| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr, San Antonio, TX 78284-7800. Received August 4, 1998; revision requested September 10; final revision received July 13, 1999; accepted July 26. Address reprint requests to S.C. (e-mail: chopra@uthscsa.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Three radiologists retrospectively assessed the contrast-enhanced abdominal CT scans of 77 patients with cirrhosis and 100 patients without cirrhosis for gastrointestinal wall thickening from the stomach through the descending colon. The frequencies of wall thickening were determined in the cirrhosis and in the control groups and were compared by using the Fisher exact test. The segmental distribution, symmetry, and enhancement pattern were evaluated in all patients with cirrhosis and gastrointestinal wall thickening.
RESULTS: Gastrointestinal wall thickening was seen in 49 (64%) patients with cirrhosis and in seven (7%) control subjects (P < .005). The jejunum and ascending colon were the most common sites of gastrointestinal wall thickening; each was involved in 32 (42%; 95% CI: 30%, 53%) patients. The scans of 30 (61%; 95% CI: 47%, 75%) patients with gastrointestinal wall thickening showed multisegmental distribution. Gastrointestinal wall thickening was concentric and homogeneous in all patients with cirrhosis.
CONCLUSION: Gastrointestinal wall thickening is common on contrast-enhanced abdominal CT scans in patients with cirrhosis. It frequently involves multiple segments. The jejunum and ascending colon are the most common sites of involvement.
Index terms: Abdomen, CT, 70.12114 • Gastrointestinal tract, abnormalities, 70.799 • Liver, cirrhosis, 761.794
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In many instances, gastrointestinal wall thickening appears severe enough to lead to a misdiagnosis of ischemia, inflammation, hemorrhage, or neoplasm of the involved segments of the bowel. At review of the literature, we could find only a few references to the radiologic appearance of gastrointestinal wall thickening in patients with cirrhosis (1,2). These publications describe the appearance of gastrointestinal edema in patients with cirrhosis on barium studies. To our knowledge, there are no published reports describing the frequency and patterns of gastrointestinal wall thickening on CT scans in patients with cirrhosis.
Prompted by our observations, we undertook this study to determine the frequency and to describe the CT imaging patterns of gastrointestinal wall thickening in patients with cirrhosis.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patient Selection
In our institution's pathology department database, 183 patients had hepatic biopsy results positive for cirrhosis between January 1995 and March 1998. From this group, we selected the first 100 patients who had CT scans of the abdomen available and who met the criterion of having no evidence of malignancy or other conditions that may cause gastrointestinal wall thickening, as specified in the next paragraph.
Malignancy was ruled out with preliminary review of the CT images for hepatic or extrahepatic masses and with subsequent laparotomy, follow-up CT, or review of the patient's chart after a clinic visit at least 6 months after the original CT examination. Conditions that may cause gastrointestinal wall thickening—specifically, cardiac failure; nephrotic syndrome; acute pancreatitis; and inflammatory, ischemic, and neoplastic diseases of the gastrointestinal tract—were excluded with a review of the patients' clinical records.
The CT images of all patients were also initially reviewed by the principal investigator (O.I.K.) for adequate bowel wall visualization. This was defined as clear visualization of the bowel wall in a well-distended loop because of the presence of luminal gas, fluid, or orally administered contrast material in most of the length of each section of the gastrointestinal tract.
Of the 100 patients, 23 did not have adequate bowel wall distention and therefore were excluded. The remaining 77 patients formed the test group that will be referred to as the "cirrhosis group" throughout this article. This group consisted of 48 men and 29 women, with an age range of 21–70 years (mean age, 48 years).
We selected the control group from patients in our daily clinical caseload who underwent CT for nonspecific abdominal symptoms or for trauma. We included in the control group the first 100 patients who had adequate bowel wall visualization at initial review of the abdominal CT images and who met the criterion at chart review of having no known evidence of cirrhosis, of malignancy, or of other conditions capable of causing gastrointestinal wall thickening. This group consisted of 44 men and 56 women, with an age range of 18–80 years (mean age, 42 years).
CT Scanning
Of the 77 patients with cirrhosis, 52 patients underwent dual-phase, contrast-enhanced hepatic scanning for pretransplantation evaluation, pre–transjugular intrahepatic portosystemic shunt evaluation, or examination to rule out hepatocellular carcinoma and 25 had undergone single-phase, contrast-enhanced, abdominal-pelvic scanning for miscellaneous indications. All patients in the control group underwent single-phase, contrast-enhanced, abdominal-pelvic scanning.
