| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiology (S.H.P., H.K.H., M.J.K., K.W.K., A.Y.K., D.H.Y., M.G.L., P.N.K., Y.M.S.) and Internal Medicine (S.K.Y., S.J.M., Y.I.M.), University of Ulsan College of Medicine, Asan Medical Center, 388–1 Poongnap-Dong, Songpa-Gu, 138–040 Seoul, Korea. Received April 2, 2004; revision requested June 8; revision received June 26; accepted July 27. Address correspondence to H.K.H. (e-mail: hkha@amc.seoul.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Institutional review board approval and informed consent were obtained. The study included 394 colonic segments in 31 men and 25 women at high risk for colorectal cancer (mean age ± standard deviation, 60.2 years ± 9.3 for men and 56.8 years ± 13.3 for women). Multi–detector row CT colonography and colonoscopy (reference standard) were performed in a blinded manner, and the results were compared. CT colonographic findings were interpreted in consensus by two readers using a primary two-dimensional with three-dimensional problem-solving approach. Adequacy of colonic preparation and distention was recorded. Sensitivity and specificity were obtained with 95% confidence intervals (CIs). Lesions missed at CT colonography were retrospectively reassessed to identify why they were missed, and, if the causes were not apparent, logistic regression analysis was performed to determine the presumptive causes.
RESULTS: Colonic preparation and distention were optimal in 17 patients (30%) but suboptimal in 37 (66%) and poor enough to make the results nondiagnostic in two (4%). Twenty-nine of 63 lesions were missed at CT colonography. When all flat, sessile, and pedunculated lesions (n = 63) were included, sensitivities were 75% (nine of 12; 95% CI: 48%, 100%) for lesions 10 mm or larger and 79% (19 of 24; 95% CI: 65%, 93%) for those 6 mm or larger. When only sessile and pedunculated lesions (n = 60) were included, corresponding sensitivities were 100% (nine of nine; 73%, 100%) and 90% (19 of 21; 78%, 100%), respectively. All three missed lesions larger than 10 mm were flat, and all three flat lesions were missed. Two 3-mm high lesions, including one invasive adenocarcinoma, were misinterpreted as feces at blinded image review; one 1-mm high tubular adenoma with adenocarcinoma foci could not be visualized even in retrospect. Sessile or pedunculated polyps 5 mm or smaller were significantly more likely to be missed than those 6 mm or larger (adjusted odds ratio, 11.6; P = .027).
CONCLUSION: Aside from inadequate bowel preparation and/or distention, flat lesions and small polyps are the two main causes for missed lesions at multi–detector row CT colonography.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
The size of a lesion is regarded as the most important factor determining lesion detectability at CT colonography, which is more sensitive to larger lesions (4–7,9,12). The causes for false-negative findings or missed lesions at CT colonography have been briefly but descriptively addressed in retrospective reviews of false-negative results (4,6,8,9,13). According to these reports, the main causes for missing lesions were small lesion size and poor bowel preparation and/or distention (4,6,8,9,13). To our knowledge, however, not much attention has yet been given to other possible causes for missing lesions, and, with one exception (14), systematic multivariate analysis to determine factors affecting lesion detection has not been attempted. A few reports have mentioned that the flat morphology of the lesions has impeded their detection with CT colonography (4,7,11,14,15). However, two of those reports were based on results acquired with single–detector row spiral CT scanners (7,11). Therefore, the purpose of our study was to determine causes of false-negative results at multi–detector row CT colonography and to determine presumptive causes by means of logistic regression analysis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Patient Population
Between May 24 and December 27, 2003, 113 consecutive patients in whom colon cancer was suspected or had been newly diagnosed were referred to our institution. There was no statistically significant difference in age distribution between the 68 men (mean age ± standard deviation, 58.5 years ± 11.2; range, 19–77 years) and the 45 women (mean age, 57.1 years ± 12.5; range, 31–81 years) in this series. The patients were scheduled to undergo optical colonoscopy and prospectively enrolled to undergo comparative multi–detector row CT colonography. Of these 113 patients, 103 underwent colonoscopy and 10 refused the examination.
