Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
989
MR Imaging of Thoracic and Abdominal Wall Infections: Comparison Procedures
Hassan S. Sharif1 David C. Clark Mohamed Y. Aabed Osarugue A. Aideyan Maurice C. Haddad Tor A. Mattsson
Infections
can occur or trauma. is difficult on early
of the
with Other
chest
and
abdominal
Imaging
wall
are
rare
but potentially
fatal
disorders
that
spontaneously or in association with diabetes mellitus, immunosuppression, The condition (either in the form of necrotizing fasciitis and/or pyomyositis) to diagnose clinically because of poor localizing signs. Prognosis depends recognition,
extent
the infections
can manifest
To
the
determine
value
of disease,
and
as cellulitis, of
MR
type
of causative
abscess,
imaging
in
the
and/or
organism.
Pathologically,
granulation
assessment
of
tissue
these
formation.
infections,
we
compared the findings of MR with those of CT, sonography, scintigraphy, and plain radiography in 13 patients with proved thoracic and/or abdominal wall infection. The imaging findings were correlated with microbiological, pathologic, and/or surgical data. The isolated pathogens were Staphylococcus aureus (n = 6), Klebsiella pneumoniae (n = 1), Mycobacterium tuberculosis (n = 4), and Streptomyces somaliensis (n = 2). In 10 of 13 patients,
the nature
MR
imaging
and extent
and
CT
were
comparable
of the inflammatory
was useful in guiding percutaneous
process.
and
proved
In seven
accurate
in detecting
of the patients,
biopsy and/or partial drainage
procedures.
CT also
Coronal
and sagittal MR images were helpful for planning surgery. Rib osteomyelitis was missed with both techniques in one patient; in two other patients who did not have CT, MR imaging missed osteomyelitis of the ribs, the spinous process of a vertebral body, and the iliac bone. Sonography underestimated the extent of the disease in all 13 patients, but detected fluid collections in six. Findings on scintigraphy and plain radiography
were the least contributory Our results suggest thoracoabdominal wall
percutaneous AJR
Received August 4, 1 989; accepted
after
revi-
20, 1989.
Presented Society
in part at the annual meeting of the of Magnetic Resonance Imaging, Boston,
February 1988. 1
All authors:
Department
of Radiology,
Riyadh
Armed Forces Hospital, P.O. Box 7897 (979), Riyadh 11159, Kingdom of Saudi Arabia. Address reprint
requests
to H. S. Sharif.
0361 -803X/90/1 545-0989 © American Roentgen Ray Society
and treatment
of these patients.
drainage
are required.
May 1990
of the chest
and
abdominal
wall
is an uncommon
but
serious
condition
can occur spontaneously or in association with diabetes mellitus, immunosuppression, sepsis, surgery, trauma, atherosclerosis, alcoholism, obesity, and malnutrition [1-3]. Clinical recognition of such lesions can be difficult and patients may present with fever, pain, and skin changes ranging from minor discoloration to frank cutaneous necrosis [1 ]. The extent of underlying tissue involvement is frequently underestimated by physical examination, and at surgery abscesses are often found to extend fan beyond the visualized area of cutaneous abnormality [1,3]. The advantages and limitations of plain radiography, sonography, and CT scanning in these cases have been addressed [3-8]. The role of MR imaging in assessing these lesions has not been evaluated. To determine the relative merit of MR imaging in detecting the presence and extent of thonacoabdominal wall infection, we compared the findings of MR with those of other imaging techniques (CT, sonography, scintigraphy, and plain radiography) in 1 3 patients with proved infections. We also correlated the imaging findings with microbiological, pathologic, and surgical data. that
sion November
biopsy and/or
154:989-995,
Infection
to the diagnosis
that MR imaging and CT are the best methods for detecting infection. CT appears to be the technique of choice whenever
TABLE
1: Summary
of Imaging
Findings
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
Radiography
in Patients
with
Bone
Thoracic
No abnormality
No abnormality
2
Increased bone density of T12 and Li on left; loss of psoas
Minimal increased in T12, Li
3
Wall
Infection
CT and/or
Sonography No abnormality
uptake
Small hypoechoic in left psoas
mass
mass
thoracic
MR Imaginga
L: Midline posterior abdominal wall collection; E: Li -Si ; 0: subcutaneous tissue posterior to spine L: Left-sided posterolateral thoracoabdominal
on left
Extrapleural right
Abdominal
Scintigraphy
1
shadow
and/or
wall;
E: T8-
L4; 0: diaphragm, muscles, two vertebrae, epidural space, subcutaneous tissue (Fig. 2) in
cavity
Minimal increased uptake in pedicle of Ti 1 on right
Small hypoechoic lection posterior liver
colto
L: Right-sided posterolateral thoracoabdominal
wall;
small
amount
4
No abnormality
Minimal
increased
of 10th,
uptake
of gas seen on CT; E: T4-L3; 0: pleura, diaphragm, muscles, epidural space, subcutaneous tissue (Fig. 1) L: Left-sided posterolateral thoracoabdominal wall; E: T8L2; 0: diaphragm, muscles, subcutaneous tissue L: Midline posterior thoracoabdominal wall; gas seen best on CT; E: T4-T8; 0:
No abnormality
11th ribs on
left
5
No abnormality
Not done
No abnormality
subcutaneous
6
No abnormality
No abnormality
No abnormality
7
No abnormality
No abnormality
Small hypoechoic lection posterior
colto
liver
tissue
posterior to spine L: Midline posterior abdominal wall; E: Ti2coccyx; 0: subcutaneous tissue postenor to spine (Fig. 3) L: Right-sided posterolateral abdominal wall; E: Ti2-L4; 0: intercostals, psoas, subcutaneous
8
9
No abnormality
Small area struction,
of bone deright iliac
Increased
uptake
right
sternoclavicular
joint
and
right
C6-T7; sixth
and seventh ribs anteriorly and Li -L3 Minimal increased uptake in right iliac bone
bone
No abnormality
Large hypoechoic collection in right psoas and iliac fossa
tissue
L: Right-sided anterior chest wall; CT not done; E: T8-Ti2; 0: intercostals, subcutaneous tissue L: Midline and rightsided
posterolateral
abdominal wall; CT not done; E: L3-midshaft of right femur;
0: psoas, iliacus, glutei, thigh subcutaneous
10
No abnormality
No abnormality
No abnormality
11
No abnormality
No abnormality
Small hypoechoic collection in posterior thoracic wall
muscles, tissue
(Fig. 4) L: Midline posterior abdominal wall; CT not done; E: Li -Si ; 0: subcutaneous tissue posterior to spine
L: Left-sided lateral dominal
posteroabwall; E: T8-
thoracic
L3; 0: pleura, diaphragm, abdominal wall muscles, subcutaneous
tissue
TABLE
1: Continued Case
Radiography
12
Large
posterior
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
diastinal
Bone Scintigraphy me-
Minimal
mass at
increased
Sonography
uptake
CT and/or MR Imaginga
Not done
L: Midline
posterior
from lower three ribs
thoracoabdominal
bilaterally
wall; E: T4-coccyx; 0: posterior mediastinum, epidural space, all posterior thoracoabdominal
T4-T12; irregular lower ribs; slender vertebral osteophytes
wall
muscles,
upper
glutei,
muscles
thighs,
of
subcutaneous
fat (Fig. 6) 13
Multiple rounded cal-
No abnormality
Small hypoechoic
culi in right renal bed; excretory urography: nonfunctioning right
col-
L: Right-sided
lections in region of renal bed on right
postero-
lateral thoracoabdominal mass; E: Ti i -L5; 0: kidney and
kidney
adrenal,
muscles
of posterolateral abdominal waIl, subcutaneous tissue (Fig. a
CT and MR imaging
involvement
findings
were
long TA/TE
did not occur
and STIR sequences.
