| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, University of Manitoba, Health Sciences Centre, 820 Sherbrook St, Winnipeg, Manitoba, Canada R3A 1R9. From the 2001 RSNA scientific assembly. Received November 26, 2001; revision requested February 11, 2002; final revision received June 17; accepted June 27. Address correspondence to I.D.C.K. (e-mail: umkirkp4@cc.umanitoba.ca).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Abdominal CT scans obtained during a 6-year period were reviewed retrospectively to identify 76 neutropenic patients with radiologic bowel abnormalities. Scans were analyzed for wall thickening, pneumatosis, wall nodularity, mucosal enhancement, bowel dilatation, ascites, and mesenteric stranding. The location and extent of abnormalities were noted. Independent chart and pathology report reviews were used to determine the patients’ final diagnoses: neutropenic enterocolitis (n = 53), Clostridium difficile colitis (n = 14), graft-versus-host disease (n = 7), cytomegaloviral colitis (n = 1), and ischemic bowel (n = 1). Results were assessed with the Student t test for quantitative wall thickness and the 
2 test for the number of patients with each diagnosis who demonstrated each CT finding.
RESULTS: Mean bowel wall thickening was greatest in C difficile colitis (12 mm) and least in graft-versus-host disease (5 mm). Pneumatosis was limited to neutropenic enterocolitis (21% [11 of 53 patients]) and bowel ischemia. Wall nodularity was significantly more common (P < .01) in C difficile colitis (36% [five of 14 patients]). In graft-versus-host disease, the rates of mucosal enhancement and bowel dilatation were highest (P < .05) (71% [five of seven patients] and 86% [six of seven patients], respectively). In C difficile colitis, the rates of ascites and mesenteric stranding were highest (57% [eight of 14 patients] and 71% [10 of 14 patients], respectively). Although findings in neutropenic enterocolitis and graft-versus-host disease could involve any bowel segment, C difficile colitis was always limited to the colon.
CONCLUSION: Several CT findings can help differentiate specific gastrointestinal complications in neutropenic patients.
© RSNA, 2003
Index terms: Colitis, 75.2043, 75.206 • Colitis, pseudomembranous, 75.263 • Enteritis, 73.69, 74.69 • Intestines, ischemia, 73.266, 74.266
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Often, the gastrointestinal complications seen in neutropenic patients are difficult to diagnose on the basis of clinical findings alone. Symptoms such as abdominal pain, fever, diarrhea, and gastrointestinal bleeding are nonspecific, and blood and stool samples take time to collect and process for laboratory assays (4–8). Endoscopy with biopsy carries an increased risk in these patients, who are often also thrombocytopenic (1–3). Thus, computed tomography (CT) is an attractive noninvasive option to use to quickly diagnose intestinal pathologic conditions in these patients.
Although some investigators have reported a few case series in the literature and have described some CT findings of commonly seen gastrointestinal diseases that occur in neutropenic patients, many of these series have small numbers of cases, and to the best of our knowledge, no comparative studies in this population exist to date (7,9–13). Thus, the purpose of this study was to characterize the CT findings of gastrointestinal complications seen in neutropenic patients and to identify CT features that can help differentiate these complications.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Of these 289 patients, 76 demonstrated imaging abnormalities related to the bowel. These 76 patients ranged from 15 to 79 years old, with two patients younger than 18 years old. The causes of neutropenia are shown in Table 1. Reasons documented in the charts for ordering the CT scans of the 76 patients included abdominal pain in 68 patients (89%), fever of an unknown source in 39 patients (51%), diarrhea in 27 patients (36%), and liver function test abnormalities in 10 patients (13%).
|
The CT scans were analyzed for the presence of bowel wall thickening (>3 mm), pneumatosis intestinalis, wall nodularity (nodular projections into the lumen that were >3 mm in width), mucosal enhancement, bowel dilatation (defined as a single region of small bowel with a diameter >2.5 cm, or large bowel >8 cm), ascites, and mesenteric stranding. Only patients who received contrast material intravenously were assessed for mucosal enhancement, and this was taken into consideration for the purposes of calculating percentages and statistics. The locations and extent of any of these abnormalities within the bowel (wall thickening, nodularity, mucosal enhancement, dilatation, pneumatosis, or adjacent mesenteric stranding) were also noted. Hard-copy scans were reviewed independently by two radiologists. Any discrepancies between interpretations were resolved with consensus review. This was required in three cases in which there was initial disagreement with regard to the extent of disease within the bowel. Quantitative wall measurements were obtained by one radiologist (I.D.C.K.) from the areas of greatest thickness by using electronic calipers for those patients imaged with the multidetector scanner and by using conventional calipers for those imaged with the helical scanner.
