Barbara Edward
C. Dangman, O’Rourke,
MD2 MD
J.
Fredric
#{149}
Osteomyelitis Gadolinium-enhanced Fifteen opsy-
pediatric
patients
osteomyelitis netic resonance teomyelitis
acute
M
in seven
pa-
tients, subacute in three, and chronic in five. Four patients had subperiosteal abscesses, one had a large associated soft-tissue abscess, and one had an intraosseous (Brodie) abscess. Areas
of active
inflammation
had
de-
creased marrow signal intensity on Ti-weighted images, increased signal intensity on T2-weighted images, and enhancement on Ti-weighted images obtained after gadopentetate dimeglumine administration (n = 10). Abscesses were rim enhancing (n = 3) or not (n = 2) with gadolinium-enhanced MR imaging. Nonenhancing areas presumably represented necrotic material. Gadolinium-enhanced MR imaging assisted in definition of the presence and extent of nonvascularized fluid collections within the bone and/or adjacent soft tissues
and
volvement
the
disease neous
It also
helped
debridement (n = 7) and drainage
soft-tissue
of bone
with
osteomyelitis.
surgical
extent
in patients
I
From
F.A.H.),
1992;
the
or
(n
5).
=
182:743-747
Departments
Orthopedics
Imaging’
resonance (MR) imaging has been found to be valuable in the diagnostic evaluation of acute and chronic osteomyelitis in adults and children (1-8). Previously described signal characteristics indude decreased signal intensity of the involved marrow space on Tiweighted images and increased signal intensity of inflammatory tissue on T2-weighted images (5-8). In animals, gadolinium-enhanced MR imaging has been shown to delineate areas of cellulitis (9) and to produce rim enhancement of abscesses (10). There have been a few reports of the use of gadopentetate dimeglumine in the evaluation of patients with vertebral AGNETIC
osteomyelitis
(11-13),
but
none
have
focused on nonspinal or pediatric osteomyelitis. This study was performed to determine whether gadoliniumenhanced MR imaging was valuable for orthopedic surgeons and interventional radiologists in caring for young patients with nonspinal osteomyelitis.
MATERIALS
Index terms: Bones, infection, 45.2121, 45.2132, 45.2141, 40.2111, 40.2112, 40.2113 #{149} Gadolinium #{149} Magnetic resonance (MR), contrast enhancement Radiology
MD
AND
METHODS
of Radiology
(B.C.D.,
(F.F.R.), and Medicine
(E.J.O.), Children’s Hospital, 300 Longwood Aye, Boston, MA 02115-5747. Received June 7, 1991; revision requested July 2; revision received October 14; accepted October 30. Address reprint requests to F.A.H. 2 Current address: Department of Diagnostic Imaging, Rhode Island Hospital, Providence, RI. e RSNA, 1992
Twelve patients who underwent imaging to rube out osteomyelitis were found to have other diagnoses, which included cellulitis (n = 3), avascular necrosis (ii = 2), tendinitis (n = 2), septic arthritis, fasciitis, fibrosis, osteoid osteoma, and occult fractune (n = I each). Patients with osteomyelitis that underwent only postoperative MR imaging were also excluded.
MR imaging
was performed
with
a 1.5-T
unit (Signa; GE Medical Systems, Milwaukee). Body, extremity, or 5-inch (12.5-cm) surface coils were employed as appropriate for the body part being imaged. Our
routine
imaging
tudinal
spin-echo
600/15-20
protocol
included
a longi-
(SE) TI-weighted
sequence
(repetition
time
400msec/ SE T2-
echo time msec), and longitudinal weighted (2,000/20,80), gradient-recalled echo (GRE) (33/13; flip angle, 30#{176}), or multiplanar GRE (250-450/13; flip angle, 30#{176}) sequences. Thirteen patients underwent T2-weighted sequences, and GRE and mubtiplanar GRE sequences were used in six patients each. Whenever possible, a
repeated was
longitudinal
performed
nous
Ti-weighted
immediately
administration
dimeglumine tories, Wayne,
series
after
intrave-
of gadopentetate
(Magnevist; Beniex LaboraNJ), to help distinguish be-
tween vasculanzed inflammatory tissue.
guide
of intraosseous open or percuta-
collections
F. Rand,
in-
chronic
of subperiosteal
fluid
Frank
#{149}
MR
were studied with mag(MR) imaging. Os-
was
MD
in Children:
with binonspinal
or culture-proved
A. Hoffer,
and nonvasculanized The dosage of gado-
From July 1988 to December 1990, we performed MR imaging in 15 pediatric patients with osteomyelitis confirmed by means of positive culture results (from blood, abscess drainage, or tissue aspirate), positive biopsy results, or both (Table 1).
linium employed was 0.2 mL/kg, with a maxiumum of 10 mL per dose. Five patients were not given gadopentetate
Patients
ranged
in age
19 years
(mean,
8.4 years).
dation, Berlex Laboratories, Wayne, NJ); one because of age less than 2 years; and three because they were studied before the contrast medium was approved by the U.S. Food and Drug Administration for pediatric usage. Surgical, bacteriologic, and histopatho-
from
14 months
Seven
to
patients
had cases five The (n = (n =
acute symptoms (2-14 days), three were subacute (1-3 months), and were chronic (7 months to 5 years). involved bones included the femur 5), pelvis (n = 3), tibia (n = 3), rib 2), humerus (n = 1), and radius (n = 1). Sites for imaging were chosen on the basis of clinical symptoms, in combination with abnormal plain radiographs
(n = 10) and/or abnormal findings on bone scans (n = 9) or gaffium scans (n 1). The mean interval between posi=
tive blood culture or pathologic evaluation and MR imaging was 5.67 days. Fourteen patients underwent imaging within 13 days or less (Table 1), and one underwent imaging 23 days after positive blood cubtune but only 3 days after a positive gab-
hum
scan.
dimeglumine: disease,
linium
usage
logic reports aging
one because
a stated
findings
of sickle cell
contraindication
(manufacturer’s
were
correlated
in each
to gado-
recommen-
with
MR im-
patient.