Materials
(CT was
performed
in 1 0 cases).
Findings
are presented
The most appropriate
planes
of imaging
were
the coronal
obtained
imaging
infection
by various and/or
were
analyzed
findings
(n
Riyadh
Armed
Forces
patient
group
The
=
radiologic
pathologically retrospectively
1 2). All patients Hospital
were
and
in 1 3 patients
with
thoracoabdominal correlated
examined
with
and treated
wall surgical
at the
and May 1989. 1 0 men and three women 18-75
between
comprised
methods verified
January
1 987
The pathogens were Staphylococcus aureus (n = 6), Klebsiella pneumoniae (n = 1 ), Mycobacterium tuberculosis (n = 4), and Streptomyces somallensis (n = 2). Signs, symptoms, and duration of symptoms were different for the different types of pathogens; those with pyogenic infections (n = 7) had a more acute course. Skin changes suggestive of an underlying infection were detected in two patients with pyogenic infection, one patient with tuberculosis, and two patients with mycetoma. Four of the seven years
old (mean age, 42 years).
patients
with
pyogenic
on immunosuppressive
infections
had
diabetes
mellitus
and
one
was
therapy.
In all 1 3 patients, frontal and lateral chest radiographs and frontal and lateral radiographs of the affected body region had been obtained on the basis of the clinical signs and symptoms. Two patients were examined further with excretory urography. Real-time sonography of the abdomen was performed with special emphasis on areas in the thoracoabdominal
wall
(L), extent
(E), and organ
and sagittal.
findings.
that
produced
clinical
signs
and
symptoms.
When
spinal
involvement
in the
(3 mm) were
was
suspected,
additional
The CT studies of seven patients were repeated after IV injection of 1 00 ml (300 mg I/mI) of Omnipaque (Nyegaard, Oslo, Norway). Appropriate sagittal and coronal reconstructed images were obtained from the 3-mm-thick slices contiguous
Images
to location
in cases
and Methods
microbiologically
with respect
1, 6, and 7; all pyogenic and tuberculous abscesses were hypo- or isodense relative to muscle and after IV contrast. On MA, pyogenic and tuberculous abscesses were isointense relative to muscle on short TRITE and hyperintense on
(0). CT enhancement
showed rim enhancement
comparable
5)
sections
spine.
All
images
were
scanned.
evaluated
on
bone
and
soft-tissue
windows.
All 13 patients superconductive
underwent unit
ages were obtained were imaged in the of 1 mm in a 256 x Eight patients had of therapy. The
following
multislice
imaging Philips,
on a 0.5-T
Best,
the
whole-body
Netherlands).
Im-
in sagittal and coronal planes, and seven patients axial plane also. Sections 7-mm thick with a gap 256 matrix were obtained in the affected region. MR imaging studies repeated during the course
scanning
sagittal
MR
(Gyroscan,
and
parameters coronal
were
images
were
used:
(1
)
spin-echo
obtained
in four
(SE) to
six
500/30 (TR/TE); (2) SE multislice coronal, sagittal, and/or axial images in two acquisitions at 2000/50,1 00; and (3) multislice coronal and sagittal images for two patients with short-tau inversion recovery (STIR), 1 500/i 20/30 (TR/Tl/TE), with two measurements. The area of the body evaluated by the MR study was determined on the basis of clinical findings as well as from the abnormalities acquisitions
detected
at
on previous imaging studies.
Patchy or diffuse loss of signal
On the sonograms, areas of abnormal echogenicity in the thoracic or abdominal wall were identified and were labeled as cystic if they were echo free with increased through transmission and a thin rim, as abscesses if they were relatively echo free with debris and a thick wall, and as a solid mass if they were echogenic. Single-phase static bone scintigraphy was done in all 13 patients by using a gamma camera with a low-energy high-resolution cdlimator and 600 MBq of ‘Tc methylene diphosphonate given IV. Anterior and posterior views (500,000 counts/image) of the entire
from subcutaneous fat and muscle planes on the short TR/TE sequence was considered evidence of involvement. Anatomic distortion
skeleton
in seven
were obtained
3 hr after injection.
Ten patients were examined with CT by using a G.E. 9800 scanner (General Electric, Milwaukee, WI). Contiguous sections (i 0 mm) were scanned through what was suspected as the affected body region from the clinical signs and symptoms and with the guidance of earlier
of
skin,
muscles,
and
other
organs
(in
comparison
with
a normal
contralateral side) was also considered evidence of involvement on this sequence. High signal on long TRITE and STIR sequences in these or other areas was considered indicative of inflammation. All initial radiologic investigations were done within 1 0 days of
admission
to the hospital; all were done preoperatively.
Percutaneous invariably
patients, loculated.
drainage was
of abscess unsuccessful
Complete
surgical
under
CT guidance,
because drainage
the
attempted
abscesses of
the
were
abscesses
(containing 70-i 500 ml of pus) and removal of necrotic infected tissue were achieved in all pyogenic and tuberculous lesions. Only partial resection was possible in one of the mycetoma cases because
992
SHARIF
of adherence
of the
inflammatory
mass
to the
inferior
vena
cava
and
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
aorta. The patient with the more extensive mycetoma had several skin biopsies that showed characteristic histopathologic changes. The different studies were evaluated separately and independently by three radiologists blinded to the final diagnosis. For each imaging technique
and
location
for
and
every
extent
patient,
of the
the
organ
pathology (example, inflammatory was disagreement, the consensus All
imaging
findings
following
lesion,
were
points
were
involvement,
assessed: and
type
vs neoplastic). Whenever opinion was accepted.
subsequently
correlated
with
of
there
ET AL.
AJR:154,
case (Fig. 6) was
surgical
imaging.
two
patients
with
imaging
findings
tions
was
similar
tuberculous
are
summarized
the signal for
all the
infections.
in Table
behavior
In the
The
1 1 patients abscesses
1.
MR
of the abscess with were
collec-
pyogenic
and
isointense
rela-
tive to skeletal muscle on the short TR/TE SE images and became hypenintense on the long TR/TE and STIR images (Figs. 1-4). Epidural extension and associated vertebral Osteomyelitis high
were
signal
proved ribs,
appeared
isointense
exhibited
TE images right
ogenic
only
(Figs.
kidney
(Fig.
abscesses.
Fig. 1.-Case A, Postcontrast
was
process, patients
in two
on long
osteomyelitis
a spinous
In the two and
detected
intensity
and with
patients
(areas
of
images), but pathologically with MR imaging in three
an iliac
mycetoma,
bone
(Fig.
