DOI:10.2214/AJR.05.1179
AJR 2006; 187:1179-1183
© American Roentgen Ray Society
Accuracy of MDCT in Predicting Site of Gastrointestinal Tract Perforation
Bernard Hainaux1,2,
Emmanuel Agneessens2,
Raphael Bertinotti1,
Viviane De Maertelaer3,
Erika Rubesova1,
Elie Capelluto4 and
Constantin Moschopoulos1
1 Department of Radiology, Centre Hospitalier Universitaire St.-Pierre,
Université Libre de Bruxelles, 322 Rue Haute, Brussels 1000,
Belgium.
2 Department of Radiology, HIS Site Etterbeek-Ixelles, Brussels, Belgium.
3 Statistical Unit, Institut de Recherche Interdisciplinaire en Biologie Humaine
et Moléculaire, Université Libre de Bruxelles, Brussels,
Belgium.
4 Department of Gastrointestinal Surgery, Centre Hospitalier Universitaire
St.-Pierre, Université Libre de Bruxelles, Brussels, Belgium.
Received July 8, 2005;
accepted after revision September 18, 2005.
Address correspondence to B. Hainaux
(hainauxb{at}yahoo.fr).
Abstract
OBJECTIVE. The purpose of this study was to prospectively evaluate
the accuracy of MDCT for preoperative determination of the site of surgically
proven gastrointestinal tract perforations and to determine the most
predictive findings in this diagnosis.
SUBJECTS AND METHODS. We prospectively studied 85 consecutive
patients with extraluminal air on MDCT who had surgically proven
gastrointestinal tract perforations. All patients underwent surgery within 12
hours after MDCT was performed. Two experienced radiologists, blinded to the
surgical diagnosis, reached a consensus prediction of the site of the
perforation using the following eight MDCT findings: concentration of
extraluminal air bubbles adjacent to the bowel wall, free air in
supramesocolic or inframesocolic compartments, extraluminal air in both
abdomen and pelvis, focal defect in the bowel wall, segmental bowel-wall
thickening, perivisceral fat stranding, abscess, and extraluminal fluid. MDCT
imaging results were compared with surgical and pathologic findings. Logistic
regression analyses were performed to assess the significance of the different
radiologic criteria.
RESULTS. Analysis of MDCT images was predictive of the site of
gastrointestinal tract perforation in 73 (86%) of 85 patients. Logistic
regression showed that concentration of extraluminal air bubbles (p
< 0.001), segmental bowel wall thickening (p < 0.001), and
focal defect of the bowel wall (p = 0.007) were strong predictors of
the site of bowel perforation.
CONCLUSION. MDCT is highly accurate for predicting the site of
gastrointestinal tract perforations. Three of eight CT findings significantly
correlate with surgical diagnosis.
Keywords: abdominal imaging • bowel • CT • gastrointestinal radiology • gastrointestinal tract perforation • MDCT
Introduction
Free intraperitoneal air in association with acute abdominal pain is a
major finding in the diagnosis of gastrointestinal tract perforation, usually
leading to surgical treatment
[1]. A laparoscopic procedure
is currently performed instead of open surgery, even for acute conditions such
as gastroduodenal perforated ulcer and acute perforated appendicitis. As a
consequence, it is beneficial for surgeons to know in advance where the bowel
tract is perforated. During the past few years, numerous investigations have
established CT as the imaging technique of choice for diagnosing
pneumoperitoneum [2,
3]. However, few authors have
studied the accuracy of CT to localize the site of the gastrointestinal tract
perforation, and no large prospective series focusing on this point has been
reported. Because an accurate preoperative diagnosis is helpful for surgeons,
we undertook this prospective study to evaluate the accuracy of MDCT in the
identification of the site of bowel perforation and to objectively assess the
CT findings that accurately lead to diagnosis.
Subjects and Methods
Patients
Between January 2003 and June 2004, we prospectively studied 85 consecutive
patients presenting with acute abdominal pain and pneumoperitoneum on CT. All
patients had surgically proven gastrointestinal tract perforation. The study
included 43 men and 42 women with an age range of 16 to 94 years (mean age, 56
years). Patients younger than 16 years were not included. Surgery was
performed on all patients within 12 hours after detection of extraluminal air
and provided the exact location of the perforation site. Institutional review
board approval was obtained for this study.
CT Examinations
Abdominal examinations were performed on a 4-MDCT scanner (Somatom Plus
Volume Zoom, Siemens Medical Solutions) with the following parameters:
collimation, 4 x 2.5 mm; table speed, 15 mm per 0.5 second of scanner
rotation; 120 kV; and 130 effective milliampere seconds. All the patients
underwent scanning in the supine position starting from the diaphragm down to
the pubic symphysis. None of the patients received oral or rectal contrast
material; unenhanced scanning was performed in all patients. In 46 patients,
additional contrast-enhanced images were acquired with the same parameters
after injection of 120 mL of a low-osmolarity contrast medium (iobitridol,
Xenetix 350 mg I/mL, Guerbet) at a rate of 2.5 mL/s via a power injector.