RESULTS Seven patients had acute elitis and eight had subacute chronic osteomyelitis. With
Abbreviations: SE = spin echo.
GRE
=
gradient-recalled
osteomyon unen-
echo,
743
Table 1 History and
Patient Age
Findings
in 15 Pediatric
Patients
Osteomyelitis
with
Duration of
No./
(y)/Sex
Culture
Results Pathologic
Symptoms
Site of Infection
Blood
Abscess
or Tissue
Acute 1/13/M 2/7/F 3/6/M 4/13/M 5/1/M 6/13/M 7/6/M
2 d 4 d 4 d 5 d lOd 2 d 2 wk
Distal nt femur Lt ischium Lt tibia Lt ilium Rtrib Lt ischium Lt tibia
S aureus S aureus Negative S aureus
S aureus ND S aureus
Saureus
ND S aureus S aureus,
Negative S aureus
I mo 3 mo 1 mo
Distal Distal Distal
It femur lt femur lt radius
Negative ND ND
11/12/M
4 y
Lt humerus
12/19/M 13/5/F
5 y 1 y
Lt femur Lt ischium
14/4/F 15/8/M
3#{189} y 7 mo
Lt rib Lt tibia
and
ilium
(hip
effusion)
Kiebsiella
Evaluation
ND
0
ND
10
ND ND
2 6
ND ND
23* 2
Osteomyelitis
0
Osteomyelitis Osteomyelitis Osteomyelitis
0 7 5
Osteomyelitis
Negative Negative Mixed flora
Chronic
Findings
Osteomyelitis
ND
Subacute 8/3/F 9/2/M 10/14/M
Interval (d) between MR Imaging and Positive Culture or Pathologic
(contaminants)
Osteomyelitis
ND
S aureus
Chronic osteomyelitis osteonecrosis
ND Negative
Mycobacterium
S aureus
tuberculosis
Osteomyelitis Granulomatous
and
6
osteomyeli-
1 8
tis
Note.-Lt * Three
=
days
left, ND = not done, after positive gallium
ND ND
S aureus
Escherichia
coli
Chronic osteomyelitis Osteomyelitis
b. 1.
2
Rt = right. scanning.
a. Figure
13
c.
MR images
obtained in a 7-year-old girl with fever and left hip pain of 4 days duration and Staphylococcus aureus septicemia. Plain radiographs of the pelvis were negative, hut a hone scan revealed evidence of osteomyelitis in the left ischium. (a) il-weighted SE 600/15 cornnal image of the pelvis shows decreased signal intensity in the left ischium (arrow) and thickening of the obturator internus muscle, indicating osteomyelitis and adjacent soft-tissue reaction. (b) 12-weighted SE 2,000/80 coronal image of the left hip obtained by using a 5-inch surface coil shows high signal intensity in the corresponding portion of the left ischium (arrow) and in the obturator internus muscle, consistent with a phlegmon or abscess. The elongated high-signal-intensity fluid collection (arrowhead) presumably represents a suhpeniosteal abscess. (c) Tiweighti.d SE 600/2() gadolinium-enhanced coronal MR image shows a nonenhancing area with a thin rim of enhancement, representing the prestimt.’d suhpeniosteal abscess (arrow). Fhe involved ischium and muscle are enhanced, suggesting no intraosseous or soft-tissue abscess. The patient responded well to medical managenient and did not require drainage of the abscess.
hanced MR imaging had findings typical including decreased intensity
on TI-weighted
increased signal active inflammation 744
sequences, all of osteomyelitis, marrow signal
#{149} Radiology
intensity on
images in areas T2-weighted
and of
SE or GRE images (Fig 1). Subpeniosteal abscesses were present in four patients (Figs 1-3), intraosseous abscess in one (Fig 4), and soft-tissue abscess in one (Fig 5). The subperiosteal and soft-tissue abscesses ap-
peared
as well-defined
homogeneous
areas of low signal intensity on Tiweighted images and as high signal intensity on T2-weighted one intraosseous abscess
to distinguish
from
images. The was difficult
vasculanized March
in1992
Figure
2.
T2-weighted
tamed in a 6-year-old left pretibial pain and
SE 2,000/80 boy with swelling,
axial images
sickle cell and fever
of both
disease, of 1-day
a 4-day duration.
tibias
ob-
history High
of sig-
nal intensity in the marrow space of the left tibia, swelling of the sunrounding soft tissue, and high signal intensity in the muscle and subcutaneous tissue represent edema and inflammation. A high-intensity focus (arrow) represents a subperiosteal abscess that was drained sungically; bacteriologic culture revealed S aurcus. The right leg is displayed on the left side; the left leg is on the right.
In the eight patients with subacute or chronic osteomyebitis, preoperative MR imaging allowed accurate localization of the extent of inflamed tissue (Figs 4-6). In two patients with chronic osteomyelitis, there were sequestra, seen as focal areas of low signab intensity on all images and with no enhancement after administration of gadopentetate dimeglumine (Fig 6c, Table 2). Two patients had draining sinus tracts extending from the infected bone into the soft tissues; these were of high signal intensity on T2-weighted SE on GRE sequences and showed enhancement on Tiweighted SE images obtained after administration of gadopentetate dimeglumine (Figs 5a, 6c). Surgical debndement was performed in all eight of these patients. Three required repeated debridement; in each of
these,
the
surgeon
had
not
performed
as extensive an initial debnidement as was suggested by findings at preoperative MR imaging.
a. Figure
b. 3.
