| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Suita Municipal Hospital, Suita, Osaka, Japan.
2 Department of Radiology, Osaka University Graduate School of Medicine, 2-2,
Yamadaoka, Suita, Osaka 565-0871, Japan.
3 Department of Radiology, Minoh Municipal Hospital, Minoh, Japan.
4 Department of Radiological Sciences, University of Rome, Rome, Italy.
5 Department of Surgery, Minoh Municipal Hospital, Minoh, Japan.
Received November 9, 2004;
accepted after revision December 23, 2004.
Address correspondence to T. Murakami.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty consecutive patients underwent both CT colonography and conventional colonoscopy. Three radiologists independently analyzed the CT colonographic examinations of each patient using a primary 3D method. All examinations were analyzed using two techniques: navigation from rectum to cecum only (one-way) and navigation from rectum to cecum and vice versa (two-way). Sensitivity and positive predictive value were calculated on both a per-polyp basis and a per-patient basis. Alternative free-response receiver operating characteristic (ROC) curve analysis was estimated, and image interpretation time was documented.
RESULTS. One hundred fifty-five polyps were depicted in 45 patients by colonoscopy. The mean sensitivity of CT colonography for polyp detection with two-way (88.4%) was significantly superior to that with one-way (78.1%) (p < 0.01). The mean positive predictive value of each observer with one-way was 66.7%, whereas that with two-way was 65.8%. The mean area under the alternative free-response ROC curve (Az value) with two-way (0.827) was higher than that with one-way (0.816), but there was not a statistically significant difference. The average interpretation time of each observer with two-way (39 min) was statistically significantly longer than that with one-way (25 min) (p < 0.01).
CONCLUSION. When using a primary 3D interpretation technique at CT colonography, complete 3D navigation from rectum to cecum and from cecum to rectum is mandatory to maximize polyp detection. The image interpretation time for two-way interpretation is statistically significantly longer than that with one-way interpretation.
Keywords: colorectal cancer • CT colonography • MDCT • 3D navigation
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To our knowledge, no study to date has formally compared a one-way (rectum to cecum) interpretation with a two-way (rectum to cecum and vice versa) interpretation technique. Thus, we undertook this study, using MDCT colonography, to assess the value of two-way interpretation compared with one-way interpretation in terms of polyp detection and interpretation time.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Inclusion criteria were screening of a patient at average risk for colorectal cancer; a personal or family history of colorectal polyps; a family history of colorectal cancer; follow-up of an abnormal screening test (positive fecal occult blood test or barium enema); or evaluation of iron deficiency anemia, hematochezia, abdominal pain, or weight loss. Exclusion criteria included history of familial adenomatous polyposis or hereditary nonpolyposis cancer syndromes; prior colorectal surgery; suspected diagnosis of inflammatory bowel disease, bowel obstruction, or acute diverticulitis; a medical condition that precluded the use of bowel preparation; rejection for CT colonography or optimal colonoscopy for any reason; and pregnancy.
Twenty-four hours before examination, all patients underwent standard bowel preparation using magnesium citrate. In addition, 6 hr before examination, they received polyethylene glycol solution diluted in 2 L of water. To reduce bowel peristalsis and colonic spasm, 20 mg of scopolamine butylbromide (Buscopan, Boehringer Ingelheim) was routinely administered intramuscularly to patients immediately before air insufflation, and if contraindicated (e.g., for prostate hyperplasia, ischemic heart disease, or glaucoma), 1 mg of glucagon (Glucagon G Novo, Novo Nordisk) was intramuscularly administered.
CT Colonographic Technique
Patients were placed in the right lateral decubitus position on the CT
table, a balloon-tipped rectal tube was inserted, and room air was gently
insufflated into the colon to patient tolerance. A standard CT scout image was
obtained with the patient in the supine position to assess the degree of colon
distention. All patients underwent scanning in the supine position, and a
single breath-hold acquisition was used to obtain images from the top of the
colon through the rectum, as determined from the scout image. A prone
acquisition was not performed to reduce patient exposure to ionizing
radiation. CT was performed using an 8-MDCT scanner (LightSpeed Ultra, GE
Healthcare). The following scanning parameters for CT colonography were used:
120 kV; 120 mA; rotation time, 0.7 sec; collimation, 8 x 1.25 mm;
helical pitch, 13.75; and table speed, 13.5 mm per rotation. Transverse images
were reconstructed at 1.25-mm slice thickness and 0.62-mm section overlap. The
matrix size of the resulting transverse images was 512 x 512.
Conventional Colonoscopy
Conventional colonoscopy was performed with a videotape colonoscope
(Olympus 140, Olympus Optical) by three gastroenterologists with comparable
colonoscopic experience (1,000 colonoscopies before this study) without
knowledge of the CT colonography results. The presence of polyps was confirmed
by the colonoscopic findings, which served as the reference standard.
