of the lationship of the tumor to the adjacent 18 cases. Computed tomography and
Involvement
with MRI spatial rejoint in all MRI dem-
Vascuiar With MRI it was sels and to determine in all cases without (Figure 1 1). Normal tumor and nearby excluded vascular
tumors
possible to visualize large yestheir proximily to the tumor the use of a contrast medium tissue situated between the large vessels correctly involvement by MRI and en-
onstrated joint involvement correctly in nine cases (Figure I 1), but CT wrongly suggested joint involvement in two normal knee joints (Figure 4). All normal joints were correctly identified.
invoivement hanced CT (Figure 12). In the absence ofa clear interface, MRI and CT did not permit one to differentiate tumor invasion of the vessel wall from vessel compression without direct invasive growth (Figure 13).
Chemotherapy Three of four patients treated with chemotherapy (cis-platinum, vincristin, and adriamycin) improved clinically. In these three patients, CT and MRI demonstrated a reduction in tumor volume. In addition, a decrease in signal intensify on the spin echo MR image was observed in Iwo patients (Figures 14 and 15). In one patient, the decrease in signal intensify could be explained by the development oftumor calcifications (Figure 14). In the other patient, the change was thought to be due to a Change in the spin densily or relaxation time ofthe tissue following chemotherapy, or both (Figure I5).The possibiliiythatthedecrease in signal intensify was caused by accumulated cisplatinum in the tumor acting as a paramagnetic substance was considered less likely on the basis
Volume
5, Number
of in vitro experiments. The experiments consisted of measuring TI and T2 relaxation times of liver and muscle in WAG/RY rats treated with I mg platinum and comparing them with the values obtamed in a normal control group. No differences in the relaxation times ofthe specified tissues were observed. These findings were supported by determinations ofthe relaxation times of I millimolar cis-platinum solutions. Between water and the cisplatinum solutions, the difference in TI was 1000 msec and the difference in T2 was 997 msec. Although the cis-platinum in the solution decreased relaxation times, the paramagnetic effectfor such concentrations is low. Therefore, it is not likely that the much lower concentrations in tumors could produce a measurable effect.
6
‘
November,
1985
RadioGraphics
#{149}
859
MRI of malignant
tumors
Bioem
Case
Figure 4A Radiographs femur.
ofa 16 year old boy show an osteolytic fibrosarcoma
19
in the distal
Figure 4B A CT scan atthe level ofthe patella demonstrates (arrows).
860
RadioGraphics
November,
#{149}
1985
Volume
#{149}
possible joint involvement
5, Number
6
et ai.
Bloem
MRI of malignant
et al.
tumors 0 a
U,
0
‘0
FIgure4C Sagiffal view, IRtechnique, 0.15 1. Viable tumor emits a low intensity signal (open arrows) relative to that of normal bone marrow. The area of necrosis (solid arrows) has an even lower signal intensity.
The central
area of high signal
intensity (curved arrow) represents hemorrhage.
Figure 4D Sagiffal view, SEtechnique, 0.15 1. Tumor, necrosis, and hemorrhage (arrows) have a high signal intensity almost identical to the signal intensity of normal bone marrow. Note the normal retropatellar space (arrowhead) which excludes the joint involvement suggested on CT.
Figure 4E The gross specimen sliced in the sagittal plane. The extent of viable tumor and the area of necrosis correspond well with the MR images. The bony ridge (arrows) atthe proximal border of the tumor can be identified on the MR images as an area of decreased signal intensity on IRand SE. The joint was not involved.
Volume
5, Number
6
‘
November,
1985
‘
RadioGraphics
861
MRI of malignant
tumors
Bloem
Case 2
Figure 5A These are the radiographs of an 18 year old man with a mixed osteosclerotic and osteocytic osteosarcoma of the distal femur.
862
RadioGraphics
‘
November,
1985
VolumeS,
#{149}
Number
6
et ai.
Bioem
et al
MRI of malignant
tumors 0 a U,
0 Pt)
B
Figures 5B & C Sagittal and transverse views, SEtechnique, teosclerotic
part ofthe
0.15 T.The os-
tumor is seen as an area of low signal
intensity (arrows) extending
into the medial condyle.
Figure 5D Comparison with the gross specimen. The area of calcifications in the tumor corresponds
to the darker zone on the
spin echo image. Volume
5, Number
6
‘
November,
1985
‘
RadioGraphics
863
MRI of malignant
tumors
Bloem
Case
7
Figure 6A Radiographs of a 15 year old girl show an osteosarcoma of the distal femur.