Colonoscopy was performed between 21 days before and 24 days after CT colonography (mean interval between examinations, 4.4 days ± 5.6). In 40 patients, colonoscopy was incomplete because of colonic obstruction by colon cancer (n = 23), poor bowel preparation (n = 7), poor patient cooperation or intolerance of the procedure (n = 2), technical failure in manipulation of the colonoscope (n = 2), or unspecified reasons (n = 6). The 10 patients who did not undergo colonoscopy and the 40 in whom colonoscopy was incomplete were excluded from the study. In five other patients, polypectomy was performed during colonoscopy, 1–5 days before CT colonography. These patients were also excluded from the study rather than being included as patients with negative cases, because mucosal changes after recent polypectomy, such as localized mucosal edema, could mimic true lesions and their inclusion could be construed as a bias. Two other patients were excluded because innumerable polyps made complete comparison of CT colonography and colonoscopy impossible.
The remaining 56 patients who underwent both CT colonography and complete colonoscopy formed the population for this study; this group included 31 men (mean age, 60.2 years ± 9.3; range, 39–77 years) and 25 women (mean age, 56.8 years ± 13.3; range, 31–81 years). They underwent colonoscopy between 11 days before and 24 days after CT colonography (mean interval between examinations, 3.7 days ± 5.2). Sixteen patients underwent CT colonography and colonoscopy on the same day, with CT colonography performed first and colonoscopy performed 3–4 hours later. Finally, colon cancer was pathologically proved in 55 of the 56 patients (44 rectal and 11 sigmoid adenocarcinomas); one patient had no pathologic evidence of colon cancer.
CT Colonography
Study technique.—All patients underwent bowel preparation with a low-residue diet followed by an oral magnesium carbonate preparation (Magcorol Soln; Taejoon Pharmaceutical, Seoul, Korea) and 10 mg of bisacodyl (Dulcolax; Boehringer Ingelheim, Seoul, Korea) the evening before CT colonography. CT colonography was performed after intramuscular injection of 20 mg of scopolamine N-butylbromide (Buscopan; Boehringer Ingelheim) in the absence of contraindications (history of glaucoma or arrhythmia or symptoms of urinary outflow obstruction).
Before the examination, the patient was placed in a left lateral decubitus position, and the colon was gently insufflated with room air through a flexible rubber tube placed in the rectum, according to the patient’s tolerance. On average, about 2 L of room air was insufflated into the colon. Then the patient was scanned in the prone position. With the patient prone, an anteroposterior CT scout image was obtained to ensure adequate bowel distention. Further air insufflation was performed when collapsed bowel segments were identified. Before the patient was scanned in the supine position, colonic distention was checked with a second scout image, and the colon was insufflated with additional air when collapsed segments were identified.
CT scanning was performed with a multi–detector row CT scanner (LightSpeed QX/i; GE Medical Systems, Milwaukee, Wis). Scanning parameters for CT colonography were as follows: beam collimation, 4 x 2.5 mm (four detectors with 2.5-mm section thickness); reconstruction interval, 1.25 mm; pitch, 6; gantry rotation time, 0.8 second; table speed, 18.75 mm/sec; field of view to fit; 120 kV; and 160 mAs. Scanning was performed from the top of the diaphragm down to the anus, and the acquisition time ranged from 21 to 27 seconds. Patients were instructed to suspend respiration during scanning. Images were reconstructed with a standard body reconstruction algorithm available on the CT scanner and routinely used for abdominal CT scanning.
No oral contrast agent was used. Patients were scanned in the prone position without contrast material enhancement and in the supine position after intravenous injection of 150 mL of iopromide (Ultravist 370; Schering, Berlin, Germany). Iopromide was injected with a power injector at a rate of 3 mL/sec through an 18-gauge angiographic catheter inserted in the antecubital vein. Contrast material enhancement was performed to detect possible liver metastasis, because the patients were either suspected of having or newly diagnosed as having colon cancer. Therefore, patients were scanned in the supine position 72 seconds after initiation of contrast material injection, which corresponded to the venous phase for the liver.