MR images
obtained
in a 13-year-old
boy with
night knee
pain
and
fever
of 2 days
duration. Response to the initial 48 hours of antibiotic therapy had been poor. (a) T2-weighted SE 2,000/80 sagittal image of the distal right femur shows high signal intensity in the marrow space and adjacent muscles. The elongated high-intensity subpeniosteal fluid collection (annow) is an abscess. (b) TI-weighted SE 600/15 gadolinium-enhanced sagittal MR image shows the subpeniosteal abscess (arrow) to be nonenhancing. Some enhancement is evident in the marrow and adjacent muscles. The many small nonenhancing areas in the marrow space may represent edema or intraosseus collections of pus (arrowheads). The condition of the patient improved within 24 hours of percutaneous drainage of the subpeniosteal abscess and continuation of antibiotic therapy. Cultures of blood and abscess were positive for S aureus.
flammatony tissue on the unenhanced images (Fig 4a, 4b). After gadopentetate dimeglumine administration, areas of vascularized inflammatory tissue showed marked enhancement on Ti-weighted images in all 10 patients (Table 2). Five of the six patients with abscesses received gadopentetate dirneglumine; in two of these, the abscesses did not enhance (Fig 3b), and in three, there was rim enhancement (Figs ic, 4c, 5b). Three of the subpeniosteab abscesses and the one soft-tissue abscess were Volume
182
#{149} Number
3
barge was
enough
to require
drained
ultrasound
drainage;
percutaneously
guidance
by
one using
(Fig 3), and
three were drained surgically. One patient with S aureus osteomyelitis and negative plain radiographs but a positive bone scan (Fig 1) had a small
rim-enhancing collection.
well
subperiosteal The
to medical
the fluid represent
though required.
patient
fluid
responded
management
collection was a subperiosteal
no drainage
alone; presumed abscess,
procedure
to al-
was
DISCUSSION Our experience supports previous reports of the usefulness of MR imaging in studying patients with osteo-
myelitis,
whether
the disease
is acute,
subacute, or chronic (1-8), but in addition demonstrates the value of gadolinium enhancement in evaluating nonspinal osteomyelitis, particularly in young patients. In acute osteomyelitis, involved marrow is of low signal intensity on Ti-weighted images and of high signab intensity on T2-weighted images (i,5-8), because of edema, hyperemia, and exudate (7). The margins of affected bone and associated soft-tissue edema are rather poorly defined (2,7). We have found that, following administration of gadopentetate dimeglumine, areas of vascularized inflammatory tissue will enhance, but nonvascularized abscess collections will be nonenhancing or show rim Radiology
#{149} 745
Figure 4. MR images obtained in a 3year-old girl with i-month history of left knee pain, difficulty walking, and recent onset of low-grade fevers. Plain radiography showed a 1-cm lytic lesion in the bateral part of the distal left femoral epiphysis. (a) Ti-weighted SE 600/17 coronal image of the left knee shows abnormally decreased signal intensity in the entire distal femoral epiphysis, in comparison with normal signal intensity in the proximal tibial epiphysis. The decrease is most marked laterally. (b) GRE (38/17; flip angle, 30#{176}) coronal image shows high signal intensity in the lateral portion of the femoral epiphysis and an irregular low-signal-intensity rim (arrows). (c) Ti-weighted SE 600/17 gadolinium-enhanced image shows enhancement of the entire epiphysis, except for a wellcircumscribed nonenhancing area with slight rim enhancement (arrow), consistent with an intraosseous abscess. The abscess was surgically debnided. (d) Ti-weighted SE 600/17 gadolinium-enhanced image, obtained 6 weeks after debnidement and antibiotic therapy, shows resolution of the abscess cavity.
enhancement of chronic nent rim
(Table osteornyelitis enhancement
2). Our one case with promiof a soft-tissue
abscess appears similar to that described by Paajanen et al (10) in animals, in which the rim of enhancement was shown to correspond to the cellular inflammatory zone, and the nonenhancing center to necrotic tissue. Our experience with a Brodie abscess was limited to one patient. Tang et al (7) described the Brodie abscess as a well-defined intnaosseous area of low signal intensity on Ti-weighted images that demonstrates high signal intensity on T2-weighted images, with a rim of low intensity due to sclerotic bone. This is consistent the appearance of our Brodie on TI- and T2-weighted images. our case, however, gadopentetate dimeglumine defined the intnaosseous abscess much more due to rim enhancement (Fig The subpeniosteal abscesses patients with acute osteomyelitis were well demonstrated with
with abscess In
Table
was
found
the
Ti-
to be partic-
in evaluation of chronic Areas of inhomoge-
neous
intensity
signal
746
#{149} Radiology
were
with being
of Tissues
in Patients
with
Osteomyelitis
seen
Image
in
areas of of low sig-
Type Gadolinium-enhanced
Tissue
ularby useful osteomyelitis.
the medullary cavity, active inflammation
Intensities
cleanly, 4). in our
and T2-weighted sequences. After administration of gadopentetate dimeglumine, however, two of these also showed multiple small nonenhancing areas in the involved marrow space (Fig 3b), which may represent areas of edema on possibly small intraosseous pus collections, although we have no pathologic confirmation available.
MR imaging
2
Signal
Vascularized
inflammatory
Ti-weighted
T2-weighted
Ti-weighted
Low
High
High
Low Low
High Low
Low Low
tissue
Abscess Sequestrum
nab intensity on Ti-weighted images and of high signal intensity on T2weighted images. Pathologically, this
cluded
appearance
Sequestra appeared as areas of low to intermediate signal intensity on Tiand T2-weighted images (2,6,7) and did not enhance with gadolinium.
has been
shown
to come-
late with the presence of granulation tissue (6). Intervening areas of chronic fibrosis were of low signal intensity with Ti- and T2-weighted sequences. Other chronic
findings observed in our osteomyelitis patients in-
cortical sequestra
bone (Fig
thickening 6), and sinus
(Figs tracts
to the adjacent
soft
tissues
5, 6).