Scanning was started 80 seconds after initiating contrast material
injection.
Image Analysis
The images were stored on compact disks and reviewed on a workstation
(Wizard, Siemens Medical Solutions). Two experienced abdominal radiologists
evaluated axial and multiplanar reconstruction images. They were blinded to
all records but were informed that all patients presented with acute abdominal
pain and their abdominal CT scans showed pneumoperitoneum. Together, the
radiologists completed a questionnaire on the presence or absence of eight CT
findings indicative of acute bowel disease and gastrointestinal perforation.
The eight CT findings evaluated were few bubbles of extraluminal air
concentrated in close proximity to the bowel wall, extraluminal air
exclusively in supramesocolic or inframesocolic compartments, extraluminal air
in both abdomen and pelvis, focal defect in the bowel wall, segmental
bowel-wall thickening, perivisceral inflammatory fat stranding, abscess, and
extraluminal fluid. Bowel-wall thickening was a subjective criterion evaluated
without measurement of the bowel wall. After reviewing these eight findings,
the two radiologists were asked to determine by consensus the site of bowel
perforation. The predicted sites of perforation were categorized as follows:
stomach and duodenum, jejunum and ileum, appendix, cecum and ascending colon,
transverse colon, descending colon, sigmoid colon, and rectum.
Statistical Analysis
For each of the eight CT findings, the percentage of patients with positive
results for the finding among the correct diagnoses and the percentage of
patients with negative results for the finding among the 12 patients where the
perforation site could not be determined—no diagnosis was
false—were calculated. The percentage of correct diagnoses when each
finding was present and the percentage of correct diagnoses when each finding
was absent were also calculated. Logistic regression analyses were performed
to evaluate the probability of a correct diagnosis as a function of the
presence or absence of any of the eight findings. A first analysis was
performed using the enter procedure. Rao's efficient scores
[4] for each of the eight
findings were calculated to detect which findings were statistically
associated with a correct diagnosis. A second analysis was performed with a
stepwise forward procedure to detect the successive most relevant findings for
correctly predicting the localization of the perforation. Statistical analyses
were performed using SPSS version 13.0 (Statistical Package for the Social
Sciences) for Windows (Microsoft).
The link between the location of free intraabdominal air and the site of
the perforation was assessed with a chi-square test. Exact p values
were reported. Statistical analysis software (StatXact version 5.0, Cytel
Software) was used.
Results
The perforation sites identified by surgery were the following: 37 stomach
or duodenum perforations (34 gastroduodenal peptic ulcers, one perforated
gastric carcinoma, two anastomotic dehiscences after gastric bypass procedures
for morbid obesity), nine perforations of the small intestine (six ischemia,
two perforating tumors, one blunt abdominal trauma), six acute perforated
appendicitis cases, seven cecum or ascending colon perforations (three
ischemia, one cecal volvulus, one diverticulitis, one iatrogenic complication
after polypectomy, one idiopathic perforation), two transverse colon
perforations (one polypectomy, one postoperative leak), six descending colon
perforations (five postoperative leaks, one blunt abdominal trauma), 12
sigmoid colon perforations (10 diverticulitis, one ischemia, one sigmoid
volvulus), and six rectum perforations (one carcinoma, one perforation after
polypectomy, two perforations after improper cleansing enema, two
postoperative leaks).
The prediction of the site of gastrointestinal tract perforation using the
eight CT findings was correct in 73 (86%) of 85 patients. The diagnostic
values for each CT finding are listed in
Table 1. Rao's scores for each
finding (Table 2) showed that
three findings—concentration of extraluminal air bubbles (Figs.
1A and
1B), focal defect of the bowel
wall (Figs. 2A and
2B), and segmental bowel-wall
thickening (Fig. 3)—were
statistically significant predictors of correct localization of the
perforation. The logistical analysis regression with the stepwise procedure
also showed that the best predictors for perforation localization were the
presence of concentrated extraluminal air bubbles (p < 0.001)
followed by the presence of segmental bowel-wall thickening (p =
0.046). The three most predictive findings of the site of perforation were not
present in any of the 12 patients for whom the site was undetermined by the
radiologists. We also detected a link (p < 0.001) between the
location of free intraabdominal air and the site of perforation, which is
summarized in Table 3. The 37
patients with upper gastroduodenal perforation all had free intraabdominal air
around the liver or stomach. Eight of these patients also had free air in the
pelvis. None of them presented with pelvic pneumoperitoneum alone. In
contrast, all of the 39 patients with colon perforation had free air in the
pelvis. For 15 of these patients, free air was only observed in the pelvis,
and no patient had free air only in the supramesocolic compartment. Of the
nine patients with small-bowel perforation, six had free air in both the
abdomen and pelvis.
