| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Emergency Radiology |
1 From the Departments of Radiology (J.W.S., J.A.S., B.C.L.), Emergency Medicine (A.U., N.K.R.), and Surgery (P.A.B., E.F.H.), Boston University Medical Center, One Boston Medical Center Pl, Boston, MA 02118. From the 2003 RSNA scientific assembly. Received December 5, 2003; revision requested February 12, 2004; revision received March 30; accepted May 17. Address correspondence to J.W.S. (e-mail: joshua.stuhlfaut@bmc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The investigational review board approved the study. Informed consent was waived. CT reports for October 2001 to September 2003 were reviewed and 1082 patients were identified who had undergone abdominopelvic CT with a multi–detector row scanner and without oral contrast material. Findings were divided into four categories: negative, solid organ injury with or without hemoperitoneum, free fluid only, and suspected bowel or mesenteric injury. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated by comparing CT findings with laparotomy reports and hospital course.
RESULTS: CT findings were no intraabdominal injury (n = 932), solid organ injury only (n = 102), free fluid only (n = 34), and suspected bowel or mesenteric injury (n = 14). CT findings in patients suspected of having bowel or mesenteric injury were pneumoperitoneum with other secondary findings (n = 4), mesenteric hematoma and bowel wall abnormality (n = 2), mesenteric hematoma only (n = 4), and bowel wall thickening only (n = 4). In 11 patients, bowel or mesenteric injury was proved surgically. Thus, the study included 1066 true-negative, nine true-positive, two false-negative, and five false-positive results. Based on these data, sensitivity was 82% (95% confidence interval [CI]: 52%, 95%), specificity was 99% (95% CI: 98%, 99%), positive predictive value was 64% (95% CI: 39%, 83%), and negative predictive value was 99% (95% CI: 98%, 99%) for depiction of bowel and mesenteric injuries.
CONCLUSION: Multi–detector row CT without oral contrast material is adequate for depiction of bowel and mesenteric injuries that require surgical repair. Results are comparable with previously reported data for single–detector row helical CT with oral contrast material.
© RSNA, 2004
Index terms: Abdomen, CT, 70.12112 • Abdomen, injuries, 70.41, 70.49 • Computed tomography (CT), contrast media • Intestines, injuries, 74.71, 75.71 • Trauma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
CT diagnosis of bowel and mesenteric injury is one of the more challenging tasks that face the radiologist interpreting trauma CT scans. In the absence of gross bowel wall disruption, one relies on other signs of possible injury. Reported CT signs of bowel and mesenteric injury include bowel wall discontinuity, pneumoperitoneum, focal bowel wall thickening, mural hematoma, pneumatosis, mesenteric hematoma, triangle-shaped fluid collections in the mesentery, abnormal mural enhancement, extravasation of contrast material from mesenteric vessels, the sentinel clot sign, and extravasation of oral contrast material (7–10). Most prior studies were based on protocols in which conventional or single–detector row helical CT technology was used with both intravenous and oralcontrast material. Of the CT findings just mentioned, the presence of extravasated oral contrast material was determined to be among the least sensitive for bowel and mesenteric injury (7,9,10). Thus, controversy exists about the use of oral contrast material, because of the poor sensitivity of this finding, the small yet definite risk of aspiration, and the potential delay in patient care. Some authors have suggested that use of oral contrast material may not be necessary at initial evaluation of patients with blunt abdominal trauma (11–13). This suggestion also is based on the results of studies in which conventional protocols and single–detector row helical technology were used.
At our institution, a level I trauma center, we currently perform all CT examinations with a multi–detector row CT scanner without oral contrast material in patients suspected of having abdominal injury due to blunt trauma. The purpose of our study, therefore, was to retrospectively evaluate multi–detector row CT without oral contrast material for depiction of bowel and mesenteric injuries that require surgical repair in patients with blunt abdominal trauma.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
During the study period, 1102 patients were evaluated with multi–detector row CT for blunt abdominal trauma. Nine patients were subsequently excluded because they had received oral contrast material prior to CT scanning. All of the patients who were excluded for this reason had received oral contrast material at an outside institution before their transfer to our medical center. None of these patients had CT findings suspicious for bowel or mesenteric injury, and none were diagnosed with bowel injury prior to discharge. In addition, 11 patients were excluded because of death prior to laparotomy. Thus, our study included 1082 patients (765 men and 317 women) with a mean age of 39.8 years (men, 39.2 years; women, 41.7 years) and a range of 18–98 years (men, 18–97 years; women, 18–98 years). Age distribution between the sexes was analyzed with the Student t test, and no significant difference in age was found between men and women (P = .08). Mechanisms of injury were collision of motor vehicles (n = 652), fall from a height (n = 179), collision of motor vehicle with pedestrian (n = 101), assault (n = 79), and others (n = 71).
