| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Hartford Hospital, Hartford, CT.
2 Department of Radiology, Duke University Medical Center, Durham, NC.
3 Present address: Department of Radiology, University of Connecticut–St.
Francis Hospital and Medical Center, 1000 Asylum Ave., 3201E, Hartford, CT
06105.
4 Department of Surgery, University of Connecticut School of Medicine,
Farmington, CT.
Received January 8, 2008;
accepted after revision April 22, 2008.
Address correspondence to S. I. Zink
(choizink{at}cox.net).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Over 17 months, 55 patients (32 men, 23 women; age range, 21–92 years) were evaluated prospectively with contrast-enhanced MDCT using 100 mL of iopromide 300 mg I/mL. Technetium-99m-labeled RBC scans were obtained on 41 of 55 patients and select patients underwent angiography for attempted embolization. Each imaging technique was reviewed in a blinded fashion for sensitivity for detection of active bleeding as well as the active lower gastrointestinal bleeding location.
RESULTS. Findings were positive on both examinations in eight
patients and negative on both examinations in 20 patients. Findings were
positive on contrast-enhanced MDCT and negative on 99mTc-labeled
RBC in two patients; findings were negative on contrast-enhanced MDCT and
positive on 99mTc-labeled RBC in 11 patients. Statistics showed
significant disagreement, with simple agreement = 68.3%, 
= 0.341, and
p = 0.014. Sixteen of 60 (26.7%) contrast-enhanced MDCT scans were
positive prospectively, with all accurately localizing the site of bleeding
and identification of the underlying lesion in eight of 16 (50%). Nineteen of
41 (46.3%) 99mTc-labeled RBC scans were positive. Eighteen of 41
matched patients went on to angiography. In four of these 18 (22.2%) patients,
the site of bleeding was confirmed by angiography, but in 14 of 18 (77.8%),
the findings were negative.
CONCLUSION. Contrast-enhanced MDCT and 99mTc-labeled RBC scanning show significant disagreement for evaluation of active lower gastrointestinal bleeding. Contrast-enhanced MDCT appears effective for detection and localization in cases of active lower gastrointestinal bleeding in which hemorrhage is active at the time of CT.
Keywords: colon • gastrointestinal • hemorrhage • MDCT • nuclear medicine
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Often, gastrointestinal bleeding will stop spontaneously, but in approximately 25% of patients, bleeding is massive or recurrent, requiring imaging localization and directed therapy [3]. Active lower gastrointestinal bleeding may be diagnosed and localized by endoscopy; angiography; 99mTc-labeled RBC scanning; and, in rare instances, sulfur colloid scanning. Each of these techniques has limitations.
Technetium-99m-labeled RBC scanning is considered by many to be the reference standard for detection of active lower gastrointestinal bleeding. Strengths of this technique include noninvasiveness, sensitivity for low rates of bleeding (0.1 mL/min) [4], and potential for a prolonged period of observation enabling detection of intermittent bleeding. However, there are several drawbacks. Even in settings in which there is an on-call nuclear medicine technologist, 99mTc-labeled RBC remains a difficult examination to perform in a timely manner at off-hours. In addition, there is wide variability in the sensitivity of the procedure, ranging from 20% to greater than 90% [3]. Furthermore, 99mTc-labeled RBC scanning has shown inaccuracies in determining bleeding location [3].
Contrast-enhanced MDCT has recently been described as feasible, safe, and useful for investigation of active lower gastrointestinal bleeding. In a swine model, colonic bleeding rates below 0.4 mL/min are detectable [5], which is comparable to the quoted detectable bleeding rates for angiography and 99mTc-labeled RBC scanning (0.5 mL/min and 0.1 mL/min, respectively) [4]. A recent in vitro study found the threshold for detecting bleeding with contrast-enhanced MDCT to be 0.35 mL/min [6]. To date, there have been few prospective studies of contrast-enhanced MDCT used for detection of active lower gastrointestinal bleeding and no study has attempted to prospectively compare contrast-enhanced MDCT to 99mTc-labeled RBC scanning or angiography with the intent of measuring agreement between these techniques. The purpose of this study was to compare contrast-enhanced MDCT and 99mTc-labeled RBC scanning for the evaluation of active lower gastrointestinal bleeding.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patients meeting any of the following criteria were excluded from the study: serum creatinine > 1.5 mg/dL, history of IV contrast allergy, less than 18 years old, pregnant or breastfeeding, or having received an upper gastrointestinal series or CT with oral contrast within the past week.