5,
6),
Sinus
tracts
are linear
(Figs
or curvilinear
areas of high signal intensity weighted images, communicating
on
March
T2be-
1992
tween the medullary cavity and soft tissues (4,6). One of our patients with chronic osteomyelitis had gadolinium enhancement of a sinus tract to a softtissue abscess (Fig 5), and another had an enhancing sinus tract to granulation tissue at the skin surface (Fig 6). Because of the ability to separate active inflammation from chronic fibrosis and to detect noncontiguous areas of involvement, MR imaging is particularly helpful in guiding surgical debnidement in patients with chronic or subacute osteomyelitis (46). In addition, we have found gadolinium enhancement to be especially
a.
useful
b.
Figure 5. MR images obtained in a i9-year-old man with a 5-year history of recurrent softtissue mass of the left thigh, treated in Central America with drainage and antibiotics. He had recent thigh swelling of 3 weeks duration. Signs on plain nadiographs were consistent with chronic osteomyelitis. (a) T2-weighted multiplanan GRE (400/13; flip angle, 30#{176}) coronal image of both femora shows cortical thickening in the left femur, high signal intensity in the marrow space, presumably due to granulation tissue (long arrows), and a large ovoid fluid collection (short arrows) in the soft tissue, consistent with abscess. A portion of a sinus tract (arrowhead) is seen adjacent to the cortex, connecting the marrow space with the abscess. (b) Ti-weighted SE 600/20 gadolinium-enhanced image obtained by using fat-suppression technique shows enhancement of the granulation tissue and the sinus tract and marked rim enhancement of the soft-tissue abscess (arrows). Note that field inhomogeneity resulted in lack of fat suppression in the lateral portion of the left thigh. The patient did well for 1 year after surgical drainage of the abscess, debnidement of the lower third of the intramedullary granulation tissue in the femur, and antibiotic therapy. Recurrent symptoms that required more extensive debridement then developed.
tions,
i.
6.
7.
8.
9.
6. MR images obtained in a 12-year-old boy with a 4-year history of chronic osteomyelitis of the left humerus following an immunization. (a) T2-weighted multiplanar GRE (350/ 13; flip angle, 30#{176}) coronal image of the left humerus (obtained with the arm elevated above the head) shows cortical thickening and high-signal-intensity granulation tissue in the marnow space (arrows). (b) Ti-weighted SE 600/20 gadolinium-enhanced coronal image shows enhancement of granulation tissue extending to the distal metaphysis. (c) Ti-weighted SE 600/20 gadolinium-enhanced image posterior to b shows two low-signal-intensity sequestra (short arrows) and enhancement of a sinus tract (long arrows) that extends to granulation tissue at the skin surface (curved arrow).
10.
Figure
182
#{149} Number
3
fluid
as associated
sinus
of incolbec-
tracts
Gadobininum-en-
MR imaging
thus
offers
fur-
Beltran J, Noto AM, McGhee RB, Freedy RM, McCalla MS. Infections of the musculoskeletal system: high-field-strength MR imaging. Radiology 1987; 164:449-454. Cohen MD, Cory DA, Kleiman M, Smith JA, Broderick NJ. Magnetic resonance differentiation of acute and chronic osteomyelitis in children. Clin Radiol 1990; 41: 53-56. Fletcher BD, Scoles PV, Nelson AD. Osteomyelitis in children: detection by magnetic resonance. Radiology 1984; 150:57-60. Mason MD, Zlatkin MB, Esterhai JL, Dalinka MK, Velchik MG, Kressel HY. Chronic complicated osteomyelitis of the lower extremity: evaluation with MR imaging. Radiology 1989; 173:355-359. Modic MT. Pflanze W, Feiglin DH, Belhobek G. Magnetic resonance imaging of musculoskeletal infections. Radiol Clin North Am 1986; 24:247-258.
Quinn SF, Murray W, Clark RA, Cochran C. MR imaging of chronic osteomyelitis. Comp Assist Tomogr i988; 12:113-117. Tang JSH, Gold RH, Bassett LW, Seeger LL. Musculoskeletal infection of the extremities: evaluation with MR imaging. Radiology 1988; 166:205-209. Unger E, Moldofsky P, Gatenby R, Hartz W, Broder C. Diagnosis of osteomyelitis by MR imaging. AJR 1988; 150:605-610. Paajanen H, Brasch RC, Schmiedl U, Ogan M. Magnetic resonance imaging of local soft tissue inflammation using gadoliniumDTPA. Acta Radiol 1987; 28:79-83. Paajanen H, Grodd W, Revel D, Engelstad B, Brasch RC. Gadolinium-DTPA enhanced MR imaging of intramuscular abscesses. Magn Reson Imaging 1987; 5:109115.
ii.
J, Chandnani Gadopentetate hanced MR imaging Beltran
al.
system.
12.
13.
Volume
presence
References
5.
c.
the soft-tissue
then assistance in the planning of sungicab debridement of bone on abscess drainage by surgeons or interventional radiologists. U
4.
A’1
sequestra.
hanced
3.
b.
and
as well
and
2.
a.
in defining
tnaosseous
AIR
V. McGhee RA Jr. et dimeglumine-enof the musculoskeletal
i99i; 156:457-466.
Donovan Post MJ, Sze G, Quencer RM, Eismont FJ, Green BA, Gahbauer 1-1. Gadolinium-enhanced MR in spinal infection. Comp Assist Tomogr 1990; 14:721-729. deRoos A, van Persijn van Meerten EL, Bloem JL, et al. MRI of tuberculous spondylitis. AJR 1986; 146:79-82.