Chart and Pathology Report Reviews
Chart and pathology report reviews, which were conducted independently of the CT analysis, were used to determine the patients’ final diagnoses: neutropenic enterocolitis (n = 53), Clostridium difficile colitis (n = 14), gastrointestinal graft-versus-host disease (n = 7), cytomegaloviral colitis (n = 1), and ischemic bowel (n = 1). All patients with a diagnosis of gastrointestinal graft-versus-host disease had received bone marrow transplants. The single radiologist (I.D.C.K.) who performed the chart and pathology report reviews was blinded at the time of the review to the results of the CT examinations for that patient and was provided only the patient chart numbers and the dates of their corresponding CT examination or examinations. The review of CT scans by that radiologist had been performed several months previously.
Neutropenic enterocolitis was confirmed by the pathologic analysis findings in 11 patients. In the other 42 patients, neutropenic enterocolitis was diagnosed by using previously published criteria of abdominal pain, tenderness, and fever in a neutropenic patient, with resolution after administration of appropriate treatment and with restoration of neutrophil counts (6,14,15). For the current study, the diagnosis of neutropenic enterocolitis was also one of exclusion, and all patients had negative results of three serial C difficile toxin assays, stool cultures negative for other infectious enteritides, negative results of blood and stool assays for cytomegalovirus, and no evidence of any other abdominal finding that could explain the clinical symptoms and signs. All of these 42 patients had been followed up with subsequent CT imaging to confirm the resolution of abnormal gastrointestinal findings with standard treatment for neutropenic enterocolitis. All patients with a diagnosis of C difficile colitis had positive results of stool toxin assays. Patients with gastrointestinal graft-versus-host disease, cytomegaloviral colitis, and ischemic bowel all had pathologic proof of diagnosis, either at autopsy or with biopsy.
All patients in the study who survived the neutropenic episode (n = 62) had clinical follow-up, which was reviewed until approximately 1 year from the date of the scan (55 patients were followed up for at least 1 year, three patients for 3 months, and four patients for 6 weeks). Follow-up documentation was reviewed for any evidence of misdiagnosis at any repeat imaging examinations, biopsies, laboratory tests, or on the basis of ongoing symptoms and signs. At this review, no patients were found to have an initial misdiagnosis. The clinical course of each patient after scanning was also noted.
Statistical Analysis
Results were analyzed with the Student t test for comparing mean bowel wall thickness data and with the 2 test for comparing the number of patients with each diagnosis demonstrating wall thickening, pneumatosis, wall nodularity, mucosal enhancement, bowel dilatation, ascites, and mesenteric stranding. All P values less than .05 were considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Neutropenic enterocolitis is perhaps the most poorly understood gastrointestinal complication of neutropenia. Pathologically, neutropenic enterocolitis is seen as compromise of bowel wall integrity with subsequent bacterial or fungal invasion. The cause remains unclear and in fact may be multifactorial, potentially as a result of altered immune function, toxic effects of chemotherapy, bowel ischemia, thrombocytopenic intramural hemorrhage, or even infiltration of the wall with neoplastic cells (4,16,17). Although pathologic abnormalities can involve any segment of the small or large bowel, the cecum is the most common location of abnormality (16,17). Surgical intervention was favored in the past, but conservative therapy with antibiotics and gut rest is now preferred in all but the most serious cases (6,18). Some evidence indicates that patients with a clinical diagnosis of neutropenic enterocolitis who have consistent radiologic abnormalities have a more serious form of the disease, and this may prompt consideration of more aggressive therapy (19). Note that of the 11 patients with neutropenic enterocolitis and pneumatosis in the current study, seven recovered with conservative therapy alone.
