HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Johnson, C. D. Articles by Dachman, A. H. Search for Related Content PubMed PubMed Citation Articles by Johnson, C. D. Articles by Dachman, A. H.

CT Colonography: The Next Colon Screening Examination?¹

¹ From the Department of Radiology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (C.D.J.), and the Department of Radiology, the University of Chicago, Ill (A.H.D.). Received July 2, 1999; revision requested August 9; revision received October 19; accepted November 10. Address correspondence to C.D.J.

	ABSTRACT

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
Computed tomographic (CT) colonography is a new-generation technique for detecting colorectal neoplasms by using volumetric CT data combined with specialized imaging software. Patient examinations require full colonic preparation, insufflation, and data acquisition with the patient in the supine and prone positions. Current CT technology allows a single image of the colon to be acquired in as little as 20 seconds with a minimum of patient discomfort. Specialized computer software for interpretation usually combines transverse, multiplanar reformation, and three-dimensional endoluminal images for the optimal visualization of the colon and rectum. As of the time this article was written, CT colonography was competitive as a full structural colonic examination for the detection of polyps and cancer. To the authors’ knowledge, no study results have yet been reported in a screening population. The unique capabilities of CT colonography include the display of the proximal colon that is inaccessible at colonoscopy because of obstructing colonic lesions or because of incomplete endoscopic examinations and the assessment of extracolonic abdominal and pelvic organs. This abdominopelvic survey potential provides radiologists with an opportunity to discover other potentially life-threatening, asymptomatic conditions. Further technologic developments and validation studies are in progress. CT colonography is an exciting and promising technique with an enormous potential for colorectal screening in the future.

Index terms: Colon, abnormalities, 75.311, 75.32 • Colon, CT, 75.12117 • Colon, diseases, 75.311, 75.32

	INTRODUCTION

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
Colorectal cancer screening in the United States is being promoted by major medical organizations, which include the American Cancer Society (1,2), and now is being reimbursed by Medicare for eligible recipients. These new public health policies are based on the widely accepted opinion that most colorectal cancers arise from preexistent adenomatous polyps. The detection and removal of these precursor adenomas will result in a decrease in the incidence of and in the mortality rate of colorectal cancer (3–11). Patients and physicians have a broad range of screening choices available, which include fecal occult blood testing, flexible sigmoidoscopy, barium enema radiography, and colonoscopy. The investigation of promising new-generation colorectal screening examinations is indicated, since, in our opinion, none of the currently available examinations are optimal in terms of performance, safety, or patient acceptance.

Few radiologists conceived of detecting colorectal polyps by using computed tomography (CT) when this technology was first introduced into practice more than 20 years ago. Major advances in CT technology have facilitated the rapid acquisition of thin sections and volumes of body tissue. New-generation multichannel CT scanners currently can be used to acquire all of the data for an abdominopelvic examination in a single breath hold. Advances in software and computer technology enable unique image displays of the colon in ways that previously were not possible. A comprehensive structural colon examination now is possible with less than 1 minute of scanning. This rapid acquisition translates into an easier and more comfortable examination than other available structural examinations. The purpose of this article is to review the current status of CT colonography, to compare it with other colorectal screening examinations, and to provide insight into its future.

	CT COLONOGRAPHY: DEFINITION

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
In 1994, the use of volumetric CT data produced by a spiral CT scanner, "virtual colonoscopy," or CT colonography, was reported by Vining and Gelfand (12). The three-dimensional (3D), endoluminal-perspective images displayed in a cine loop, which simulated the endoluminal views at colonoscopy, excited many investigators in the field of gastrointestinal imaging. In 1995, Coin (13) was granted a patent for imaging the colon by creating cross sections orthonormal to the longitudinal axis of the colon.

After several years of development, CT colonography today refers to a CT examination of the fully prepared and air-distended colon. Volumetric CT data in the entire colon are acquired with only a few seconds to minutes of scanning and with a total of 15 minutes or less of examination time. By combining these data with advanced imaging software, the colon is examined at an off-line workstation by using the combination of two-dimensional (2D) and 3D images. The 2D and 3D images are complementary and are both required for the thorough assessment of the colon and rectum.

Terms used previously to describe this technique, such as virtual colonoscopy, virtual endoscopy, 3D endoscopy, and CT colography, have been abandoned in favor of CT colonography (14).

	PATIENT PREPARATION AND EXAMINATION

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
Thorough bowel preparation is ideal for an accurate examination. Residual stool presents many of the same problems encountered at barium enema radiography. Since most centers have performed CT colonography in conjunction with a research study, with colonoscopy serving as the standard of reference, a majority of patients have been prepared with polyethylene glycol electrolyte solution (Go-Lytely; Braintree Laboratories, Braintree, Mass). The colonic lavage preparation often results in excess residual fluid in the colon. In theory, any preparation that results in a clean colon will suffice. A common barium enema preparation (magnesium citrate, bisacodyl tablets, cleansing enemas, or suppositories) resulting in a colon with less residual fluid might be preferred, but our anecdotal experience indicates that the colon is not cleansed as well. Study of the optimal preparation for this examination is needed.

The use of spasmolytic agents is controversial. Many investigators today use glucagon to prevent unwanted collapse and spasm, a problem encountered most commonly in the sigmoid colon. One milligram of glucagon can be administered either subcutaneously 10 minutes prior to the examination or intravenously just prior to colon insufflation. Yee et al (15) have reported no beneficial effect from routine glucagon administration. As a result of its effect on the ileocecal valve, glucagon can cause the unwanted reflux of air into the small bowel and can secondarily reduce colonic distention.