4).
the inflammatory
relative
to muscle
moderate
increase
on short in signal
TR/TE
tissue
images
on long
TR/
5 and 6). The cystic lesions replacing the 5) showed behavior similar to that of pyExtensive epidunal extension in the second
3: 50-year-old diabetic axial CT scan shows
with large extrapleural
of these
TR/TE missed
collection
starting
body
on disk
epidural
extension,
patient
with
the
and
rib
osteomyelitis
(Fig. 2). A small amount of two patients (Fig. 1). night-sided
mycetoma,
of gas
changes
in
the renal bed similar to those of pyogenic abscesses were seen (Fig. 5). The affected areas in both mycetoma patients enhanced minimally after administration of contrast material. Destruction
On MR images
vertebral
On CT, pyogenic and tuberculous abscesses were better visualized on poStcontrast scans and appeared as areas of low attenuation surrounded by an enhancing rim (Figs. 1 and 2). Vertebral osteomyelitis was detected in only one of the
extensive, bral body
Results
with
involvement.
was missed in one patient was seen in the abscesses
data with regard to the size (estimated volume), number (loculations), and extent of the abscesses. The findings at surgery of bone, muscle, and abdominal organ involvement also were compared with those of
The
not associated
May 1990
of
the
muscles
and
replacement
imaging
and
CT findings
were
lated well with surgery, particularly scess collections. In patients who incisions
by
fat
were
and, despite severe epidural extension, no verteor disk involvement was detected (Fig. 6).
for better
drainage,
surgical
comparable
with were
and
come-
respect to the abexplored by large
and
pathologic
proof
of
muscle group involvement correlated with the MR and CT appearances. The presence of osteomyelitis in flat bones (nibs, spinous process, iliac bone) was missed by CT in one and by MR in three cases. Epidural extension in two patients, cleanly noted by both techniques, could not be verified sungically because the area was not explored (Figs. 1 and 2). Two radiologists correctly identified the nature of the disease as being inflammatory on both CT (n 1 0) and MR. The third was in doubt about the MR study of two patients =
(unilateral
mycetoma,
anterior
chest
wall
tuberculosis),
as to
whether the changes were caused by a neoplastic process. No attempts were made by the three radiologists to suggest the possible pathogen except in the case of the unilateral
man with fever, right flank pain, and Staphylococcus aureus septicemia. small fluid collection extending from extrapleural space and dissecting posterior
from level of T4. Note gas within collection,
small abscesses
within erector
to diaphragm.
spinae muscles
This is in continuity (straight arrow), and
inflamed dura suggesting epidural extension (curved arrow). B, Right parasagittal MR image (SE 2000/50) shows extrapleural collection extending from thoracic down to abdominal cavity. Pleura, lung, diaphragm, and liver are displaced anteriorly. C, Midsagittal MR image (SE 2000/100) shows epidural extension (arrowheads) posterior to spinal cord. Signal is higher than that of CSF, presumably because of higher protein content of pus and/or because of its limited mobility. Note high signal intensity in subcutaneous tissue (arrow).
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
May 1990
THORACIC
AND
ABDOMINAL
WALL
INFECTIONS
993
Fig. 2.-Case 2: 47-year-old man with fever and left flank pain of 12 days duration. A, Axial postcontrast CT scan at level of inferior endplate of L3 shows multiloculated abscess involving left psoas muscle extending into neural foramen and involving lateral abdominal wall muscles and subcutaneous fat (asterisk). Staphylococcus aureus was isolated from pus obtained by CT-guided percutaneots biopsy. B, Coronal MR image (SE 500/30) at level of kidneys shows replacement of left perinephric fat by isointense collection that extends into subcutaneous fat (arrow). C, Coronal MR image (SE 2000/100) at same level as B shows multiple hyperintense collections corresponding to isointense areas in B.
Fig. 3.-Case 6: 70-year-old diabetic man with fever and low back pain. A, Axial CT scan at level of L3 shows isodense collection between latissimus dorsi muscle and skin (asterisk). B, Midsagittal MR image (SE 2000/50) shows marked increase in signal of soft tissues pasterior to spinous processes (stars). Staphylococcus aureus was isolated from removed subcutaneous collection.
mycetoma, where two radiologists thought the changes were compatible with tuberculous autonephrectomy (Fig. 5). On sonography, abnormal fluid collections were detected in six patients, but delineation of the extent was obscured by neighboring bone. On scintigraphy, moderate increased uptake of nadionuclide was detected in six patients from ribs, the transverse process, vertebrae, and an iliac bone. The spinous process of L5, which was involved by the infection, showed normal activity. Plain radiognaphs showed abnormal bone changes in three patients and abnormal soft-tissue shadowing in two.
Discussion Infections of the thoracic and abdominal wall are potentially fatal. The prognosis depends on early clinical recognition, the type of causative organism (pyogenic infections being the
more severe), and the extent 1 1 cases in this series with tions
had
quick
and
of tissue pyogenic
complete
involvement [1 -3]. The and tuberculous infec-
recovery
because
of
early
diagnosis and adequate therapy. In this study, CT and MR examinations provided the most useful information about the nature and extent of the infections. Findings with both techniques were similar and correlated well with surgery. MR offered the advantage of direct multiplanar imaging without ionizing radiation and the use of IV contrast agents; CT helped in guiding percutaneous biopsy and/or partial drainage procedures in seven patients. Abscess loculation prevented complete percutaneous drainage. Of the different planes used, sagittal and coronal MR images were found to be most helpful for surgical planning because they clearly displayed the extent and communication sites of the different abscess loculations (Figs. 1 -4). The MR signal of the pyogenic and tuberculous abscesses was com-
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
994
SHARIF
Fig. 4.-Case 9: 23-year-old man with back pain of 6 months duration A, Coronal MR image (SE 500/30) shows large mass inferior to kidney, B, Right parasagittal MR image (STIR 1500/120/30) at level of femoral upper thigh muscles. C, Midsagittal MR image (STIR 1500/120/30) shows large collection canal. At surgery, acid-fast bacilli were isolated together with 1500 ml of
ET AL.
AJR:i54,
radiating to right lower extremity. isointense relative to psoas muscle (asterisk). head (asterisk) shows psoas collection extending
from
iliopsoas
May 1990
muscle
involving skin, subcutaneous tissue, and posterior elements but sparing pus. Spinous process of L5 also was removed because of involvement.
Fig. 5.-Case ness in right upper
13: 43-year-old abdominal
into spinal
man with full-
quadrant
of a few
years duration. A, Postenhancement L3 shows inflammatory
axial CT scan at level of process affecting lateral and posterior abdominal wall (asterisks). Note loss of demarcation between abdominal wall muscles and extension into subcutaneous fat. There is enhancement in residual tissue of destroyed right kidney with rim enhancement around small collection (large arrowhead). Infenor vena cava is displaced to left (small arrow#{149} head). B, Coronal MR image (SE 500/30) shows destroyed right kidney with inflamed pennephric fat displacing inferior vena cava to left (arrow. head). Note thickened lateral abdominal wall together with subcutaneous fat, which has multiple linear areas of echo void, probably representing subcutaneous edema (arrow).
Fig. 6.-Case 12: 34-year-old man with extensive mycetoma involving entire posterior thoracoabdominal wall. A, Axial postcontrast CT scan at level of Ti I
shows large posterior mediastinal mass displacing aorta anteriorly (stars). Mass extends posteriorly
to involve
all muscles
of back.