Study Protocol
The study protocol is outlined in Figure
1. All CT examinations were performed on a 64-MDCT scanner
(LightSpeed VCT, GE Healthcare) or an 8-MDCT scanner (LightSpeed QX/i, GE
Healthcare). Contrast material, 100 mL total of iopromide (Ultravist 300,
Bayer HealthCare Pharmaceuticals) 300 mg I/mL, was injected at 4 mL/s with out
administration of oral contrast. Unenhanced and contrast-enhanced CT was
performed through the abdomen and pelvis with 5-mm-thick images reconstructed
to 2.5-mm thickness; pitch, 1.35:1; scan field of view, large; displayed field
of view, 36 (patient dependent); kVp, 120; and mA, 440. A timing run was then
performed with 15 mL of contrast material at 4 mL/s at the celiac axis with a
maximum-intensity region of interest (ROI) on the abdominal aorta to determine
the time of peak arterial enhancement. Calculation of delay involv ed
multiplying the time of peak aortic enhancement by two and adding a delay of
35 seconds for the 8-MDCT and a delay of 40 seconds for the 64-MDCT.
Unenhanced and contrast-enhanced 2.5-mm- and 5-mm-thick images as well as un
enhanced and contrast-enhanced 2.5-mm-thick coronal and sagittal reformat
images were archived for interpretation.
|
Contrast-Enhanced MDCT
Contrast-enhanced MDCT cases received a prospective interpretation by one
of eight experienced interventional radiologists holding a certificate of
additional qualification as well as more than 10 years of experience in both
angiography and CT. All CT readers were blinded to the nuclear medicine
results. A second reader examined equivocal findings. For all cases,
side-by-side comparison of unenhanced and contrast-enhanced images was
performed. Any new visible focal density in the bowel lumen seen on
contrast-enhanced images was interpreted as a positive bleed. ROI
maximum-attenuation differences between unenhanced and contrast-enhanced
images were measured in Hounsfield units
[7] at the site of the
suspected bleed. Moreover, all scans were reviewed in retrospect by the
principal investigator after completion of the clinical algorithm. In
retrospect, six of 60 (10%) scans were found to have limited visualization of
the bowel, whereas 54 of 60 (90%) of the scans were technically adequate. The
six subjective limitations were dilution of contrast pooling at a site of
bleeding, respiratory motion artifact, high-density material in the bowel
(n = 2), suboptimal bolus timing, and suboptimal bolus timing with
high-density bowel material. Three of these six limited-visual ization cases
were confirmed false-negatives because sites of active bleeding were seen on
CT only in retrospect.
Attenuation of maximum aortic enhancement, maximal superior mesenteric vein enhancement, and maximal mucosal enhancement of bowel (small bowel, hepatic flexure, splenic flexure, and sigmoid) were determined by placing an ROI on the structure in question. There were 16 CT scans with detected active hemorrhage, for which maximal density on unenhanced as well as contrast-enhanced CT was measured at the site of the bleed (Table 1). Moreover, all study patients were monitored through the hospital's clinical database for out comes, readmissions, creatinine value changes, and results of any endoscopy examinations or surgeries through June 2007.
|
Technetium-99m-Labeled RBC Scanning
After contrast-enhanced MDCT, patients underwent 99mTc-labeled
RBC scanning. We used a standard nuclear medicine in vitro UltraTag (Covidien)
technique. Examinations used standard 25-mCi (925-MBq)
99mTc-labeled RBC scanning over a total observation time of 120
minutes, with inter pretation performed by one of four physicians with
additional certification in nuclear medicine. All readers were blinded to the
contrast-enhanced MDCT results. Seventeen of 41 (41.4%) patients underwent
abbreviated scanning times with 13 of 17 scans discontinued because bleeding
was seen early. Two unstable patients were taken early to angiography and were
imaged for only 31 and 30 minutes, respectively. Two studies were terminated
early, at 100 minutes, without explanation. Because of an industrial supply
shortage of technetium, the final six nuclear medicine scans included in our
study were performed using an in vivo labeling technique. This inconsistency
did not appear to introduce obvious bias because three of six (50%) of these
99mTc-labeled RBC scans were positive. None of the negative cases
was positive on contrast-enhanced MDCT.
Blood samples for 99mTc-labeled RBC scans were obtained on patient arrival in CT. Blood tagging preparation requires 20–30 minutes and began during CT. Cases that showed bleeding on contrast-enhanced MDCT or 99mTc-labeled RBC scanning then underwent angiography for attempted embolization therapy. In cases in which the contrast-enhanced MDCT and the 99mTc-labeled RBC scan results disagreed, patients still underwent angiography.