Colon insufflation is performed with the patient in a lateral decubitus position after an enema tip or a catheter (with or without a retention cuff) is placed in the rectum. By using either room air or carbon dioxide, the colon is distended fully, without causing undue patient discomfort. The end point for insufflation may be a set volume of gas or patient tolerance. Most patients will retain 1.5–2.0 L. Supine and prone images are obtained at most centers. A single acquisition takes three to four 20-second breath holds (60–80 seconds total of CT scanning) with a single-section helical scanner. As an alternative, clinicians at some centers scan continuously for 60 seconds; patients are asked to breathe either quietly or slowly after their best breath hold. Multisection scanners can be used to acquire the entire data set in a 20-second breath hold. This rapid examination without the use of sedation, intervention, or compression is well tolerated and is assessed by patients to be more comfortable when compared with other full colonic examinations such as barium enema radiography or colonoscopy (16). Patients appreciate that the interrogation of the colon is performed on the CT data at the physician’s workstation and not on them, which spares them time and discomfort. Patients can return to work immediately after the procedure, and to our knowledge no complications had been reported at the time this article was written.

	TECHNICAL ISSUES

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
CT Scanning Parameters
Hara et al (17) reported on the first study that we know of in which scanning parameters were evaluated. By using a phantom composed of vitamin E capsules and Tygon tubing, the collimation (3–10 mm) and reconstruction intervals (1–10 mm) were varied by using a pitch of 1. The best image quality was obtained by using a collimation of 5 mm, a pitch of 1, and a reconstruction interval of 1 mm. In a subsequent publication by investigators in the same laboratory (18), 100% sensitivity for the detection of polyps 5 mm or larger by using 5-mm collimation, a 3-mm reconstruction interval, and a pitch of 1.3 was reported. This protocol had the desired effect of reducing the amount of acquired data by 60% without affecting diagnostic performance.

Beaulieu et al (19) studied proper scanning collimation and pitch by using a sophisticated phantom and various-sized beads that simulated polyps. All 4-mm or larger "polyps" could be detected by using 5-mm collimation, with a pitch of up to 2. Geometric distortion of the polyp developed with increasing collimation (the major factor) and with increasing pitch. The theoretic advantages of thinner collimation include improved surface feature delineation and depiction of the junction between the lesion and the colon wall. These features may be helpful in characterizing some lesions.

Dachman et al (20) studied simulated polyps in a pig colon and created polyps by puckering the mucosa of a normal colon and suturing its base. All of the 3-, 7- and 10-mm polyps were detected by using a collimation of either 5 or 7 mm, with a pitch of 1–2.

Hara et al (18) reported that the radiation dose can be reduced substantially compared with that of standard body CT settings because of the high contrast between the air-filled colon and the soft-tissue-attenuation colon wall. They reported no change in the diagnostic efficacy between CT settings of 140 mA and 70 mA. By using the lower-dose setting of 70 mA (5-mm collimation, 3-mm reconstruction interval, and a 1.3 pitch), the effective dose equivalent for the acquisition of a single supine image was 1.87 mGy for men and was 2.85 mGy for women. This dose would double by performing both prone and supine scanning. The dose from combined supine and prone scanning is approximately 20% lower than the dose at double-contrast barium enema radiography by using 5 minutes of fluoroscopy and acquisition of seven spot radiographs and five overhead radiographs (4.53 mGy for men and 7.45 mGy for women).

Several investigators have studied the use of prone imaging in addition to routine supine imaging. Chen et al (21) reported that better distention of the colon was achieved in 28% of patients by using prone and supine imaging (in 20 [87%] of 23 patients with adequate distention) when compared with either supine or prone imaging alone (27 [59%] of 46 patients). Fletcher et al (22,23) reported that substantially more patients with polyps of 5 mm or greater received correct diagnoses with prone and supine imaging (sensitivity, 88% [114 of 130]) than with supine imaging alone (sensitivity, 75% [97 of 130]). Similar clinically important findings favoring prone and supine scanning also were demonstrated for polyps 1 cm or greater. Most experts now agree that the added yield of polyps detected by using prone imaging is worth the extra radiation dose and time (Figs 1, 2) (22). Clinicians at some centers scan only the pelvis with the patient in a prone position, while others review the CT scan online and perform repeat scanning in problematic areas with the patient in the position best suited to that colon segment (Hopper K, oral communication, 1999).

View larger version (124K): [in this window] [in a new window] [Download PPT slide]   Figure 1a. Prone and supine reformation transverse 2D CT colonographic images. (a) Supine image. The descending colon (curved arrows) is collapsed, and no lesions are visible in this segment. (b) Prone image. The same colonic segment as in a is identified; a proved polyp (arrow) is seen on the nondependent wall of this now air-distended colonic segment. Prone and supine imaging can be very helpful to optimally depict segments that were collapsed or that were fluid filled.

View larger version (97K): [in this window] [in a new window] [Download PPT slide]   Figure 1b. Prone and supine reformation transverse 2D CT colonographic images. (a) Supine image. The descending colon (curved arrows) is collapsed, and no lesions are visible in this segment. (b) Prone image. The same colonic segment as in a is identified; a proved polyp (arrow) is seen on the nondependent wall of this now air-distended colonic segment. Prone and supine imaging can be very helpful to optimally depict segments that were collapsed or that were fluid filled.

View larger version (107K): [in this window] [in a new window] [Download PPT slide]   Figure 2a. A pedunculated polyp, colonoscopically proved, is shown by using prone and supine CT colonography. (a) Supine transverse image. A pedunculated polyp in the mid sigmoid colon can be identified. The attachment of the polyp to the colon wall (black arrow), the pedicle of the polyp (curved white arrow), and the polyp head (straight white arrow) are identified. (b) Prone transverse image. The stalk (arrow) and attachment of the polyp to the colonic wall are well demonstrated. The head of the polyp lies adjacent to a haustral fold and might be overlooked without the associated pedicle. Prone and supine images are complementary and can aid in lesion detection and improved confidence in a diagnosis.