CT Technique and Report Analysis
All CT scans were acquired by using a four–detector row scanner (MX8000; Philips Medical Systems, Andover, Mass). CT scanning parameters were as follows: 120–140 kVp; 200–250 mAs; pitch, 6; field of view, 240–350 mm; collimation, 2.5 mm (3.2 mm effective); and reconstruction interval, 3 mm. All patients received 100 mL of intravenous contrast material containing 320 mg of iodine per milliliter (iohexol, Optiray; Mallinckrodt Imaging, Hazelwood, Mo), administered with a power injector (Liebel-Flarsheim CT 9000 Adv; Mallinckrodt) at a rate of 2–3 mL/sec. CT of the abdomen and pelvis was performed from the superior surface of the diaphragm through the lesser trochanters and began 60 seconds after the initiation of intravenous contrast material injection. CT of the chest, abdomen, and pelvis was performed from the thoracic inlet to the lesser trochanters and began 30 seconds after intravenous contrast material injection. No oral contrast material was administered.
All CT scans were initially interpreted by a staff radiologist in the body imaging department at our institution. During the study period, reports were generated by nine radiologists with various levels of experience (2–20 years) in the interpretation of trauma CT scans. Each report was printed and reviewed by one of three investigators (J.W.S., J.A.S., B.C.L.). On the basis of the CT findings recorded in the initial report, the study population was divided into the following groups: patients with negative results (no intraabdominal injury); patients with solid organ injury with or without hemoperitoneum and without findings suspicious for bowel or mesenteric injury; patients with free fluid collection; and patients with findings suspicious for bowel or mesenteric injury requiring surgical repair and with or without associated solid organ injury. Findings recorded in the initial CT report that were considered suspicious for bowel or mesenteric injury requiring surgical repair were the presence of bowel discontinuity, pneumoperitoneum, focal bowel wall abnormality (thickening, abnormal enhancement, pneumatosis, disruption, or hematoma), active contrast material extravasation within the mesentery, sentinel clot within the mesentery, or mesenteric hematoma. If free fluid was reported at interpretation of the initial CT scan, we recorded whether measurement in Hounsfield units or any description of the attenuation of that fluid (eg, "attenuation similar to that of water" or "hyperattenuating fluid consistent with hemorrhage") was included in the report.
Standards of Reference
Hospital course, follow-up CT reports, and laparotomy reports were used as standards of reference to determine the presence or absence of bowel or mesenteric injury requiring surgical repair. Bowel or mesenteric injury was considered present in patients who required bowel repair or resection. Bowel or mesenteric injury was considered absent in patients with negative findings at laparotomy or in patients discharged without surgical intervention. Isolated serosal injuries found at laparotomy but not requiring bowel resection or repair were also considered negative findings for the purpose of this study, since the trauma surgeons at our institution generally consider such injuries incidental and not life-threatening.
Statistical Analysis
Reports of initial CT scanning were compared with the standards of reference. If there were discrepancies between the initial report and the outcome, CT scans were reviewed by one of two experienced investigators (J.A.S., B.C.L.). Sensitivity, specificity, positive predictive value, and negative predictive value (along with 95% confidence intervals [CIs]) were then calculated for depiction of bowel and mesenteric injuries.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Group 1
Among the 932 patients with negative CT scans, one patient subsequently developed rectal bleeding and required sigmoidoscopy. A rectal perforation was identified that required laparotomy and resection. Retrospective review of the initial CT scan in this patient did not show any evidence of rectal injury in a location that corresponded to the site of perforation seen at sigmoidoscopy. The CT result in this case therefore was considered false-negative.
