Abdominal Alec Steven
J.
Megibow,
MD
W. Medwid,
Bowel
Emil J. Balthazar, MD #{149} Bernard A. Bimbaum,
Obstruction:
histories
determined
the
pres-
ence or absence of bowel obstruction. Sixty-four patients ultimately proved to have intestinal obstruction, and 20 did not. Diagnosis was established by means of surgery (n = 39), barium studies (n = 17), and clinical course (n = 28). Causes of obstruction included adhesions (n = 37), metastases (n = 6), primary tumor (n = 7), Crohn disease (n = 4), hernia (n = 3), hematoma (n = 2), colonic diverticulitis (n = 2), and other (n = 3). In addition, 83 CT examinations in patients with no history or indication of intestinal obstruction were simultaneously reviewed. The overall sensitivity was 94%, specificity was 96%, and accuracy was 95%. The cause of obstruction was correctly predicted in 47 of 64 cases (73%). Intestinal obstruction was not diagnosed in any of the 83 control patients. CT is most useful in patients with a history of abdominal malignancy
and
in patients
who
have
not been operated on and who have signs of infection, bowel infarction, or a palpable abdominal mass. Index terms: Intestinal neoplasms, CT, 758.32, 758.33 #{149}Intestines, CT, 758.1211 #{149} Intestines, diseases, 758.262 #{149} Intestines, stenosis or obstruction, 74.7, 758.7291 Radiology
1991;
180:313-318
#{149} Kyunghee
#{149}
MD
Eighty-four computed tomographic (CT) scans from patients referred for bowel obstruction between January 2, 1988, and December 31, 1989, were retrospectively evaluated. A pair of radiologists without knowledge of patient
and MD
Gastrointestinal
C. Cho, MD Marilyn E. Noz, PhD
Evaluation
T
diagnosis is usually
with
of bowel made
the
basis
(CT)
has
been
shown
of
to be useful
in revealing the site, level, and cause of obstruction, and in displaying signs of threatened bowel viability (5-18). To our knowledge, no formal study attempting to assess the use of CT in evaluating patients with bowel obstruction has been published. With this study, we attempted to determine the usefulness of CT in patients with suspected bowel obstruction. MATERIALS Records
diagnosis struction
AND
of 169 patients of small bowel
seen
and December
METHODS with a clinical or colonic ob-
betweenjanuary
31, 1989,
1,1988,
were
CT’ ative
obstruction
on
clinical signs and patient history. Radiologic examination, most often by means of plain abdominal radiography, is confirmatory (1,2). If this examination is not diagnostic, or if the clinical picture is confusing, barium studies are performed to show the site, level, and seventy of the obstruction and to attempt to determine whether strangulation is present or imminent (3,4). Computed tomography
Radiolofv
reviewed.
Ninety-five patients were referred for CT, and 84 scans were available for review. Eighty-three CT examinations in patients with no history or indication of intestinal obstruction were randomly chosen to serve as controls. This group comprised 78 inpatients and five outpatients. In this group, 38 had known neoplasms (colorectal, 11; lung, bladder, gynecologic, and lymphoma, four each; pancreas and melanoma, three; renal cell, two; breast, one), and 27 were evaluated for acute abdominal inflammatory processes. (Seventeen of these 27 were studied within the perioper-
‘ From the Department of Radiology, New York University Medical Center, 560 First Aye, New York, NY 10016 (A.J.M., E.J.B., B.A.B., M.E.N.); the Department of Radiology, Albert Einstein College of Medicine, New York (K.C.C.); and the Department of Radiology, St Luke’s-Roosevelt Hospital, New York (S.W.M.). From the 1990 RSNA scientific assembly. Received December 3, 1990; revision requested January 8,1991; final revision received February 25; accepted March 11. Address reprint requests to A.J.M. ORSNA, 1991 See also the editorial by Rubesin (pp 307-308) in this issue.
period
following
gery.) The remaining variety of indications, and
eurysms
Medical ips
were
performed
anwith
and CT/T 8800 (GE Milwaukee; 159 and four
respectively)
Medical
for a
aortic
9800
Systems,
examinations,
sur-
studied
including cysts.
renal
CT examinations three units: CTfF
abdominal
18 were
Systems,
and T60 (Phil-
Shelton,
Conn;
four
examinations). Seventy-six patients were given contrast medium intravenously by use of a dynamic sequential bolus technique, administering approximately 45 g of iodine
by means
of power
injector.
A
rapid (1.5-2.5 mLlsec) 50-mL bolus injection was given before the commencement of scanning and was immediately followed by a sustained infusion (0.8 mLlsec) of the remaining
volume.
Section
thick-
with 12-mm spacing over the liver and 15-mm spacing through the remainder of the abdomen. All studies were monitored by either an attending physician or imaging fellow. In eight ness
was
cases,
10 mm,
repeat
est were
images
over
obtained
with
the
area
5-mm
of inter-
collimation
and interleaving section indexing. In all cases, either 1.2% barium or 2% iodinated water-soluble contrast medium was administered orally 30-45 minutes before the scanning actually commenced. In many cases, oral contrast medium did not reach
the level
of obstruction
(Figs
1-3). Because time of contrast medium administration differed in relation to the onset of the scanning, and because of varying patient
compliance
full recommended propagation
in consuming
volume,
of contrast
the
the distal
material
was
not
evaluated. Many patients were unable to drink, and contrast medium was administered by the clinical staff through nasogastic tubes while the patients were still in the intensive care unit. It is unknown how much
contrast
medium
ultimately
was
received. None of the patients received any rectally administered contrast agent. The 167 CT examinations were randomized and interpreted by either of two pairs of readers consisting of two experienced gastrointestinal radiologists. Each pair was unfamiliar
with
final diagnosis.
Abbreviation:
the patient
history
The readers
were
CI
=
confidence
and
asked
to
intervaL
313
evaluate whether obstruction was present and, if present, the level and cause of the obstruction. The presence or absence of obstruction was determined by means of consensus of the two readers with a fivepoint confidence scale (0 = none, 1 = probably no obstruction, 2 = indeterminate, 3 = probable obstruction, 4 = definite obstruction). readings were
used
Only the consensus for analysis.
Bowel obstruction was considered to be present at CT when distended bowel loops were seen proximal to collapsed loops. If this transition zone could not be defined, obstruction was considered indeterminate or absent. Diagnoses were established study course sions
at surgery (n = 39) and barium 17), and by means of clinical (n = 28). Causes included adhe(ti = 37), metastases (n = 6), primary tumor (n = 7), Crohn disease (n = 4), hernia (n = 3), hematoma (n = 2), colonic di(i
=
verticulitis
(n = 2), and intussusception, gallstone ileus, and appendicitis (one each). When a point of transition from dilated bowel to normal caliber without apparent cause was identified, adhesions
were
presumed
to be the cause
of obstruc-
lion. Sixteen patients clinically suspected of bowel obstruction were scanned and were proven not to have bowel obstruclion. For purposes of statistical analysis, any response of 3 or 4 was considered positive, and any response less than 3 was considered negative. The results of the readings were analyzed by using a x2 test with Yeats correction, and 95% confidence intervals (CIs) were calculated for specificity, sensitivity,
Berry
and
method
accuracy,
according
to the
(19).
a.
b.