Patient
was scanned in prone position because he was unable to lie supine. B, Coronal
MR image
(SE 500/30) shows pos(asterisks) extending paraspinally into abdominal cavity. There is al most complete destruction of psoas muscles, but vertebrae and disks appear normal. Spleen is markedly enlarged.
terior mediastinal
mass
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
THORACIC
May 1990
AND
ABDOMINAL
parable to what has been described for abdominal collections [9-1 1 1. The abscesses appeared hypointense on short TR/ TE SE pulse sequences and became hypenintense on the long TR/TE and STIR sequences (Figs. 1 -4). Although the short TR/TE study provided optimal contrast for diffenentiating normal from abnormal structures in the thoracoabdominal wall, it was the long TR/TE and STIR images that helped in better delineation of the location and extent of the abscesses. This concurs with observations made by others in relation to musculoskeletal infections and other types of disease affecting this anatomic region [1 2-1 5]. Rib osteomyelitis was missed with CT and MR imaging, probably because of the relatively small size of these flat bones and also because of the difficulty of imaging the ribs in one plane with either technique. Furthermore, MR imaging failed to depict the presence of osteomyelitis in a spinous process of a vertebral body and in the iliac bone of the patient with the extensive tuberculous abscess, presumably because they were obscured by the hypenintense signal of the overlying abscesses (Fig. 4). Sonography, bone scintigmaphy, and radiography provided limited help in the overall assessment of these lesions. Infections of the thonacoabdominal wall are usually divided into necrotizing fasciitis if localized to the subcutaneous fat and superficial fascia and pyomyositis if they involve the muscles alone. Such a differentiation could be made in only four of our cases where the disease was localized to the skin and superficial fascia (Fig. 3). The nest of the patients had involvement of the skin, subcutaneous fat, superficial fascia, and muscles. Tuberculous abscesses predominantly affecting the soft tissues are infrequent but nevertheless well recognized [7, 1 6, 17]. In the four patients with thonacoabdominal wall tuberculosis, imaging findings were indistinguishable from those in the cases of pyogenic infections (Fig. 4). Mycetoma is a chronic exogenous granulomatous infection that affects the skin and extends to involve deeper tissue [18, 1 9]. True fungi and aerobic actinomycetes are the causative agents, introduced into the skin after trauma usually by a sharp thorn [1 8]. Both types of organisms tend to induce an identical histopathologic picture; diagnosis is achieved by isolation of the organism and/or by characteristic histologic findings [19]. The foot is the most frequently affected site, but in endemic countries the hand, the back, and the metroperitoneum can be involved [1 8-20]. The two cases in this study with actinomycotic mycetoma represent a severe form of this complication (Figs. 5 and 6). The long duration of the illness is not unusual in this disease. In conclusion, this study has shown that MR imaging and CT are equally adequate for evaluating patients with thonacoabdominal
wall
infection.
They
both
provide
useful
infon-
mation about the nature and extent of these lesions, and hence have important impact on their management. CT is the preferred technique whenever percutaneous biopsy and/on drainage is required. MR imaging can be used for the initial
WALL
INFECTIONS
995
diagnosis and is most appropriately used in follow-up therapy because it offers direct multiplanar imaging absence of ionizing radiation and IV contrast agents.
after in the
ACKNOWLEDGMENTS We
thank
Studies ment
the
and
support;
their
cases;
leen
Harrison,
and
this
manuscript.
of Medical
Al Kharj
George
colleagues
radiologic ferring
Departments
of the Riyadh
Hospital
Green
for
for cooperation; and Julie
the
Department
Mcore
for
Administration
and
Programme reading
Medical
for encourage-
the
manuscript;
our
our clinical
colleagues
for re-
of Medical
Illustration,
Kath-
assistance
in the
preparation
of
REFERENCES 1 . Stamenkovic I, Lew PD. Early recognition of potentially fatal necrotizing fasciitis. N EngI J Med 1984;31 0:1689-1693 2. Cave DR. Robinson WA, Brotschi EA. Necrotizing fasciitis following percutaneous endoscopic gastrostomy. Gastrointest Endosc 1986;32: 294-296 3. Fisher JR. Conway MJ, Takeshita AT, Sandoval MR. Necrotizing fasciitis, importance of roentgenographic studies for soft tissue gas. JAMA 1979; 241 :803-806 4. Weiner Cl, Diaconis JN. Primary abdominal wall abscess diagnosed by ultrasound. Arch Surg 1975;1 10:341-342 5. Yeh HC, Aabinowitz JG. Ultrasonography and computed tomography of inflammatory abdominal wall lesions. Radiology 1982; 144 : 859-863 6. Gibson AK, Resenthal SJ, Lukert BP. Pyomyositis, increasing recognition in temperate climates. Am J Med 1984;77 :768-772 7. Coppola J, Muller NL, Connell 0G. Computed tomography of musculoskeletal tuberculosis. J Can Assoc Radiol 1987;38: 199-203 8. Pandolfo I, Balandino A, Gaeta M, Racchiusa 5, Chirico G. CT findings in palpable lesions of the anterior abdominal wall. J Comput Assist Tomogr 1986:10:629-633 9. Wall SD, Fisher MA, Amparo EG, Hricak H, Higgins CB. Magnetic resonance imaging in the evaluation of abscesses. AJR 1985;1 44:1217-1221 10. Beltran J, Noto AM, McGhee RB, Freedy AM, McCalla MS. Infections of the musculoskeletal system. High-field strength MR imaging. Radiology
1987;164:449-454 1 1 . Beltran J, McGhee RB, Shaffer PB, et al. Experimental infections musculoskeletal system. Evaluation with MR imaging and Tc-99m and Ga-67 scintigraphy. Radiology 1988;167: 167-1 72
12. Weinreb JC, Cohen JM, Maravilla 13.
normal anatomy and disease. Lee JK, Glazer HS. Psoas 1986;160:683-687
KR. lliopsoas
Radiology
muscles:
of the MDP
MR study of
1985;156:435-440
muscle disorders:
MR imaging.
Radiology
14. De Roes A, Van Meerten EP, Bloem JL, Bluem AG. MRI of tuberculous spondylitis. AJR 1986;146:79-82 1 5. Haggar AM, Pearlberg carcinoma of the lung: 1087
16. Chafetz
JL, Froelich JW, et al. Chest wall invasion by detection by MR imaging. AJR 1987;148:1075-
N, Genant HK, Hoaglund
with progressive
bony destruction.
FT. lschiogluteal J Can Assoc
tuberculous bursitis 1982:33:119-120 disease in the muscu-
Radiol
17. Butorac A, Littlejohn GO, Hcoper J. Mycobacterial loskeletal system. Med J Aust I987;147:388-391
1 8. Mariat F, Desfombes P, Segretain G. The mycetomas: clinical features, pathology, etiology and epidemiology. Contrib Microbiol Immunol 1977; 4:1-39 19. Lewall DB, Ofale 5, Bendl B. Mycetoma. Skeletal Radio! 1985;14:
257-262 20.