Group 2
Of the 102 patients with solid organ injury with or without hemoperitoneum and without initial CT findings suspicious for bowel or mesenteric injury, 26 required surgical management. In 25 of these patients, no injury to the viscera or mesentery was found at laparotomy. The remaining patient had an intraperitoneal urinary bladder perforation and an extensive serosal injury in the sigmoid colon that required partial sigmoid resection. Retrospective review of the CT scan for this patient demonstrated multiple pelvic fractures, evidence of bladder rupture with intra- and extraperitoneal fluid accumulation, and diffuse enhancement and thickening of the wall of the small and large bowel, including the sigmoid colon (Fig 1). The bowel wall abnormalities were interpreted as secondary to profound hypovolemia (shock bowel). The CT result in this case also was considered false-negative.
|
|
Group 4
In 14 patients, findings at initial CT were considered suspicious for bowel or mesenteric injury. Specific findings in these 14 patients were free air with other signs of bowel or mesenteric injury (n = 4); mesenteric hematoma and focal bowel wall thickening (n = 1); mesenteric hematoma, bowel wall thickening, and mural enhancement (n = 1); mesenteric hematoma only (n = 4); and bowel wall thickening only (n = 4). At laparotomy, bowel or mesenteric injuries were confirmed in nine of the 14 patients (true-positive CT results). Five patients recovered without surgical intervention, and the CT findings in these five cases were considered false-positive. Other CT signs, such as bowel disruption, wall discontinuity, pneumatosis, mural hematoma, active contrast material extravasation within the mesentery, and the presence of sentinel clot within the mesentery, were not seen in our study population.
Bowel perforation was found in all four patients in whom initial CT scans showed free air and other signs of bowel or mesenteric injury (Fig 2). Surgical findings in this group included ileal perforation (n = 2), sigmoid colon perforation (n = 1), and rectal perforation (n = 1). All four patients were successfully treated with surgical resection. Another patient, who had initial CT findings of mesenteric hematoma and bowel wall thickening in the transverse colon, had a middle colic artery laceration with associated colonic ischemia that required subtotal colectomy. No contrast material extravasation from the arterial injury was seen at retrospective review of this patient’s CT scan. A sixth patient had findings at initial CT of focal concentric wall thickening in a proximal segment of the ileum with abnormal mural enhancement and a focal hematoma in the leaflets of the adjacent mesentery. After close clinical observation of this patient, repeat CT was requested because of the patient’s increasing acidosis and worsening clinical status. Repeat CT was performed 12 hours later with oral contrast material and showed new free intraperitoneal air with persistent bowel wall thickening and mesenteric hematoma. No extraluminal oral contrast material was noted on the repeat CT scan (Fig 3). A 1-cm-long perforation of the ileum was identified at surgery.
|
|
|
|
|
Sensitivity, Specificity, and Predictive Values
In summary, the results of our study include 1066 true-negative, nine true-positive, two false-negative, and five false-positive CT interpretations. The sensitivity of multi–detector row CT for identification of bowel or mesenteric injury requiring surgical repair was 82% (nine of 11 patients; 95% CI: 52%, 95%), specificity was 99% (1066 of 1071; 95% CI: 98%, 99%), positive predictive value was 64% (nine of 14; 95% CI: 39%, 83%), and negative predictive value was 99% (1066 of 1068; 95% CI: 98%, 99%).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
At our institution, all CT examinations for trauma are performed with a multi–detector row scanner and without oral contrast material, at the request of the trauma service, because of a death that occurred within 24 hours after aspiration during oral administration of contrast material. We reserve the use of oral contrast material for follow-up CT in patients in whom bowel injury is indicated by the results of clinical examination or prior CT findings. It is in this setting that we report our experience in detecting bowel or mesenteric injury without the administration of oral contrast material.
The results of our study suggest that multi–detector row CT performed without oral contrast material is adequate for evaluation of patients with blunt abdominal trauma. Our results accord with data previously reported in the literature for single–detector row helical CT scanners (10,16,20). This similarity is likely related to the increased awareness of signs of bowel or mesenteric injury that were previously described in the literature (7–10). In addition, the routine use of multi–detector row CT, with its ability to acquire image data with thinner collimation and higher spatial resolution, allows improved detection of small mesenteric hematomas and small bubbles of free air. This improvement in technology allows adequate depiction of subtle bowel and mesenteric injuries without the use of oral contrast material. The performance values measured in our study are not as high as those found in a retrospective study by Killeen et al (20). In their study population, the positive predictive value of CT for detection of bowel and mesenteric injuries was 90% and 97%, respectively. The two studies should not be compared directly, however, because Killeen et al considered CT scans true-positive regardless of whether abnormal findings were confirmed at surgery (44% of their patients did not undergo laparotomy). In our study, we considered unproved findings false-positive because they could lead to nontherapeutic laparotomies. Our performance results were negatively affected by the use of these strict criteria.