Figure 1. Bowel obstruction due to sigmoid carcinoma. (a) Scan through midpelvis reveals two dilated small bowel loops (short solid arrows). Note the presence of air-fluid levels. The sigmoid colon (long solid arrows) is dilated and filled with soft fecal material. The rectum
(open
96%
(CI
was
95%
out
84 patients
obstruction,
examined
64 ultimately
to rule proved
to have bowel obstruction (55, small bowel; nine, colon). In 16 of the remaining 20 cases, an absence of obstruction was correctly diagnosed. Of the 167 total cases considered in the study, 60 true-positive, tive, four false-negative,
was
mistakenly
obstruction. patients with were correctly
Within
diagnosed
this
nonobstructive identified
group,
=
89.3%-96.7%);
314
#{149}
Radiology
specificity
no
was
and
four
bowel
were
as 2). Ten
study
days
later,
showed
right
per-
as normal
sions
were
lysed
third
case,
the
nal loops interpreted lion, but
(rated finding
ob-
0). Although
this
further, as the
false-positive case, no surgery ium studies were performed, sults of CT were interpreted mal (rated 0), but the clinical
diagnosis was obstruction.resulting
believed
This
signs
patient’s
jeju-
patient the barium standard
a
or barand reas nor-
to represent from adhesions. ameliorated
was
was
not
distended recognized
considered
of an
obstructing
A fifth
case
was
correctly
cephalic colos-
(all rated 4), a the dilated small
a fluid-filled,
colon.
nosed
during a barium study was as causing partial obstrucCT scans were read as nor-
(rated
into
as the
(Fig
to be evilesion
in the
of nonobstructive diagnosed.
cause
in 27 of 37 cases
In a
of kinked
administered
The cause of the obstruction was correctly predicted in 47 of 64 cases (73%). Adhesions were correctly diag-
1), yet adhe-
at laparotomy.
orally
dence ileus
struction. At surgery, adhesions were lysed. In a second case, CT scans obtamed 2 weeks before surgery were
read
though
colon
4), which
a second
high-grade
even
(b) Scan at slightly more patient underwent diverting
other three patients transition zone from
false-negative
examinations
is possible
of obstruction. (arrow). The
accuracy
of reference, findings rendering this true CT-negative case. In the fourth
as
obstruction. Ninety-four scans were interpreted as normal, eight as probably normal, two as indeterminate, nine as probable obstruction, and 54 as definite obstruction. Of the false-positive diagnoses, three were rated 4 and one was rated 3. Two false-negative cases were rated as 0, one was rated as 1, and one as 2. The overall sensitivity was 94% (CI
of the
was not examined study results served
four ileus
as having
the point carcinoma
90.6%-97.4%).
=
CT
The diagnosis
formed 5 and 14 days before surgery to lyse adhesions. In one case, results of a barium study performed 5 days prior to CT were normal, and results of CT also were interpreted as normal
mat
and four diagnoses of obstruction None of the 83 control
false-positive were made. cases
99 true-nega-
caliber.
91.9%-98.1%) (CI
the
barium
RESULTS
is of normal
=
In two cases,
(rated
Of the
arrow)
contrast medium has not reached position reveals the obstructing tomy and subsequent resection.
fol-
of the
(73%).
obstruction
When
other
causes were evident (27 cases), the presence of obstruction was correctly established with CT in all, and the cause was correctly predicted in 20 of the 27 (74%). Within this group, 13 patients had neoplastic disease (seven primary, six metastatic), and the presence of obstruction was correctly established with CT in all. Neoplasm was correctly predicted with CT as the cause of the obstruction in 12 cases. In one patient with diffuse carcinomatosis, obstruction was attributed to neoplastic disease, but was actually caused by a histologically benign adhesion. In the four patients
with Crohn recognized cific
cause
disease, obstruction in all cases, but the was
correctly
was spe-
identified
in
lowing Cantor tube placement. The four patients with false-positive findings all had nonobstructive ileus. One of these patients had fo-
only one. In each case, however, abnormal bowel was commented on. Readers failed to specify an etiologic factor for obstruction in one case each
cally dilated bowel mesenteric abscess, preted as obstruction
of acute sigmoid diverticulitis, dicitis, and gallstone-induced (Table).
loops adjacent to a which was inter(rated 3). In the
appenileus
August
1991
Figure 2. Bowel obstruction due to adhesions. (a) Scan through middle area of kidneys reveals dilated duodenum (D), jejunum (I), and ileum (I). The ascending and descending colon (C) are normal in caliber. These findings are diagnostic of bowel obstruction. (b) At the “transition zone,” a dilated, fluid-filled loop is seen in the immediate proximity of a collapsed loop of distal ileum (curved arrow). Several other collapsed loops are evident (straight arrows). No mass is seen. Adhesions were lysed at surgery. Note, as in Figure 1, oral contrast medium at the site of obstruction is not necessary to either diagnose obstruction or identify its cause.
Figure
3.
dergone
Bowel
obstruction
due
an appendectomy
to adhesions
approximately
with
strangulation.
1 year previously.
(a) The
patient
presented
gle straight arrow) are evident. The ascending and descending colon (curved arrows) are between the leaves of the small bowel mesentery (double straight arrows). (b) At the level tal ileal loops (curved arrows). Proximal to this are dilated loops of small bowel; the more
intramural
edema
(straight
gery, a strangulated required.
arrows).
6-inch
(15-cm)
This finding small
intestinal
The possibility of ischemic changes in bowel was commented on in three cases. In one case, the patient had long-standing radiation enteritis, but obstruction was due to adhesions. In the
other
two
edematous
Volume
patients,
bowel
180
Number
#{149}
ischemic,
associated
2
with
lo-
is associated volvulus
with
Scan at the level of the iliac crests.
with was
high-grade
reduced.
obstruction
Cloudy
cal cloudy pentoneal exudate found at surgery. These two had closed loop obstructions
fluid
was
seen
the
data
confirmed
the validity
of the
=
data
jejunum
of bowel
(open
obstruction;
he
arrow)
and
had
ileum
un-
(sin-
of normal caliber. A small amount of fluid is trapped of the middle sacroiliac joints, note the collapsed disanterior loop has a “target” appearance indicative of with vascular compromise (strangulation). At sur-
was patients (Fig 3).
Results of a x2 analysis showed achieved statistical significance at P < .001. A power level of w
symptoms
Dilated
0.72
at
in the
99% (n
peritoneal
cavity.
CI for a sample =
Resection
of this
was
not
size
167).
DISCUSSION Intestinal obstruction tered virtually every
day
is encounin hospital-
Radiology
315
#{149}
Figure
4.