Gumaa SA, Mahgoul ES, El Sid MA. Mycetoma J Trop Med Hyg 1986;35:594-600
of the head and neck. Am
989
MR Imaging of Thoracic and Abdominal Wall Infections: Comparison Procedures
Hassan S. Sharif1 David C. Clark Mohamed Y. Aabed Osarugue A. Aideyan Maurice C. Haddad Tor A. Mattsson
Infections
can occur or trauma. is difficult on early
of the
with Other
chest
and
abdominal
Imaging
wall
are
rare
but potentially
fatal
disorders
that
spontaneously or in association with diabetes mellitus, immunosuppression, The condition (either in the form of necrotizing fasciitis and/or pyomyositis) to diagnose clinically because of poor localizing signs. Prognosis depends recognition,
extent
the infections
can manifest
To
the
determine
value
of disease,
and
as cellulitis, of
MR
type
of causative
abscess,
imaging
in
the
and/or
organism.
Pathologically,
granulation
assessment
of
tissue
these
formation.
infections,
we
compared the findings of MR with those of CT, sonography, scintigraphy, and plain radiography in 13 patients with proved thoracic and/or abdominal wall infection. The imaging findings were correlated with microbiological, pathologic, and/or surgical data. The isolated pathogens were Staphylococcus aureus (n = 6), Klebsiella pneumoniae (n = 1), Mycobacterium tuberculosis (n = 4), and Streptomyces somaliensis (n = 2). In 10 of 13 patients,
the nature
MR
imaging
and extent
and
CT
were
comparable
of the inflammatory
was useful in guiding percutaneous
process.
and
proved
In seven
accurate
in detecting
of the patients,
biopsy and/or partial drainage
procedures.
CT also
Coronal
and sagittal MR images were helpful for planning surgery. Rib osteomyelitis was missed with both techniques in one patient; in two other patients who did not have CT, MR imaging missed osteomyelitis of the ribs, the spinous process of a vertebral body, and the iliac bone. Sonography underestimated the extent of the disease in all 13 patients, but detected fluid collections in six. Findings on scintigraphy and plain radiography
were the least contributory Our results suggest thoracoabdominal wall
percutaneous AJR
Received August 4, 1 989; accepted
after
revi-
20, 1989.
Presented Society
in part at the annual meeting of the of Magnetic Resonance Imaging, Boston,
February 1988. 1
All authors:
Department
of Radiology,
Riyadh
Armed Forces Hospital, P.O. Box 7897 (979), Riyadh 11159, Kingdom of Saudi Arabia. Address reprint
requests
to H. S. Sharif.
0361 -803X/90/1 545-0989 © American Roentgen Ray Society
and treatment
of these patients.
drainage
are required.
May 1990
of the chest
and
abdominal
wall
is an uncommon
but
serious
condition
can occur spontaneously or in association with diabetes mellitus, immunosuppression, sepsis, surgery, trauma, atherosclerosis, alcoholism, obesity, and malnutrition [1-3]. Clinical recognition of such lesions can be difficult and patients may present with fever, pain, and skin changes ranging from minor discoloration to frank cutaneous necrosis [1 ]. The extent of underlying tissue involvement is frequently underestimated by physical examination, and at surgery abscesses are often found to extend fan beyond the visualized area of cutaneous abnormality [1,3]. The advantages and limitations of plain radiography, sonography, and CT scanning in these cases have been addressed [3-8]. The role of MR imaging in assessing these lesions has not been evaluated. To determine the relative merit of MR imaging in detecting the presence and extent of thonacoabdominal wall infection, we compared the findings of MR with those of other imaging techniques (CT, sonography, scintigraphy, and plain radiography) in 1 3 patients with proved infections. We also correlated the imaging findings with microbiological, pathologic, and surgical data. that
sion November
biopsy and/or
154:989-995,
Infection
to the diagnosis
that MR imaging and CT are the best methods for detecting infection. CT appears to be the technique of choice whenever
TABLE
1: Summary
of Imaging
Findings
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
Radiography
in Patients
with
Bone
Thoracic
No abnormality
No abnormality
2
Increased bone density of T12 and Li on left; loss of psoas
Minimal increased in T12, Li
3
Wall
Infection
CT and/or
Sonography No abnormality
uptake
Small hypoechoic in left psoas
mass
mass
thoracic
MR Imaginga
L: Midline posterior abdominal wall collection; E: Li -Si ; 0: subcutaneous tissue posterior to spine L: Left-sided posterolateral thoracoabdominal
on left
Extrapleural right
Abdominal
Scintigraphy
1
shadow
and/or
wall;
E: T8-
L4; 0: diaphragm, muscles, two vertebrae, epidural space, subcutaneous tissue (Fig. 2) in
cavity
Minimal increased uptake in pedicle of Ti 1 on right
Small hypoechoic lection posterior liver
colto
L: Right-sided posterolateral thoracoabdominal
wall;
small
amount
4
No abnormality
Minimal
increased
of 10th,
uptake
of gas seen on CT; E: T4-L3; 0: pleura, diaphragm, muscles, epidural space, subcutaneous tissue (Fig. 1) L: Left-sided posterolateral thoracoabdominal wall; E: T8L2; 0: diaphragm, muscles, subcutaneous tissue L: Midline posterior thoracoabdominal wall; gas seen best on CT; E: T4-T8; 0:
No abnormality
11th ribs on
left
5
No abnormality
Not done
No abnormality
subcutaneous
6
No abnormality
No abnormality
No abnormality
7
No abnormality
No abnormality
Small hypoechoic lection posterior
colto
liver
tissue
posterior to spine L: Midline posterior abdominal wall; E: Ti2coccyx; 0: subcutaneous tissue postenor to spine (Fig. 3) L: Right-sided posterolateral abdominal wall; E: Ti2-L4; 0: intercostals, psoas, subcutaneous
8
9
No abnormality
Small area struction,
of bone deright iliac
Increased
uptake
right
sternoclavicular
joint
and
right
C6-T7; sixth
and seventh ribs anteriorly and Li -L3 Minimal increased uptake in right iliac bone
bone
No abnormality
Large hypoechoic collection in right psoas and iliac fossa
tissue
L: Right-sided anterior chest wall; CT not done; E: T8-Ti2; 0: intercostals, subcutaneous tissue L: Midline and rightsided
posterolateral
abdominal wall; CT not done; E: L3-midshaft of right femur;
0: psoas, iliacus, glutei, thigh subcutaneous
10
No abnormality
No abnormality
No abnormality
11
No abnormality
No abnormality
Small hypoechoic collection in posterior thoracic wall
muscles, tissue
(Fig. 4) L: Midline posterior abdominal wall; CT not done; E: Li -Si ; 0: subcutaneous tissue posterior to spine
L: Left-sided lateral dominal
posteroabwall; E: T8-
thoracic
L3; 0: pleura, diaphragm, abdominal wall muscles, subcutaneous
tissue
TABLE
1: Continued Case
Radiography
12
Large
posterior
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
diastinal
Bone Scintigraphy me-
Minimal
mass at
increased
Sonography
uptake
CT and/or MR Imaginga
Not done
L: Midline
posterior
from lower three ribs
thoracoabdominal
bilaterally
wall; E: T4-coccyx; 0: posterior mediastinum, epidural space, all posterior thoracoabdominal
T4-T12; irregular lower ribs; slender vertebral osteophytes
wall
muscles,
upper
glutei,
muscles
thighs,
of
subcutaneous
fat (Fig. 6) 13
Multiple rounded cal-
No abnormality
Small hypoechoic
culi in right renal bed; excretory urography: nonfunctioning right
col-
L: Right-sided
lections in region of renal bed on right
postero-
lateral thoracoabdominal mass; E: Ti i -L5; 0: kidney and
kidney
adrenal,
muscles
of posterolateral abdominal waIl, subcutaneous tissue (Fig. a
CT and MR imaging
involvement
findings
were
long TA/TE
did not occur
and STIR sequences.