For unstable patients, we reserved the option of performing a contrast-enhanced MDCT before going directly to angiography. Hemodynamic instability was defined as orthostatic hypotension, systolic blood pressure less than 100 mm Hg, or patients determined to be unstable after direct discussion with the interventional radiologist. Because contrast-enhanced MDCT is a rapid test, patients could feasibly undergo contrast-enhanced MDCT before the time the angiography suite would be ready to accept them. Preangiography contrast-enhanced MDCT offered valuable road-mapping and anatomic description that could potentially shorten and further focus the time spent in angiography. These cases were not included in the statistical analysis of agreement between contrast-enhanced MDCT and 99mTc-labeled RBC scanning, but they provide meaningful additional cases to help determine the clinical value of contrast-enhanced MDCT in cases of active hemorrhage. Ultravist 300 mg I/mL was used as the institutional angiographic standard and the angiographic approach was according to the angiographer's discretion. Moreover, mesenteric angiography was not performed on patients without evidence of bleeding on contrast-enhanced MDCT or 99mTc-labeled RBC scanning. Exceptions were allowed at the discretion of the interventional radiologist.
According to the results from Pennoyer et al. [8], the diagnostic accuracy of 99mTc-labeled RBC scanning has an estimated sensitivity of 84% and a specificity of 80%. In this study, we assumed a similar ability to detect bleeding and a prevalence of 50% for those who are targeted for the test and whose bleed is identified by the test. However, in accordance with ethical considerations, a small portion of patients did not receive invasive angiography and another portion of hemodynamically unstable individuals did not undergo 99mTc-labeled RBC scanning, thus eliminating the ability to calculate sensitivity and specificity. Rather, for this study, we directly compared the performance of the tagged RBC test with MDCT through a variant of the kappa statistic to evaluate the similarity in results of the two tests. The demographics of the group were reported using means, SD, and 95% CI. Cohen's kappa was used for analysis. All statistical analyses were performed with SPSS version 14.0 (SPSS) software.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Of the 55 of 94 (58.5%) patients with active lower gastrointestinal bleeding who were included, 32 (58.2%) were men and 23 (41.8%) were women with an age range of 20–91 years (mean age [± SD], 73.8 ± 13.9 years) (Fig. 1). Sixty contrast-enhanced MDCT examinations were performed on these 55 patients with 31 of 60 performed on an 8-MDCT scanner and 29 of 60 performed on a 64-MDCT scanner. Twenty-three of the 55 patients received both contrast-enhanced MDCT and angiography contrast loads with no subsequent elevation in creatinine.
Of the 60 CT examinations performed, 41 patients also underwent contemporaneous 99mTc-labeled RBC studies. Of the other 19, five patients went directly to angiography from CT because of hemodynamic instability. Moreover, 14 nonmatched examinations were not successfully coordinated, primarily because they occurred after hours.
In the 41 patients for whom both contrast-enhanced MDCT and 99mTc-labeled RBC scanning were performed, results included 20 who were negative on both contrast-enhanced MDCT and 99mTc-labeled RBC scanning, 11 who were negative on contrast-enhanced MDCT and positive on 99mTc-labeled RBC scanning, two who were positive on contrast-enhanced MDCT and negative on 99mTc-labeled RBC scanning, and eight who were positive on both contrast-enhanced MDCT and 99mTc-labeled RBC scanning (Table 2). These results showed a simple agreement of 68.3%, yet the resulting kappa value of 0.341 yielded a statistically significant disagreement (p = 0.014).
|
In the 11 patients with hemorrhage not seen on contrast-enhanced MDCT and only visualized on 99mTc-labeled RBC scanning, only two of these bleeds could be detected angiographically (Table 3). One bleed was intermittent during angiography and one was a nonbleeding focus of angiodysplasia. In all, three of 11 of these cases were treated with definitive in vasive therapy (surgery, angioembolization, or endoscopic electrocautery). The third case was angiographically negative but showed a tiny focus of radiotracer activity in the left upper quadrant. Retrospective comparison with CT images localized the focus to a site of prior surgical small-bowel anastomosis. Ulceration was detected at this site on push enteroscopy, and surgical revision was performed. Notably, this site was not actively bleeding at the time of enteroscopy or surgery.
|
Sixteen of 60 (26.7%) contrast-enhanced MDCT scans were positive prospectively (Fig. 2A, 2B, 2C). Nineteen of 60 (31.7%) were positive in total, including three false-negative cases (Figs. 3A, 3B, 3C and 4A, 4B). Nine of 16 (56.3%) and 10 of 19 (52.6%) cases required definitive invasive therapy. All positive contrast-enhanced MDCT scans accurately localized the site of bleeding. We identified no false-positive cases on contrast-enhanced MDCT. It should be noted that high-attenuation stool was not a confounder in our study because both unenhanced and contrast-enhanced images were reviewed.