All scans were obtained with a HiSpeed Advantage (GE Medical Systems, Milwaukee, Wis) or model 2000 scanner (Picker International, Cleveland, Ohio) by using 250–300 mA, 120 kVp, and 7- or 8-mm collimation with a pitch of 1 from the dome of the diaphragm to the iliac crest.
Dual-phase hepatic scanning consisted of obtaining an arterial phase scan 25 seconds and a portal venous phase scan 70 seconds after the initiation of an intravenous power injection of 150 mL of iothalamate meglumine (Conray 60; Mallinckrodt Medical, St Louis, Mo; iodine concentration, 282 mg/mL) or ioversol (Optiray 300; Mallinckrodt Medical; iodine concentration, 300 mg/mL) at 5 mL/sec. As is the routine in our department, the pelvis was not included in dual-phase imaging.
Single-phase, contrast-enhanced abdominal-pelvic imaging consisted of obtaining a portal venous phase scan 70 seconds after the initiation of an intravenous power injection of 120 mL of contrast medium at 3 mL/sec, which was followed at 2 minutes by obtaining 5-mm-thick sections at 10-mm intervals through the pelvis. In all patients, only the images obtained during the portal venous phase were used for this study. All images were obtained at a window width and level of 400 and 40 HU, respectively.
Image Analysis
The CT scans of the patients with cirrhosis and those of the patients without cirrhosis were mixed together randomly. The scans of each patient were then read by a group of three readers (G.D.D., K.N.C., S.C.), all board-certified radiologists, who were not involved in patient selection and who were blinded to the diagnosis. Instances of disagreement among the readers were resolved by majority opinion.
The stomach, duodenum, jejunum, ileum, ascending colon (which included the cecum), transverse colon, and descending colon were considered as separate anatomic sections of the gastrointestinal tract and were assessed for wall thickening in each patient. The jejunum was identified by the location of its loops in the left upper quadrant and by its abundance of folds. The ileum was identified by the location of its loops in the right lower quadrant and pelvis and by its relative paucity of folds. The descending colon was identified as the colonic segment from the summit of the splenic flexure to the pelvic brim. Because the majority of CT scans in patients with cirrhosis did not include the pelvis, the sigmoid colon (the portion of the colon below the pelvic brim) and the rectum were excluded from the assessment.
To diagnose gastrointestinal wall thickening, we used criteria obtained or modified from the literature. The gastric wall was considered to be thickened if its thickness exceeded 1 cm at comparison with the measurement scale on the CT image (3). The small-intestinal and colonic walls were considered thickened if their thickness was more than just perceptible. This criterion was modified from that of Scanlon et al (4), who considered the bowel wall to be thickened if it was thick enough to be measured. The colonic wall was also considered thickened if there was haustral thickening. To our knowledge, there are no specific criteria for the diagnosis of haustral thickening in the literature. Therefore, we diagnosed haustral thickening if the haustra had a nodular appearance.
Whenever gastrointestinal wall thickening was identified in the cirrhosis group, its anatomic distribution, symmetry, and enhancement pattern were assessed. The anatomic distribution was read as single segmental if the thickening was confined to only one anatomic section and as multisegmental if it involved more than one anatomic section of the gastrointestinal tract.
The symmetry of wall thickening was described as concentric if the whole circumference of the gastrointestinal wall was uniformly thickened and as eccentric if the wall of the involved gastrointestinal segment was asymmetrically thickened.
The enhancement pattern of wall thickening was described as ringlike if it showed alternating rings of high and low attenuation within the thickened wall and as homogeneous if it showed homogeneous attenuation within the thickened wall, as described by Balthazar (5).
Data Analysis
The frequency of gastrointestinal wall thickening was calculated in the cirrhosis group and in the control group, and the Fisher exact test was applied to determine the significance of their difference. The numbers and percentages of patients with cirrhosis and with single-segmental or with multisegmental distribution were determined. In patients with single-segmental wall thickening, the number and percentages of involvement of different sections of the gastrointestinal tract were determined. In patients with multisegmental wall thickening, the number of sections involved and their combinations were determined. Ninety-five-percent CIs were calculated for all percentages. Statistical analysis was performed by a qualified statistician with a commercially available statistical software package (SAS SOFTWARE; SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In the various sections of the small intestine, the jejunum showed wall thickening in 32 (42%; 95% CI: 30%, 53%) patients; the duodenum, in 15 (19%; 95% CI: 12%, 28%) patients; and the ileum, in 10 (13%; 95% CI: 6%, 19%) patients with cirrhosis.