Nonobstructive
ileus
secondary
to pelvic
disease and cramping abdominal pain. Multiple pelvis. Dilated, fluid-filled small bowel is seen. struction. If careful search of the hepatic flexure was seen in two of our cases. Further prospective
based
radiologic
practices.
ogist’s role is to establish the diagnosis, determine site(s),
establish
to predict
radiol-
Results
or exclude the level and
a cause,
whether
The
inflammatory
and
ple causes (1). The radiologic diagnosis of intestinal obstruction can be made by using a variety of imaging techniques, including plain radiographic studies, medium
studies
(oral
or rec-
tal), enteroclysis, and cross-sectional imaging. The use of these procedures depends and the
on the particular
through
midabdomen
of Obstruction
for Bowel
Obstruction
Total
TP(C+)
in patient
FN
FP
37
27
6
0
4
0
6 7 4
6 6 1
Hernia Hematoma
3 2
3 2
Diverticulitis
2
Intussusception Gallstone ileus
1
1 1 0 0 0
0 1 3 0 0 1 0 1 1 0
0 0 0 0 0 0 0 0 0 99
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 4
47
13
99
4
4
Appendicitis Noobstruction
1 1 103
Total
167
Note.-FN = obstruction not diagnosed, obstruction present; FP = obstruction diagnosed, no obstruction present; TN = obstruction not diagnosed, obstruction not present; TP(c+) = obstruction diagnosed, cause specified; TP(C-) = obstruction diagnosed, cause not specified.
clinical presentation knowledge sought
questions that should be answered in any case of suspected intestinal obstruction: (a) Is the bowel obstructed? (b) What is the level, cause, and severity? (c) Is strangulation present? (d) Is
with neoplasms, obstruction was diagnosed in 25 (49%), equivocally suggested in 10 (20%), present but not detected in eight (16%), and not doc-
derwent In 68%
umented
In cases
conservative
tients with hernias, obstruction was diagnosed in 20 (43%), equivocally suggested in 17 (37%), and present but not detected in 10 (20%) (1).
plain
50%-60% tients
sufficient
radiographic
radiologic intestinal estimated
of cases; may
evidence 154
patients
tion
was
radiography suggested
316
therapy
or is
required?
The
up to 20%
no plain of obstruction
diagnosed
is the
of pa-
radiographic (1,20-22).
adhesions,
by use
mented
of plain not
in 14 (9%),
in five
and
(3%)
in eight
not
docu-
ries,
(1). In 51 patients
(15%)
(1). In 47 pa-
Barium studies are the standard of reference in diagnosing and characterizing intestinal obstruction. In the Mayo
In
obstruc-
in 89 (58%), equivocally in 46 (30%), present but
Radiology
#{149}
study
technique used to obstruction, and to be conclusive in
have
with
inflammatory
TN
detected
simplest diagnose has been
pelvic
TP(C-)
in an individual case. Herlinger and Maglinte (3) have identified four key
surgery
with
Metastasis Primary neoplasm Crohndisease
Adhesions
is
present. Eighty percent of all bowel obstructions are caused either by adhesions (50%), hernias (15%), or neoplasms (15%), and 35.5% of cases of bowel obstruction result from multi-
contrast
(a) Scan
of CT Examination
Cause
attempt
strangulation
disease.
loops are dilated and fluid filled, as is the right colon (RC). (b) Scan through midcolon is of normal caliber (arrow). This case was mistakenly diagnosed as oban obstructing lesion, nonobstructive ileus should be considered. This pattern should be attempted before using this sign toward making a diagnosis.
small-bowel The descending fails to reveal evaluation
Clinic
series,
23%
of patients
with adhesions, 27% with neoplasm, and 4% with hernia underwent contrast material-enhanced supplement plain clinical evaluation
studies radiographic (1). In another
to and Se-
of 327 patients
21%
cause
of
small
be-
obstruction
contrast-enhanced of peroral and
grade tional
75%
un-
studies. of retro-
colonic studies, useful addiinformation was obtained (22).
struction ema study nal ileum opacification diagnosis out colonic ministered revealing
there
admitted
bowel
in which
distal
is suspected, with reflux and retrograde
bowel a contrast into the small
obentermibowel
is a fast, safe method of (23). This method also rules obstruction. Barium adorally is relatively rapid in early obstructions, when
is exaggerated
peristalsis;
how-
ever, ability
as motor activity diminishes, the to propagate the bolus decreases. Since most patients delay medical evaluation until late in their
August
1991
course
(1),
the
examinations
are
ally performed when intestinal ity has significantly decreased, the examination may require Enteroclysis
is becoming
usu-
motiland
hours. the
recom-
mended method for evaluating small bowel disease (3) and has been shown to be diagnostic in patients with closed loop obstruction (24). Herlinger and Maglinte tabulated data from small bowel enema studies in 92 patients with small bowel obstruction. The overall accuracy was 85% (3). Enteroclysis has several disadvantages that
have
limited
its acceptance.
Sev-
eral workers utilize intravenous sedation for patients (3), yet this may not be feasible in critically ifi patients in the perioperative period. If the patient already has a nasogastric tube in place, a second tube must be placed to reach the proximal jejunum. Enteroclysis
is contraindicated
in patients
with complete obstruction or in those likely to have bowel infarction. The large amount of barium necessary may delay cross-sectional imaging in patients with fever, malignancy, or diseases
with
prominent
perientenc
components, such as Crohn disease. Several articles, mostly case reports, have demonstrated the usefulness of CT in a wide variety of obstructive processes
(5-18).
The
use
of CT
in
evaluating patients with bowel obstruction has several advantages. On the basis of findings of a discrepancy in the caliber of proximal and distal bowel, we were able to correctly identify obstruction at a true-positive rate of 94% and a true-negative rate of 96%. The overall accuracy was 95%. Most cases were read at the extremes of the CI scale. This clearly indicates that CT can aid in excluding a diagnosis of bowel
obstruction.
Furthermore,
the cause of obstruction dicted in 73% of cases.
was preIn the present
study,
not
the
readers
did
know
the
case histories, but in clinical practice, the radiologist more often has some background information and may be better able to determine the cause of obstruction. By virtue of its ability to demonstrate the bowel wall, CT may possibly aid in predicting which patients are in danger of bowel strangulation (Fig 3) (25). Although we did not specifically test this hypothesis in this study, several reports have described the CT appearance of intestinal volvulus and closed loop obstruction with vascular compromise (15-18). The specificity of CT in revealing bowel infarction is controversial. Smerud et al retrospectively reviewed 23 cases of mesenteric infarction and Volume
180
Number
#{149}
2
found CT was not significantly better than plain radiography in detection of bowel infarction (26). However, the time between the radiographic examination and surgery was not controlled. Alpern et al reviewed 23 cases and found specific signs of bowel infarction in 13%-26% of them (27). Mesenteric infarction can occur rapidly, and unless CT scans and plain radiographs are obtained simultaneously and within hours of surgery, it may be difficult to draw any significant conclusions. The finding of thickened, regionally enhancing bowel wall with submucosal edema (target sign) and localized ascites within the leaves of the small bowel mesentery must be considered
highly
suggestive
of intes-
tinal ischemia (Fig 3) and may be an indication for rapid surgical decompression. CT enabled correct identification of both cases of edematous (probably ischemic) bowel in our series,
and
the
patients
underwent
rapid surgical decompression without bowel resection, possibly before infarction. Patients with a history of abdominal malignancy represent a significant number of those admitted for bowel obstruction and may maximally benefit from CT. Surgical versus conservative management is based on the cause of the obstruction: adhesive bands, peritoneal carcinomatosis, radiation enteritis, and so on. The percentage of cancer patients in whom bowel obstruction is due to adhesions is 21%-38% (28,29). CT can be used to determine the presence of peritoneal carcinomatosis by directly imaging peritoneal surfaces (30). Mass must be evident along the serosa of the bowel at the transition zone; otherwise, adhesions may need to be considered as a possible cause. We did not test the ability of CT to differentiate between these two factors in a large number of patients. In this series, 13 patients had obstruction secondary to neoplastic disease. In all, obstruction was corredly identified; in 12, the correct diagnosis
of malignant
obstruction
was
made. In the case in which a benign adhesion was the cause, the etiologic diagnosis was not made. In one patient with known ovarian carcinoma, the prediction of small bowel obstruction secondary to adhesions was correctly made based on the lack of peritoneal mass. In patients with a dilated right colon, nonobstructive ileus was mistaken for mechanical bowel obstruction (Fig 4). Five patients had nonobstructive ileus, and the four false-positive diagnoses occurred in
this
patient
group.