Materials
(CT was
performed
in 1 0 cases).
Findings
are presented
The most appropriate
planes
of imaging
were
the coronal
obtained
imaging
infection
by various and/or
were
analyzed
findings
(n
Riyadh
Armed
Forces
patient
group
The
=
radiologic
pathologically retrospectively
1 2). All patients Hospital
were
and
in 1 3 patients
with
thoracoabdominal correlated
examined
with
and treated
wall surgical
at the
and May 1989. 1 0 men and three women 18-75
between
comprised
methods verified
January
1 987
The pathogens were Staphylococcus aureus (n = 6), Klebsiella pneumoniae (n = 1 ), Mycobacterium tuberculosis (n = 4), and Streptomyces somallensis (n = 2). Signs, symptoms, and duration of symptoms were different for the different types of pathogens; those with pyogenic infections (n = 7) had a more acute course. Skin changes suggestive of an underlying infection were detected in two patients with pyogenic infection, one patient with tuberculosis, and two patients with mycetoma. Four of the seven years
old (mean age, 42 years).
patients
with
pyogenic
on immunosuppressive
infections
had
diabetes
mellitus
and
one
was
therapy.
In all 1 3 patients, frontal and lateral chest radiographs and frontal and lateral radiographs of the affected body region had been obtained on the basis of the clinical signs and symptoms. Two patients were examined further with excretory urography. Real-time sonography of the abdomen was performed with special emphasis on areas in the thoracoabdominal
wall
(L), extent
(E), and organ
and sagittal.
findings.
that
produced
clinical
signs
and
symptoms.
When
spinal
involvement
in the
(3 mm) were
was
suspected,
additional
The CT studies of seven patients were repeated after IV injection of 1 00 ml (300 mg I/mI) of Omnipaque (Nyegaard, Oslo, Norway). Appropriate sagittal and coronal reconstructed images were obtained from the 3-mm-thick slices contiguous
Images
to location
in cases
and Methods
microbiologically
with respect
1, 6, and 7; all pyogenic and tuberculous abscesses were hypo- or isodense relative to muscle and after IV contrast. On MA, pyogenic and tuberculous abscesses were isointense relative to muscle on short TRITE and hyperintense on
(0). CT enhancement
showed rim enhancement
comparable
5)
sections
spine.
All
images
were
scanned.
evaluated
on
bone
and
soft-tissue
windows.
All 13 patients superconductive
underwent unit
ages were obtained were imaged in the of 1 mm in a 256 x Eight patients had of therapy. The
following
multislice
imaging Philips,
on a 0.5-T
Best,
the
whole-body
Netherlands).
Im-
in sagittal and coronal planes, and seven patients axial plane also. Sections 7-mm thick with a gap 256 matrix were obtained in the affected region. MR imaging studies repeated during the course
scanning
sagittal
MR
(Gyroscan,
and
parameters coronal
were
images
were
used:
(1
)
spin-echo
obtained
in four
(SE) to
six
500/30 (TR/TE); (2) SE multislice coronal, sagittal, and/or axial images in two acquisitions at 2000/50,1 00; and (3) multislice coronal and sagittal images for two patients with short-tau inversion recovery (STIR), 1 500/i 20/30 (TR/Tl/TE), with two measurements. The area of the body evaluated by the MR study was determined on the basis of clinical findings as well as from the abnormalities acquisitions
detected
at
on previous imaging studies.
Patchy or diffuse loss of signal
On the sonograms, areas of abnormal echogenicity in the thoracic or abdominal wall were identified and were labeled as cystic if they were echo free with increased through transmission and a thin rim, as abscesses if they were relatively echo free with debris and a thick wall, and as a solid mass if they were echogenic. Single-phase static bone scintigraphy was done in all 13 patients by using a gamma camera with a low-energy high-resolution cdlimator and 600 MBq of ‘Tc methylene diphosphonate given IV. Anterior and posterior views (500,000 counts/image) of the entire
from subcutaneous fat and muscle planes on the short TR/TE sequence was considered evidence of involvement. Anatomic distortion
skeleton
in seven
were obtained
3 hr after injection.
Ten patients were examined with CT by using a G.E. 9800 scanner (General Electric, Milwaukee, WI). Contiguous sections (i 0 mm) were scanned through what was suspected as the affected body region from the clinical signs and symptoms and with the guidance of earlier
of
skin,
muscles,
and
other
organs
(in
comparison
with
a normal
contralateral side) was also considered evidence of involvement on this sequence. High signal on long TRITE and STIR sequences in these or other areas was considered indicative of inflammation. All initial radiologic investigations were done within 1 0 days of
admission
to the hospital; all were done preoperatively.
Percutaneous invariably
patients, loculated.
drainage was
of abscess unsuccessful
Complete
surgical
under
CT guidance,
because drainage
the
attempted
abscesses of
the
were
abscesses
(containing 70-i 500 ml of pus) and removal of necrotic infected tissue were achieved in all pyogenic and tuberculous lesions. Only partial resection was possible in one of the mycetoma cases because
992
SHARIF
of adherence
of the
inflammatory
mass
to the
inferior
vena
cava
and
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
aorta. The patient with the more extensive mycetoma had several skin biopsies that showed characteristic histopathologic changes. The different studies were evaluated separately and independently by three radiologists blinded to the final diagnosis. For each imaging technique
and
location
for
and
every
extent
patient,
of the
the
organ
pathology (example, inflammatory was disagreement, the consensus All
imaging
findings
following
lesion,
were
points
were
involvement,
assessed: and
type
vs neoplastic). Whenever opinion was accepted.
subsequently
correlated
with
of
there
ET AL.
AJR:154,
case (Fig. 6) was
surgical
imaging.
two
patients
with
imaging
findings
tions
was
similar
tuberculous
are
summarized
the signal for
all the
infections.
in Table
behavior
In the
The
1 1 patients abscesses
1.
MR
of the abscess with were
collec-
pyogenic
and
isointense
rela-
tive to skeletal muscle on the short TR/TE SE images and became hypenintense on the long TR/TE and STIR images (Figs. 1-4). Epidural extension and associated vertebral Osteomyelitis high
were
signal
proved ribs,
appeared
isointense
exhibited
TE images right
ogenic
only
(Figs.
kidney
(Fig.
abscesses.
Fig. 1.-Case A, Postcontrast
was
process, patients
in two
on long
osteomyelitis
a spinous
In the two and
detected
intensity
and with
patients
(areas
of
images), but pathologically with MR imaging in three
an iliac
mycetoma,
bone
(Fig.