Figure 6B On CT, the extent of marrow involvement (arrows) is difficult to
ascertain.
864
RadloGraphics
‘
November,
1985
VolumeS,
#{149}
Number
6
et al.
MRI of malignant
Bioemetai.
tumors 0 a
UI
0
Figure 6C SE,O.5T.The osteolytic part of the tumor has a high signal intensity (small arrows) and is, therefore, difficult to separate from the normal bone marrow which has a similarly high signal intensity.
Transverseview,
Figure 6D Transverse
view, IR, 0.5 1. There is a clear delinea-
tion of intramedullary sequence,
appears
tumor which, with this pulse as a zone of low signal inten-
sity (small arrows). Normal marrow is clearly dentified as a residual rim producing a signal of high intensity (long arrows).
Figure 6E Transverse section ofthe gross specimen. There is a striking similarity in the spatial distribution of normal bone marrow, cortical bone, and tumor between the MR image and the gross specimen.
Volume
5, Number
6
‘
November,
1985
RadioGraphics
#{149}
865
MRI of malignant
Bloem
tumors
Case
9
Figure7A Radiographs of this 13 year old girl show an osteosarcoma of the fibula.
Figure lB CT at the level of the proximal diaphysis demonstrates fibular destruction and an ill defined mass containing flecks of calcium (arrows).
Figure 7C This transverse SEimage, 0. 15 T, at the level of the proximal diaphysis outlines the tumor more clearly (arrows).
866
5, Number
RadioGraphics
November,
#{149}
1985
Volume
#{149}
6
et al.
Bloem
MRI of malignant
et al.
tumors 0 a 0
b Figure 7D The low attenuation of the proximal fibular metaphysis shown on this CT scan suggests tumor involvement.
Figure 7E This technetium
(arrow)
thetumorarea
scintigram
shows high uptake in
extendingtothelevelofthe
proxi-
mal metaphysis. This supports the suggestion of metaphyseal involvementfrom the CT scan.
Flgure7F MR images,
SE and IR, atthe
site ofthe
(arrows) without replacement existence
of osteoporosis
metaphysis
show normal
by tumor. The surgical specimen
but absence
oftumor
involvement
Volume
bone marrow
confirmed atthis
5, Number
the
level.
6
‘
November,
1985
‘
RadioGraphics
867
MRI of malignant
tumors
Bioem
Case
Figure8A Radiographs
868
of a I 7 year old girl show an osteosarcoma
RadioGraphics
November,
#{149}
I 985
‘
8
in the fibula.
Volume
5, Number
6
et ai.
Bioem
MRI of malignant
et al.
tumors 0 0 U,
0
Figure 8B SEimage, sagittal view, 0.15 1. Normal low intensity lines (white arrows) representing cortical bone can be identified atthe level ofthe diaphysis and epiphysis. Atthe level of the metaphysis, tumor (open arrows) can be identified and interruption of cortical bone is present.
Figure 8C Sagittally sectioned gross specimen. Tumor at the level of the metaphysis destroys cortical bone. At the level of the normal epiphysis and diaphysis, normal cortical bone can be seen.
Volume
5, Number
6
‘
November,
1985
‘
RadioGraphics
869
MRI of malignant
tumors
Bioem
Case
10
Figure 9A This isthe radiograph ofa 44 year old woman who had a fibrosarcoma of the right iliac bone.
8 70
RadioGraphics
November,
#{149}
I 985
Volume
#{149}
S. Number
6
(arrows)
et al.
Bioem
MRI of malignant
et al.
tumors 0
a U,
0 .
a
Figure 9B The x-ray attenuation ofthe tumor equals the affenuation of normal muscle. Therefore, on CT, only destruction ofcortical bone (arrows) and distortion (open arrow) of normal contours can be seen.
Volume
5, Number
6
‘
November,
I 985
#{149}
RadioGraphics
871
MRI ofmalignant
tumors
Bioem
a 0 U,
a
Figure 9C Coronal view, SE,0.5 1. Contrast between the high signal intensity of the lobulated tumor (arrows) and surrounding structures allows accurate definition of softtissue extension. Tumor infiltrates the spinous muscle comportment. The high intensity artefact is caused by a metallic suture.