Radiology
#{149} 747
C. Dangman, O’Rourke,
MD2 MD
J.
Fredric
#{149}
Osteomyelitis Gadolinium-enhanced Fifteen opsy-
pediatric
patients
osteomyelitis netic resonance teomyelitis
acute
M
in seven
pa-
tients, subacute in three, and chronic in five. Four patients had subperiosteal abscesses, one had a large associated soft-tissue abscess, and one had an intraosseous (Brodie) abscess. Areas
of active
inflammation
had
de-
creased marrow signal intensity on Ti-weighted images, increased signal intensity on T2-weighted images, and enhancement on Ti-weighted images obtained after gadopentetate dimeglumine administration (n = 10). Abscesses were rim enhancing (n = 3) or not (n = 2) with gadolinium-enhanced MR imaging. Nonenhancing areas presumably represented necrotic material. Gadolinium-enhanced MR imaging assisted in definition of the presence and extent of nonvascularized fluid collections within the bone and/or adjacent soft tissues
and
volvement
the
disease neous
It also
helped
debridement (n = 7) and drainage
soft-tissue
of bone
with
osteomyelitis.
surgical
extent
in patients
I
From
F.A.H.),
1992;
the
or
(n
5).
=
182:743-747
Departments
Orthopedics
Imaging’
resonance (MR) imaging has been found to be valuable in the diagnostic evaluation of acute and chronic osteomyelitis in adults and children (1-8). Previously described signal characteristics indude decreased signal intensity of the involved marrow space on Tiweighted images and increased signal intensity of inflammatory tissue on T2-weighted images (5-8). In animals, gadolinium-enhanced MR imaging has been shown to delineate areas of cellulitis (9) and to produce rim enhancement of abscesses (10). There have been a few reports of the use of gadopentetate dimeglumine in the evaluation of patients with vertebral AGNETIC
osteomyelitis
(11-13),
but
none
have
focused on nonspinal or pediatric osteomyelitis. This study was performed to determine whether gadoliniumenhanced MR imaging was valuable for orthopedic surgeons and interventional radiologists in caring for young patients with nonspinal osteomyelitis.
MATERIALS
Index terms: Bones, infection, 45.2121, 45.2132, 45.2141, 40.2111, 40.2112, 40.2113 #{149} Gadolinium #{149} Magnetic resonance (MR), contrast enhancement Radiology
MD
AND
METHODS
of Radiology
(B.C.D.,
(F.F.R.), and Medicine
(E.J.O.), Children’s Hospital, 300 Longwood Aye, Boston, MA 02115-5747. Received June 7, 1991; revision requested July 2; revision received October 14; accepted October 30. Address reprint requests to F.A.H. 2 Current address: Department of Diagnostic Imaging, Rhode Island Hospital, Providence, RI. e RSNA, 1992
Twelve patients who underwent imaging to rube out osteomyelitis were found to have other diagnoses, which included cellulitis (n = 3), avascular necrosis (ii = 2), tendinitis (n = 2), septic arthritis, fasciitis, fibrosis, osteoid osteoma, and occult fractune (n = I each). Patients with osteomyelitis that underwent only postoperative MR imaging were also excluded.
MR imaging
was performed
with
a 1.5-T
unit (Signa; GE Medical Systems, Milwaukee). Body, extremity, or 5-inch (12.5-cm) surface coils were employed as appropriate for the body part being imaged. Our
routine
imaging
tudinal
spin-echo
600/15-20
protocol
included
a longi-
(SE) TI-weighted
sequence
(repetition
time
400msec/ SE T2-
echo time msec), and longitudinal weighted (2,000/20,80), gradient-recalled echo (GRE) (33/13; flip angle, 30#{176}), or multiplanar GRE (250-450/13; flip angle, 30#{176}) sequences. Thirteen patients underwent T2-weighted sequences, and GRE and mubtiplanar GRE sequences were used in six patients each. Whenever possible, a
repeated was
longitudinal
performed
nous
Ti-weighted
immediately
administration
dimeglumine tories, Wayne,
series
after
intrave-
of gadopentetate
(Magnevist; Beniex LaboraNJ), to help distinguish be-
tween vasculanzed inflammatory tissue.
guide
of intraosseous open or percuta-
collections
F. Rand,
in-
chronic
of subperiosteal
fluid
Frank
#{149}
MR
were studied with mag(MR) imaging. Os-
was
MD
in Children:
with binonspinal
or culture-proved
A. Hoffer,
and nonvasculanized The dosage of gado-
From July 1988 to December 1990, we performed MR imaging in 15 pediatric patients with osteomyelitis confirmed by means of positive culture results (from blood, abscess drainage, or tissue aspirate), positive biopsy results, or both (Table 1).
linium employed was 0.2 mL/kg, with a maxiumum of 10 mL per dose. Five patients were not given gadopentetate
Patients
ranged
in age
19 years
(mean,
8.4 years).
dation, Berlex Laboratories, Wayne, NJ); one because of age less than 2 years; and three because they were studied before the contrast medium was approved by the U.S. Food and Drug Administration for pediatric usage. Surgical, bacteriologic, and histopatho-
from
14 months
Seven
to
patients
had cases five The (n = (n =
acute symptoms (2-14 days), three were subacute (1-3 months), and were chronic (7 months to 5 years). involved bones included the femur 5), pelvis (n = 3), tibia (n = 3), rib 2), humerus (n = 1), and radius (n = 1). Sites for imaging were chosen on the basis of clinical symptoms, in combination with abnormal plain radiographs
(n = 10) and/or abnormal findings on bone scans (n = 9) or gaffium scans (n 1). The mean interval between posi=
tive blood culture or pathologic evaluation and MR imaging was 5.67 days. Fourteen patients underwent imaging within 13 days or less (Table 1), and one underwent imaging 23 days after positive blood cubtune but only 3 days after a positive gab-
hum
scan.
dimeglumine: disease,
linium
usage
logic reports aging
one because
a stated
findings
of sickle cell
contraindication
(manufacturer’s
were
correlated
in each
to gado-
recommen-
with
MR im-
patient.