When
a dilated
right colon is visualized, particularly with a normal-caliber transverse and descending colon, one could rescan the region of the hepatic flexure to look for an obstructing lesion; however, the simplest approach is to perform a contrast enema study to assess colonic patency. Until more experience is gained with CT appearance of ileus, in equivocal cases, the contrast enema study is recommended. In using CT to aid in diagnosing obstruction, the evaluation of the level of obstruction must be carefully analyzed. It is easy to assume that more distal bowel loops are displayed on more distal scans. As the obstructed loops distend, they assume a linear orientation, roughly aligning along the long axis of the small bowel mesentery.
Thus,
proximal
jejunal
loops may be visualized in the pelvis and ileal loops in the upper abdomen. Care should also be taken to identify a transition zone. The easiest sequence of observations is to evaluate the case in a retrograde fashion, following the course of the distal colon proximally until dilated loops are encountered. If the patient is still on the scanner table, acquisition of thin (5mm), closely spaced sections through the transition zone may clarify the cause of obstruction. CT is a relatively expensive examination; therefore, guidelines for its rational use should be developed. Plain radiographic examination was diagnostic of bowel obstruction in many of these patients referred for CT, but clinicians will require further evaluation (possibly with CT) in certam definable situations, estimated at approximately 20%-30% of cases (1,20,22). We believe that CT is the procedure of choice in patients with a history of abdominal malignancy and clinical symptoms suggestive of bowel obstruction. CT has a secondary role in postsurgical patients without cancer who most likely have adhesive obstruction. In patients who have no history of abdominal surgery and who present with symptoms of bowel obstruction, CT is most valuable when there are systemic signs suggestive of infection, bowel infarction, or an associated palpable mass. In these cases, CT may help confirm the presumptive diagnosis
or reveal
other
causes,
such
as appendicitis or diverticulitis. Effects of CT versus barium enema studies on management of simple clinical symptoms of bowel obstruction remain to be explored. Furthermore, prospective analyses of the findings in ileus versus mechanical obstruction and Radiology
#{149} 317
the accuracy of CT in predicting intestinal ischemia remain to be performed. U References 1. 2
3.
4.
5.
6.
Mucha
7.
computed 1984;
12.
13.
736. Berry bowel
Obturator
tomography.
DF, Willing obstruction
SJ, Beers GJ.
CT 1987;
11:707-708. Passas V, Karavias D, Grilias D, Birbas A. Computed tomography of left paraduodenal hernia.J Comput Assist Tomogr 1986;
14.
15.
18.
Grumbach stone
ileus
K, Levine diagnosed
MS.
WexlerJA.
by computed
Obstructive Z Gastroenterol
M, Hoffman R, ileus and acute 1990;
28:206-
P, Segal I, Epstein B, Essop R, C. Total obstruction of the ascolon complicating acute pancreatitis. AmJ Gastroenterol 1983; 78:28-30. Shaff MI, Himmelfarb E, Sacks GA, Burks DD, Kullcrani MV. The whirl sign: a CT finding in volvulus of the large bowel. J Comput Assist Tomogr 1985; 9:410. FischerJK Computed tomographic diagnosis of volvulus in intestinal mairotation. Radiology 1981; 140:145-146. Baithazar EJ, BaumanJS, Megibow AJ. CT diagnosis of dosed loop obstruction. Comput Assist Tomogr 1985; 9:953-955. Cho KC, Hoffman-TretinJC, Alterman DD. Closed-loop obstruction of the small bowel: CT and sonographic appearance. J Comput Assist Suzuki
Tomogr 1989; 13:256-258. M, Takashima T, Funaki of strangulated obstruction
computed 19.
Gall-
tomog-
Am J Gastroenterol
20.
24.
Marcerollo
A case
10:542-543.
9.
F.
tomography.
EurJ
25.
Dunn
YF,
Halls
JM, Berne
26.
studies Arch
101:367-378. TV. Roentgeno-
in acute
Surg
small 119:
1984;
1305-1308. Miller RE, Miller WJ. Small bowel obstruction: complete reflux small bowel examination. AmJ Surg 1985; 10:572-577. PriceJ, Nolan DJ. Closed loop obstruction: diagnosis by enteroclysis. Gastrointest Radiol 1989; 14:251-254. Lund EC, Han SY, Holley HC, Berland LL. Intestinal ischemia: comparison of plain
findings
graphic findings. 8:1083-1108.
Smerud
MJ,Johnson
and computed
RadioGraphics
tomo-
1988;
CD, Stephens
DH.
of bowel infarction: a comparison of plain films and CT scans in 23 cases. Diagnosis
AJR 1990; 27.
28.
29.
154:99-103. MB, Glazer GM, Francis IL Ischemic or infarcted bowel: CT findings. Radiology 1988; 166:149-152. Ketcham AS, Hoye RC, Pilch YH, Morton DL. Delayed intestinal obstruction following treatment for cancer. Cancer 1970; 25:406-410. Osteen RT, Guyton 5, Steele C Jr, Wilson Alpern
Malignant intestinal obstruction. gery 1980; 87:611-615. Walkey MM, Friedman AC, Sohotra
Sur-
RE.
with
1983;
sulfate suspensions
radiographic
H, et al.
Radiol
SW, Christoforides
barium
graphic contrast bowel obstruction.
presen-
the value of 23.
FriCk TW, Schrettzenmaier
AJ. The use of in the study of obstruction of the
Nelson
suspected mechanical small intestine. AJR 1967; 22.
Elisseos cending
Small
due to an enterolith: Comput Assist Tomogr
tomography.
21.
L, et aL wall
79:55-58.
Largiader pancreatitis. 207.
17.
appearance.J 8.
11.
hernia diagnosed by AJR 1983; 140:735.-
E.
AB, Needleman
tations of acute pancreatitis:
10:
1986;
Cross sectional imaging of abdominal hernias. AJR 1989; 153:517-521. Epstein BM, Hertzanu Y. Clinical
16.
computed
Tomogr
Wechsler
Computed tomography and magnetic resonance imaging of the acute abdomen. Surg Clin North Am 1988; 68:233-254. Cubillo
RJ, Kurtz
Assist
10.
P.
Small intestinal obstruction. Surg Clin North Am 1987; 67:597-620. TIbblin S. Diagnosis of intestinal obstruction with special regard to plain roentgen examinations of the abdomen. Ada Chir Scand 1969; 135:249-252. Herlinger H, Maglinte DDT. Small bowel obstruction. In: Herlinger H, Maglinte DDT, eds. Clinical radiology of the small boweL Philadelphia: Saunders, 1989; 479509. Smith SM, Kubicka RA, Smith C. Small bowel obstruction. In: Thompson WM, ed Common problems in gastrointestinal radiology. Chicago: Year Book Medical, 1989; 56-65. Shaff MI, Tarr RW, Partain CL, James AEJr.
J Comput
raphy. 146-148.
30.
P. CT manifestations neal cardnomatosis. AJR 1988;
13:256-258.
Radecki
Berry CC. Perspective: a tutorial on confidence intervals for proportions in diagnos-
1041.
P,
of perito150:1035-
tic radiology. AJR 1990; 154:477-481. Lo AM, Evans WE, Carey LC. Review of small bowel obstruction at Milwaukee County General Hospital. Am J Surg 1986; 11:884-887.
318
#{149} Radiology
August
1991
J.