View larger version (90K): [in this window] [in a new window] [Download PPT slide]   Figure 2b. A pedunculated polyp, colonoscopically proved, is shown by using prone and supine CT colonography. (a) Supine transverse image. A pedunculated polyp in the mid sigmoid colon can be identified. The attachment of the polyp to the colon wall (black arrow), the pedicle of the polyp (curved white arrow), and the polyp head (straight white arrow) are identified. (b) Prone transverse image. The stalk (arrow) and attachment of the polyp to the colonic wall are well demonstrated. The head of the polyp lies adjacent to a haustral fold and might be overlooked without the associated pedicle. Prone and supine images are complementary and can aid in lesion detection and improved confidence in a diagnosis.

Three-dimensional Image Display
The role of 3D images in interpreting CT colonographic scans varies widely. Although most investigators rely on 3D endoluminal images to confirm the presence of a lesion and to improve diagnostic confidence, few use it as the primary method of evaluating the colon.

Two 3D rendering techniques exist: surface rendering and volume rendering. Both techniques have been implemented successfully for CT colonography and can be used to display the colon anatomy accurately (24). Surface rendering is based on a preprocessing step that identifies isointense surfaces from an endoluminal perspective and that reduces the data to a set of surface triangles. Data that are deep in relation to the identified surface are discarded. Since the quantity of data that remains is reduced markedly, computations can be performed quickly. Volume rendering in its traditional form does not discard any of the information within the volume of interest. Extraluminal soft tissues and attenuation data (opacity mapping) can be displayed with volume rendering. Because of the large amount of data that must be managed by the computer, volume rendering computationally is more demanding, expensive, and time-consuming. Advances in computer speed and lower costs have allowed many medical centers and CT manufacturers to adopt volume rendering for their workstations that are equipped with CT colonographic software. Although the theoretic advantages favor volume rendering, to our knowledge there are no findings to suggest that one method is superior to the other diagnostically.

The use of volume rendering requires the optimization of several variables. McFarland et al (25) examined the effect of processing parameters on perspective volume-rendered images. Reconstruction intervals with an overlap of 60% or greater resulted in images that were equivalent to those obtained with equal voxel dimensions (87%–90% overlap). Transition zones occurring between areas of complete transparency and complete opacity that are positioned at values below -700 HU resulted in anatomic distortion. Opacity functions mapped with a sigmoidal function curve versus a step opacity function resulted in preferred images that were smoother, without a substantial loss in edge sharpness.

Reed and Johnson (26) and Reed (27) have reported several methods of optimizing and improving volume-rendering speed at CT colonography without sacrificing image quality. A "hyperlist" technique renders voxels to a limited depth beyond the mucosal surface, generally 12 voxels, with the effect of reducing the data set markedly. This technique provides many of the advantages of traditional volume rendering but allows the interactive inspection of the data with much less expensive computers.

The use of 3D images for the primary evaluation of the entire colon requires the user to navigate through the lumen of the colon. Because of the tortuous nature of the colon, the definition of a colon centerline ahead of time is a helpful preliminary step that can reduce interpretation time markedly. This centerline acts as a virtual path for the endoluminal camera and allows the observer to traverse the colon while concentrating on the endoluminal anatomy and disease rather than on navigation. Centerline determination can be made either manually or semiautomatically. Methods are being developed for automatic midline determination (28–33). To our knowledge, none of these methods has been evaluated in a large-scale clinical trial, and their future role is uncertain, since, as far as we know, complete 3D endoluminal assessment of the colon no longer is performed by clinicians in most centers.

Other unique displays of 3D colon data are under investigation at several academic centers. Reed and Johnson (34) published the first description of these techniques, which was followed by reports from Paik et al (35) and others (36–40). Most of this work is aimed at reducing diagnostic interpretation time and at improving the amount of colon visualized. Several methods can be used to display large segments of the colon in a single image that can be inspected rapidly. Other methods can be used to display the colonic mucosa either as a flat surface or as similar to a Mercator map. A number of terms have been applied to each unique imaging method, including virtual pathology (Fig 3), Mercator map projection, and panoramic projection. Thompson et al (36), Fletcher et al (37,38), and Beaulieu et al (39,40) have reported on the feasibility of these techniques, but to our knowledge no large clinical series have been reported.

View larger version (154K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Virtual pathology. CT colonographic transverse longitudinal image displays the anterior (top of image) and posterior (bottom of image) surfaces of the colon as though the colon were opened along its longitudinal axis. A pedunculated polyp (arrow) is clearly visible.

View larger version (134K): [in this window] [in a new window] [Download PPT slide]   Figure 4a. Coarsened mucosal texture pattern on 3D endoluminal CT transverse longitudinal colonographic images because of image noise. Identical rectal folds are present on (a) supine and (b) prone images. a was acquired by using a setting of 50 mA, while b was acquired by using a setting of 20 mA on a multichannel CT scanner. All other imaging parameters were identical between images. The textured mucosal surface in b is due to additional image noise inherent in low-dose images.

View larger version (133K): [in this window] [in a new window] [Download PPT slide]   Figure 4b. Coarsened mucosal texture pattern on 3D endoluminal CT transverse longitudinal colonographic images because of image noise. Identical rectal folds are present on (a) supine and (b) prone images. a was acquired by using a setting of 50 mA, while b was acquired by using a setting of 20 mA on a multichannel CT scanner. All other imaging parameters were identical between images. The textured mucosal surface in b is due to additional image noise inherent in low-dose images.

