| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology and Research Institute of Radiology, University of
Ulsan College of Medicine, Asan Medical Center, 388-1 Poongnap2-Dong,
Songpa-Gu, Seoul 138-040, Korea.
2 Weill Medical College of Cornell University, New York, NY.
3 Department of Radiology, Seoul National University College of Medicine, Seoul
National University Bundang Hospital, Seongnam-si, Korea.
Received February 22, 2007;
accepted after revision March 26, 2007.
Address correspondence to S. H. Park
(seongho{at}amc.seoul.kr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Our institutional review board approved this study and waived patient informed consent. Forty-eight colonoscopy-proven polyps (6–20 mm) and 41 polypoid tagged feces (6–19 mm) were selected from contrast-enhanced CTC performed without (n = 37 examinations) and with (n = 10 examinations) fecal tagging, respectively. Scanning was performed 72 seconds after IV injection of 150 mL of contrast material at a rate of 2.5 mL/s. Fecal tagging consisted of three doses of 200 mL of 5% weight/volume (w/v) barium sulfate suspension taken at each meal the day before CTC. Attenuation of the polyps and tagged feces was measured. Four independent blinded radiologists reviewed the polyps and tagged feces at both wide (width, 1,500 H; level –400 H) and soft-tissue (width, 400 H; level, 20 H) window settings to distinguish them by using subjective visual assessment.
RESULTS. Polyp attenuation on the portal phase was not correlated with size (R = –0.003; p = 0.99) and was not different between histologic types (p = 0.884). Enhanced polyps (mean ± SD, 119.9 ± 25.3 H; range, 50–173 H) showed significantly lower attenuation than did tagged feces (1,521.4 ± 683.6 H; range, 495–2,683 H) without any overlap (p < 0.0005). An 8-mm sessile adenomatous polyp was misinterpreted as tagged feces by one reviewer. The rest of the lesions were correctly interpreted by all reviewers, resulting in high interobserver agreement (kappa value, 0.988).
CONCLUSION. Polyp attenuation on portal phase contrast-enhanced CTC ranges from 50 to 173 H. Contrast-enhanced polyps are clearly and consistently distinguished from barium-tagged polypoid feces.
Keywords: attenuation • colorectal polyps • CT colonography • fecal tagging
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although IV contrast enhancement is not routinely used with screening CTC because of its risk and uncertain cost-effectiveness [8], it has been shown to be helpful in differentiating polyps from fecal residue [9, 10] and improving the detection of polyps in suboptimally prepared colons [11]. Because of the use of fecal tagging, however, the use of IV contrast enhancement for the sole purpose of improving the ability to distinguish polyps from fecal residue may not be necessary. IV contrast enhancement is also important for the detection and characterization of clinically significant extracolonic abnormalities [12]. It therefore is necessary for use in patients with known or suspicious colorectal cancer and those who are seen for follow-up after curative surgery for colorectal cancer in whom IV contrast-enhanced CTC can provide simultaneous evaluation of colonic lesions and extracolonic metastasis [8, 13, 14]. In such cases, the portal phase, which is better for extracolonic evaluation including the detection of hepatic metastasis, is presumably more appropriate than the arterial, or mucosal phase in which bowel-wall enhancement is maximized [15] and polyps are better visualized [16].
Both fecal tagging and IV contrast enhancement increase the attenuation of luminal protrusions (either polyps or residual feces). Because differentiation of polyps from tagged feces is largely based on the attenuation differences, there exist potential concerns of the simultaneous use of fecal tagging and IV contrast enhancement complicating the interpretation. To our knowledge, there has been limited information on the attenuation of colorectal polyps at portal phase contrast-enhanced CTC [9, 10] and no information on comparison of polyp attenuation values at contrast-enhanced CTC with those of tagged feces. Therefore, we performed this study to determine the attenuation of colorectal polyps on portal phase contrast-enhanced CTC and evaluate whether those enhanced polyps can be clearly distinguished from tagged feces during CTC review.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CTC
No dietary fecal tagging was performed in the enhanced polyp group, whereas
fecal tagging was achieved in the tagged feces group by instructing the
ingestion of three doses of 200 mL of 5% weight/volume (w/v) barium sulfate
suspension immediately after each meal the day before the examination. Every
other preparation and procedure for CTC except fecal tagging was the same
between the two groups of patients. All patients underwent colonic preparation
the day before CTC by ingesting a low-residue semifluid diet and 4 L (246 mg)
of polyethylene glycol (PEG) solution. Colonic insufflation was performed with
carbon dioxide (CO2) gas using an automated insufflator
(PROTOCO2L, E-Z-EM). CT was performed with a 16-MDCT scanner
(Somatom Sensation 16, Siemens Medical Solutions). Scanning parameters were as
follows: beam collimation, 0.75 x 16; gantry rotation time, 0.5 second;
table feed, 12 mm per gantry rotation; field of view to fit; 120 kV;
50–150 mAs depending on anatomic location for the supine scan (i.e., the
use of automated dose reduction system [CARE Dose 4D, Siemens Medical
Solutions]) and 50 mAs for the prone scan; slice thickness, 1 mm; and
reconstruction interval, 0.7 mm. Supine images were acquired at 72 seconds
after the initiation of IV administration of 150 mL of iopromide (Ultravist
370, Schering) at a rate of 2.5 mL/s through a 20-gauge angiographic catheter
inserted in an antecubital vein. The prone imaging was performed after the
supine imaging.