RESULTS Seven patients had acute elitis and eight had subacute chronic osteomyelitis. With
Abbreviations: SE = spin echo.
GRE
=
gradient-recalled
osteomyon unen-
echo,
743
Table 1 History and
Patient Age
Findings
in 15 Pediatric
Patients
Osteomyelitis
with
Duration of
No./
(y)/Sex
Culture
Results Pathologic
Symptoms
Site of Infection
Blood
Abscess
or Tissue
Acute 1/13/M 2/7/F 3/6/M 4/13/M 5/1/M 6/13/M 7/6/M
2 d 4 d 4 d 5 d lOd 2 d 2 wk
Distal nt femur Lt ischium Lt tibia Lt ilium Rtrib Lt ischium Lt tibia
S aureus S aureus Negative S aureus
S aureus ND S aureus
Saureus
ND S aureus S aureus,
Negative S aureus
I mo 3 mo 1 mo
Distal Distal Distal
It femur lt femur lt radius
Negative ND ND
11/12/M
4 y
Lt humerus
12/19/M 13/5/F
5 y 1 y
Lt femur Lt ischium
14/4/F 15/8/M
3#{189} y 7 mo
Lt rib Lt tibia
and
ilium
(hip
effusion)
Kiebsiella
Evaluation
ND
0
ND
10
ND ND
2 6
ND ND
23* 2
Osteomyelitis
0
Osteomyelitis Osteomyelitis Osteomyelitis
0 7 5
Osteomyelitis
Negative Negative Mixed flora
Chronic
Findings
Osteomyelitis
ND
Subacute 8/3/F 9/2/M 10/14/M
Interval (d) between MR Imaging and Positive Culture or Pathologic
(contaminants)
Osteomyelitis
ND
S aureus
Chronic osteomyelitis osteonecrosis
ND Negative
Mycobacterium
S aureus
tuberculosis
Osteomyelitis Granulomatous
and
6
osteomyeli-
1 8
tis
Note.-Lt * Three
=
days
left, ND = not done, after positive gallium
ND ND
S aureus
Escherichia
coli
Chronic osteomyelitis Osteomyelitis
b. 1.
2
Rt = right. scanning.
a. Figure
13
c.
MR images
obtained in a 7-year-old girl with fever and left hip pain of 4 days duration and Staphylococcus aureus septicemia. Plain radiographs of the pelvis were negative, hut a hone scan revealed evidence of osteomyelitis in the left ischium. (a) il-weighted SE 600/15 cornnal image of the pelvis shows decreased signal intensity in the left ischium (arrow) and thickening of the obturator internus muscle, indicating osteomyelitis and adjacent soft-tissue reaction. (b) 12-weighted SE 2,000/80 coronal image of the left hip obtained by using a 5-inch surface coil shows high signal intensity in the corresponding portion of the left ischium (arrow) and in the obturator internus muscle, consistent with a phlegmon or abscess. The elongated high-signal-intensity fluid collection (arrowhead) presumably represents a suhpeniosteal abscess. (c) Tiweighti.d SE 600/2() gadolinium-enhanced coronal MR image shows a nonenhancing area with a thin rim of enhancement, representing the prestimt.’d suhpeniosteal abscess (arrow). Fhe involved ischium and muscle are enhanced, suggesting no intraosseous or soft-tissue abscess. The patient responded well to medical managenient and did not require drainage of the abscess.
hanced MR imaging had findings typical including decreased intensity
on TI-weighted
increased signal active inflammation 744
sequences, all of osteomyelitis, marrow signal
#{149} Radiology
intensity on
images in areas T2-weighted
and of
SE or GRE images (Fig 1). Subpeniosteal abscesses were present in four patients (Figs 1-3), intraosseous abscess in one (Fig 4), and soft-tissue abscess in one (Fig 5). The subperiosteal and soft-tissue abscesses ap-
peared
as well-defined
homogeneous
areas of low signal intensity on Tiweighted images and as high signal intensity on T2-weighted one intraosseous abscess
to distinguish
from
images. The was difficult
vasculanized March
in1992
Figure
2.
T2-weighted
tamed in a 6-year-old left pretibial pain and
SE 2,000/80 boy with swelling,
axial images
sickle cell and fever
of both
disease, of 1-day
a 4-day duration.
tibias
ob-
history High
of sig-
nal intensity in the marrow space of the left tibia, swelling of the sunrounding soft tissue, and high signal intensity in the muscle and subcutaneous tissue represent edema and inflammation. A high-intensity focus (arrow) represents a subperiosteal abscess that was drained sungically; bacteriologic culture revealed S aurcus. The right leg is displayed on the left side; the left leg is on the right.
In the eight patients with subacute or chronic osteomyebitis, preoperative MR imaging allowed accurate localization of the extent of inflamed tissue (Figs 4-6). In two patients with chronic osteomyelitis, there were sequestra, seen as focal areas of low signab intensity on all images and with no enhancement after administration of gadopentetate dimeglumine (Fig 6c, Table 2). Two patients had draining sinus tracts extending from the infected bone into the soft tissues; these were of high signal intensity on T2-weighted SE on GRE sequences and showed enhancement on Tiweighted SE images obtained after administration of gadopentetate dimeglumine (Figs 5a, 6c). Surgical debndement was performed in all eight of these patients. Three required repeated debridement; in each of
these,
the
surgeon
had
not
performed
as extensive an initial debnidement as was suggested by findings at preoperative MR imaging.
a. Figure
b. 3.