Two-dimensional Image Display
Two-dimensional transverse images in the abdomen are obtained routinely as part of every CT examination. Radiologists are very familiar with this type of image display and interpretation. In the early development of CT colonography, two assumptions regarding image display techniques were made that have since been shown to be erroneous. The first assumption was that polyps would be detected more accurately by using the 3D endoluminal images, since this image display simulated the colonoscopic examination, the standard of reference for colonic imaging. The second assumption was that the large amount of CT data that would be generated with this examination could be reviewed more efficiently by using the 3D endoluminal "fly-through" approach. It has now been shown clearly that polyp detection can be just as effective by using 2D images as by using 3D endoluminal images, but each is complementary to the other. The 3D endoluminal fly through of the colon often is a tedious and time-consuming approach that requires more time than the 2D approach.

The general consensus among most researchers indicates that 2D images, either traditional transverse images in the body or optimized 2D images (zoomed or perpendicular to the colon lumen), are adequate for the primary evaluation of the colon. Three-dimensional images are helpful to confirm or improve the observer’s confidence that a lesion is present and to display complex anatomic relationships (complex or bulbous folds, the ileocecal valve region, or the inner contour of a flexure).

Dachman et al (42) studied a method of initially interpreting the transverse images, then using 3D endoluminal-perspective imaging as a limited problem-solving procedure to evaluate the colon. The sensitivity and specificity of this technique for the detection of polyps 8 mm or greater were 83% (five of six polyps) and 100% (six of six polyps), respectively. This approach also is preferred by the Mayo Clinic laboratory (43).

Two-dimensional techniques must be optimized for accurate colon evaluation. It is imperative that the 2D images are reviewed by using lung window settings. Small polyps cannot be detected if soft-tissue window settings are used. As a corollary, soft-tissue windows must routinely be used, especially in regions of suspected colonic collapse. Colon cancers that narrow the lumen may appear as regions of collapse with the lung window settings. Thickening of the colon wall (flat lesions, lesions that thicken folds, and some cancers) and pericolonic soft-tissue stranding also are easier to detect by using soft-tissue window settings, which can also be used to look for incidental extracolonic abnormalities (Fig 5). The internal fat attenuation of lipomas is visible clearly on 2D images and is appreciated most easily by using soft-tissue window settings (Fig 6).

View larger version (69K): [in this window] [in a new window] [Download PPT slide]   Figure 5a. Transverse longitudinal CT colonographic images show complementary wide and soft-tissue window settings for detecting colorectal neoplasms. (a) Wide window settings are optimal for detecting small polyps, but circumferential tumors (arrows) on this transverse image could be misinterpreted as a region of bowel collapse. (b) Soft-tissue window settings are helpful in optimally displaying the bowel wall thickening (arrows) on this transverse image and in detecting other extraluminal abnormalities.

View larger version (93K): [in this window] [in a new window] [Download PPT slide]   Figure 5b. Transverse longitudinal CT colonographic images show complementary wide and soft-tissue window settings for detecting colorectal neoplasms. (a) Wide window settings are optimal for detecting small polyps, but circumferential tumors (arrows) on this transverse image could be misinterpreted as a region of bowel collapse. (b) Soft-tissue window settings are helpful in optimally displaying the bowel wall thickening (arrows) on this transverse image and in detecting other extraluminal abnormalities.

View larger version (119K): [in this window] [in a new window] [Download PPT slide]   Figure 6a. CT colonographic images show colonic lipoma, colonoscopically proved. (a) Three-dimensional endoluminal image shows a nearly obstructing intraluminal mass (arrows). The surface characteristics of the mass are not helpful for characterization. (b) The fat internal attenuation (arrow) of a lipoma is visible on the 2D reformation image (Reprinted, with permission, from reference 44.) A lipoma has homogeneous fat attenuation and is best evaluated by using soft-tissue window settings.

View larger version (116K): [in this window] [in a new window] [Download PPT slide]   Figure 6b. CT colonographic images show colonic lipoma, colonoscopically proved. (a) Three-dimensional endoluminal image shows a nearly obstructing intraluminal mass (arrows). The surface characteristics of the mass are not helpful for characterization. (b) The fat internal attenuation (arrow) of a lipoma is visible on the 2D reformation image (Reprinted, with permission, from reference 44.) A lipoma has homogeneous fat attenuation and is best evaluated by using soft-tissue window settings.

Two-dimensional images (combined traditional transverse and multiplanar reformation images) of the colon are the "display engines" of CT colonography today. Used with proper window settings and linked interactively with optimized transverse views, they are powerful methods of quickly examining the colon for abnormalities. Complementary 3D endoluminal views in regions of concern can be used to improve lesion detection and characterization.

Two- versus Three-dimensional Display
The complementary nature of 2D and 3D image displays was predicted by Hara et al (45) by using a phantom model constructed from the data set of a patient with a normal sigmoid colon that contained 11 computer-simulated polyps 1–10 mm in diameter. Blinded readers reviewed the same data sets by viewing only 2D images, only 3D images, and a display combining 2D and 3D images. The best polyp detection occurred with the combined 2D and 3D image display. All polyps greater than 2 mm were identified. The consensus opinion of experts today confirms this finding.

Hardware and Software
Most computer platforms without the software required for CT colonography can be obtained for $8,000–$20,000. CT manufacturers all have software for creating 3D endoluminal and multiplanar reformatted views. Independent software companies (eg, Vital Images, St Paul, Minn) also have applications for performing the advanced imaging required for CT colonography. Several academic centers have developed their own CT colonographic software or are working in conjunction with a commercial vendor. Each type of software varies in the method it implements: image display options and features, interactivity, and time required from the radiologist. Buyers should familiarize themselves with the necessary components of a user-friendly workstation before deciding on a purchase. Most authorities believe that 2D and 3D images should be available seamlessly on a single interactive screen.