0
Figure 9D Transverse view, SE,0.5 T.Tumor invades the iliac bone, gluteus, and spinous muscle compartments (arrows).
Figure 9E The transversely correlates well image. Gluteus compartments
8 72
RadioGraphics
sliced gross specimen with the transverse SE and spinous muscle are involved.
November,
#{149}
1985
Volume
#{149}
5, Number
6
et ai.
Bloem
et ai.
MRI ofmalignant
Case
tumors
I
Figure lOC Transverse view, SE,0.15 T.The softtissue extension (high intensity signal) containing calcifications (low intensity signals) is well defined. In addition to the clearly demarcated tumor outline, the edematous compartment iswell shown as a zone of high signal intensity.
Figure IOA This radiograph of a 36 year old woman shows an osteosarcoma in the femoral diaphysis. A large soft tissue mass is present. Figure lOB Abundant calcifications make it possible to delineate the tumor (white arrows) on CT. The zone of low affenuation (curved black arrows) within the tumor is due to an edematous compartment.
Volume
5, Number
6
‘
November,
1985
‘
RadioGraphics
873
MRI ofmalignant
Bloem
tumors
Case
4
Figure hA These radiographs of a 16 year old girl show an extensive lesion that proved to be a teleangiectatic osteosarcoma ofthe femur. Displacement ofthe patella is demonstrated.
Figure IIB Transverse view, SE,0.15 T.A large softtissue mass of high signal intensity is seen to be invading the suprapatellar pouch (arrows). Note, also, thatthe femoral artery (curved arrow) is dorsally displaced.
874
RadioGraphics
November,
#{149}
1985
.
Volume
5, Number
#{243}
et al.
Bioem
MRI ofmalignant
et al.
tumors 0 a U,
0
Figure hlC Sagiffal view, SE,0. 15 T.A large mass is seen (white arrows) that is displacing the patella anteriorly and is infiltrating the retropatellar space (long arrow). The femoral vessels, represented by a line ot decreased signal intensity (small arrows), is contiguous with the tumor mass.
Figure I ID Sagittal view, IR, 0.15 T. The large hemorrhagic compartment of the tumor (arrows) is responsible for the relatively high signal intensity seen in this IR image. Extension of the lesion into the knee joint is confirmed.
Figure lIE This sagitfal section ofthe gross specimen demonstrates a large hemorrhagictumor involving thejoint. Notethe sclerotic part ofthe tumor which is visible as an area of decreased signal intensity on SEand IR.
Volume
5, Number
6
‘
November,
1985
‘
RadioGraphics
875
MRI ofmalignant
tumors
Bioem
Case 21 Figures 12A& B This 20 year old woman had a non-Hodgkin’s lymphoma of the scapula. Axillary DSA (A) and arteriography (B) show the arterial walls to be smooth and regular.
876
RadioGraphics
November,
#{149}
1 985
Volume
#{149}
5, Number
6
et al.
Bloem
MRI ofmalignant
et al.
tumors 0 a U,
0
Figurel2C
On CT, afatplane can be identified between the major vessels (v) and the tumor surface (t). Note the destruction ofthe scapula (arrowheads).
Figure Coronal view, SE,0.5 1. MRI demonstrates rows) encasing the joint. -__
a large tumor (ar-
Volume
5, Number
Figure I__ Sagiffal view, SE,0.5 T.Tumor (arrows) is destroying the scapula (arrowheads) and extending into the subscapular muscle (curved arrows). The artery and vein (open arrows) are separated from the tumor margin by interposed axillary fat.
6
November,
#{149}
1985
‘
RadloGraphics
877
MRI of malignant
tumors
Bloem
Case
14
Figure 13A This 30 year old man with multiple exostoses had a recurrence of chondrosarcoma after the resection of a tumor located in the scapula. The axillary artery does not show involvement.
878
RadioGraphics
#{149}
November,
1 985
Volume
#{149}
5, Number
6
et al.
Bloem
et al.
MRI ofmalignant
tumors 0 a U, 0
Figure 13B Sagiffal view, SE,0.5 T.The tumor (arrows) is clearly defined. No clear fat interface is seen between the tumor and the axillary artery (arrowhead), and,therefore, a definite statement aboutvascular involvement cannot be made. A = anterior; P = posterior
Figure 13C Sagittal sections ofthe gross specimen. At surgery, a close relationship between the axillary vessels and the tumor was found. The tumor could be easily separated from the vascular bundle, however. M = tumor mass; S = scapula; ss = subscapularis; is = infraspinatus.