4).
the inflammatory
relative
to muscle
moderate
increase
on short in signal
TR/TE
tissue
images
on long
TR/
5 and 6). The cystic lesions replacing the 5) showed behavior similar to that of pyExtensive epidunal extension in the second
3: 50-year-old diabetic axial CT scan shows
with large extrapleural
of these
TR/TE missed
collection
starting
body
on disk
epidural
extension,
patient
with
the
and
rib
osteomyelitis
(Fig. 2). A small amount of two patients (Fig. 1). night-sided
mycetoma,
of gas
changes
in
the renal bed similar to those of pyogenic abscesses were seen (Fig. 5). The affected areas in both mycetoma patients enhanced minimally after administration of contrast material. Destruction
On MR images
vertebral
On CT, pyogenic and tuberculous abscesses were better visualized on poStcontrast scans and appeared as areas of low attenuation surrounded by an enhancing rim (Figs. 1 and 2). Vertebral osteomyelitis was detected in only one of the
extensive, bral body
Results
with
involvement.
was missed in one patient was seen in the abscesses
data with regard to the size (estimated volume), number (loculations), and extent of the abscesses. The findings at surgery of bone, muscle, and abdominal organ involvement also were compared with those of
The
not associated
May 1990
of
the
muscles
and
replacement
imaging
and
CT findings
were
lated well with surgery, particularly scess collections. In patients who incisions
by
fat
were
and, despite severe epidural extension, no verteor disk involvement was detected (Fig. 6).
for better
drainage,
surgical
comparable
with were
and
come-
respect to the abexplored by large
and
pathologic
proof
of
muscle group involvement correlated with the MR and CT appearances. The presence of osteomyelitis in flat bones (nibs, spinous process, iliac bone) was missed by CT in one and by MR in three cases. Epidural extension in two patients, cleanly noted by both techniques, could not be verified sungically because the area was not explored (Figs. 1 and 2). Two radiologists correctly identified the nature of the disease as being inflammatory on both CT (n 1 0) and MR. The third was in doubt about the MR study of two patients =
(unilateral
mycetoma,
anterior
chest
wall
tuberculosis),
as to
whether the changes were caused by a neoplastic process. No attempts were made by the three radiologists to suggest the possible pathogen except in the case of the unilateral
man with fever, right flank pain, and Staphylococcus aureus septicemia. small fluid collection extending from extrapleural space and dissecting posterior
from level of T4. Note gas within collection,
small abscesses
within erector
to diaphragm.
spinae muscles
This is in continuity (straight arrow), and
inflamed dura suggesting epidural extension (curved arrow). B, Right parasagittal MR image (SE 2000/50) shows extrapleural collection extending from thoracic down to abdominal cavity. Pleura, lung, diaphragm, and liver are displaced anteriorly. C, Midsagittal MR image (SE 2000/100) shows epidural extension (arrowheads) posterior to spinal cord. Signal is higher than that of CSF, presumably because of higher protein content of pus and/or because of its limited mobility. Note high signal intensity in subcutaneous tissue (arrow).
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
May 1990
THORACIC
AND
ABDOMINAL
WALL
INFECTIONS
993
Fig. 2.-Case 2: 47-year-old man with fever and left flank pain of 12 days duration. A, Axial postcontrast CT scan at level of inferior endplate of L3 shows multiloculated abscess involving left psoas muscle extending into neural foramen and involving lateral abdominal wall muscles and subcutaneous fat (asterisk). Staphylococcus aureus was isolated from pus obtained by CT-guided percutaneots biopsy. B, Coronal MR image (SE 500/30) at level of kidneys shows replacement of left perinephric fat by isointense collection that extends into subcutaneous fat (arrow). C, Coronal MR image (SE 2000/100) at same level as B shows multiple hyperintense collections corresponding to isointense areas in B.
Fig. 3.-Case 6: 70-year-old diabetic man with fever and low back pain. A, Axial CT scan at level of L3 shows isodense collection between latissimus dorsi muscle and skin (asterisk). B, Midsagittal MR image (SE 2000/50) shows marked increase in signal of soft tissues pasterior to spinous processes (stars). Staphylococcus aureus was isolated from removed subcutaneous collection.
mycetoma, where two radiologists thought the changes were compatible with tuberculous autonephrectomy (Fig. 5). On sonography, abnormal fluid collections were detected in six patients, but delineation of the extent was obscured by neighboring bone. On scintigraphy, moderate increased uptake of nadionuclide was detected in six patients from ribs, the transverse process, vertebrae, and an iliac bone. The spinous process of L5, which was involved by the infection, showed normal activity. Plain radiognaphs showed abnormal bone changes in three patients and abnormal soft-tissue shadowing in two.
Discussion Infections of the thoracic and abdominal wall are potentially fatal. The prognosis depends on early clinical recognition, the type of causative organism (pyogenic infections being the
more severe), and the extent 1 1 cases in this series with tions
had
quick
and
of tissue pyogenic
complete
involvement [1 -3]. The and tuberculous infec-
recovery
because
of
early
diagnosis and adequate therapy. In this study, CT and MR examinations provided the most useful information about the nature and extent of the infections. Findings with both techniques were similar and correlated well with surgery. MR offered the advantage of direct multiplanar imaging without ionizing radiation and the use of IV contrast agents; CT helped in guiding percutaneous biopsy and/or partial drainage procedures in seven patients. Abscess loculation prevented complete percutaneous drainage. Of the different planes used, sagittal and coronal MR images were found to be most helpful for surgical planning because they clearly displayed the extent and communication sites of the different abscess loculations (Figs. 1 -4). The MR signal of the pyogenic and tuberculous abscesses was com-
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
994
SHARIF
Fig. 4.-Case 9: 23-year-old man with back pain of 6 months duration A, Coronal MR image (SE 500/30) shows large mass inferior to kidney, B, Right parasagittal MR image (STIR 1500/120/30) at level of femoral upper thigh muscles. C, Midsagittal MR image (STIR 1500/120/30) shows large collection canal. At surgery, acid-fast bacilli were isolated together with 1500 ml of
ET AL.
AJR:i54,
radiating to right lower extremity. isointense relative to psoas muscle (asterisk). head (asterisk) shows psoas collection extending
from
iliopsoas
May 1990
muscle
involving skin, subcutaneous tissue, and posterior elements but sparing pus. Spinous process of L5 also was removed because of involvement.
Fig. 5.-Case ness in right upper
13: 43-year-old abdominal
into spinal
man with full-
quadrant
of a few
years duration. A, Postenhancement L3 shows inflammatory
axial CT scan at level of process affecting lateral and posterior abdominal wall (asterisks). Note loss of demarcation between abdominal wall muscles and extension into subcutaneous fat. There is enhancement in residual tissue of destroyed right kidney with rim enhancement around small collection (large arrowhead). Infenor vena cava is displaced to left (small arrow#{149} head). B, Coronal MR image (SE 500/30) shows destroyed right kidney with inflamed pennephric fat displacing inferior vena cava to left (arrow. head). Note thickened lateral abdominal wall together with subcutaneous fat, which has multiple linear areas of echo void, probably representing subcutaneous edema (arrow).
Fig. 6.-Case 12: 34-year-old man with extensive mycetoma involving entire posterior thoracoabdominal wall. A, Axial postcontrast CT scan at level of Ti I
shows large posterior mediastinal mass displacing aorta anteriorly (stars). Mass extends posteriorly
to involve
all muscles
of back.