Megibow,
MD
W. Medwid,
Bowel
Emil J. Balthazar, MD #{149} Bernard A. Bimbaum,
Obstruction:
histories
determined
the
pres-
ence or absence of bowel obstruction. Sixty-four patients ultimately proved to have intestinal obstruction, and 20 did not. Diagnosis was established by means of surgery (n = 39), barium studies (n = 17), and clinical course (n = 28). Causes of obstruction included adhesions (n = 37), metastases (n = 6), primary tumor (n = 7), Crohn disease (n = 4), hernia (n = 3), hematoma (n = 2), colonic diverticulitis (n = 2), and other (n = 3). In addition, 83 CT examinations in patients with no history or indication of intestinal obstruction were simultaneously reviewed. The overall sensitivity was 94%, specificity was 96%, and accuracy was 95%. The cause of obstruction was correctly predicted in 47 of 64 cases (73%). Intestinal obstruction was not diagnosed in any of the 83 control patients. CT is most useful in patients with a history of abdominal malignancy
and
in patients
who
have
not been operated on and who have signs of infection, bowel infarction, or a palpable abdominal mass. Index terms: Intestinal neoplasms, CT, 758.32, 758.33 #{149}Intestines, CT, 758.1211 #{149} Intestines, diseases, 758.262 #{149} Intestines, stenosis or obstruction, 74.7, 758.7291 Radiology
1991;
180:313-318
#{149} Kyunghee
#{149}
MD
Eighty-four computed tomographic (CT) scans from patients referred for bowel obstruction between January 2, 1988, and December 31, 1989, were retrospectively evaluated. A pair of radiologists without knowledge of patient
and MD
Gastrointestinal
C. Cho, MD Marilyn E. Noz, PhD
Evaluation
T
diagnosis is usually
with
of bowel made
the
basis
(CT)
has
been
shown
of
to be useful
in revealing the site, level, and cause of obstruction, and in displaying signs of threatened bowel viability (5-18). To our knowledge, no formal study attempting to assess the use of CT in evaluating patients with bowel obstruction has been published. With this study, we attempted to determine the usefulness of CT in patients with suspected bowel obstruction. MATERIALS Records
diagnosis struction
AND
of 169 patients of small bowel
seen
and December
METHODS with a clinical or colonic ob-
betweenjanuary
31, 1989,
1,1988,
were
CT’ ative
obstruction
on
clinical signs and patient history. Radiologic examination, most often by means of plain abdominal radiography, is confirmatory (1,2). If this examination is not diagnostic, or if the clinical picture is confusing, barium studies are performed to show the site, level, and seventy of the obstruction and to attempt to determine whether strangulation is present or imminent (3,4). Computed tomography
Radiolofv
reviewed.
Ninety-five patients were referred for CT, and 84 scans were available for review. Eighty-three CT examinations in patients with no history or indication of intestinal obstruction were randomly chosen to serve as controls. This group comprised 78 inpatients and five outpatients. In this group, 38 had known neoplasms (colorectal, 11; lung, bladder, gynecologic, and lymphoma, four each; pancreas and melanoma, three; renal cell, two; breast, one), and 27 were evaluated for acute abdominal inflammatory processes. (Seventeen of these 27 were studied within the perioper-
‘ From the Department of Radiology, New York University Medical Center, 560 First Aye, New York, NY 10016 (A.J.M., E.J.B., B.A.B., M.E.N.); the Department of Radiology, Albert Einstein College of Medicine, New York (K.C.C.); and the Department of Radiology, St Luke’s-Roosevelt Hospital, New York (S.W.M.). From the 1990 RSNA scientific assembly. Received December 3, 1990; revision requested January 8,1991; final revision received February 25; accepted March 11. Address reprint requests to A.J.M. ORSNA, 1991 See also the editorial by Rubesin (pp 307-308) in this issue.
period
following
gery.) The remaining variety of indications, and
eurysms
Medical ips
were
performed
anwith
and CT/T 8800 (GE Milwaukee; 159 and four
respectively)
Medical
for a
aortic
9800
Systems,
examinations,
sur-
studied
including cysts.
renal
CT examinations three units: CTfF
abdominal
18 were
Systems,
and T60 (Phil-
Shelton,
Conn;
four
examinations). Seventy-six patients were given contrast medium intravenously by use of a dynamic sequential bolus technique, administering approximately 45 g of iodine
by means
of power
injector.
A
rapid (1.5-2.5 mLlsec) 50-mL bolus injection was given before the commencement of scanning and was immediately followed by a sustained infusion (0.8 mLlsec) of the remaining
volume.
Section
thick-
with 12-mm spacing over the liver and 15-mm spacing through the remainder of the abdomen. All studies were monitored by either an attending physician or imaging fellow. In eight ness
was
cases,
10 mm,
repeat
est were
images
over
obtained
with
the
area
5-mm
of inter-
collimation
and interleaving section indexing. In all cases, either 1.2% barium or 2% iodinated water-soluble contrast medium was administered orally 30-45 minutes before the scanning actually commenced. In many cases, oral contrast medium did not reach
the level
of obstruction
(Figs
1-3). Because time of contrast medium administration differed in relation to the onset of the scanning, and because of varying patient
compliance
full recommended propagation
in consuming
volume,
of contrast
the
the distal
material
was
not
evaluated. Many patients were unable to drink, and contrast medium was administered by the clinical staff through nasogastic tubes while the patients were still in the intensive care unit. It is unknown how much
contrast
medium
ultimately
was
received. None of the patients received any rectally administered contrast agent. The 167 CT examinations were randomized and interpreted by either of two pairs of readers consisting of two experienced gastrointestinal radiologists. Each pair was unfamiliar
with
final diagnosis.
Abbreviation:
the patient
history
The readers
were
CI
=
confidence
and
asked
to
intervaL
313
evaluate whether obstruction was present and, if present, the level and cause of the obstruction. The presence or absence of obstruction was determined by means of consensus of the two readers with a fivepoint confidence scale (0 = none, 1 = probably no obstruction, 2 = indeterminate, 3 = probable obstruction, 4 = definite obstruction). readings were
used
Only the consensus for analysis.
Bowel obstruction was considered to be present at CT when distended bowel loops were seen proximal to collapsed loops. If this transition zone could not be defined, obstruction was considered indeterminate or absent. Diagnoses were established study course sions
at surgery (n = 39) and barium 17), and by means of clinical (n = 28). Causes included adhe(ti = 37), metastases (n = 6), primary tumor (n = 7), Crohn disease (n = 4), hernia (n = 3), hematoma (n = 2), colonic di(i
=
verticulitis
(n = 2), and intussusception, gallstone ileus, and appendicitis (one each). When a point of transition from dilated bowel to normal caliber without apparent cause was identified, adhesions
were
presumed
to be the cause
of obstruc-
lion. Sixteen patients clinically suspected of bowel obstruction were scanned and were proven not to have bowel obstruclion. For purposes of statistical analysis, any response of 3 or 4 was considered positive, and any response less than 3 was considered negative. The results of the readings were analyzed by using a x2 test with Yeats correction, and 95% confidence intervals (CIs) were calculated for specificity, sensitivity,
Berry
and
method
accuracy,
according
to the
(19).
a.
b.