Volume
5, Number
6
‘
November,
1985
RadioGraphics
#{149}
879
MRI of malignant
tumors
Bloem
Case
16
oronaI view, , u. I I , ewing’s sarcoma ol ..e femur s. .is a zone of slightly increased signal intensity (open arrow) along the thickened cortex following chemotherapy. Diminished signal intensity is seen in the marrow (arrows). This finding is consistent with calcification. p..
880
.
RadioGraphics
..
November,
#{149}
1985
Volume
#{149}
5, Number
6
et al.
Bioem
MRI of malignant
et al.
tumors 0
a U, 0 0’
Figure I4B The recalcification in the area of the diaphysis after chemotherapy firmed by these plain radiographs.
VolumeS,
Number#{243}
is con-
November,
#{149}
1985
‘
RadioGraphics
881
MRI of malignant
tumors
Bioemetai.
Case
Figure h5A Sagiffal view, SE,0. 15 T. Osteosarcoma prior to chemotherapy. The softtissue extension ofthe tumor (arrows) shows high signal intensity, while the intramedullary sclerotictumor (arrowheads) shows low signal intensity. Also note the joint involvement.
882
RadioGraphics
November,
#{149}
I 985
Volume
#{149}
3
Figure I5B Sagittal view, SE0.15 T. Following chemotherapy, a reduction oftumor volume as well as a decrease in signal intensity can be observed.
5, Number
6
Bloem
MRI ofmalignant
et ai.
tumors 0
a U, 0 ‘3
Figures 15C & D Lateral conventional radiographs before (C) and after (D) chemotherapy firm the absence of extensive soft tissue calcifications.
Volume
5, Number
con-
6
‘
November,
1985
‘
RadioGraphics
883
MRI of malignant
tumors
Bloem
Case
et al.
15
Figure 16A This is the radiograph of a l#{243} year old girl with Ewing’s sarcoma in the tibial diaphysis.
(See opposite
page
for Figures
16B and C.)
t._
1
‘
.
.
.
. -.-
..i.
. . -.,
Figure 16D Transverse section of gross specimen. In the softtissue, fibrosis and some viable tumor was present. In bone marrow, only fibrosis without viable tumor was found.
884
RadioGraphics
#{149}
November,
1985
‘
Volume
5, Number
6
Bloem
MRI ofmalignant
et ai.
tumors 0
a U,
0 UI
Figure 16B Transverse view, SE,0.15 T and CT. Prior to chemotherapy, the softtissue extensian (open arrows) and the marrow involvement (arrow) are better appreciated on the MR image (left) than on CT (right).
Figure 16C After chemotherapy, MRI shows essentially normal signal intensity in the soft tissue as well as in the marrow (arrow) compared to the examination before chemotherapy. This indicates a satisfactory response.
Volume
5, Number
6
November,
#{149}
1985
‘
RadioGraphics
885
MRI ofmalignant
tumors
Bioem
et al.
Conclusions I Although a prospective study is not yet available, MRI seems to be superior to CT for the preoperative staging of primary bone tumors because of its superior delineation of intra- and extraosseous tumor extension. 2. Involvement of coritcal bone and soft tissues including vascular structures and joint spaces is best shown with the SE (spin echo) technique (TR I 000, TE 50 msec). 3. The extent of marrow involvement is best shown with the IR (inversion recovery) technique (TI 400, TR I 400, TE 50 msec). .
4. Osteopenia owing to tumor invasion of the bone marrow can be differentiated from osteoporosis caused by tumor related hyperemia with MRI. 5. The decrease in signal intensify of a bone tumor seen on SE images after chemotherapy may be due to calcifications or may be related to a change in the MR characteristics of the tissue. The change seems to be a function of tumor regression. Therefore, MRI may be useful in monitoring the effect of chemotherapy.
References I
.
2.
3.
4,
5.
6.
7.
8.
9.