Optical Colonoscopy
Optical colonoscopy was performed on the same day or up to 15 days after
CTC. Three board-certified gastroenterologists, each with experience of more
than 1,000 cases, performed optical colonoscopy with a standard
videocolonoscope (CF series, Olympus Optical). The gastroenterologists were
not blinded to the results of CTC when performing the optical colonoscopic
examinations.
Review of CTC: Lesion Attenuation Measurement and Visual Assessment
Two board certified radiologists, one with experience of approximately 100
cases and the other with experience of approximately 500 cases,
retrospectively interpreted in consensus all CTC examinations using a
commercial CTC system (syngo Colonography, Siemens Medical Solutions).
Selection criteria for enhanced polyps were five or fewer polyps per patient;
clear visualization on the supine scan—location within a well-distended,
well-cleansed CO2-filled portion of the colon, no submergence in
the fluid, and absence of any image artifacts; 6–20 mm measured by CTC
using an optimized multiplanar reformatted (MPR) plane—an arbitrary MPR
plane that allows view of the maximum polyp diameter; sessile or pedunculated
morphology; unambiguous match at optical colonoscopy; and available histologic
diagnosis. To avoid any ambiguous matches, we excluded patients with more than
five polyps. Flat polyps were excluded because of the technical infeasibility
of obtaining an accurate region of interest (ROI) for measurement of lesion
attenuation. Selection criteria for polypoid tagged feces were clear
visualization on the supine scan—location within a well-distended
CO2-filled portion of the colon, no submergence in the fluid, and
absence of any image artifacts; 6–20 mm measured by CTC using an
optimized MPR plane; and sessile or pedunculated morphology.
The longest dimension of the polyps and tagged feces was measured on an optimized MPR plane at window width and level settings of 1,500H and –400 H using an electronic ruler after appropriate magnification. Attenuation value was measured on 2D images at window width and level settings of 400 H and 20 H by manually drawing the ROI to encompass as much of the lesion as possible. We used standard soft-tissue window settings rather than colonic window settings for attenuation measurement to avoid erroneous inclusion of colonic air within the ROI due to partial volume averaging effect. All selected polyps and tagged polypoid feces were bookmarked on supine data sets for later blind independent review.
Four board-certified abdominal radiologists, who were not involved in the patient selection, consensus review, and lesion attenuation measurement and were blinded to the CTC techniques and true identity of the polyp versus feces groups, independently reviewed all the selected polyps and tagged feces that were bookmarked during the initial consensus review session. Two reviewers had case experience of approximately 500, whereas the remaining two reviewers each had 5 years of experience with abdominal CT but limited experience with CTC. Each reviewer interpreted the bookmarked polypoid structures on a colon window setting (width,1,500 H; level, –400 H) and a soft-tissue window setting (width, 400 H; level, 20 H) with adequate magnification and scrolling of images using the same CTC system (syngo Colonography) as was used in the initial consensus review. Reviewers were asked to decide whether the polypoid structure represented a true polyp or tagged feces by using visual assessment.
Data and Statistical Analysis
The attenuation values of the enhanced polyps and tagged feces were
summarized. For enhanced polyps, correlation between the attenuation value and
polyp size was analyzed by using the Spearman's correlation coefficient. The
polyp attenuation of the four histologic types of nonadenoma, adenoma, adenoma
with high-grade dysplasia, and adenocarcinoma was compared using the linear
mixed model to account for data clustering—that is, multiple lesions per
subject. For tagged feces, the correlation between the attenuation value and
size was analyzed using Spearman's correlation coefficient. The attenuation
value was compared between the enhanced polyps and tagged feces using a linear
mixed model to account for data clustering. Interobserver agreement of the
visual assessment (i.e., determination of enhanced polyp vs tagged feces) was
assessed using kappa statistics. In case of misinterpretation (i.e., an
enhanced polyp misinterpreted as a tagged feces or vice versa), the relation
of its occurrence with lesion attenuation and reviewer experience was
analyzed.
