J Gastroenterol 2002; 37[Suppl XIII]:92-96
Journal of
Gastroenterology
9
Springer-Verlag 2002
Colorectal cancer screening: the potential role of virtual colonoscopy JOHN H. BOND Gastroenterology Section (111D), VA Medical Center, One Veterans Drive, Minneapolis, MN 55417, USA; and University of Minnesota, Minneapolis, MN, USA
Virtual colonoscopy is a promising new technique that combines rapid spiral CT scanning of the abdomen with advanced computer programs capable of re-creating two- and three-dimensional views of the colon and rectum. Recent studies comparing this method with conventional colonoscopy show that virtual colonoscopy already is more accurate than barium enema X-ray studies for the detection of colorectal polyps, and that it approaches the accuracy of colonoscopy for diagnosing advanced lesions. Before virtual colonoscopy can be promoted for population-based screening for colorectal cancer, a number of issues discussed in this review need to be addressed. These include questions of accuracy, availability, acceptability, and cost-effectiveness.
Key words: virtual colonoscopy, colorectal cancer, screening
Introduction Evidence-based guidelines, recently developed or revised in the U.S., recommend that all average-risk men and women in high-incidence countries undergo screening for colorectal cancer. 1-3 Asymptomatic individuals without other special risk factors for the disease should be offered fecal occult blood tests (FOBT) yearly and flexible sigmoidoscopy every 5 years, beginning at age 50. A positive screening test result is an indication for colonoscopy. A number of recently completed scientific studies plus accumulating clinical experience indicate that widespread compliance with these screening recommendations could reduce the mortality from this major cancer killer by over half. Although the current guidelines also include the options of direct screening
Reprint requests to: J.H. Bond
with a barium enema X-ray study every 5 years or conventional colonoscopy every 10 years, these alternatives currently are supported by indirect evidence only. According to the guidelines, for a new method to be substituted for an established screening recommendation, it first should be shown to be as safe, acceptable, available, and cost-effective as the method it might replace. Several recent studies indicate that double-contrast barium enema may be insufficiently accurate for the detection of early cancer and advanced adenomatous polyps to be used either for screening or as a first-choice diagnostic examination. For example, a retrospective study in 20 medical centers in Indiana by Rex et al. reported that the sensitivity of barium enema for diagnosing colorectal cancer was only 84%, compared with 95% for colonoscopy.4 When colonoscopy was performed by a gastroenterologist, the accuracy was 98%. The National Polyp Study conducted a controlled comparison between back-to-back colonoscopy and doublecontrast barium enema examinations performed in the same patient for the detection of polyps.5 A total of 862 paired examinations were done in 580 patients undergoing postpolypectomy surveillance. Examinations were performed by experienced study coinvestigators blinded to the results of the alternative evaluation. Barium enema detected a polyp in only 39% of cases in which one was subsequently found during colonoscopy. Even when a patient had an advanced adenoma - 1 cm in diameter, the paired barium enema result was negative in 52% of cases. False-positive barium enema results occurred in 14% of cases. Colonoscopy accurately detects polyps ->lcm in diameter. Tandem studies in which two back-to-back colonoscopies were performed in the same patients showed that the procedure missed up to 24% of polyps <1 cm in diameter. 6,7 However, larger, more advanced adenomas were seldom missed. Virtual colonoscopy (CT colography) is a promising new method that appears to have greater potential than
J.H. Bond: Virtual colonoscopy screening
93
does barium enema as a radiological screening test for colorectal polyps and cancer? ,9 This review will outline the advantages and limitations of this new technique, present data on its sensitivity and specificity for detecting colorectal polyps, and discuss questions and issues that still need to be resolved before multicenter screening trials are carried out.
Contracted or spastic colonic segments during imaging can interfere with detection of lesions, causing both false-negative and false-positive results. Therefore, prior to scanning, the patient's colon is distended with air or carbon dioxide insufflated through a rectal tube according to patient tolerance. Glucagon also may be given in some cases to help eliminate spastic colonic segments.