MR images
obtained
in a 13-year-old
boy with
night knee
pain
and
fever
of 2 days
duration. Response to the initial 48 hours of antibiotic therapy had been poor. (a) T2-weighted SE 2,000/80 sagittal image of the distal right femur shows high signal intensity in the marrow space and adjacent muscles. The elongated high-intensity subpeniosteal fluid collection (annow) is an abscess. (b) TI-weighted SE 600/15 gadolinium-enhanced sagittal MR image shows the subpeniosteal abscess (arrow) to be nonenhancing. Some enhancement is evident in the marrow and adjacent muscles. The many small nonenhancing areas in the marrow space may represent edema or intraosseus collections of pus (arrowheads). The condition of the patient improved within 24 hours of percutaneous drainage of the subpeniosteal abscess and continuation of antibiotic therapy. Cultures of blood and abscess were positive for S aureus.
flammatony tissue on the unenhanced images (Fig 4a, 4b). After gadopentetate dimeglumine administration, areas of vascularized inflammatory tissue showed marked enhancement on Ti-weighted images in all 10 patients (Table 2). Five of the six patients with abscesses received gadopentetate dirneglumine; in two of these, the abscesses did not enhance (Fig 3b), and in three, there was rim enhancement (Figs ic, 4c, 5b). Three of the subpeniosteab abscesses and the one soft-tissue abscess were Volume
182
#{149} Number
3
barge was
enough
to require
drained
ultrasound
drainage;
percutaneously
guidance
by
one using
(Fig 3), and
three were drained surgically. One patient with S aureus osteomyelitis and negative plain radiographs but a positive bone scan (Fig 1) had a small
rim-enhancing collection.
well
subperiosteal The
to medical
the fluid represent
though required.
patient
fluid
responded
management
collection was a subperiosteal
no drainage
alone; presumed abscess,
procedure
to al-
was
DISCUSSION Our experience supports previous reports of the usefulness of MR imaging in studying patients with osteo-
myelitis,
whether
the disease
is acute,
subacute, or chronic (1-8), but in addition demonstrates the value of gadolinium enhancement in evaluating nonspinal osteomyelitis, particularly in young patients. In acute osteomyelitis, involved marrow is of low signal intensity on Ti-weighted images and of high signab intensity on T2-weighted images (i,5-8), because of edema, hyperemia, and exudate (7). The margins of affected bone and associated soft-tissue edema are rather poorly defined (2,7). We have found that, following administration of gadopentetate dimeglumine, areas of vascularized inflammatory tissue will enhance, but nonvascularized abscess collections will be nonenhancing or show rim Radiology
#{149} 745
Figure 4. MR images obtained in a 3year-old girl with i-month history of left knee pain, difficulty walking, and recent onset of low-grade fevers. Plain radiography showed a 1-cm lytic lesion in the bateral part of the distal left femoral epiphysis. (a) Ti-weighted SE 600/17 coronal image of the left knee shows abnormally decreased signal intensity in the entire distal femoral epiphysis, in comparison with normal signal intensity in the proximal tibial epiphysis. The decrease is most marked laterally. (b) GRE (38/17; flip angle, 30#{176}) coronal image shows high signal intensity in the lateral portion of the femoral epiphysis and an irregular low-signal-intensity rim (arrows). (c) Ti-weighted SE 600/17 gadolinium-enhanced image shows enhancement of the entire epiphysis, except for a wellcircumscribed nonenhancing area with slight rim enhancement (arrow), consistent with an intraosseous abscess. The abscess was surgically debnided. (d) Ti-weighted SE 600/17 gadolinium-enhanced image, obtained 6 weeks after debnidement and antibiotic therapy, shows resolution of the abscess cavity.
enhancement of chronic nent rim
(Table osteornyelitis enhancement
2). Our one case with promiof a soft-tissue
abscess appears similar to that described by Paajanen et al (10) in animals, in which the rim of enhancement was shown to correspond to the cellular inflammatory zone, and the nonenhancing center to necrotic tissue. Our experience with a Brodie abscess was limited to one patient. Tang et al (7) described the Brodie abscess as a well-defined intnaosseous area of low signal intensity on Ti-weighted images that demonstrates high signal intensity on T2-weighted images, with a rim of low intensity due to sclerotic bone. This is consistent the appearance of our Brodie on TI- and T2-weighted images. our case, however, gadopentetate dimeglumine defined the intnaosseous abscess much more due to rim enhancement (Fig The subpeniosteal abscesses patients with acute osteomyelitis were well demonstrated with
with abscess In
Table
was
found
the
Ti-
to be partic-
in evaluation of chronic Areas of inhomoge-
neous
intensity
signal
746
#{149} Radiology
were
with being
of Tissues
in Patients
with
Osteomyelitis
seen
Image
in
areas of of low sig-
Type Gadolinium-enhanced
Tissue
ularby useful osteomyelitis.
the medullary cavity, active inflammation
Intensities
cleanly, 4). in our
and T2-weighted sequences. After administration of gadopentetate dimeglumine, however, two of these also showed multiple small nonenhancing areas in the involved marrow space (Fig 3b), which may represent areas of edema on possibly small intraosseous pus collections, although we have no pathologic confirmation available.
MR imaging
2
Signal
Vascularized
inflammatory
Ti-weighted
T2-weighted
Ti-weighted
Low
High
High
Low Low
High Low
Low Low
tissue
Abscess Sequestrum
nab intensity on Ti-weighted images and of high signal intensity on T2weighted images. Pathologically, this
cluded
appearance
Sequestra appeared as areas of low to intermediate signal intensity on Tiand T2-weighted images (2,6,7) and did not enhance with gadolinium.
has been
shown
to come-
late with the presence of granulation tissue (6). Intervening areas of chronic fibrosis were of low signal intensity with Ti- and T2-weighted sequences. Other chronic
findings observed in our osteomyelitis patients in-
cortical sequestra
bone (Fig
thickening 6), and sinus
(Figs tracts
to the adjacent
soft
tissues
5, 6).