Patient
was scanned in prone position because he was unable to lie supine. B, Coronal
MR image
(SE 500/30) shows pos(asterisks) extending paraspinally into abdominal cavity. There is al most complete destruction of psoas muscles, but vertebrae and disks appear normal. Spleen is markedly enlarged.
terior mediastinal
mass
Downloaded from www.ajronline.org by 109.161.234.174 on 11/14/15 from IP address 109.161.234.174. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
THORACIC
May 1990
AND
ABDOMINAL
parable to what has been described for abdominal collections [9-1 1 1. The abscesses appeared hypointense on short TR/ TE SE pulse sequences and became hypenintense on the long TR/TE and STIR sequences (Figs. 1 -4). Although the short TR/TE study provided optimal contrast for diffenentiating normal from abnormal structures in the thoracoabdominal wall, it was the long TR/TE and STIR images that helped in better delineation of the location and extent of the abscesses. This concurs with observations made by others in relation to musculoskeletal infections and other types of disease affecting this anatomic region [1 2-1 5]. Rib osteomyelitis was missed with CT and MR imaging, probably because of the relatively small size of these flat bones and also because of the difficulty of imaging the ribs in one plane with either technique. Furthermore, MR imaging failed to depict the presence of osteomyelitis in a spinous process of a vertebral body and in the iliac bone of the patient with the extensive tuberculous abscess, presumably because they were obscured by the hypenintense signal of the overlying abscesses (Fig. 4). Sonography, bone scintigmaphy, and radiography provided limited help in the overall assessment of these lesions. Infections of the thonacoabdominal wall are usually divided into necrotizing fasciitis if localized to the subcutaneous fat and superficial fascia and pyomyositis if they involve the muscles alone. Such a differentiation could be made in only four of our cases where the disease was localized to the skin and superficial fascia (Fig. 3). The nest of the patients had involvement of the skin, subcutaneous fat, superficial fascia, and muscles. Tuberculous abscesses predominantly affecting the soft tissues are infrequent but nevertheless well recognized [7, 1 6, 17]. In the four patients with thonacoabdominal wall tuberculosis, imaging findings were indistinguishable from those in the cases of pyogenic infections (Fig. 4). Mycetoma is a chronic exogenous granulomatous infection that affects the skin and extends to involve deeper tissue [18, 1 9]. True fungi and aerobic actinomycetes are the causative agents, introduced into the skin after trauma usually by a sharp thorn [1 8]. Both types of organisms tend to induce an identical histopathologic picture; diagnosis is achieved by isolation of the organism and/or by characteristic histologic findings [19]. The foot is the most frequently affected site, but in endemic countries the hand, the back, and the metroperitoneum can be involved [1 8-20]. The two cases in this study with actinomycotic mycetoma represent a severe form of this complication (Figs. 5 and 6). The long duration of the illness is not unusual in this disease. In conclusion, this study has shown that MR imaging and CT are equally adequate for evaluating patients with thonacoabdominal
wall
infection.
They
both
provide
useful
infon-
mation about the nature and extent of these lesions, and hence have important impact on their management. CT is the preferred technique whenever percutaneous biopsy and/on drainage is required. MR imaging can be used for the initial
WALL
INFECTIONS
995
diagnosis and is most appropriately used in follow-up therapy because it offers direct multiplanar imaging absence of ionizing radiation and IV contrast agents.
after in the
ACKNOWLEDGMENTS We
thank
Studies ment
the
and
support;
their
cases;
leen
Harrison,
and
this
manuscript.
of Medical
Al Kharj
George
colleagues
radiologic ferring
Departments
of the Riyadh
Hospital
Green
for
for cooperation; and Julie
the
Department
Mcore
for
Administration
and
Programme reading
Medical
for encourage-
the
manuscript;
our
our clinical
colleagues
for re-
of Medical
Illustration,
Kath-
assistance
in the
preparation
of
REFERENCES 1 . Stamenkovic I, Lew PD. Early recognition of potentially fatal necrotizing fasciitis. N EngI J Med 1984;31 0:1689-1693 2. Cave DR. Robinson WA, Brotschi EA. Necrotizing fasciitis following percutaneous endoscopic gastrostomy. Gastrointest Endosc 1986;32: 294-296 3. Fisher JR. Conway MJ, Takeshita AT, Sandoval MR. Necrotizing fasciitis, importance of roentgenographic studies for soft tissue gas. JAMA 1979; 241 :803-806 4. Weiner Cl, Diaconis JN. Primary abdominal wall abscess diagnosed by ultrasound. Arch Surg 1975;1 10:341-342 5. Yeh HC, Aabinowitz JG. Ultrasonography and computed tomography of inflammatory abdominal wall lesions. Radiology 1982; 144 : 859-863 6. Gibson AK, Resenthal SJ, Lukert BP. Pyomyositis, increasing recognition in temperate climates. Am J Med 1984;77 :768-772 7. Coppola J, Muller NL, Connell 0G. Computed tomography of musculoskeletal tuberculosis. J Can Assoc Radiol 1987;38: 199-203 8. Pandolfo I, Balandino A, Gaeta M, Racchiusa 5, Chirico G. CT findings in palpable lesions of the anterior abdominal wall. J Comput Assist Tomogr 1986:10:629-633 9. Wall SD, Fisher MA, Amparo EG, Hricak H, Higgins CB. Magnetic resonance imaging in the evaluation of abscesses. AJR 1985;1 44:1217-1221 10. Beltran J, Noto AM, McGhee RB, Freedy AM, McCalla MS. Infections of the musculoskeletal system. High-field strength MR imaging. Radiology
1987;164:449-454 1 1 . Beltran J, McGhee RB, Shaffer PB, et al. Experimental infections musculoskeletal system. Evaluation with MR imaging and Tc-99m and Ga-67 scintigraphy. Radiology 1988;167: 167-1 72
12. Weinreb JC, Cohen JM, Maravilla 13.
normal anatomy and disease. Lee JK, Glazer HS. Psoas 1986;160:683-687
KR. lliopsoas
Radiology
muscles:
of the MDP
MR study of
1985;156:435-440
muscle disorders:
MR imaging.
Radiology
14. De Roes A, Van Meerten EP, Bloem JL, Bluem AG. MRI of tuberculous spondylitis. AJR 1986;146:79-82 1 5. Haggar AM, Pearlberg carcinoma of the lung: 1087
16. Chafetz
JL, Froelich JW, et al. Chest wall invasion by detection by MR imaging. AJR 1987;148:1075-
N, Genant HK, Hoaglund
with progressive
bony destruction.
FT. lschiogluteal J Can Assoc
tuberculous bursitis 1982:33:119-120 disease in the muscu-
Radiol
17. Butorac A, Littlejohn GO, Hcoper J. Mycobacterial loskeletal system. Med J Aust I987;147:388-391
1 8. Mariat F, Desfombes P, Segretain G. The mycetomas: clinical features, pathology, etiology and epidemiology. Contrib Microbiol Immunol 1977; 4:1-39 19. Lewall DB, Ofale 5, Bendl B. Mycetoma. Skeletal Radio! 1985;14:
257-262 20.
Gumaa SA, Mahgoul ES, El Sid MA. Mycetoma J Trop Med Hyg 1986;35:594-600
of the head and neck. Am