Previously noted findings in neutropenic enterocolitis at CT include right-sided colon wall thickening, pericolonic stranding, ascites, and cecal pneumatosis (9–12). In many of the series of cases in the published literature on this topic, the numbers of cases are small, and they lack pathologic correlation (9,11,12). Furthermore, it is frequently unclear what criteria were used to make the diagnosis clinically and to what extent other possible causes of the CT findings were ruled out. C difficile colitis, for example, can affect the cecum and right-sided colon in isolation and cause many similar findings at CT (20,21). Although only 11 of the 53 cases in the current study had pathologic correlation, we have used the most rigid criteria of any study to date in making the diagnosis in the remaining 42 patients. Patients not only had to meet previously published clinical diagnostic criteria but also underwent extensive laboratory testing and clinical and radiologic follow-up to exclude other confounding disease processes such as C difficile colitis (6,14,15).
Bowel wall thickening was present in all 53 cases of neutropenic enterocolitis, but although the mean thickness was significantly less than that seen in C difficile colitis, the overlap of measurements limits the diagnostic value of this finding. A wall thickness of greater than 7 mm would help exclude graft-versus-host disease, however. Unfortunately, mesenteric stranding and ascites were also not significantly different in their rates of occurrence among the three most common diseases. Pneumatosis intestinalis was a much more specific finding, and the location of the abnormalities is also helpful. The combined involvement of the small and large bowel points toward a more likely diagnosis of neutropenic enterocolitis than C difficile colitis, and graft-versus-host disease was never limited to the right-sided colon in its involvement. Cecal involvement alone was also limited to neutropenic enterocolitis in the current study, although the results of a previous study have shown that C difficile colitis can occur in this focal fashion (20).
Neutropenic enterocolitis has often been cited in the CT literature as a disease of the right-sided colon and is sometimes referred to simply as neutropenic colitis (9–12,22,23). The findings of this study, however, demonstrated numerous cases in which other portions of both small and large bowel were involved. The results in the original article describing this entity, which was published in the radiology literature, documented that changes can occur pathologically in both the small and large bowel (16), and this observation was further confirmed by the findings of the most extensive postmortem series to date (17).
The results of the current study support this finding, and the diagnosis of neutropenic enterocolitis should thus be considered even in the absence of isolated right-sided colon involvement or in the presence of isolated small-bowel involvement. To our knowledge, this is the first study of this entity performed since the advent of helical scanning, and subsequent improved depiction of the small bowel may in part explain these findings. We did still find that the right-sided colon was the most common site of CT abnormality, however, and that even in patients who showed relatively normal cecal walls at CT, pathologic changes were seen with microscopy in two cases.
C difficile colitis is another common complication in neutropenic patients, sometimes occurring after chemotherapy or antibiotic treatments, both of which are nearly ubiquitous in this population. Alteration of normal gut flora by the medications allows colonization by C difficile and production of both a cytotoxin and enterotoxin, resulting in colonic inflammation, diarrhea, and characteristic pseudomembranous exudates (5,8).
The CT findings of C difficile colitis were characterized by the greatest degree of wall thickening, the most common occurrence of a pancolitis, and significantly more common wall nodularity. The rates of mesenteric stranding and ascites, while high in this disease, were not significantly higher than those in neutropenic enterocolitis or graft-versus-host disease. No cases of small-bowel involvement were identified. These findings are in keeping with those of previously published studies of C difficile colitis in the general population (20,21). In a previous study, we proposed that patients in whom the findings at CT are suggestive of this diagnosis should begin treatment while confirmatory stool test results are pending (20).
The acute form of gastrointestinal graft-versus-host disease was the most distinctive of the three most common complications. It can occur after bone marrow transplantation while patients are still neutropenic, caused by early graft reaction against the immunosuppressed host recipient (4,5). It is important to differentiate this complication from other gastrointestinal diseases because treatment involves further immunosuppression, and a CT scan suggestive of gastrointestinal graft-versus-host disease should prompt attempts at tissue confirmation (3). Wall thickening was the least pronounced in this disease, and the characteristic findings were extensive areas of bowel dilatation and mucosal enhancement. Mesenteric stranding and ascites occurred somewhat less frequently but again were not significantly different in their rates of occurrence from those in neutropenic enterocolitis and C difficile colitis. No characteristic location of involvement was identified.