|
|
|
|
|
|
|
|
In total, after noninvasive evaluation, 23 patients underwent angiography for attempted embolization. Eighteen of 41 matched patients went on to angiography. Seven of 18 contrast-enhanced MDCT and 14 of 18 99mTc-labeled RBC scans were positive, respectively. In four of 18 (22.2%) patients the site of bleeding was confirmed by angiography, but in 14 of 18 (77.8%), the findings were negative. In eight of 18 (44.4%) cases, the underlying lesion was definitely identified by contrast-enhanced MDCT including six cases of bleeding diverticulosis, one case of bleeding rectal–sigmoid mass, and one case of angiodysplasia. Moreover, in a ninth patient, a case of probable angiodysplasia was identified.
Similarly, results for the five patients who went directly from contrast-enhanced MDCT to angiography were as follows: positive on contrast-enhanced MDCT and angiography for three (Fig. 5A, 5B, 5C), positive on contrast-enhanced MDCT and negative on angiography for one, and negative on contrast-enhanced MDCT and positive on angiography for one.
|
|
|
In total, five gastrointestinal carcinomas were identified in our 55 patients (two new colon, one new rectal, one recurrent rectal, and one new rectal–sigmoid) (Fig. 6A, 6B, 6C). Of these five cases, one bleeding mass was identified on contrast-enhanced MDCT with bleeding also seen on 99mTc-labeled RBC scanning, one nonbleeding mass lesion was identified on contrast-enhanced MDCT, one bleeding site was identified on contrast-enhanced MDCT and 99mTc-labeled RBC scanning with underlying mass later identified at colonoscopy, one nonbleeding mass not seen on contrast-enhanced MDCT or 99mTc-labeled RBC scanning was later identified on colonoscopy, and one nonbleeding mass not seen on contrast-enhanced MDCT was not evaluated with 99mTc-labeled RBC scanning and was later detected on colonoscopy.
|
|
|
To minimize the chance of intermittent bleeding between diagnostic tests, we made an effort to carefully coordinate test scheduling. Time intervals for matched contrast-enhanced MDCT, 99mTc-labeled RBC scanning, and angiography examinations were as follows: median time for contrast-enhanced MDCT to 99mTc-labeled RBC scanning: 30 minutes (range: 5 minutes to 28 hours 19 minutes), median time for 99mTc-labeled RBC scanning to angiography: 54 minutes (range: 10 minutes to 28 hours 46 minutes), and median time for contrast-enhanced MDCT to angiography: 56 minutes (range: 21 minutes to 1 hour 57 minutes).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our protocol, modified from one used at Harvard Medical School and Beth Israel Deaconess Medical Center (R. G. Sheiman, October 2004, personal communication), was designed to maximize the conspicuity of contrast pooling within the bowel lumen, comparing the unenhanced and delayed contrast-enhanced phases. Arterial or late delayed series were not included so as to minimize radiation dose. Also, we designed our protocol with the intent of minimizing IV contrast dose because patients showing active hemorrhage would undergo traditional angiography to minimize risk of contrast-induced nephropathy.
Moreover, we avoided low-attenuation oral contrast agents so as to simplify coordination of cases, minimize possible dilutional effects at locations of hemorrhage, and minimize oral intake in patients who might undergo surgical intervention. Timing of contrast-enhanced scanning to the portal venous phase allowed maximum exposure time of the bleeding site to densely contrast-enhanced blood; all segments of the bowel to be perfused with contrast-enhanced blood, given the different circulation times in different vascular distributions; and maximum time for the relative washout of bowel mucosa enhancement; thus, theoretically decreasing the likelihood of false-negative or false-positive studies. Moreover, this relatively brief delay after the contrast bolus left minimal time for dispersion or dilution of extravasated contrast material in the bowel lumen, improving conspicuity and spatial localization.
Because of its long, continuous sampling time, 99mTc-labeled RBC scanning will always be able to detect intermittent bleeding better than any test that samples for a short period of time, whether it be angiography, colonoscopy, or CT. In fact, nuclear medicine sulfur colloid bleeding studies are sometimes used specifically because of the short testing time but have widely been replaced with labeled RBC scanning because of recognition of the advantages of longer sampling times. Because of the nature of the tests, simultaneous testing and comparison will likely never be possible. Given our experience, none of these tests is perfect, with the sensitivity and the specificity of the test dependent more on the dynamics of the bleeding and the timing of the test rather than on the test itself.