Image analysis.—Two board-certified gastrointestinal radiologists, who had 7 (A.Y.K.) and 5 (M.J.K.) years of experience in gastrointestinal radiology and experience with more than 50 CT colonography cases with colonoscopic correlation, interpreted the CT colonographic images in consensus at a dedicated workstation by using a software package with volume-rendering capabilities (V-works 5.0; CyberMed, Seoul, Korea). Both observers were blinded to the colonoscopic findings but knew that the patients were suspected of having or were newly diagnosed as having colon cancer.
The colon was divided into eight segments: cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid colon, and rectum. Colonic segments were included in the analysis unless they had obviously detectable, fungating, or luminal encircling masses. In 44 patients with rectal cancer and 10 with sigmoid colon cancer, colon cancer masses were obviously detected as fungating or luminal encircling masses larger than 3 cm. Therefore, the 44 rectums and 10 sigmoid colons of these patients were excluded from analysis, leaving 394 colonic segments in 56 patients for the analysis. The rationale for excluding segments with large masses was that these large, advanced lesions are obviously much more detectable at CT colonography than are smaller lesions. If they were included, the performance of the technique would be overestimated, especially in segments with a high prevalence for advanced colon cancer, resulting in study bias.
For the image analysis, observers were asked to use a "time-efficient" technique (2,16,17). The analysis primarily consisted of a review of the two-dimensional transverse images. When a suspected lesion was identified on two-dimensional transverse images and coronal and sagittal images, three-dimensional endoluminal views were evaluated to characterize the lesion better. Specifically, coronal and sagittal reformations and endoluminal images were used to differentiate hypertrophied haustral folds or residual stool from colonic lesions.
The CT colonographic images were reviewed with a window width and level of 1500 HU and –200 HU, respectively. In addition, when a suspected lesion was detected, the observers carefully examined the internal attenuation of the lesion by modifying the window width and level settings to depict small gas bubbles, a finding consistent with stool. To further differentiate residual stool from polyps, the images from both the prone and the supine data sets were reviewed, to see whether the suspected lesion changed position when the patient moved from a prone to a supine position, another finding consistent with stool. Attention was paid to distinguishing the true mobility of polyps from positional changes of the colon in the mesentery (1,18).
The presence of respiratory and motion artifacts and the adequacy of colonic preparation and distention were recorded. The overall adequacy of colonic preparation and distention was evaluated by visually estimating the approximate percentage of the colonic wall accessible for evaluation. The presence, location, size, and morphology of all suspected lesions were recorded. The location of each lesion was specified as one of the eight segments (cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid colon, and rectum). Locations other than the cecum were further recorded as the proximal, middle, or distal portion within each segment.
Before each lesion was measured, its overall shape was checked on the three-dimensional endoluminal view to see whether it was symmetric or asymmetric. Lesions were measured with an electronic ruler after appropriate magnification; the longest dimension on two-dimensional images was measured for symmetric lesions, and both long- and short-axis diameters and height were measured for asymmetric lesions. The morphology of the lesion was described as flat, sessile, or pedunculated. Flat lesions were defined as mucosal elevations with a height less than half the lesion diameter (19). All lesions seen at CT colonography were photographed and stored in digital Joint Photographic Experts Group format.