|
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A total of 41 polypoid tagged feces in 10 of the 15 patients in the tagged feces group fulfilled the selection criteria: two patients each with seven, one patient with six, two patients each with five, three patients each with three, and two patients each with one. The regional distribution of tagged feces was as follows: sigmoid colon, n = 12; descending colon, n = 5; transverse colon, n = 9; and ascending colon, n = 15. Mean size ± SD of tagged feces was 10 ± 3.4 mm (size range, 6–19 mm). All 41 polypoid tagged feces were homogeneously tagged, and the attenuation values ranged from 495 H (Figs. 3A, 3B, and 3C) to 2,683 H with a mean ± SD of 1,521.4 ± 683.6 H. There was no significant correlation between the size of tagged feces and the attenuation values (R = –0.29; p = 0.07).
|
|
|
|
|
|
Visual Assessment
The kappa statistic for the visual interpretation (i.e., determination of
enhanced polyp vs tagged feces) of the four reviewers was 0.988 (SE = 0.043),
indicating an excellent interobserver agreement. Inaccurate interpretation
occurred in only one case. An 8-mm sessile adenomatous polyp (attenuation
value, 114 H; 28th highest attenuation value of the 48 enhanced polyps) in the
transverse colon was misinterpreted as tagged feces by one experienced
reviewer by misinterpreting the lesion attenuation on the soft-tissue window
as artificially high (i.e., tagging agent) (Figs.
4A,
4B, and
4C). No tagged feces was
misinterpreted as an enhanced polyp.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mean attenuation value of polyps at a 72-second delay of contrast-enhanced CTC in our study was 119.9 ± 25.3 H. This was higher than the attenuation of benign colorectal polyps (78.9 ± 16.9 H) and carcinomas (90.7 ± 9.6 H) at a 70-second delay reported by Oto et al. [10]. It was also higher than the attenuation of benign polyps (90 ± 18 H) but similar to the attenuation of colorectal cancers (124 ± 18 H) at an 80-second delay in a study by Neri et al. [9]. The maximum attenuation value of polyps of our study (173 H) was also higher than that reported by Oto et al. (106 H for both benign polyps and carcinomas) [10]. The reasons for the discrepancies between the studies are not completely understood; however, several factors can be considered. The study of Oto et al. used a 5-mm slice thickness as opposed to the 1-mm thickness used in our study, resulting in some degree of inaccuracy due to partial volume averaging effects in cases of small polyps. The Oto et al. study also used a single-detector scanner with considerably slower table speed compared with our 16-MDCT scanner, most likely resulting in a longer scanning time delay for each lesion because of the longer scanner travel time from the diaphragm to the anus despite similar scanning start times. In comparison with our study, Neri et al. used a smaller amount of contrast material volume (i.e., 140 mL) and a greater timing delay (i.e., 80 seconds). All studies with a moderate number of lesions may also be subject to sampling bias. Given the variations between studies, a definitive conclusion about the attenuation of colorectal polyps on portal phase contrast-enhanced CTC may require further studies. Our study, however, clearly shows that polyp attenuation can reach as high as 173 H on portal phase contrast-enhanced CTC.
Detailed information on polyp attenuation is a prerequisite for accurate application of advanced postprocessing techniques of CTC such as computer-aided detection, electronic cleansing [20], or translucency rendering [21] that use attenuation values of intraluminal structures. Considering the difference in polyp attenuation between unenhanced and contrast-enhanced CTC, the accuracy and usefulness of those techniques will be improved with further adjustment for contrast-enhanced cases. Our study results may serve as reference data for the adjustment and improvement of such techniques. Specifically, our data suggest that the attenuation threshold for electronic cleansing should at least be higher than 173 H for portal phase contrast-enhanced CTC.
In our study, the attenuation value of polyps at a 72-second delay of IV contrast enhanced CTC did not correlate with size or histologic type, which is consistent with the results of Sosna et al. [22], showing the absence of correlation between the degree of contrast enhancement on 45-second delay CTC and the size or degree of histologic differentiation. In contrast, Oto et al. [10] showed that the mean attenuation value of carcinoma was significantly higher than that of benign polyps. The relationship of polyp enhancement and histology is still uncertain. Despite possible differences in polyp attenuation on the basis of histologic type, we do not expect contrast enhancement to significantly convey the clinical significance of polyps in the way polyp size does in current practice [23, 24].