In the various sections of the colon, the ascending colon showed wall thickening in 32 (42%; 95% CI: 30%, 53%) patients; the transverse colon, in 11 (14%; 95% CI: 5%, 18%) patients; and the descending colon, in nine (12%; 95% CI: 5%, 18%) patients with cirrhosis.
As shown in the Table, in all segments of the colon, haustral thickening was seen more commonly than colonic wall thickening. The lowest frequency of gastrointestinal wall thickening was seen in the stomach, where it occurred in five (6%; 95% CI: 1%, 12%) patients. Examples of normal and thickened gastrointestinal segments are shown in Figures 1 and 2–5, respectively.
|
|
|
|
|
|
|
Among the patients who showed a multisegmental anatomic distribution of wall thickening, both the small intestine and the colon were involved in 14 (47%; 95% CI: 28%, 65%) patients; multiple sections of only the small intestine, in nine (30%; 95% CI: 13%, 47%); and multiple sections of only the colon, in seven (23%; 95% CI: 8%, 34%).
Single-segmental distribution was seen in 19 (39%; 95% CI: 25%, 53%) of the 49 patients with wall thickening. Ten (53%; 95% CI: 29%, 77%) of these 19 patients had involvement limited to the ascending colon; eight (42%; 95% CI: 18%, 66%), to the jejunum; and one (5%; 95% CI: 0%, 16%), to the stomach.
Duodenal and/or ileal wall thickening was seen only in the presence of jejunal wall thickening, and transverse and/or descending colonic wall thickening was seen only in the presence of ascending colonic wall thickening.
Other Characteristics of Gastrointestinal Wall Thickening in the Cirrhosis Group
In the 49 patients with cirrhosis and gastrointestinal wall thickening, all thick-walled gastrointestinal segments showed concentric wall thickening and homogeneous enhancement. No patient had eccentric wall thickening or ringlike enhancement.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
For correct patient treatment, it is important to recognize conditions other than cirrhosis when interpreting abdominal CT scans that show gastrointestinal wall thickening in patients with cirrhosis. This is, however, difficult without the knowledge of the patterns of gastrointestinal wall thickening in patients with cirrhosis.
Although the appearance of gastrointestinal edema on barium studies in patients with cirrhosis has been described (1,2), to our knowledge, there are no details regarding the frequency and patterns of CT findings of gastrointestinal wall thickening in patients with cirrhosis in the literature.
Different criteria have been used by different authors for the diagnosis of bowel wall thickening. Three millimeters has been considered as the upper limit of normal for small-intestinal or colonic wall thickness (6,7). However, in our opinion, it is not possible to make such small measurements accurately on the CT images printed on film. Also, in practice, most radiologists depend on their visual impression of the gastrointestinal wall thickness to determine the presence of gastrointestinal wall thickening on CT images.
Therefore, for our study, we used a visual criterion modified from the work of Scanlon et al (4), who called the wall thickness abnormally increased if it was measurable. We considered the small intestinal and colonic walls to be thickened if their thickness was more than just perceptible. By using this criterion, gastrointestinal wall thickening was seen to occur in 64% of patients with cirrhosis. Multisegmental distribution (in 61% of patients) was more common than single-segmental distribution (in 39% of patients). The jejunum and the ascending colon were the most common sections of the gastrointestinal tract to show wall thickening. This is in agreement with the earlier literature (1), which described more common involvement of the jejunum and the ascending colon in edema due to cirrhosis.
We believe that our study has some findings of clinical importance. First, if the jejunum was normal, no wall thickening was seen in the duodenum or in the ileum; if the ascending colon was normal, no wall thickening was seen in the transverse or descending colon. To our knowledge, ours is the first study in which this pattern of gastrointestinal wall thickening in patients with cirrhosis is reported. The clinical implication of this finding is that gastrointestinal wall thickening should not be ascribed to cirrhosis alone if it involves the duodenum or the ileum without involvement of the jejunum or involves the transverse or descending colon without involvement of the ascending colon. Such findings on the CT scan of a patient with cirrhosis should prompt a search for coexistent, noncirrhotic potential causes of gastrointestinal wall thickening. These would include inflammatory, ischemic, and neoplastic diseases of the small or large bowel.