Optical Colonoscopy
Colonoscopy was considered the reference standard. Bowel preparation was the same as for CT colonography. A single board-certified gastroenterologist (S.K.Y.) who had performed more than 1000 examinations performed colonoscopy with a standard videocolonoscope (CF series; Olympus Optical, Tokyo, Japan) and without knowledge of the CT colonographic findings. The instrument tip was inserted into the cecum and was subsequently withdrawn for the detection of lesions. Lesions were photographed, and their presence, location, size, and morphology were documented. Lesion location was estimated by using endoscopic landmarks and insertion distances at withdrawal and was specified as one of the eight segments. Locations other than the cecum were further recorded as the proximal, middle, or distal portion within each segment. Lesion size was determined by comparing the lesion with open endoscopic biopsy forceps pushed against the lesion or by measuring the resected specimen directly with a ruler when the lesion was retrieved in total. The morphology of the lesions was described as flat, sessile, or pedunculated. Flat lesions were defined as mucosal elevations with a height less than half the lesion diameter (19). Histopathologic diagnosis was obtained in all the lesions found at colonoscopy, by means of either colonoscopic polypectomy or biopsy.
CT Colonographic and Colonoscopic Data Comparison and Review of Lesions Missed at CT Colonography
All the lesions identified at CT colonography were compared with those identified at colonoscopy by the two radiologists (A.Y.K., M.J.K.) who reviewed the CT colonographic images and the gastroenterologist (S.K.Y.) who performed the optical colonoscopy. For a given lesion to be considered a true-positive match between CT colonography and colonoscopy, it had to be in the same colonic segment or an adjacent portion of an adjacent segment (eg, a lesion seen in the hepatic flexure at CT colonography was considered matched with one seen in the distal ascending colon, hepatic flexure, or proximal transverse colon at colonoscopy), the two recorded sizes had to be the same within a 30% margin of error, and the lesion had to show similar morphologic features at colonography and colonoscopy. All the lesions depicted with CT colonography that were not seen at colonoscopy or did not match a colonoscopic finding were considered false-positive findings. Lesions that were missed at the blinded review of the CT colonographic images were reassessed in retrospect by the same two radiologists with knowledge of the colonoscopic findings, and a consensus judgment was made as to the cause of the false-negative results. Unlike in the blinded review, window width and level setting was modified to visualize contrast enhancement of the colonic lesions when necessary, and an intermediate soft-tissue window (width, 400 HU; level, 20 HU) (20) was reviewed to detect flat lesions that had been missed at blinded review.
Data and Statistical Analysis
The numbers of missed lesions and correctly detected lesions at the blinded review of CT colonographic images were counted and further analyzed according to their histopathologic diagnosis, size, and morphology. The sensitivities in lesion detection for lesions 10 mm or larger, 8 mm or larger, 6 mm or larger, and for all lesions were calculated on a per-lesion basis, along with 95% confidence intervals. The specificities in lesion detection (with 95% confidence intervals) were calculated on a per-patient basis. Data clustering (ie, more than one lesion per patient) was accounted for with the method of Rao and Scott (21). For the lesions that could not be visualized with CT colonography for reasons that were not apparent at retrospective image review, presumptive causes were assessed with logistic regression analysis. Those lesions were compared with those identified at either blinded or retrospective review of the CT colonographic findings with regard to histopathologic diagnosis, size, morphology, and location. Variables with P values less than .25 in the univariate simple logistic regression analysis were chosen as the variables for multiple logistic regression analysis. In both univariate simple and multiple logistic regression analysis, P values less than .05 were considered to indicate a statistically significant difference. Statistical analysis was performed with commercially available software (SPSS 10.0 for Windows; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Blinded Review
At colonoscopy, 68 lesions were found in 49 segments of 28 patients. Of those lesions, five were demonstrated at histopathologic evaluation to be chronic nonspecific inflammation, which was not considered a truly pathologic finding; these lesions were therefore excluded from the analysis. The remaining 63 lesions in 48 segments of 28 patients included three flat, 46 sessile, and 14 pedunculated lesions. Thirty-four (54%) of 63 lesions were correctly detected at blinded review of the CT colonographic images, but 29 (46%) were missed. There were 41 false-positive findings in 20 patients at CT colonography (size range, 3–10 mm; mean size, 5.0 mm ± 1.5). The numbers of missed lesions and correctly detected lesions at blinded review of the CT colonographic images are listed in Table 1 according to the histopathologic diagnosis, size, and morphology of the lesions.