Our results show that portal phase contrast-enhanced CTC can be performed with barium-based fecal tagging without interfering with the discrimination of polyps from tagged feces. On ROI attenuation measurement, there was a statistically significant attenuation difference between tagged feces (mean, 1,521.4 ± 683.6 H; range, 495–2,683 H) and enhanced polyps (mean, 119.9 ± 25.3 H; range, 50–173 H) without any overlap. Subjective visual assessment also led to correct differentiation of tagged feces from polyps in all cases and reviewers except for one polyp that was misinterpreted as tagged feces by an experienced reviewer. These results show a few notable points.
Unlike lesion detection on CTC that is heavily dependent on reviewer experience [25], discrimination of enhanced polyps from tagged feces is straightforward to the point of being independent of reviewer experience. Contrary to our expectation that polyps with higher attenuation values would be more likely to be misinterpreted as tagged feces, the one misinterpreted polyp had an attenuation value of 114 H, the 28th highest attenuation value of the 48 polyps in our study.
Visual perception of a reviewer may be influenced by various factors other than the attenuation of the intraluminal polypoid structure. Such detailed analysis, however, was not possible in our study because of the presence of only one misinterpretation. Despite the slight possibility of misinterpretation between enhanced polyp and tagged feces on subjective visual assessment, it is most likely not a critical problem in practice. Given the clear attenuation difference between enhanced polyps and barium-tagged feces, ROI measurement of lesion attenuation can lead to a clear distinction between the two in ambiguous cases on visual assessment.
Our study has limitations. First, the results may only be applicable to the specific tagging methods used in our study. We used 200 mL of 5% w/v barium sulfate suspension at each meal the day before the examination. Various tagging agents including barium, iodinated agents, and their combinations are, however, currently being used [3–5]. Moreover, the specific methods of administration of the tagging agents and their combined use with laxatives may also vary across institutions. Although all the tagged feces in our series were homogeneously tagged, the effectiveness of fecal tagging may vary depending on the tagging method used. In fact, difficulty in distinction between a polyp (either contrast-enhanced or unenhanced) and feces often occurs in cases of incomplete tagging. Our results, therefore, may not be generalized to every contrast-enhanced CTC performed with various tagging methods. Nevertheless, our results may provide guidelines regarding appropriate tagging methods in cases where simultaneous IV contrast material is used. An optimal fecal tagging method should produce homogeneous fecal tagging of significantly higher attenuation than enhanced polyps (i.e., 173 H on portal phase imaging according to our results) such that the simultaneous use of IV contrast material does not pose difficulty in polyp differentiation from feces.
Second, the effect of IV contrast enhancement on the detection of polyps submerged in tagged fluid and flat polyps was not evaluated. Although IV contrast enhancement has been reported to improve the detection of polyps submerged in untagged fluid [11] and to facilitate the differentiation of flat polyps from untagged feces [26], the effect of simultaneous use of IV contrast material and fecal and fluid tagging on the detection of such lesions requires further studies.
Third, the visual assessment in our study in distinguishing a polyp from tagged feces did not completely follow the standard CTC review protocol. Reviewers made the decision of polyp versus tagged feces largely based on the attenuation characteristics of lesions seen on supine images. In clinical practice, however, both supine and prone data sets are used for interpretation. Therefore, movability of a lesion with positional change of the patient (supine vs prone) can also contribute to the interpretation, possibly allowing for easier distinction between enhanced polyps and tagged feces.
Last, we did not perform unenhanced imaging. For research purposes, it would have been ideal to acquire both unenhanced and contrast-enhanced scans to assess absolute levels of enhancement. Such calculation, however, may not be practical, and it is not always possible to obtain reliable attenuation measurement of a polyp on both supine and prone (i.e., unenhanced and contrast-enhanced) scans. In our practice, we initially performed the supine imaging followed by the prone imaging because that is the typical examination sequence used with automated CO2 insufflation [27] (colonic insufflation with an automated insufflator is not performed efficiently in the prone position). In addition, for clinical practice, portal phase contrast-enhanced images needed to be obtained at the supine position.
In conclusion, the attenuation values of colorectal polyps on portal phase contrast-enhanced CTC range from 50 to 173 H without significant difference between histologic types. Contrast-enhanced polyps are clearly and consistently distinguished from barium-tagged polypoid feces with rare occurrence of misinterpretation. Such error can be avoided by using the ROI measurement of lesion attenuation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. H. Park, S. S. Lee, J. K. Kim, M.-J. Kim, H. J. Kim, S. Y. Kim, M.-Y. Kim, A. Y. Kim, and H. K. Ha Volume Rendering with Color Coding of Tagged Stool during Endoluminal Fly-through CT Colonography: Effect on Reading Efficiency Radiology, September 1, 2008; 248(3): 1018 - 1027. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