Figure 1. Bowel obstruction due to sigmoid carcinoma. (a) Scan through midpelvis reveals two dilated small bowel loops (short solid arrows). Note the presence of air-fluid levels. The sigmoid colon (long solid arrows) is dilated and filled with soft fecal material. The rectum
(open
96%
(CI
was
95%
out
84 patients
obstruction,
examined
64 ultimately
to rule proved
to have bowel obstruction (55, small bowel; nine, colon). In 16 of the remaining 20 cases, an absence of obstruction was correctly diagnosed. Of the 167 total cases considered in the study, 60 true-positive, tive, four false-negative,
was
mistakenly
obstruction. patients with were correctly
Within
diagnosed
this
nonobstructive identified
group,
=
89.3%-96.7%);
314
#{149}
Radiology
specificity
no
was
and
four
bowel
were
as 2). Ten
study
days
later,
showed
right
per-
as normal
sions
were
lysed
third
case,
the
nal loops interpreted lion, but
(rated finding
ob-
0). Although
this
further, as the
false-positive case, no surgery ium studies were performed, sults of CT were interpreted mal (rated 0), but the clinical
diagnosis was obstruction.resulting
believed
This
signs
patient’s
jeju-
patient the barium standard
a
or barand reas nor-
to represent from adhesions. ameliorated
was
was
not
distended recognized
considered
of an
obstructing
A fifth
case
was
correctly
cephalic colos-
(all rated 4), a the dilated small
a fluid-filled,
colon.
nosed
during a barium study was as causing partial obstrucCT scans were read as nor-
(rated
into
as the
(Fig
to be evilesion
in the
of nonobstructive diagnosed.
cause
in 27 of 37 cases
In a
of kinked
administered
The cause of the obstruction was correctly predicted in 47 of 64 cases (73%). Adhesions were correctly diag-
1), yet adhe-
at laparotomy.
orally
dence ileus
struction. At surgery, adhesions were lysed. In a second case, CT scans obtamed 2 weeks before surgery were
read
though
colon
4), which
a second
high-grade
even
(b) Scan at slightly more patient underwent diverting
other three patients transition zone from
false-negative
examinations
is possible
of obstruction. (arrow). The
accuracy
of reference, findings rendering this true CT-negative case. In the fourth
as
obstruction. Ninety-four scans were interpreted as normal, eight as probably normal, two as indeterminate, nine as probable obstruction, and 54 as definite obstruction. Of the false-positive diagnoses, three were rated 4 and one was rated 3. Two false-negative cases were rated as 0, one was rated as 1, and one as 2. The overall sensitivity was 94% (CI
of the
was not examined study results served
four ileus
as having
the point carcinoma
90.6%-97.4%).
=
CT
The diagnosis
formed 5 and 14 days before surgery to lyse adhesions. In one case, results of a barium study performed 5 days prior to CT were normal, and results of CT also were interpreted as normal
mat
and four diagnoses of obstruction None of the 83 control
false-positive were made. cases
99 true-nega-
caliber.
91.9%-98.1%) (CI
the
barium
RESULTS
is of normal
=
In two cases,
(rated
Of the
arrow)
contrast medium has not reached position reveals the obstructing tomy and subsequent resection.
fol-
of the
(73%).
obstruction
When
other
causes were evident (27 cases), the presence of obstruction was correctly established with CT in all, and the cause was correctly predicted in 20 of the 27 (74%). Within this group, 13 patients had neoplastic disease (seven primary, six metastatic), and the presence of obstruction was correctly established with CT in all. Neoplasm was correctly predicted with CT as the cause of the obstruction in 12 cases. In one patient with diffuse carcinomatosis, obstruction was attributed to neoplastic disease, but was actually caused by a histologically benign adhesion. In the four patients
with Crohn recognized cific
cause
disease, obstruction in all cases, but the was
correctly
was spe-
identified
in
lowing Cantor tube placement. The four patients with false-positive findings all had nonobstructive ileus. One of these patients had fo-
only one. In each case, however, abnormal bowel was commented on. Readers failed to specify an etiologic factor for obstruction in one case each
cally dilated bowel mesenteric abscess, preted as obstruction
of acute sigmoid diverticulitis, dicitis, and gallstone-induced (Table).
loops adjacent to a which was inter(rated 3). In the
appenileus
August
1991
Figure 2. Bowel obstruction due to adhesions. (a) Scan through middle area of kidneys reveals dilated duodenum (D), jejunum (I), and ileum (I). The ascending and descending colon (C) are normal in caliber. These findings are diagnostic of bowel obstruction. (b) At the “transition zone,” a dilated, fluid-filled loop is seen in the immediate proximity of a collapsed loop of distal ileum (curved arrow). Several other collapsed loops are evident (straight arrows). No mass is seen. Adhesions were lysed at surgery. Note, as in Figure 1, oral contrast medium at the site of obstruction is not necessary to either diagnose obstruction or identify its cause.
Figure
3.
dergone
Bowel
obstruction
due
an appendectomy
to adhesions
approximately
with
strangulation.
1 year previously.
(a) The
patient
presented
gle straight arrow) are evident. The ascending and descending colon (curved arrows) are between the leaves of the small bowel mesentery (double straight arrows). (b) At the level tal ileal loops (curved arrows). Proximal to this are dilated loops of small bowel; the more
intramural
edema
(straight
gery, a strangulated required.
arrows).
6-inch
(15-cm)
This finding small
intestinal
The possibility of ischemic changes in bowel was commented on in three cases. In one case, the patient had long-standing radiation enteritis, but obstruction was due to adhesions. In the
other
two
edematous
Volume
patients,
bowel
180
Number
#{149}
ischemic,
associated
2
with
lo-
is associated volvulus
with
Scan at the level of the iliac crests.
with was
high-grade
reduced.
obstruction
Cloudy
cal cloudy pentoneal exudate found at surgery. These two had closed loop obstructions
fluid
was
seen
the
data
confirmed
the validity
of the
=
data
jejunum
of bowel
(open
obstruction;
he
arrow)
and
had
ileum
un-
(sin-
of normal caliber. A small amount of fluid is trapped of the middle sacroiliac joints, note the collapsed disanterior loop has a “target” appearance indicative of with vascular compromise (strangulation). At sur-
was patients (Fig 3).
Results of a x2 analysis showed achieved statistical significance at P < .001. A power level of w
symptoms
Dilated
0.72
at
in the
99% (n
peritoneal
cavity.
CI for a sample =
Resection
of this
was
not
size
167).
DISCUSSION Intestinal obstruction tered virtually every
day
is encounin hospital-
Radiology
315
#{149}
Figure
4.