Apple JS, Martinez 5, Nelson PA, Rosenberg ER, Harrelson JM. Bowie JD. Sonographic correlation in extremity soft tissue masses. Noninvas Med Imag 1984; 1:75-81. BradyTJ, Gebhardt MC, Pykeff IL, etal. Imaging of forearms in healthy volunteers and patients with giant cell tumor of bone. Radiology 1982; 144:549-552. Brady TJ, Rosen BR. Pykeff IL, McGuire MH, Mankin HJ. Rosenthal Dl. NMR imaging of leg tumors. Radiology 1983; 149:181-187. Zimmer WD, Berquist TH. McLeod PA, et al. Bone tumors: Magnetic resonance imaging versus computed tomography. Radiology 1985; 155:709-7 18. Campbell CJ. Place of resection in the management of primary bone tumors. Can J Surg 1977; 6:518-521. Cohen MD, Klaffe EC, Baehner R, Smith JA, Martin-Simmerman P. et al. Magnetic resonance imaging of bone marrowdisease in children. Radiology 1984; 151:715-718. DeSantos LA, Goldstein HM, Murray JA, Wallace S. Computed tomography in the evaluation of musculoskeletal neoplasms. Radiology 1979; 131 :431-437. Destouet JM, Gilula LA, Murphy WA Computed tomography of long-bone osteosarcoma. Radiology 1979; I 31:439-445. Falke THM, Bloem JL, Taminiau AHM, et aI. 1. Application of MRI in the diagnosis and followup during treatment of bone tumors. Society of Magnetic Resonance in Medicine ThirdAnnual Meeting 1984:275.
10. Forfner JG. Kim DK, Shin MH. Limb preserving vascular surgery for malignant tumors of the lower extremily. Arch Surg 1977; 112:391-394. I I . Genant HK, Cann CE, Dalinka MK, DeSmet Aft,, Henrix P. Musculoskeletal radiology. Invest Padiol 1984; 19:530-542. I 2. Goldsmith M, Koutcher JA, Damadian P. NMR in cancer XIII. Application ofthe NMR malignancy indexto human mammary tumors. Br J Cancer I 978; 38:547-554. 13. Heelan PT, Caird Watson P. Smith J. Computed tomography of lower extremity tumors. MR 1979; 132:933-937.
14. Hinshaw WS, Andrew ER, Boffomley PA Holland GN, Moore WS, Worthington BS. An in vivo study of the forearm and hand by thin section NMR imaging. Br J Radiol 1979;
52:3#{243}-43. 15. Levine E. Lee KR, Neff JR. Maklad NF, Robeson PG. Preston DF. Comparison of computed tomography and other imaging modalities in the evaluation of musculoskeletal tumors. Radiology 1979; 131:431-437. 16. Miller TR. Surgical management of malignant bone tumors. Can J Surg 1977; 20:513-517. I 7. Moon KL, Genant HK Helms CA Chafetz NI, Crooks LE, Kaufman L. Musculoskeletal applications of nuclear magnetic resonance. Radiology I 983; 1 47:1 #{243}l -171. I 8. Morton DL. Eilber FR, Townsend CM, Grant TT, Mirra J, Weisdenburger TH. Limb salvage from a multidisciplinary treatment approach for the skeletal and soft tissue sarcomas of the extremily. Ann Surg 1976; 184:268-278. 19. Rosen G. Surwasirikul S. Kwon C, Tan C, Wu SJ, Beattie EJ, Murphy ML. High dose methotrexate with citrovorum factor-rescue and adriamycin in childhood osteogenic sarcoma. Cancer 1974; 33:1151-1163. 20. Simon MA Enneking WF. The management of soft tissue sarcomas ofthe extremities. J Bone Joint Surg (Am) 1976; 58:31 7-327. 21 . Smith J, Heelan PT, Huvos AG, Caparros B, Rosen G, et al. Radiographic changes in primary osteogenic sarcoma following intensive chemotherapy: Radiological-pathological correlation in 63 patients. Radiology I 983;
143:355-3#{243}0.
22. Yiu-Chiu
V. Chiu L. Multiple imaging modalities in the evaluation of musculoskeletal masses. J Comput Assist Tomogr 1982; 6:201-223. 23. Doornbos J, Booijen P, Grimbergen H. Falke THM. An improved saddle coil for resistive magnet systems. Magn Res Med 1985; Vol 2(5):(in press). 24. Ziedses des Plantes BG, Falke THM. De Boer JA. Pulse Sequences and contrast in magnetic resonance imaging. RadioGraphics 1984; 4(6):869-883.
The authors gratefully acknowledge the assistance of D. Ruiter, M.D., C. Ruygrok, Popkes, F. Noorderljk, M. Henry, and J. Fields in the preparation ofthis manuscript.
886
RadioGraphics
#{149}
November,
1985
Volume
#{149}
5, Number
6
M.