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —69-year-old man presenting with acute abdominal pain few
hours after sigmoid polypectomy. Image at level of sigmoid shows concentration
of extraluminal air bubbles in close proximity to sigmoid wall
(arrows).
|
|
View larger version (120K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3 —42-year-old man with perforated sigmoid colon diverticulitis.
Transverse CT image shows segmental bowel-wall thickening
(arrowheads), mild pericolic fat stranding (asterisk), and
few extraluminal bubbles of air adjacent to the bowel wall
(arrow).
|
|
Discussion
Several authors have shown that CT is the most valuable imaging technique
for detecting free intraperitoneal air
[2,
3]. However, few authors have
investigated the sensitivity of CT in determining the site of gastrointestinal
tract perforation. To our knowledge, only retrospective studies have been
reported. Previous reports have emphasized the limits of CT for determining
the site of perforation. In a retrospective series of 14 patients with
perforated peptic ulcers reported by Chen et al.
[5], the sensitivity of CT for
showing free air was 100%, but the site of perforation could only be
determined by CT in five patients (36%). In the series reported by Fultz et
al. [6] of 11 patients with
perforated peptic ulcers who all received oral contrast material, the site of
perforation could be detected only in the three patients who showed
extravasation of ingested contrast material. The small number of patients with
leakage of oral contrast material was thought to be caused by the rapid
sealing of a large percentage of perforated ulcers. In addition, the supine
position used for CT provides less chance for contrast material to extravasate
with an anterior perforation. As a result, we believe that CT may be better
than gastrointestinal opacification for the diagnosis of gastrointestinal
tract perforation and the determination of the site of perforation. Moreover,
the recent introduction of MDCT has allowed high-speed acquisition, thin-slice
collimation, and reformatting of images in any plane with high spatial
resolution, making this technique particularly suitable for the assessment of
abdominal abnormalities
[7].
In our study, multiplanar reconstruction images were routinely performed,
which we found to be useful for determining the precise location of bubbles of
extraluminal air and defects in the bowel wall. Several articles stress the
value of intraluminal gastrointestinal contrast material in the evaluation of
patients with blunt abdominal trauma. The detection of free extraluminal
contrast material is considered a direct sign of bowel perforation. However,
in a literature review, the reported sensitivity of extravasation of oral
contrast material varies from 19-42%
[8]. The study of
Shanmuganathan et al. [9] shows
that triple-contrast helical CT (CT with administration of rectal, oral, and
IV contrast material) is highly accurate for evaluating patients with
penetrating injuries into the thoracoabdominal region. Nevertheless, only 15%
of patients with bowel injury showed oral or rectal contrast material
extravasation. Our study concerns mainly patients with nontraumatic bowel
perforations who present with peritonitis. In these patients, it is often
difficult to obtain oral or rectal opacification. One of the aims of our study
was to determine the accuracy of CT to diagnose the site of bowel perforation
using CT findings other than extravasation of contrast material.
The overall accuracy in predicting the site of bowel perforation was 86% in
our study. These results are similar to the three previously published
retrospective series
[10-12].
Maniatis et al. [10] reported
a sensitivity of 85.5% in a series of 76 patients presenting with bowel
perforation secondary to various causes. Kim et al.
[11] reported an overall
accuracy of 82% in a series of 57 patients with gastrointestinal tract
perforation after blunt abdominal trauma. The same result, 82%, was also
reported by Catalano [12] in a
series of 38 patients.
Previous reports have emphasized CT manifestations of bowel perforation
secondary to various causes. However, no previous reports have compared CT
findings to statistically evaluate the most useful findings. Ongolo-Zogo et
al. [13] reported on a series
of 10 perforated gastroduodenal peptic ulcers in which two important CT
findings were indicative of the site of perforation: discontinuity in the
bowel wall in six patients and tiny extraluminal air bubbles in close
proximity to the bowel wall in two patients. Miki et al.
[14] also reported on the
direct visualization of a ruptured colonic wall in four of six patients with
colonic perforation. However, these two findings did not occur in the series
of 42 proximal and distal gastrointestinal tract perforations reported by
Yeung et al. [15]. In our
series of 85 perforations, free air bubbles in close proximity to the bowel
wall indicated the site of perforation in 65 patients, with segmental
bowel-wall thickening observed in 42 patients. Those two findings were strong
predictors of the site of perforation (p <0.001). The third
predictor of the site of perforation was the direct visualization of a focal
defect in the bowel wall, seen in 29 patients (p =0.007). Moreover,
in the 12 patients where CT was unable to predict the site of the perforation,
none of these three findings was present.