	PERFORMANCE

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
The performance data of CT colonography are detailed in the Table. It is accepted widely that, for screening purposes, 1 cm or larger should be the target lesion size (1).

The largest series at the time this article was written, 180 patients with 420 colonoscopically proved polyps, was reported on by Fletcher et al (22). These patients were known to have or were suspected to have colorectal neoplasms, except for 20 patients recruited from a surveillance population with a low prevalence of large polyps. By using both supine and prone patient positioning, the sensitivity and specificity for the identification of patients with polyps 1 cm or greater were 85% (82 of 96 patients) and 93% (78 of 84 patients), respectively. The sensitivity and specificity for the detection of polyps 5 mm or greater were 88% (114 of 130 patients) and 72% (36 of 50 patients), respectively.

Royster et al (49) reported a sensitivity of 100% (20 of 20 patients) in detecting large colorectal lesions 2 cm or greater. In addition, the value of CT colonography was demonstrated in patients with obstructing carcinomas (49,55,56). In many patients, lesions above a colonic obstruction were identified correctly by using CT colonography. These lesions were inaccessible at colonoscopy. Fenlon et al (56) and Barish et al (57) reported, in patients who underwent incomplete colonoscopy, that previously unexamined portions of the colon were visualized at CT colonography in more than 26 (90%) of 29 patients. Macari et al (53) also have confirmed the use of CT colonography after incomplete colonoscopy.

All of the patients from all of the published reports at the time this article was written were recruited from either symptomatic patients or those known to have or highly suspected to have lesions. To our knowledge, no investigators to date have reported on the performance of CT colonography in an asymptomatic screening population—the patient population most likely to benefit from this procedure. To our knowledge, all of the current studies arise from individual institutions. Several multicenter trials now are being considered, but none have been completed as far as we are aware. Investigators in nearly all of these studies have compared CT colonography with colonoscopy, but to our knowledge no investigators have compared CT colonography with the other colorectal screening examinations.

An acceptable performance statistic for CT colonography has not been published. As a full structural colorectal examination, it must compete with other total colon examinations, barium enema and colonoscopy. Other variables, including patient acceptance, cost, and safety, are important factors that also will influence referring physicians’ and patients’ decisions regarding their choice of a colorectal examination. To our knowledge, no investigators at the time this article was written had evaluated the performance of barium enema radiography in a screening population. Steine et al (58) reported on 190 patients who were referred by general practitioners for double-contrast barium enema radiography and who were selected randomly to undergo colonoscopy. The sensitivity and specificity at double-contrast barium enema radiography were 81% and 96%, respectively. In a report that included guidelines for colorectal cancer screening, the American Gastroenterological Association estimated the sensitivity at double-contrast barium enema radiography for depicting polyps 1 cm or greater to be 70%–90%, with a false-positive rate of 5%–10% (1). It seems reasonable to expect CT colonography to at least match the performance of barium enema radiography if it is to be considered a feasible colon imaging modality.

Extraintestinal Findings
The ability to evaluate the extracolonic organs of the entire abdomen and pelvis, in addition to assessing the colon, is an important benefit inherent in CT colonography. No other colorectal screening examination has this use. Since colorectal screening targets older individuals (individuals with average risk who are over the age of 50 years), other abdominal diseases are likely to be encountered incidentally. The potential for saving many lives by detecting life-threatening lesions in organs outside the colon in the course of colon screening is real and is an exciting potential benefit.

Hara et al (59) reported the extracolonic findings in 264 patients who underwent CT colonography. Approximately half of all patients had some type of extracolonic abnormality detected at CT colonography (a nonenhanced, low-dose examination). Highly clinically important findings were reported in 11% of patients and included two incidental abdominal aortic aneurysms greater than 4 cm, two asymptomatic renal adenocarcinomas, and one inguinal hernia containing bowel; all were treated surgically as a result of these findings. Other findings such as indeterminate pulmonary nodules, adrenal masses, and pneumothorax prompted further work-up. The cost of treating these findings adds to the overall cost of the examination, but when compared with many additive costs associated with colonoscopy (room charges, sedation, biopsy, histopathologic examination), these are relatively modest when averaged among all patients. Although further study is needed, it seems ironic that more clinically important disease possibly will be discovered in the extracolonic organs than in the colon, in which the prevalence of polyps 1 cm or greater is estimated to be 4%–5% in the average-risk population.

Problems and Pitfalls
Major technical issues responsible for false-negative findings are fluid-filled segments of bowel due to the lavage bowel preparation and collapsed segments of bowel from colonic spasm. It is encouraging that technical problems related to retained fluid and incompletely distended segments of the colon often can be corrected with the addition of prone imaging (Fig 1). Prone imaging in combination with supine imaging will readily move colonic fluid and often will move retained stool into opposite parts of the colon. The added benefit of prone imaging was discussed earlier in the CT scanning parameters section. Substantially more lesions are diagnosed with prone imaging combined with supine imaging versus with supine imaging alone (Fig 2) (1,21).

False-positive findings generally are due to the misinterpretation of retained stool or colonic folds as polyps. Retained stool often contains incorporated air that can be recognized at CT as a heterogeneous filling defect on 2D images (Fig 7) (44,60). A lack of wall attachment and movement of the filling defect on supine and prone images also indicate the fecal nature of the suspected abnormality (60). Colorectal neoplasms have homogeneous soft-tissue attenuation without intratumoral air. In the future, oral stool markers administered 24–48 hours prior to CT may be used to alter the attenuation of stool and to improve the differentiation of stool from soft-tissue neoplasms.