|
When all 63 lesions were considered, the sensitivities were as follows for lesions 10 mm or larger, 8 mm or larger, 6 mm or larger, and for all lesions: 75% (nine of 12; 95% confidence interval: 48%, 100%), 77% (10 of 13; 52%, 100%), 79% (19 of 24; 65%, 93%), and 54% (34 of 63; 42%, 66%), respectively. When only 60 sessile or pedunculated lesions were included, the sensitivities were 100% (nine of nine; 95% confidence interval: 73%, 100%), 100% (10 of 10; 75%, 100%), 90% (19 of 21; 78%, 100%), and 57% (34 of 60; 44%, 70%), respectively. Findings at colonoscopy were negative in 26 patients. Of these patients, 19 had negative findings at CT colonography, for a per-patient specificity of 73% (95% confidence interval: 54%, 87%).
Retrospective Review
At retrospective review of the CT colonographic results, the causes for missed lesions included flat morphology (n = 3), poor bowel preparation and/or distention (n = 6), and perceptive errors (ie, lesions not identified at the blinded review but identified retrospectively) (n = 4). For the remaining 16 missed lesions (size range, 3–6 mm; mean size, 4.1 mm ± 0.9), no obvious cause was found; that is, lesions were not identified in retrospect despite being sessile or pedunculated and despite adequate preparation and distention of relevant bowel segments. The 16 x 12 x 3-mm flat invasive adenocarcinoma (Fig 1) and the 12 x 10 x 3-mm flat tubular adenoma were noted as plaquelike elevations but were misinterpreted as adherent fecal materials at the blinded review. These two lesions showed moderate contrast enhancement during the venous phase, which enabled them to be distinguished from fecal materials (Fig 1). No other lesions were found because of intravenous contrast enhancement. The flat invasive adenocarcinoma showed a central depression within the lesion (Fig 1). The 13 x 13 x 1-mm tubular adenoma with adenocarcinoma foci (Fig 2) was also not conspicuous in retrospect on either the two- or the three-dimensional endoluminal view, despite adequate bowel preparation and distention. Three 4-mm and one 3-mm sessile tubular adenomas were missed because of perceptive errors.
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
When only sessile or pedunculated lesions were considered in our study, the sensitivity was similar to or slightly higher than that found in other studies performed in symptomatic, high-risk populations (4,6,9). In addition, all of the lesions missed at the blinded review were small (mean size, 4.1 mm ± 0.9) and generally unimportant, as none of them had high-grade dysplasia or carcinoma foci. However, when all lesions were considered, including flat lesions, the sensitivities for lesions 10 mm or larger, 8 mm or larger, and 6 mm or larger decreased to 75%, 77%, and 79%, respectively; these were lower than the sensitivities reported in the literature (4,6,9). In those previous studies there were either fewer flat lesions than in our study (two of 132 lesions [4] vs three of 63 lesions in our series) or no flat lesions (6,9).
When we included the flat lesions, two cancerous lesions (one invasive adenocarcinoma and one tubular adenoma with adenocarcinoma foci) were among those missed at blinded image review in this study. Accordingly, our results suggest that flat colonic lesions may present a critical problem in the surveillance for colon cancers or their precursors with CT colonography. However, because of our limited number of cases, the 95% confidence intervals overlapped for sensitivities with and without flat lesions included, so our results are yet to be generalized. Further studies on a larger scale are required for more definitive results.
The difficulty in detecting flat lesions with CT colonography is thought to have several causes. According to the adenoma-carcinoma hypothesis, the surveillance for colon cancers or their precursors has focused mainly on identifying sessile or pedunculated polyps. Therefore, observers were not as familiar with the morphology of flat lesions as with that of sessile or pedunculated lesions. Awareness of flat lesions and familiarity with their morphologic features would help in their diagnosis.