5,
6),
Sinus
tracts
are linear
(Figs
or curvilinear
areas of high signal intensity weighted images, communicating
on
March
T2be-
1992
tween the medullary cavity and soft tissues (4,6). One of our patients with chronic osteomyelitis had gadolinium enhancement of a sinus tract to a softtissue abscess (Fig 5), and another had an enhancing sinus tract to granulation tissue at the skin surface (Fig 6). Because of the ability to separate active inflammation from chronic fibrosis and to detect noncontiguous areas of involvement, MR imaging is particularly helpful in guiding surgical debnidement in patients with chronic or subacute osteomyelitis (46). In addition, we have found gadolinium enhancement to be especially
a.
useful
b.
Figure 5. MR images obtained in a i9-year-old man with a 5-year history of recurrent softtissue mass of the left thigh, treated in Central America with drainage and antibiotics. He had recent thigh swelling of 3 weeks duration. Signs on plain nadiographs were consistent with chronic osteomyelitis. (a) T2-weighted multiplanan GRE (400/13; flip angle, 30#{176}) coronal image of both femora shows cortical thickening in the left femur, high signal intensity in the marrow space, presumably due to granulation tissue (long arrows), and a large ovoid fluid collection (short arrows) in the soft tissue, consistent with abscess. A portion of a sinus tract (arrowhead) is seen adjacent to the cortex, connecting the marrow space with the abscess. (b) Ti-weighted SE 600/20 gadolinium-enhanced image obtained by using fat-suppression technique shows enhancement of the granulation tissue and the sinus tract and marked rim enhancement of the soft-tissue abscess (arrows). Note that field inhomogeneity resulted in lack of fat suppression in the lateral portion of the left thigh. The patient did well for 1 year after surgical drainage of the abscess, debnidement of the lower third of the intramedullary granulation tissue in the femur, and antibiotic therapy. Recurrent symptoms that required more extensive debridement then developed.
tions,
i.
6.
7.
8.
9.
6. MR images obtained in a 12-year-old boy with a 4-year history of chronic osteomyelitis of the left humerus following an immunization. (a) T2-weighted multiplanar GRE (350/ 13; flip angle, 30#{176}) coronal image of the left humerus (obtained with the arm elevated above the head) shows cortical thickening and high-signal-intensity granulation tissue in the marnow space (arrows). (b) Ti-weighted SE 600/20 gadolinium-enhanced coronal image shows enhancement of granulation tissue extending to the distal metaphysis. (c) Ti-weighted SE 600/20 gadolinium-enhanced image posterior to b shows two low-signal-intensity sequestra (short arrows) and enhancement of a sinus tract (long arrows) that extends to granulation tissue at the skin surface (curved arrow).
10.
Figure
182
#{149} Number
3
fluid
as associated
sinus
of incolbec-
tracts
Gadobininum-en-
MR imaging
thus
offers
fur-
Beltran J, Noto AM, McGhee RB, Freedy RM, McCalla MS. Infections of the musculoskeletal system: high-field-strength MR imaging. Radiology 1987; 164:449-454. Cohen MD, Cory DA, Kleiman M, Smith JA, Broderick NJ. Magnetic resonance differentiation of acute and chronic osteomyelitis in children. Clin Radiol 1990; 41: 53-56. Fletcher BD, Scoles PV, Nelson AD. Osteomyelitis in children: detection by magnetic resonance. Radiology 1984; 150:57-60. Mason MD, Zlatkin MB, Esterhai JL, Dalinka MK, Velchik MG, Kressel HY. Chronic complicated osteomyelitis of the lower extremity: evaluation with MR imaging. Radiology 1989; 173:355-359. Modic MT. Pflanze W, Feiglin DH, Belhobek G. Magnetic resonance imaging of musculoskeletal infections. Radiol Clin North Am 1986; 24:247-258.
Quinn SF, Murray W, Clark RA, Cochran C. MR imaging of chronic osteomyelitis. Comp Assist Tomogr i988; 12:113-117. Tang JSH, Gold RH, Bassett LW, Seeger LL. Musculoskeletal infection of the extremities: evaluation with MR imaging. Radiology 1988; 166:205-209. Unger E, Moldofsky P, Gatenby R, Hartz W, Broder C. Diagnosis of osteomyelitis by MR imaging. AJR 1988; 150:605-610. Paajanen H, Brasch RC, Schmiedl U, Ogan M. Magnetic resonance imaging of local soft tissue inflammation using gadoliniumDTPA. Acta Radiol 1987; 28:79-83. Paajanen H, Grodd W, Revel D, Engelstad B, Brasch RC. Gadolinium-DTPA enhanced MR imaging of intramuscular abscesses. Magn Reson Imaging 1987; 5:109115.
ii.
J, Chandnani Gadopentetate hanced MR imaging Beltran
al.
system.
12.
13.
Volume
presence
References
5.
c.
the soft-tissue
then assistance in the planning of sungicab debridement of bone on abscess drainage by surgeons or interventional radiologists. U
4.
A’1
sequestra.
hanced
3.
b.
and
as well
and
2.
a.
in defining
tnaosseous
AIR
V. McGhee RA Jr. et dimeglumine-enof the musculoskeletal
i99i; 156:457-466.
Donovan Post MJ, Sze G, Quencer RM, Eismont FJ, Green BA, Gahbauer 1-1. Gadolinium-enhanced MR in spinal infection. Comp Assist Tomogr 1990; 14:721-729. deRoos A, van Persijn van Meerten EL, Bloem JL, et al. MRI of tuberculous spondylitis. AJR 1986; 146:79-82.
Radiology
#{149} 747