View larger version (96K): [in this window] [in a new window] [Download PPT slide]   Figure 7a. CT colonographic images show retained colonic stool, which was confirmed at colonoscopy. (a) The 3D endoluminal image demonstrates a polypoid filling defect (arrow) adjacent to a colonic fold. This abnormality has the appearance of a polyp. (b) The internal attenuation of the lesion (arrow) on this 2D reformation image is heterogeneous. This heterogeneous appearance is due to air within the retained stool. A polyp should have homogeneous soft-tissue attenuation.

View larger version (111K): [in this window] [in a new window] [Download PPT slide]   Figure 7b. CT colonographic images show retained colonic stool, which was confirmed at colonoscopy. (a) The 3D endoluminal image demonstrates a polypoid filling defect (arrow) adjacent to a colonic fold. This abnormality has the appearance of a polyp. (b) The internal attenuation of the lesion (arrow) on this 2D reformation image is heterogeneous. This heterogeneous appearance is due to air within the retained stool. A polyp should have homogeneous soft-tissue attenuation.

View larger version (85K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Transverse CT colonographic image shows a large, proved polyp-simulating ileocecal valve. A large polypoid filling defect (long arrow) is seen within the ascending colon on this transverse image. In addition, a tiny polyp (short arrow) on the lateral wall of the ascending colon also is visible. At initial inspection, the larger lesion, a proved adenoma, was mistaken for a prominent ileocecal valve. The ascending colon should be inspected carefully for the ileocecal valve in each case. Other filling defects should be regarded as lesions until proved otherwise.

View larger version (125K): [in this window] [in a new window] [Download PPT slide]   Figure 9a. CT colonographic images show a large, proved polyp that simulates stool. (a) Two-dimensional, transverse supine image in the ascending colon. An irregularly surfaced filling defect (arrows) in the ascending colon simulates stool because of its heterogeneous internal attenuation. (b) The movable nature of the mass (curved arrow) is demonstrated on this transverse prone image. The ileocecal valve (straight arrow) also is visible. (c) The irregularly surfaced mass (black arrows) and the ileocecal valve (white arrow) are visible on the 3D endoluminal view. This filling defect was initially mistaken as mobile stool, but a villous neoplasm was discovered at colonoscopy. Large frondlike tumors can appear to be mobile on prone and supine images. In an otherwise clean colon, filling defects should be regarded as suspicious for neoplasm.

View larger version (107K): [in this window] [in a new window] [Download PPT slide]   Figure 9b. CT colonographic images show a large, proved polyp that simulates stool. (a) Two-dimensional, transverse supine image in the ascending colon. An irregularly surfaced filling defect (arrows) in the ascending colon simulates stool because of its heterogeneous internal attenuation. (b) The movable nature of the mass (curved arrow) is demonstrated on this transverse prone image. The ileocecal valve (straight arrow) also is visible. (c) The irregularly surfaced mass (black arrows) and the ileocecal valve (white arrow) are visible on the 3D endoluminal view. This filling defect was initially mistaken as mobile stool, but a villous neoplasm was discovered at colonoscopy. Large frondlike tumors can appear to be mobile on prone and supine images. In an otherwise clean colon, filling defects should be regarded as suspicious for neoplasm.

View larger version (105K): [in this window] [in a new window] [Download PPT slide]   Figure 9c. CT colonographic images show a large, proved polyp that simulates stool. (a) Two-dimensional, transverse supine image in the ascending colon. An irregularly surfaced filling defect (arrows) in the ascending colon simulates stool because of its heterogeneous internal attenuation. (b) The movable nature of the mass (curved arrow) is demonstrated on this transverse prone image. The ileocecal valve (straight arrow) also is visible. (c) The irregularly surfaced mass (black arrows) and the ileocecal valve (white arrow) are visible on the 3D endoluminal view. This filling defect was initially mistaken as mobile stool, but a villous neoplasm was discovered at colonoscopy. Large frondlike tumors can appear to be mobile on prone and supine images. In an otherwise clean colon, filling defects should be regarded as suspicious for neoplasm.

View larger version (98K): [in this window] [in a new window] [Download PPT slide]   Figure 10a. CT colonographic images show infiltrating adenocarcinoma, proved endoscopically, that manifests as thickened folds. (a) An isolated, thickened, irregularly surfaced haustral fold (arrows) in the transverse colon is identified on this transverse image. (b) The irregular surface contour of the fold (arrows) is confirmed on the 3D endoluminal image. The adjacent haustral folds were thin and straight. An isolated thickened and irregular haustral fold should be regarded as suspicious for infiltrating tumor.

View larger version (90K): [in this window] [in a new window] [Download PPT slide]   Figure 10b. CT colonographic images show infiltrating adenocarcinoma, proved endoscopically, that manifests as thickened folds. (a) An isolated, thickened, irregularly surfaced haustral fold (arrows) in the transverse colon is identified on this transverse image. (b) The irregular surface contour of the fold (arrows) is confirmed on the 3D endoluminal image. The adjacent haustral folds were thin and straight. An isolated thickened and irregular haustral fold should be regarded as suspicious for infiltrating tumor.

View larger version (95K): [in this window] [in a new window] [Download PPT slide]   Figure 11a. CT colonographic images show perceptive error. (a) Supine transverse image. A proved 6-mm polyp (arrow) in the splenic flexure is difficult to distinguish from other haustral folds. Optimal colonic distention and careful inspection of flexural regions of the colon can aid in their detection. (b) Three-dimensional endoluminal view is helpful in confirming the presence of the lesion (arrows) and in distinguishing it from nearby folds.