Over the course of our study period, more than 95% of CT scans were reported negative or positive only for solid organ injury. In these two groups, there were two false-negative studies. As previously mentioned, both cases involved injury to the distal colon. It is unlikely that oral contrast material would have been helpful for identifying these injuries, given their distal location. Would oral contrast material have provided added benefit in the patients whose initial CT scans showed free fluid accumulations? In our study, none of the patients in this group were subsequently found to have bowel or mesenteric injuries. However, seven patients did require surgical exploration. In each case, the decision for surgical management was based on clinical parameters and lack of other significant CT findings rather than on a specific CT finding that suggested bowel injury. Thus, the addition of oral contrast material at initial CT would not likely have altered the clinical course. In such cases, we currently recommend close clinical observation and short-term (6–12-hour) follow-up CT with oral contrast material. Change (or lack of change) in CT findings or clinical parameters helps to determine subsequent management. In one patient, 12-hour follow-up CT with oral contrast material showed persistence of apparent new duodenal wall thickening. While follow-up CT findings certainly influenced the decision to operate, the negative results at the time of surgery suggest that the results of initial CT would not have been substantially improved by the use of oral contrast material.
In the group of patients with CT findings suspicious for bowel injury, nine patients had confirmed bowel or mesenteric injury. Repeat CT was required in one patient with proved perforation of the ileum. In this patient, repeat CT with oral contrast material helped to confirm the initial CT findings and demonstrated new pneumoperitoneum but did not show extravasation of enteric contrast material. Could bowel distention produced by the use of oral contrast material at initial CT have led to leakage of gas into the peritoneum and thus to earlier surgical repair? This is possible but unlikely, given the distal location of bowel perforation. The value of repeat CT was in the confirmation of both clinical and initial CT findings and not necessarily in the addition of oral contrast material. Finally, five patients with isolated findings of bowel wall thickening or mesenteric hematoma did not require surgical exploration. These false-positive results underscore the importance of clinical assessment and integration of clinical findings with positive CT results in the evaluation of patients for bowel and mesenteric injury. In addition, these results agree with prior data that suggest that stable patients with a single positive finding at CT may warrant observation and not surgical intervention (21).
Limitations of this study include its retrospective nature and the small number of bowel injuries encountered. Prospective studies and/or further retrospective data over a longer period of time or from multiple trauma centers may help to further address this issue. The relatively low frequency of bowel or mesenteric injury in our study (approximately 1% of patients) is similar to that reported in the literature (22). In addition, all of the bowel perforations identified in our study involved distal segments of the small bowel or colon. As previously noted, the benefit from use of oral contrast material may be found in the improved visualization and detection of abnormalities in the duodenum and proximal jejunum. Oral contrast material is not likely to reach the ileum in time for initial CT and therefore is less likely to benefit evaluation of the distal small bowel and colon. The frequency of an isolated finding of extraluminal oral contrast material at multi–detector row CT in patients with bowel injury remains to be determined, but it is unlikely to be high. Future studies might also address this issue, as well as the importance of isolated findings of free fluid and bowel wall thickening and the utility of short-term follow-up CT in such patients.
In summary, we believe that the administration of oral contrast material is not routinely necessary for detecting bowel or mesenteric injury in patients with blunt abdominal trauma. Instead, the improved image quality provided by multi–detector row CT and the increased awareness of other signs of visceral and mesenteric injuries has greatly improved our ability to diagnose these injuries. Given the ever increasing number of emergency CT examinations and the increasing dependence on emergency radiology departments to perform CT safely and efficiently, we believe that the acquisition of initial CT scans without oral contrast material helps us to meet both criteria (safety and efficiency) without sacrificing diagnostic accuracy. In the small group of patients with questionable CT findings, short-term follow-up CT with oral contrast material may be warranted. This is a topic worthy of further investigation.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, J.W.S.; study concepts, all authors; study design, J.W.S., J.A.S.; literature research, J.W.S., J.A.S., N.K.R.; clinical studies, J.W.S., J.A.S., B.C.L.; data acquisition, all authors; data analysis/interpretation, J.A.S., J.W.S., B.C.L.; statistical analysis, J.W.S., J.A.S.; manuscript preparation, J.W.S., J.A.S.; manuscript definition of intellectual content and final version approval, all authors; manuscript editing, B.C.L., N.K.R., A.U., P.A.B., E.F.H.; manuscript revision/review, J.W.S., J.A.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. C. Lucey, J. W. Stuhlfaut, A. R. Hochberg, J. C. Varghese, and J. A. Soto Evaluation of Blunt Abdominal Trauma Using PACS-Based 2D and 3D MDCT Reformations of the Lumbar Spine and Pelvis Am. J. Roentgenol., December 1, 2005; 185(6): 1435 - 1440. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