The instrument was inserted to the cecum and sequentially withdrawn segment by segment for the detection of lesions. The location and size of all colorectal polyps were documented. Lesions were photographed, and size was estimated with the use of a fully open biopsy forceps pushed against the polyp. All the examinations were performed while the patients were under conscious sedation with IV administration of midazolam (Versed, Hoffmann-La Roche).
Image Analysis
CT data were transferred to a workstation (Advantage Windows 4.0, GE
Healthcare) equipped with software that has surface-rendering capabilities.
Three-dimensional evaluation required an automated midline trace before
evaluation, and in some instances in which the trace was not accurate (e.g.,
bowel segments were skipped or the trace went outside the lumen), we made
revision manually. Three radiologists independently examined CT data sets at
the workstation. Each observer had training in the interpretation of CT
colonography and had analyzed approximately 200 CT colonographic examinations
with endoscopic correlation before this study. CT colonographic images were
examined without knowledge of patient history, including whether the patient
had been recruited for screening or for evaluation of symptoms, and of the
results of conventional colonoscopy.
Transverse, coronal, and sagittal CT images and a 3D navigation tool were simultaneously available on the workstation monitor. Therefore, both 2D and 3D images were available for analysis. However, observers were asked to use a primary 3D interpretation in which 3D images were analyzed and 2D transverse and multiplanar reconstructions were used to confirm the findings. All 50 cases were analyzed twice using a one-way (i.e., from rectum to cecum) 3D navigation and a two-way 3D navigation (i.e., from rectum to cecum and vice versa). Overall, two interpretation sessions were performed. In each session, 25 cases were interpreted using a one-way 3D navigation and 25 using a two-way 3D navigation. The two interpretation sessions were performed with a time interval of 4 weeks to minimize memory bias.
Confidence in polyp detection was rated on a 5-point grade scale: 4, polyp definitely present; 3, probably present; 2, possibly present; 1, probably not present; or 0, not present. Image interpretation time for each interpretation was documented with a stop-watch for both one-way and two-way interpretations.
Statistical Analysis
Alternative free-response receiver operating characteristic (ROC) curve
analysis was performed. Although the conventional ROC method allows only one
response per image, the alternative free-response ROC method allows an
observer response for all the polyps present
[10,
11], and we analyzed all 155
polyps in this study. An alternative free-response ROC curve was fitted to
each observer's confidence rating using a maximum-likelihood estimation
(ROCKIT, version 0.9B, C. E. Metz). Diagnostic ability was evaluated with area
under the alternative free-response ROC curve (Az).
Sensitivity and positive predictive value for both one-way (rectum to cecum only) and two-way (rectum to cecum and vice versa) interpretations were also calculated. For this evaluation, the polyps that had been assigned a confidence level of 3 or higher among the 155 proven polyps were considered true-positive findings; the polyps that were not detected or that were assigned a confidence level of 2 or lower, although a lesion was actually present, were considered false-negative findings. The lesions that were not depicted with conventional colonoscopy and had been assigned a confidence level of 3 or higher at CT colonography were considered false-positive findings.
For Az values, sensitivities, positive predictive values, and image interpretation time, the statistical significance of any differences was assessed using the Student's t test for paired data. A two-tailed p value of less than 0.05 was considered significant.
To assess interobserver variability in image interpretation, the nonweighted kappa statistic with binary data was used to measure the extent of agreement among the three observers. The binary data were calculated from the confidence rating value assigned to a true lesion by each of the observers. The extent of disagreement was not factored into the calculation. Kappa values larger than zero were considered to indicate a positive correlation; values up to 0.40, positive but poor agreement; values of 0.41-0.75, good agreement; and values greater than 0.75, excellent agreement.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The Az values calculated for each observer are shown in Table 1. The mean Az values were 0.816 for polyp detection using one-way 3D interpretation and 0.827 using two-way 3D interpretation. With regard to detection, there were no statistically significant differences in the Az values among the three observers.
|
Table 2 shows the sensitivity of CT colonography for polyp detection in each size category. Two-way interpretation showed significantly higher detection sensitivity for polyps in each size group (p < 0.01). With two-way interpretation, the average sensitivity of CT colonography for polyp detection was higher for each polyp size group and each observer than those values obtained using one-way interpretation (Figs. 1A and 1B).
|
|
|
Among the polyps missed by one-way evaluation and detected by two-way evaluation, 87.5% (total, 42/48: observer A, 15/17; observer B, 14/17; observer C, 13/14) were located on a blind spot (i.e., regions of mucosa hidden behind colonic folds), and both observer A and observer B missed all 13 polyps that observer C missed. Of the missed polyps, 12.5% (total, 6/48: observer A, 2/17; observer B, 3/17; observer C, 1/14) were due to oversight with one-way interpretation. Among the false-negative lesions missed by both one-way and two-way evaluations, 57.9% (average, 11/19: observer A, 10/15; observer B, 12/26; observer C, 11/16) were due to nondistended segments or retained fluid, and all of the missed polyps were 5 mm or smaller in diameter.