As seen in two lesions in our study that were noted but misinterpreted as fecal material at blinded review, plaquelike morphology may also easily be mistaken for fecal materials. For this reason, the use of fecal tagging with an oral contrast agent (15,24–26) seems to help in detecting flat lesions. In one study, multi–detector row CT colonography with fecal tagging demonstrated four of five flat lesions 8–25 mm in diameter (15). Although contrast material enhancement helped detect two flat lesions in our study, no additional lesion was found because of intravenous contrast enhancement. Moreover, contrast enhancement is not regarded as a routine procedure for screening CT colonography.
Although our number of cases was limited, the detectability of flat lesions seemed to be related to lesion height; of the three flat lesions, two 3-mm-high lesions were detected, and one 1-mm-high lesion was not. In other studies (4,15), flat lesions 2 mm or less in height were also not detected; they included two flat lesions 1 mm high (10 and 7 mm in diameter) (4) and one lesion 2 mm high (25 mm in diameter) (15). Our results and those of others suggest that flat lesion should be at least 3 mm high to be detected with CT colonography, although further study with large numbers of cases will be needed.
To our knowledge, there have only been two large studies (14,20) on the detectability of flat lesions at CT colonography, and they showed 15%–65% sensitivities with large interreader variability (20) and 47% sensitivity (14); they did not address lesion height (14,20). We believe that the definition of flat lesions as mucosal elevations with a height less than half the lesion diameter (19) is not detailed enough (27), because it puts lesions with a wide range of heights into one category. When a flat lesion is suspected, the height of the lesion should be reported specifically (27). Given the high risk of cancer foci or high-grade dysplasia in flat lesions larger than 8–10 mm in diameter (19,22) and given that some flat lesions are large in diameter but short, CT colonography will miss some important lesions, as happened with the 13 x 13 x 1-mm tubular adenoma with adenocarcinoma foci in our study.
For the lesions that could not be detected for reasons not apparent at retrospective image review, our logistic regression analysis showed that lesion size was the only factor significantly associated with lesion detectability. Sessile or pedunculated lesions 5 mm or smaller were more difficult to visualize than those 6 mm or larger, with an adjusted odds ratio of 11.6. Nevertheless, there is still a legitimate rationale for using CT colonography in the surveillance of colon cancers and their precursors, because the malignant potential of sessile or pedunculated colonic polyps is correlated with lesion size, and, in general, diminutive polyps smaller than 5 mm are considered unimportant (1,28). If follow-up examinations are performed on an interval basis, missing these diminutive lesions is probably not of great consequence, as their dwell time is long (up to 10 years) in the development and progression of the adenoma-carcinoma sequence (1,29,30).
This study has several limitations. Although CT colonography and colonoscopy were performed prospectively, the timing for each examination was not strictly controlled; we performed the examinations according to scheduling availability. Therefore, only 16 of the 56 patients included in the study underwent CT colonography and colonoscopy on the same day. Although colonoscopy provided very high sensitivity for the detection of colonic lesions and has been used as a reference standard in many studies on CT colonography, it may not be a perfect reference standard (3,31). One advantage of same-day colonoscopy is the chance to reexamine the patient with the information from the CT colonography if blinded colonoscopy did not reveal a lesion suspected at CT colonography; this creates an enhanced reference standard (12). Another limitation of our study was that the proportion of nonneoplastic lesions (six of 63 lesions) was quite low, probably because of the high risk for colonic neoplasia in our study population.
The method of CT colonography used in our study was somewhat different from that in other studies in a few ways. A routine radiation dose for abdominal CT scanning was used, and intravenous contrast material was administered. Routine soft-tissue windows were not used for evaluation, which biased the study against detecting subtle areas of wall thickening. Three-dimensional "fly-through" scanning, which may have improved detection of some lesions, was not performed. Having two experts review images in consensus may have improved sensitivity in our study and decreased specificity.
With a four–detector row scanner with 0.8-second gantry rotation speed, we chose to use 2.5-mm-thick sections. With faster tube rotation or a higher detector number, thinner sections may have improved the detection of some flat lesions. Further study with a higher-performance scanner is needed. Finally, many false-positive findings were noted, which was probably due to both the nature of our study cohort (most of the patients had fungating or luminal encircling colon cancer masses, which made bowel preparation difficult) and the use of consensus interpretation of CT colonography by two experts.