Second, in our study, haustral thickening of the colon was far more common than wall thickening of the colon. Haustral thickening in patients with cirrhosis has been depicted previously on barium studies (1), but to our knowledge, haustral thickening without colonic wall thickening on CT scans in patients with cirrhosis has not been described before. Nodular haustral thickening has been described as a specific feature of pseudomembranous colitis (9). Although we do not have an explanation for the high frequency of haustral thickening in patients with cirrhosis, in our opinion the absence of colonic wall thickening and the predominant distribution in the ascending colon helps to differentiate the CT finding of haustral thickening due to cirrhosis from that due to colitis.
Third, eccentric gastrointestinal wall thickening and ringlike enhancement of the gastrointestinal wall did not occur in our study patients. These findings on the CT scans of a patient with cirrhosis should raise the suspicion of other diseases—for example, Crohn disease and neoplasm with eccentric wall thickening or Crohn disease and ischemia in the presence of ringlike enhancement, as described in the literature (5,6).
There are potential limitations of this study. First, because we used a visual criterion rather than a measured criterion, there was a potential for subjective bias. We tried to minimize this by using three readers and by accepting the majority opinion. To further ascertain the validity of the visual threshold of the readers for diagnosing gastrointestinal wall thickening, scans from a control group of patients without cirrhosis were included in the assessment. All scans were read without prior knowledge of whether the patient belonged to the cirrhosis group or to the control group. The frequency of gastrointestinal wall thickening diagnosed in this way was significantly higher in the cirrhosis group than in the control group.
Second, this was a retrospective study and there was no histopathologic proof that the gastrointestinal wall thickening seen on CT scans in our patients was due to edema. Any inflammation, infiltration, or neoplasm of the gastrointestinal tract can manifest as gastrointestinal wall thickening on CT scans. To minimize this limitation, we designed our inclusion and exclusion criteria to include in our study only those patients without clinical evidence of other gastrointestinal disease, of neoplastic disease, or of other potential causes of gastrointestinal wall thickening. Therefore, in our opinion, the gastrointestinal wall thickening that we observed in the patients in the cirrhosis group represented edema related to cirrhosis.
Third, we did not have available to us the Child-Pugh class of the patients' cirrhosis on the day of CT. Therefore, the findings could not be correlated with the severity of hepatic cirrhosis.
In conclusion, we found that gastrointestinal wall thickening is common in patients with cirrhosis. It is always concentric and homogeneous in enhancement. It may show single-segmental or multisegmental distribution. The jejunum and the ascending colon are the most commonly involved sections of the gastrointestinal tract and show wall thickening in patients with cirrhosis. The various patterns of gastrointestinal wall thickening should be kept in mind to correctly interpret CT findings in patients with cirrhosis. When gastrointestinal wall thickening patterns atypical for cirrhosis are encountered on contrast-enhanced CT scans, alternative diagnoses to cirrhosis should be considered.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
3 Department of Radiology, Hanyang University Hospital, Seoul, South Korea.
Author contributions: Guarantors of integrity of entire study, O.I.K., S.C., H.R.; study concepts, G.D.D.; study design, G.D.D., K.N.C.; definition of intellectual content, G.D.D.; literature research, O.I.K.; clinical studies, all authors; data acquisition, G.D.D., K.N.C., S.C., H.R.; data analysis, S.C., O.I.K.; statistical analysis, S.C., O.I.K.; manuscript preparation, S.C., K.N.C., O.I.K.; manuscript editing, S.C., O.I.K.; manuscript review, K.N.C., S.C.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. L. Ormsby, C. Duffield, F. Ostovar-Sirjani, J. P. McGahan, and C. Troppmann Colonoscopy Findings in End-Stage Liver Disease Patients with Incidental CT Colonic Wall Thickening Am. J. Roentgenol., November 1, 2007; 189(5): 1112 - 1117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. K. Chou, R. H. Wu, C.-W. Mak, and M.-P. Lin Clinical Significance of Poor CT Enhancement of the Thickened Small-Bowel Wall in Patients with Acute Abdominal Pain Am. J. Roentgenol., February 1, 2006; 186(2): 491 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Xiong, K. N. Chintapalli, G. D. Dodd III, S. Chopra, J. A. Pastrano, C. Hill, J. R. Leyendecker, R. M. Abbott, D. Grayson, and J. Feig Frequency and CT Patterns of Bowel Wall Thickening Proximal to Cancer of the Colon Am. J. Roentgenol., April 1, 2004; 182(4): 905 - 909. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Macari and E. J. Balthazar CT of Bowel Wall Thickening: Significance and Pitfalls of Interpretation Am. J. Roentgenol., May 1, 2001; 176(5): 1105 - 1116. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