Of course, only patients who have received bone marrow transplants are at risk of developing gastrointestinal graft-versus-host disease. In a previous study, investigators described bizarre patterns of bowel coating with oral contrast material in these patients (13). We saw no examples of this in the seven patients in the current study, but frequently the oral contrast material did not pass far into the small bowel before the time of scanning in these patients. The findings of the current study are otherwise in concordance with those of other studies (7,13).
Because there was only one case each of cytomegaloviral colitis and ischemic bowel, it is difficult to draw any conclusions about them other than that they were much less common causes of gastrointestinal abnormalities at CT in this study population. Note, however, that the current case of cytomegaloviral colitis shared many similar features with C difficile colitis, including marked wall thickening and nodularity in a pancolitis distribution. Fortunately, these two entities can be differentiated with noninvasive stool testing. Furthermore, the pneumatosis seen in ischemic bowel could conceivably mimic neutropenic enterocolitis, and an index of suspicion for ischemia must be maintained in high-risk patients, lest they be treated conservatively.
Although we encountered no other underlying causes for gastrointestinal abnormalities seen at CT in these patients, it is important to remember that other bowel pathologic conditions seen in the general population could still occur in these patients (eg, bacterial enteritis). No patients with cytomegaloviral infection of the small bowel were seen in this patient population, although this condition could also occur in pancytopenic patients.
The current study is unfortunately limited by its retrospective nature. It is difficult to assess at what point in the disease process these patients were when they were scanned, and this may influence the nature and severity of the CT findings. Generally, physicians at our institution image neutropenic patients early in the occurrence of their symptoms, given the potential severity of any complications. This may in fact have excluded some patients who had gastrointestinal complications but were imaged prior to the development of any radiologic abnormalities appreciated by the reviewing radiologists. However, for some conditions, such as C difficile colitis, there appears to be little association between the clinical severity of disease and the imaging findings (24).
In conclusion, to the best of our knowledge, this is the first comparative CT study to date in this population and is the largest published series of cases of neutropenic enterocolitis in the radiology literature. Neutropenic enterocolitis commonly involved the right-sided colon (with isolated cecal involvement being fairly specific but infrequent) but could occur anywhere in the small or large bowel. Neutropenic enterocolitis was characterized by pneumatosis and more frequent mesenteric stranding and ascites than graft-versus-host disease, which could also involve both the small and large bowel. C difficile colitis demonstrated the most prominent wall thickening on average and the greatest frequency of wall nodularity, was always limited to the colon in this series of cases, and was the most likely to cause a pancolitis. The acute form of gastrointestinal graft-versus-host disease was characterized by the mildest amount of wall thickening, by bowel dilatation, and by mucosal enhancement. Not only were many CT findings helpful in differentiating among specific gastrointestinal complications seen in neutropenic patients, but the results of this study also demonstrated that bowel involvement in neutropenic enterocolitis can be much more extensive than previously reported.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. d'Almeida, J. Jose, J. Oneto, and R. Restrepo Bowel Wall Thickening in Children: CT Findings RadioGraphics, May 1, 2008; 28(3): 727 - 746. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Schmit, W. Bethge, R. Beck, C. Faul, C. D. Claussen, and M. Horger CT of Gastrointestinal Complications Associated with Hematopoietic Stem Cell Transplantation Am. J. Roentgenol., March 1, 2008; 190(3): 712 - 719. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. C. Silva, S. D. Beaty, A. K. Hara, J. G. Fletcher, J. L. Fidler, C. O. Menias, and C. D. Johnson Spectrum of Normal and Abnormal CT Appearances of the Ileocecal Valve and Cecum with Endoscopic and Surgical Correlation RadioGraphics, July 1, 2007; 27(4): 1039 - 1054. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Aksoy, M. Tanriover, O Uzun, P Zarakolu, S Ercis, S Erguven, A Oto, U Kerimoglu, M Hayran, and O Abbasoglu Diarrhea in neutropenic patients: a prospective cohort study with emphasis on neutropenic enterocolitis Ann. Onc., January 1, 2007; 18(1): 183 - 189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Ash, M. E. Baker, C. M. O'Malley Jr., S. M. Gordon, C. P. Delaney, and N. A. Obuchowski Colonic Abnormalities on CT in Adult Hospitalized Patients with Clostridium difficile Colitis: Prevalence and Significance of Findings. Am. J. Roentgenol., May 1, 2006; 186(5): 1393 - 1400. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