In conclusion, except for inadequate bowel preparation and/or distention, flat lesions and the small size of polyps were the main causes for missed lesions at CT colonography in our study. Flat lesions less than 1–2 mm in height cannot be visualized with multi–detector row CT colonography, and sessile or pedunculated polyps 5 mm or smaller are 11.6 times more likely to be missed than are larger lesions.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, S.H.P., H.K.H.; study concepts and design, all authors; literature research, S.H.P., H.K.H., M.J.K., K.W.K., P.N.K., Y.M.S., S.J.M., Y.I.M.; clinical studies, A.Y.K., M.G.L., S.K.Y.; data acquisition, M.J.K., A.Y.K., D.H.Y., S.K.Y.; data analysis/interpretation, S.H.P., K.W.K.; statistical analysis, S.H.P.; manuscript preparation, S.H.P.; manuscript definition of intellectual content, S.H.P., H.K.H., K.W.K.; manuscript editing, S.H.P., H.K.H.; manuscript revision/review and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. H. Kim, J. M. Lee, C.-I. Shin, H. C. Kim, J.-G. Lee, J. H. Kim, J. Y. Choi, H. W. Eun, J. K. Han, J. Y. Lee, et al. Effects of Spatial Resolution and Tube Current on Computer-aided Detection of Polyps on CT Colonographic Images: Phantom Study Radiology, August 1, 2008; 248(2): 492 - 503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Soetikno, T. Kaltenbach, and R. V. Rouse Screening for Nonpolypoid Colorectal Neoplasms--Reply JAMA, June 18, 2008; 299(23): 2743 - 2744. [Full Text] [PDF]
|
|
|
|
 |
|
 K. K. Deshpande, R. M. Summers, R. L. Van Uitert, M. Franaszek, L. Brown, A. J. Dwyer, J. G. Fletcher, J. R. Choi, and P. J. Pickhardt Quality Assessment for CT Colonography: Validation of Automated Measurement of Colonic Distention and Residual Fluid Am. J. Roentgenol., December 1, 2007; 189(6): 1457 - 1463. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. H. Kim, J. M. Lee, H. W. Eun, M. W. Lee, J. K. Han, J. Y. Lee, and B. I. Choi Two- versus Three-dimensional Colon Evaluation with Recently Developed Virtual Dissection Software for CT Colonography Radiology, September 1, 2007; 244(3): 852 - 864. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. H. Kim, J. M. Lee, J.-G. Lee, J. H. Kim, P. A. Lefere, J. K. Han, and B. I. Choi Computer-Aided Detection of Colonic Polyps at CT Colonography Using a Hessian Matrix-Based Algorithm: Preliminary Study Am. J. Roentgenol., July 1, 2007; 189(1): 41 - 51. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. K. Choi, S. H. Park, D. Y. Kim, and H. K. Ha Malignant Rectal Polyp Overlooked on CT Colonography Because of Retention Balloon: Opposing Crescent Appearance as Sign of Compressed Polyp Am. J. Roentgenol., July 1, 2007; 189(1): W1 - W3. [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]
|
|
|
|
|
|
S. H. Park, S. S. Lee, E. K. Choi, S. Y. Kim, S.-K. Yang, J. H. Kim, and H. K. Ha Flat Colorectal Neoplasms: Definition, Importance, and Visualization on CT Colonography Am. J. Roentgenol., April 1, 2007; 188(4): 953 - 959. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Mang, A. Maier, C. Plank, C. Mueller-Mang, C. Herold, and W. Schima Pitfalls in Multi-Detector Row CT Colonography: A Systematic Approach RadioGraphics, March 1, 2007; 27(2): 431 - 454. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Halligan, S. H. Park, and H. K. Ha Causes of False-Negative Findings at CT Colonography Radiology, March 1, 2006; 238(3): 1075 - 1077. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