Nonobstructive
ileus
secondary
to pelvic
disease and cramping abdominal pain. Multiple pelvis. Dilated, fluid-filled small bowel is seen. struction. If careful search of the hepatic flexure was seen in two of our cases. Further prospective
based
radiologic
practices.
ogist’s role is to establish the diagnosis, determine site(s),
establish
to predict
radiol-
Results
or exclude the level and
a cause,
whether
The
inflammatory
and
ple causes (1). The radiologic diagnosis of intestinal obstruction can be made by using a variety of imaging techniques, including plain radiographic studies, medium
studies
(oral
or rec-
tal), enteroclysis, and cross-sectional imaging. The use of these procedures depends and the
on the particular
through
midabdomen
of Obstruction
for Bowel
Obstruction
Total
TP(C+)
in patient
FN
FP
37
27
6
0
4
0
6 7 4
6 6 1
Hernia Hematoma
3 2
3 2
Diverticulitis
2
Intussusception Gallstone ileus
1
1 1 0 0 0
0 1 3 0 0 1 0 1 1 0
0 0 0 0 0 0 0 0 0 99
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 4
47
13
99
4
4
Appendicitis Noobstruction
1 1 103
Total
167
Note.-FN = obstruction not diagnosed, obstruction present; FP = obstruction diagnosed, no obstruction present; TN = obstruction not diagnosed, obstruction not present; TP(c+) = obstruction diagnosed, cause specified; TP(C-) = obstruction diagnosed, cause not specified.
clinical presentation knowledge sought
questions that should be answered in any case of suspected intestinal obstruction: (a) Is the bowel obstructed? (b) What is the level, cause, and severity? (c) Is strangulation present? (d) Is
with neoplasms, obstruction was diagnosed in 25 (49%), equivocally suggested in 10 (20%), present but not detected in eight (16%), and not doc-
derwent In 68%
umented
In cases
conservative
tients with hernias, obstruction was diagnosed in 20 (43%), equivocally suggested in 17 (37%), and present but not detected in 10 (20%) (1).
plain
50%-60% tients
sufficient
radiographic
radiologic intestinal estimated
of cases; may
evidence 154
patients
tion
was
radiography suggested
316
therapy
or is
required?
The
up to 20%
no plain of obstruction
diagnosed
is the
of pa-
radiographic (1,20-22).
adhesions,
by use
mented
of plain not
in 14 (9%),
in five
and
(3%)
in eight
not
docu-
ries,
(1). In 51 patients
(15%)
(1). In 47 pa-
Barium studies are the standard of reference in diagnosing and characterizing intestinal obstruction. In the Mayo
In
obstruc-
in 89 (58%), equivocally in 46 (30%), present but
Radiology
#{149}
study
technique used to obstruction, and to be conclusive in
have
with
inflammatory
TN
detected
simplest diagnose has been
pelvic
TP(C-)
in an individual case. Herlinger and Maglinte (3) have identified four key
surgery
with
Metastasis Primary neoplasm Crohndisease
Adhesions
is
present. Eighty percent of all bowel obstructions are caused either by adhesions (50%), hernias (15%), or neoplasms (15%), and 35.5% of cases of bowel obstruction result from multi-
contrast
(a) Scan
of CT Examination
Cause
attempt
strangulation
disease.
loops are dilated and fluid filled, as is the right colon (RC). (b) Scan through midcolon is of normal caliber (arrow). This case was mistakenly diagnosed as oban obstructing lesion, nonobstructive ileus should be considered. This pattern should be attempted before using this sign toward making a diagnosis.
small-bowel The descending fails to reveal evaluation
Clinic
series,
23%
of patients
with adhesions, 27% with neoplasm, and 4% with hernia underwent contrast material-enhanced supplement plain clinical evaluation
studies radiographic (1). In another
to and Se-
of 327 patients
21%
cause
of
small
be-
obstruction
contrast-enhanced of peroral and
grade tional
75%
un-
studies. of retro-
colonic studies, useful addiinformation was obtained (22).
struction ema study nal ileum opacification diagnosis out colonic ministered revealing
there
admitted
bowel
in which
distal
is suspected, with reflux and retrograde
bowel a contrast into the small
obentermibowel
is a fast, safe method of (23). This method also rules obstruction. Barium adorally is relatively rapid in early obstructions, when
is exaggerated
peristalsis;
how-
ever, ability
as motor activity diminishes, the to propagate the bolus decreases. Since most patients delay medical evaluation until late in their
August
1991
course
(1),
the
examinations
are
ally performed when intestinal ity has significantly decreased, the examination may require Enteroclysis
is becoming
usu-
motiland
hours. the
recom-
mended method for evaluating small bowel disease (3) and has been shown to be diagnostic in patients with closed loop obstruction (24). Herlinger and Maglinte tabulated data from small bowel enema studies in 92 patients with small bowel obstruction. The overall accuracy was 85% (3). Enteroclysis has several disadvantages that
have
limited
its acceptance.
Sev-
eral workers utilize intravenous sedation for patients (3), yet this may not be feasible in critically ifi patients in the perioperative period. If the patient already has a nasogastric tube in place, a second tube must be placed to reach the proximal jejunum. Enteroclysis
is contraindicated
in patients
with complete obstruction or in those likely to have bowel infarction. The large amount of barium necessary may delay cross-sectional imaging in patients with fever, malignancy, or diseases
with
prominent
perientenc
components, such as Crohn disease. Several articles, mostly case reports, have demonstrated the usefulness of CT in a wide variety of obstructive processes
(5-18).
The
use
of CT
in
evaluating patients with bowel obstruction has several advantages. On the basis of findings of a discrepancy in the caliber of proximal and distal bowel, we were able to correctly identify obstruction at a true-positive rate of 94% and a true-negative rate of 96%. The overall accuracy was 95%. Most cases were read at the extremes of the CI scale. This clearly indicates that CT can aid in excluding a diagnosis of bowel
obstruction.
Furthermore,
the cause of obstruction dicted in 73% of cases.
was preIn the present
study,
not
the
readers
did
know
the
case histories, but in clinical practice, the radiologist more often has some background information and may be better able to determine the cause of obstruction. By virtue of its ability to demonstrate the bowel wall, CT may possibly aid in predicting which patients are in danger of bowel strangulation (Fig 3) (25). Although we did not specifically test this hypothesis in this study, several reports have described the CT appearance of intestinal volvulus and closed loop obstruction with vascular compromise (15-18). The specificity of CT in revealing bowel infarction is controversial. Smerud et al retrospectively reviewed 23 cases of mesenteric infarction and Volume
180
Number
#{149}
2
found CT was not significantly better than plain radiography in detection of bowel infarction (26). However, the time between the radiographic examination and surgery was not controlled. Alpern et al reviewed 23 cases and found specific signs of bowel infarction in 13%-26% of them (27). Mesenteric infarction can occur rapidly, and unless CT scans and plain radiographs are obtained simultaneously and within hours of surgery, it may be difficult to draw any significant conclusions. The finding of thickened, regionally enhancing bowel wall with submucosal edema (target sign) and localized ascites within the leaves of the small bowel mesentery must be considered
highly
suggestive
of intes-
tinal ischemia (Fig 3) and may be an indication for rapid surgical decompression. CT enabled correct identification of both cases of edematous (probably ischemic) bowel in our series,
and
the
patients
underwent
rapid surgical decompression without bowel resection, possibly before infarction. Patients with a history of abdominal malignancy represent a significant number of those admitted for bowel obstruction and may maximally benefit from CT. Surgical versus conservative management is based on the cause of the obstruction: adhesive bands, peritoneal carcinomatosis, radiation enteritis, and so on. The percentage of cancer patients in whom bowel obstruction is due to adhesions is 21%-38% (28,29). CT can be used to determine the presence of peritoneal carcinomatosis by directly imaging peritoneal surfaces (30). Mass must be evident along the serosa of the bowel at the transition zone; otherwise, adhesions may need to be considered as a possible cause. We did not test the ability of CT to differentiate between these two factors in a large number of patients. In this series, 13 patients had obstruction secondary to neoplastic disease. In all, obstruction was corredly identified; in 12, the correct diagnosis
of malignant
obstruction
was
made. In the case in which a benign adhesion was the cause, the etiologic diagnosis was not made. In one patient with known ovarian carcinoma, the prediction of small bowel obstruction secondary to adhesions was correctly made based on the lack of peritoneal mass. In patients with a dilated right colon, nonobstructive ileus was mistaken for mechanical bowel obstruction (Fig 4). Five patients had nonobstructive ileus, and the four false-positive diagnoses occurred in
this
patient
group.