View larger version (170K): [in this window] [in a new window] [Download PPT slide]   Figure 11b. CT colonographic images show perceptive error. (a) Supine transverse image. A proved 6-mm polyp (arrow) in the splenic flexure is difficult to distinguish from other haustral folds. Optimal colonic distention and careful inspection of flexural regions of the colon can aid in their detection. (b) Three-dimensional endoluminal view is helpful in confirming the presence of the lesion (arrows) and in distinguishing it from nearby folds.

	COMPARISON WITH OTHER SCREENING EXAMINATIONS

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
CT colonography represents another colorectal screening option for physicians and patients. As a full structural colorectal imaging examination, it competes directly with barium enema radiography and colonoscopy. Its diagnostic potential appears to be much greater than that of fecal occult blood testing. More than half of all colorectal cancers and a large majority of polyps will be missed by using a single screening pass with fecal occult blood testing, and most positive test results are false-positive (64–66). Since the performance of CT colonography is already markedly better than that of fecal occult blood testing, better patient outcomes can be predicted, and screening can be less frequent.

Proctosigmoidoscopy is inherently flawed by its use in examining only the descending sigmoid colon and the rectum. As a full structural colorectal screening examination, CT colonography would be anticipated to yield substantially more neoplasms. An unresolved issue is whether an adequate evaluation of the rectum is possible at CT colonography by using a balloon retention cuff. In an ideal situation, the rectum would be scanned with the cuff deflated.

CT colonography competes most directly with screening barium enema radiography; to our knowledge, formal trials in which these two techniques are compared have not yet been completed. There are some inherent advantages of CT colonography over barium enema radiography. These advantages include having no requirement for coating the colon with barium, little effect from retained fluid if prone imaging is performed, a lack of complex overlapping radiographic lines (an important source of radiographic perceptive errors at barium enema radiography), the production of an infinite number of different projections of a colonic segment on demand, and no need for a second cathartic preparation prior to colonoscopy if a polyp is discovered. In addition, because the examination is very rapid, it is more comfortable and is better tolerated by patients (16). A comparison of interpretation time and cost between barium enema radiography and CT colonography is needed.

In comparison with colonoscopy, CT colonography has the potential advantages of depicting colonic anatomy from an endoluminal perspective and in multiple cross sections. Blind spots behind colonic folds are eliminated, as the entire colon is virtually always evaluated. The accurate characterization of many lesions (polyps, stool, and lipomas) is possible, without the risks associated with sedation and biopsy. Although to our knowledge a formal study on compliance with CT colonography has not been performed, in a survey of patients after CT colonography, barium enema radiography, and colonoscopy, patients were more willing to return for CT colonography at the recommended screening interval (every 3–5 years) than for the other two examinations (16). A major disadvantage of CT colonography compared with colonoscopy is that tissue cannot be retrieved.

Magnetic resonance (MR) colonography also has been explored recently and has many of the same theoretic advantages as CT colonography. MR colonography has the advantage of having no ionizing radiation and of direct imaging of the colon in nontransverse planes. A potential disadvantage of this technique is the required gadolinium-based enema that must be retained during the imaging procedure and the extra cost of contrast material and MR imaging equipment. Preliminary performance data at MR colonography are competitive with those at CT colonography (67–71), and the potential for fecal tagging with MR imaging contrast material also is possible (72).

The advantages of examining extracolonic tissues, a unique feature of CT colonography when compared with all other colorectal screening examinations, were discussed within the performance section of this article.

	FUTURE APPLICATIONS

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 
CT colonography represents an important and exciting new-generation technique for imaging in the colon and rectum. In theory, this technique has advantages over all existing colorectal screening examinations. Early results are promising, but rigorous scientific evaluation of this examination in comparison with other colorectal examinations in a screening population is needed (64). These studies now are under way, and comprehensive data should be available within 2 years.

We can expect evolving and better ways to display the imaging data and methods that enable more efficient interpretations (73). In addition, automated methods of displaying the colon and computerized assessment of the colon wall thickness may assist radiologic interpretation (74,75). It is possible for images to be interpreted accurately within only a few minutes, and in the future some of these may be read with computer-aided diagnosis. Examination cost has not yet been determined but will depend on both resource use and interpretation time. Costs likely will be lowered as efficiencies improve with technical advances. CT colonographic costs likely will be highly competitive with those of other structural colorectal screening examinations.

Preparation for colorectal screening is a major obstacle to patient compliance. The development of "prepless" CT colonography is possible, but to our knowledge only preliminary research had been performed at the time this article was written. In theory, patients would drink a contrast agent 1–2 days prior to examination. This agent would alter stool attenuation sufficiently so that it could be recognized and removed electronically. No purgation would be required. Multiple stool markers currently are under development, and clinical trials are likely to ensue within the next several years.

In summary, CT colonography is an exciting new-generation examination that offers improved performance and better patient compliance over current colorectal imaging examinations. CT colonography is most likely to be used for colorectal cancer screening, but clinical trials in which its effectiveness is evaluated in a screening population are needed. We hope that, with the continued development of this technique, patients with polyps and cancer will be identified and that, in conjunction with colonoscopy, the incidence and mortality of this common disease will be reduced substantially.