Table 3 shows the positive predictive value of each observer. A two-tailed p value for the average positive predictive value between one-way interpretation and two-way interpretation was over 0.05, calculated with the Student's t test. The use of two-way navigation compared with one-way navigation did not improve the positive predictive value of CT colonography for polyp detection in all observers.
|
The kappa values for the three observers, calculated on the basis of each observer's confidence level for the alternative free-response ROC analysis, were 0.50 for detection with one-way interpretation and 0.47 for detection with two-way interpretation. The three observers had good agreement for both one- and two-way interpretations.
The mean image interpretation times are shown in Table 4. The average times of two-way interpretation for each of the three observers were statistically significantly longer than those with one-way interpretation (p < 0.01).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our results also show that 3D interpretation yielded good interobserver agreement. This finding supports the reproducibility of our results and is in agreement with the results of Pickhardt et al. [7] and McFarland et al. [12] but differs from the results of Johnson et al. [13]. It is likely that thin-section MDCT acquisition might have minimized interobserver variability in our study and that ultra-thin reconstructions might have brought this study a potential advantage. The use of thin sections improves lesion characterization, as shown by Lui et al. [14].
Although not statistically significant, a trend toward a decrease in positive predictive value also was noted when using two-way image analysis. This is related to the fact that an increased number of false-positive findings were present when using two-way interpretation (Table 3). A possible explanation for this result is that diminutive residual fecal material identified on 3D (and thus believed to represent a polyp) could not be correctly characterized even when using 2D images. Other methods, such as fecal tagging or scanning the patient in the prone position can reportedly reduce false-positive findings on CT colonography [15-17]. Thus, by combining our two-way technique with those methods, we found that high polyp detection sensitivity and low false-positive findings could be achieved. We think that those attempts to decrease false-positive findings are important to avoid unnecessary optical colonoscopies, and we believe that further studies are needed.
Despite the increase in polyp detection, two-way interpretation also is correlated with a statistically significant increase in image interpretation time. One potential limitation of this method can be its time-consuming nature. Indeed, as a potential screening technique, CT colonography must be performed in a clinically feasible amount of time [18]. In our study, the average interpretation time for two-way interpretation was 39 min, whereas the average interpretation time for one-way interpretation was 25 min. However, further refinements of software packages for CT colonography may render 3D image analysis faster in the near future. For example, one of several new display methods, virtual colon dissection, allows observation of colon lumen with a spread view, which may decrease interpretation time; however, Hoppe et al. [19] recently reported that the virtual colon dissection method is more time-consuming than axial interpretation and its detection rate is not superior to axial interpretation. In terms of polyp detection rate, results similar to ours have been reported using primary 2D interpretation with faster interpretation times [2, 3]. Additional studies are needed to specifically compare a primary 3D technique with a primary 2D technique.
Our study has several limitations. First, in an attempt to reduce patient exposure to ionizing radiation, CT colonography was performed with supine acquisition. The importance of obtaining both supine and prone acquisitions has been previously emphasized [16, 17]. However, the main purpose of our study was to show superiority of two-way interpretation to one-way interpretation, and we thought that the results of our data could reveal it. We think that the lack of prone images did not affect the results of our data much.
Another potential criticism of our study is that because conventional colonoscopy has a miss rate for polyp detection of up to 24% [20], some false-positive findings of CT colonography could be false-negative findings of conventional colonoscopy.
In addition, because of substantial differences among the software packages currently available for CT colonography, especially in terms of 3D capabilities [21], the reproducibility of our results using primary 3D interpretation with a software package different from that used in the present study cannot be guaranteed.
Another potential criticism of our research could be related to the
grouping of polyps in the following size groups: 
5 mm, 6-9 mm, and 
10 mm. Indeed, although this classification has been extensively adopted both
in the conventional colonoscopic
[19] and CT colonographic
[1-4,
22-25]
literature, grouping polyps in the 6- to 9-mm range can potentially obscure
important data [26]. However,
because the purpose of our study was to investigate different techniques for
3D interpretation, we believe that this cannot be construed as a major
limitation to our research.
Finally, we evaluated the performance of CT colonography in a patient population with a high prevalence of disease. Therefore, although observers in our study were blinded to indications and findings of conventional colonoscopy, further studies are needed to confirm our results in a patient population with a low prevalence of disease.
In conclusion, our study indicates that complete 3D navigation from rectum to cecum and from cecum to rectum is mandatory to maximize polyp detection at CT colonography when using a primary 3D interpretation technique.
|
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. 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]
|
|
|
|
|
|
P. J. Pickhardt, S. A. Taylor, and S. Halligan Polyp Detection at CT Colonography: Inadequate Primary 3D Endoluminal Reference Standard Precludes Meaningful Comparison Radiology, July 1, 2007; 244(1): 316 - 317. [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. WERDERMAN CT Colonography Radiol. Technol., March 1, 2007; 78(4): 309CT - 324CT. [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