When
a dilated
right colon is visualized, particularly with a normal-caliber transverse and descending colon, one could rescan the region of the hepatic flexure to look for an obstructing lesion; however, the simplest approach is to perform a contrast enema study to assess colonic patency. Until more experience is gained with CT appearance of ileus, in equivocal cases, the contrast enema study is recommended. In using CT to aid in diagnosing obstruction, the evaluation of the level of obstruction must be carefully analyzed. It is easy to assume that more distal bowel loops are displayed on more distal scans. As the obstructed loops distend, they assume a linear orientation, roughly aligning along the long axis of the small bowel mesentery.
Thus,
proximal
jejunal
loops may be visualized in the pelvis and ileal loops in the upper abdomen. Care should also be taken to identify a transition zone. The easiest sequence of observations is to evaluate the case in a retrograde fashion, following the course of the distal colon proximally until dilated loops are encountered. If the patient is still on the scanner table, acquisition of thin (5mm), closely spaced sections through the transition zone may clarify the cause of obstruction. CT is a relatively expensive examination; therefore, guidelines for its rational use should be developed. Plain radiographic examination was diagnostic of bowel obstruction in many of these patients referred for CT, but clinicians will require further evaluation (possibly with CT) in certam definable situations, estimated at approximately 20%-30% of cases (1,20,22). We believe that CT is the procedure of choice in patients with a history of abdominal malignancy and clinical symptoms suggestive of bowel obstruction. CT has a secondary role in postsurgical patients without cancer who most likely have adhesive obstruction. In patients who have no history of abdominal surgery and who present with symptoms of bowel obstruction, CT is most valuable when there are systemic signs suggestive of infection, bowel infarction, or an associated palpable mass. In these cases, CT may help confirm the presumptive diagnosis
or reveal
other
causes,
such
as appendicitis or diverticulitis. Effects of CT versus barium enema studies on management of simple clinical symptoms of bowel obstruction remain to be explored. Furthermore, prospective analyses of the findings in ileus versus mechanical obstruction and Radiology
#{149} 317
the accuracy of CT in predicting intestinal ischemia remain to be performed. U References 1. 2
3.
4.
5.
6.
Mucha
7.
computed 1984;
12.
13.
736. Berry bowel
Obturator
tomography.
DF, Willing obstruction
SJ, Beers GJ.
CT 1987;
11:707-708. Passas V, Karavias D, Grilias D, Birbas A. Computed tomography of left paraduodenal hernia.J Comput Assist Tomogr 1986;
14.
15.
18.
Grumbach stone
ileus
K, Levine diagnosed
MS.
WexlerJA.
by computed
Obstructive Z Gastroenterol
M, Hoffman R, ileus and acute 1990;
28:206-
P, Segal I, Epstein B, Essop R, C. Total obstruction of the ascolon complicating acute pancreatitis. AmJ Gastroenterol 1983; 78:28-30. Shaff MI, Himmelfarb E, Sacks GA, Burks DD, Kullcrani MV. The whirl sign: a CT finding in volvulus of the large bowel. J Comput Assist Tomogr 1985; 9:410. FischerJK Computed tomographic diagnosis of volvulus in intestinal mairotation. Radiology 1981; 140:145-146. Baithazar EJ, BaumanJS, Megibow AJ. CT diagnosis of dosed loop obstruction. Comput Assist Tomogr 1985; 9:953-955. Cho KC, Hoffman-TretinJC, Alterman DD. Closed-loop obstruction of the small bowel: CT and sonographic appearance. J Comput Assist Suzuki
Tomogr 1989; 13:256-258. M, Takashima T, Funaki of strangulated obstruction
computed 19.
Gall-
tomog-
Am J Gastroenterol
20.
24.
Marcerollo
A case
10:542-543.
9.
F.
tomography.
EurJ
25.
Dunn
YF,
Halls
JM, Berne
26.
studies Arch
101:367-378. TV. Roentgeno-
in acute
Surg
small 119:
1984;
1305-1308. Miller RE, Miller WJ. Small bowel obstruction: complete reflux small bowel examination. AmJ Surg 1985; 10:572-577. PriceJ, Nolan DJ. Closed loop obstruction: diagnosis by enteroclysis. Gastrointest Radiol 1989; 14:251-254. Lund EC, Han SY, Holley HC, Berland LL. Intestinal ischemia: comparison of plain
findings
graphic findings. 8:1083-1108.
Smerud
MJ,Johnson
and computed
RadioGraphics
tomo-
1988;
CD, Stephens
DH.
of bowel infarction: a comparison of plain films and CT scans in 23 cases. Diagnosis
AJR 1990; 27.
28.
29.
154:99-103. MB, Glazer GM, Francis IL Ischemic or infarcted bowel: CT findings. Radiology 1988; 166:149-152. Ketcham AS, Hoye RC, Pilch YH, Morton DL. Delayed intestinal obstruction following treatment for cancer. Cancer 1970; 25:406-410. Osteen RT, Guyton 5, Steele C Jr, Wilson Alpern
Malignant intestinal obstruction. gery 1980; 87:611-615. Walkey MM, Friedman AC, Sohotra
Sur-
RE.
with
1983;
sulfate suspensions
radiographic
H, et al.
Radiol
SW, Christoforides
barium
graphic contrast bowel obstruction.
presen-
the value of 23.
FriCk TW, Schrettzenmaier
AJ. The use of in the study of obstruction of the
Nelson
suspected mechanical small intestine. AJR 1967; 22.
Elisseos cending
Small
due to an enterolith: Comput Assist Tomogr
tomography.
21.
L, et aL wall
79:55-58.
Largiader pancreatitis. 207.
17.
appearance.J 8.
11.
hernia diagnosed by AJR 1983; 140:735.-
E.
AB, Needleman
tations of acute pancreatitis:
10:
1986;
Cross sectional imaging of abdominal hernias. AJR 1989; 153:517-521. Epstein BM, Hertzanu Y. Clinical
16.
computed
Tomogr
Wechsler
Computed tomography and magnetic resonance imaging of the acute abdomen. Surg Clin North Am 1988; 68:233-254. Cubillo
RJ, Kurtz
Assist
10.
P.
Small intestinal obstruction. Surg Clin North Am 1987; 67:597-620. TIbblin S. Diagnosis of intestinal obstruction with special regard to plain roentgen examinations of the abdomen. Ada Chir Scand 1969; 135:249-252. Herlinger H, Maglinte DDT. Small bowel obstruction. In: Herlinger H, Maglinte DDT, eds. Clinical radiology of the small boweL Philadelphia: Saunders, 1989; 479509. Smith SM, Kubicka RA, Smith C. Small bowel obstruction. In: Thompson WM, ed Common problems in gastrointestinal radiology. Chicago: Year Book Medical, 1989; 56-65. Shaff MI, Tarr RW, Partain CL, James AEJr.
J Comput
raphy. 146-148.
30.
P. CT manifestations neal cardnomatosis. AJR 1988;
13:256-258.
Radecki
Berry CC. Perspective: a tutorial on confidence intervals for proportions in diagnos-
1041.
P,
of perito150:1035-
tic radiology. AJR 1990; 154:477-481. Lo AM, Evans WE, Carey LC. Review of small bowel obstruction at Milwaukee County General Hospital. Am J Surg 1986; 11:884-887.
318
#{149} Radiology
August
1991