	FOOTNOTES

 
Abbreviations: 2D = two-dimensional, 3D = three-dimensional

	REFERENCES

TOP ABSTRACT INTRODUCTION CT COLONOGRAPHY: DEFINITION PATIENT PREPARATION AND... TECHNICAL ISSUES PERFORMANCE COMPARISON WITH OTHER SCREENING... FUTURE APPLICATIONS REFERENCES

 

1. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997; 112:594-642.[Medline]
2. Byers T, Levin B, Rothenberger D, et al. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update 1997. CA Cancer J Clin 1997; 47:154-160.[Abstract]
3. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med 1993; 328:1365-1371.[Abstract/Free Full Text]
4. Winawer SJ, Flehinger BJ, Schottenfeld D, et al. Screening for colorectal cancer with fecal blood testing and sigmoidoscopy. J Natl Cancer Inst 1993; 85:1311-1318.[Abstract/Free Full Text]
5. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993; 329:1977-1981.[Abstract/Free Full Text]
6. Muller AD, Sonnenberg A. Protection by endoscopy against death from colorectal cancer: a case-control study among veterans. Arch Intern Med 1995; 155:1741-1748.[Abstract]
7. Muller AD, Sonnenberg A. Prevention of colorectal cancer by flexible endoscopy and polypectomy: case control study of 32,702 veterans. Ann Intern Med 1995; 123:904-910.[Abstract/Free Full Text]
8. Selby JV, Friedman GD, Quesenberry CP, et al. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992; 326:653-657.[Abstract]
9. Selby JV, Friedman GD, Quesenberry CP, Weiss NS. Effect of fecal occult blood testing on mortality from colorectal cancer: case control study. Ann Intern Med 1993; 118:1-6.[Abstract/Free Full Text]
10. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992; 84:1572-1575.[Abstract/Free Full Text]
11. Lazovich D, Weiss NS, Stevens NG, et al. A case-control study to evaluate the effectiveness of screening for fecal occult blood. J Med Screen 1995; 2:84-89.[Medline]
12. Vining DJ, Gelfand DW. Noninvasive colonoscopy using helical CT scanning, 3D reconstruction, and virtual reality.; Presented at the 23rd Annual Meeting and Postgraduate Course of the Society of Gastrointestinal Radiologists, Maui, Hawaii, 1994..
13. Coin, inventor CG, Bond, assignee WC, Stafford, assignee TP. Computed tomographic colonoscopy. U.S. patent 5,458,111. October 17, 1995.
14. Johnson CD. Virtual endoscopy: what’s in a name?. AJR Am J Roentgenol 1998; 171:1201-1202.[Free Full Text]
15. Yee J, Akerkar GA, Hung RK, Terdiman J, McQuaid K, Wall SD. Colonic distention and colorectal polyp detection with and without glucagon on virtual colonoscopy.; Presented at the 28th Annual Meeting and Postgraduate Course of the Society of Gastrointestinal Radiologists, Palm Beach, Fla, 1999..
16. Hara A, Johnson CD, Reed JE, et al. Colorectal polyp detection with CT colography (virtual colonoscopy): a blinded, prospective study (abstr). Radiology 1996; 201(P):252.
17. Hara AK, Johnson CD, Reed JE, et al. Detection of colorectal polyps by computed tomographic colography: feasibility of a novel technique. Gastroenterology 1996; 110:284-290.[Medline]
18. Hara AK, Johnson CD, Reed JE, et al. Reducing data size and radiation dose for CT colonography. AJR Am J Roentgenol 1997; 168:1181-1184.[Free Full Text]
19. Beaulieu CF, Napel S, Daniel BL, et al. Detection of colonic polyps in a phantom model: implications for virtual colonoscopy data acquisition. J Comput Assist Tomogr 1998; 22:656-663.[Medline]
20. Dachman AH, Lieberman J, Osnis RB, et al. Small simulated polyps in pig colon: sensitivity of CT virtual colography. Radiology 1997; 203:427-430.[Abstract/Free Full Text]
21. Chen SC, Lu DSK, Hecht JR, Kadell BM. CT colonography: value of scanning in both the supine and prone positions. AJR Am J Roentgenol 1999; 172:595-599.[Abstract/Free Full Text]
22. Fletcher JG, Johnson CD, MacCarty RL, Welch TJ, Reed JE. Optimization of CT colonography technique: a prospective trial in 180 patients. Radiology; (in press).
23. Fletcher JG, Johnson CD, MacCarty RL, Welch TJ, Hara AK, Reed JE. CT colonography: overcoming the problems of collapse and colonic fluid (abstr). Radiology 1998; 209(P):296.
24. McAdams HP, Shahidi R, Argiro V, Chotas H, Kurylo W, Tapson V. Clinical evaluation of perspective surface and volume rendering techniques for performing virtual bronchoscopy. In: Lemke HU, et al., eds. Computer assisted radiology. New York, New York: Elsevier Science, 1996; 73-78.
25. McFarland EG, Brink JA, Loh J, et al. Visualization of colorectal polyps with spiral CT colography: evaluation of processing parameters with perspective volume rendering. Radiology 1997; 205:701-707.[Abstract/Free Full Text]
26. Reed JE, Johnson CD. Automatic segmentation, tissue characterization, and rapid diagnosis enhancements to the computed tomographic colonography analysis workstation. Proc SPIE 1997; 3033:433-440.
27. Reed J. Display/navigation: Mayo Clinic experience.; Presented at the First International Symposium: Virtual Colonoscopy, Boston, Mass, October 1, 1998..
28. Wang G, Dave SB, Brown BP, et al. Colon unraveling based on electrical field: recent progress and further work.; Presented at the Meeting of SPIE—International Society for Optical Engineering, 1999, San Diego, Calif, February 20–26, 1999..
29. Wyatt CL, Ge Y, Vining DJ. Automatic segmentation of the colon.; Presented at SPIE–International Society for Optical Engineering, San Diego, Calif, 1999, February 20–26, 1999..
30. Paik DS, Beaulieu CF, Jeffrey RB, Rubin GD, Napel S. Automated flight path planning for virtual endoscopy. Med Phys 1998; 25:629-637.[Medline]
31. McFarland EG, Wang G, Brink JA, Balfe DM, Heiken JP, Vannier MW. Spiral computed tomographic colonography: determination of the central axis and