Despite these limitations, there are distinct advantages to contrast-enhanced MDCT over 99mTc-labeled RBC scanning and angiography that make a provocative case for using it as the first screening study in acute lower gastrointestinal bleeding. MDCT is available in an increasing number of hospitals and is frequently available 24 hours a day. In settings where there is a lack of access to emergency angiography or emergency 99mTc-labeled RBC scanning, a positive contrast-enhanced MDCT can define with a high degree of accuracy the location, and at times the cause, of the active lower gastrointestinal bleeding. Critically important ancillary findings are commonly found on contrast-enhanced MDCT. The contrast-enhanced MDCT does not hinder or exclude the use of 99mTc-labeled RBC scanning as a subsequent study. A recent review article describes the promise of contrast-enhanced MDCT as a first-line option for the accurate diagnosis or exclusion of active gastrointestinal hemorrhage [17], further underscoring the relevance of our results.
It should be emphasized that a negative contrast-enhanced MDCT does not mean that a contemporaneous 99mTc-labeled RBC scanning or angiography will be negative. These patients still need to be assessed carefully. In our study, the overall positivity rate was higher for 99mTc-labeled RBC scanning than for contrast-enhanced MDCT. Moreover, the two cases negative on 99mTc-labeled RBC scanning and positive on CT were limited examinations, only imaged for 30 minutes each.
It should also be emphasized that a negative 99mTc-labeled RBC scan does not mean that a contemporaneous contrast-enhanced MDCT scan or angiogram will be negative. Given what is assumed to be the intermittent nature of active lower gastrointestinal bleeding, this is easily understandable.
Therefore, active lower gastrointestinal bleeding remains a challenge for diagnostic imaging. In the setting in which 99mTc-labeled RBC scanning is available, the nuclear medicine studies are still more likely to give a positive result with reasonably good anatomic localization. If that is enough, it is still the best place to start. Contrast-enhanced MDCT, either as the initial or a second study, offers clear advantages in anatomic detail and the discovery of ancillary findings that may alter therapeutic plans based on nuclear medicine or invasive angiography alone.
In the interest of time, when a patient with gastrointestinal bleeding presents, an aliquot of blood might be drawn for UltraTag labeling at the time of contrast-enhanced MDCT so as to begin the process. If contrast-enhanced MDCT is positive, and scintigraphy is not needed, then the UltraTag labeling could be terminated and discarded as radioactive waste. If the contrast-enhanced MDCT is negative, scintigraphy could begin promptly if desired. If the clinical situation dictates emergency angiography and embolization for detection and treatment, contrast-enhanced MDCT offers an alternative or replacement for nuclear medicine with little time penalty and with the added advantage of potentially localizing the site and defining vascular abnormalities such as aneurysms, occlusions, or congenital anomalies that would impact the procedure. In our test subjects, there was no adverse effect on renal function even if both contrast-enhanced MDCT and invasive angiography were performed.
In conclusion, our study has proven with statistical significance (simple
agreement = 68.3%, = 0.341, p = 0.014) that there is
disagreement between contrast-enhanced MDCT and 99mTc-labeled RBC
scanning in detect ing active lower gastrointestinal hemorrhage.
Contrast-enhanced MDCT identified an active bleeding site prospectively in 16
of 60 (26.7%) cases, 19 of 60 (31.7%) cases in total. Nine of 16 (56.3%) and
10 of 19 (52.6%) cases required definitive invasive therapy (surgery,
angioembolization, or endoscopic electrocautery). Contrast-enhanced MDCT has
limited utility in cases of intermittent hemorrhage and involves IV contrast
material as well as a relatively high radiation dose. Still, the distinct
advantages of contrast-enhanced MDCT are multifold. Contrast-enhanced MDCT is
a non invasive, nontoxic diagnostic tool with ready availability and
repeatability that is capable of performing a rapid assessment, detailing
underlying bowel pathology, and map ping anatomy for angiography or
surgery.
Acknowledgments
We are grateful for the consultation and statistical analysis provided by
David O'Sullivan and Joseph Tortora. Again, thanks to Erik Paulson of the Duke
University Medical Center department of radiology. Thank you to the Hartford
Hospital radiology technologists and physicians. Thank you to the Hartford
Hospital general surgery and colon and rectal surgery staff including Orlando
Kirton, Jeffrey Cohen, Paul Vignati, William Sardella, and Kristina Johnson
for their support and continued cooperation.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