As has been previously reported in other studies, our series also shows a
link between the location of free intraabdominal air and the site of
perforation [10,
16]. Gastroduodenal and
colonic perforation can both appear as massive pneumoperitoneum with free air
in the abdomen and in the pelvis. However, if free air is present only around
the liver and stomach and not in the pelvis, proximal gastrointestinal
perforation is likely. Conversely, if free air is present only in the pelvis,
the likely site of perforation is the colon or, less frequently, the small
bowel.
In conclusion, MDCT allows physicians not only to detect free
intraabdominal air but also to precisely determine the anatomic site of
gastrointestinal tract perforation without the use of oral or rectal contrast
material. Three CT findings—concentrated bubbles of extraluminal air in
close proximity to the bowel wall, focal defect in the bowel wall, and
segmental bowel-wall thickening—have a high predictive value for the
site of perforation.
Acknowledgments
We thank Anne Hepburn for reviewing the manuscript and Catherine Gerard and
Johan Grymonprez for editorial assistance.
References
- Jeffrey RB, Federle MP, Wall S. Value of computed tomography in
detecting occult gastrointestinal perforation. J Comput Assist
Tomogr 1983; 7:825
-827[Medline]
- Stapakis JC, Thickman D. Diagnosis of pneumoperitoneum: abdominal
CT vs upright chest film. J Comput Assist Tomogr1992; 16:713
-716[Medline]
- Earls JP, Dachman AH, Colon E, Garrett MG, Molloy M. Prevalence and
duration of postoperative pneumoperitoneum: sensitivity of CT vs left lateral
decubitus radiography. AJR 1993;161
: 781-785[Abstract/Free Full Text]
- Rao CR. Linear statistical inference and its
applications, 2nd ed. New York, NY: Wiley,1973
- Chen CH, Huang HS, Yang CC, Yeh YH. Features of perforated peptic
ulcers in conventional computed tomography.
Hepatogastroenterology 2001;48
: 1393-1396[Medline]
- Fultz PJ, Skucas J, Weiss SL. CT in upper gastrointestinal tract
perforations secondary to peptic ulcer disease. Gastrointest
Radiol 1992; 17:5
-8[CrossRef][Medline]
- Catalano C, Fraioli F, Laghi A, et al. High-resolution
multidetector CT in the preoperative evaluation of patients with renal cell
carcinoma. AJR 2003;180
: 1271-1277[Abstract/Free Full Text]
- Becker CD, Mentha G, Schmidlin F, Terrier F. Blunt abdominal trauma
in adults: role of CT in the diagnosis and management of visceral injuries.
Eur Radiol 1998;8
: 772-780[CrossRef][Medline]
- Shanmuganathan K, Mirvis SE, Chiu WC, Killeen KL, Scalea TM. Triple
contrast helical CT in penetrating torso trauma: a prospective study to
determine peritoneal violation and the need for laparotomy.
AJR 2001; 177:1247
-1256[Abstract/Free Full Text]
- Maniatis V, Chryssikopoulos H, Roussakis A, et al. Perforation of
the alimentary tract: evaluation with computed tomography. Abdom
Imaging 2000; 25:373
-379[CrossRef][Medline]
- Kim HC, Shin HC, Park SJ, et al. Traumatic bowel perforation:
analysis of CT findings according to the site and the elapsed time since
accident. Clin Imaging 2004;28
: 334-339[CrossRef][Medline]
- Catalano O. Computed tomography in the study of gastrointestinal
perforation. Radiol Med (Torino) 1996;91
: 247-252
- Ongolo-Zogo P, Borson O, Garcia P, Gruner L, Valette PJ. Acute
gastroduodenal peptic ulcer perforation: contrast-enhanced and thin-section
helical CT findings in 10 patients. Abdom Imaging1999; 24:329
-332[CrossRef][Medline]
- Miki T, Ogata S, Uto M, et al. Multidetector-row CT findings of
colonic perforation: direct visualization of ruptured colonic wall.
Abdom Imaging 2004;29
: 658-662[CrossRef][Medline]
- Yeung KW, Chang MS, Hsiao CP, Huang JF. CT evaluation of
gastrointestinal tract perforation. Clin Imaging2004; 28:329
-333[CrossRef][Medline]
- Ghahremani GG. Radiologic evaluation of suspected gastrointestinal
perforations. Radiol Clin North Am 1993;31
: 1219-1234[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