Technique of virtual colonoscopy
Advantages and disadvantages of virtual colonoscopy
CT colography or "virtual colonoscopy" uses data obtained from rapid helical CT scanning of the abdomen to create computer-reformatted two- and threedimensional images of the colon. Modern CT scanners are now capable of obtaining hundreds of 1-mm image slices in less than a minute, during a single breath-hold, eliminating motion artifacts that previously prevented high-resolution imaging of the bowel. These spiral scanners move the patient through a rotating X-ray beam, and over 500 1-mm image slices can be obtained during a single pass. The latest generation of multihead ("multislice") scanners shortens imaging time even further to only 15-20s, collecting more data points in a shorter time and thus increasing resolution. Computer software reformats these images into complex two- and three-dimensional renditions of the intact colon. Using a conventional CT workstation and a dynamic, interactive display of images, a radiologist can conduct a "flythrough" examination of the large bowel, simulating the way an endoscopist views the colon during conventional colonoscopy. This dynamic viewing of images displayed at a rate of 16-30 each second complements the radiologist's initial review of static two-dimensional images of the bowel. Computerized color and shading enhancement of images assist in the recognition of mucosal abnormalities. The radiation dose of a single CT colography examination is about half that of a barium enema, and it is substantially less than that of a standard abdominal or pelvic CT scan. Because it often is impossible to totally eliminate residual stool, fluid, and bowel spasm, most patients are scanned twice, first in the supine and then in the prone position, to reduce false-positive and falsenegative results. Once a patient is prepared for the procedure, data acquisition usually requires less than 10 min. Currently, virtual colonoscopy requires a very thorough bowel-cleansing preparation similar to that needed for a barium enema examination or a conventional colonoscopy. Residual water or fecal debris substantially hinders polyp detection. Many centers employ an oral phosphasoda preparation rather than a polyethylene glycol-based oral lavage solution, because it results in less retained luminal fluid.
There are several potential advantages of virtual colonoscopy over conventional colonoscopy.1~The examination time usually is shorter, and there is no need for preprocedure sedation. There is less risk of complications; to date no serious morbidity or mortality has been reported with virtual colonoscopy. Diagnostic conventional colonoscopy results in perforation of the colon in about 0.05% of reported cases. 11Both sides of the bowel wall and bowel folds can be scrutinized during virtual colonoscopy, and the location of lesions in the colon can be precisely determined. Unless a lesion is in the rectum or cecum, precise localization during colonoscopy often is not possible. The radiologist, using a number of static and dynamic display options, can examine and reexamine segments of the colon after a virtual colonoscopy scan has been performed. Virtual colonoscopy can be used to examine the proximal colon prior to surgery when an obstructing left-sided cancer prevents passage of a colonoscope. The disadvantages of virtual colonoscopy include the need for a thorough bowel-cleansing preparation and preprocedure gas distention of the colon. Scanning hardware is expensive, although most centers acquire these scanners for a variety of other common clinical applications. Retained stool or fluid, or contracted segments of colon that interfere with interpretation, are difficult to completely eliminate. Setting up the reconstructed colon images and interpretation of examinations currently is labor intensive and therefore relatively costly. However, a number of improvements have been made since the technique was first introduced in 1994 by Vining and colleagues?2 According to recent reports, radiologist time per examination has been reduced from 2-3 h to about 30 rain. However, no one yet has developed an interactive fly-through with which anyone wishing to review the study can repeat a virtual colonoscopy. Currently, if the person creating the initial fly-through fails to examine the site of a given lesion, those reviewing the examination will miss the same abnormality. As discussed below, another major disadvantage of this method is its low resolution for flat lesions. Some important neoplasms are flush with the colorectat mucosal contour and are difficult to detect with CT colography.
94
J.H. Bond: Virtual colonoscopy screening
Clinical studies of virtual colonoscopy The accuracy of virtual colonoscopy for the detection of colorectal polyps has been compared with that of conventional colonoscopy in several clinical series (see Table 1). The largest published U.S. direct comparison was reported by Fenlon et al. from Boston last year. 13In this excellent study, both procedures were performed in each of 100 patients at high risk for colorectal neoplasia. The endoscopists and radiologists who interpreted the findings were blinded to the results of the alternative examination. With conventional colonoscopy as the gold standard, the sensitivity of virtual colonoscopy for the detection of polyps was 91% for large polyps (->1 cm in diameter), 82% for medium-sized polyps (0.60.9 cm), and 55% for small polyps (-<0.5 cm). These authors concluded that virtual colonoscopy already is more accurate than barium enema, that it compares favorably with conventional colonoscopy for the detection of larger polyps, and that it may be an effective screening test for colorectal cancer and neoplasia. Another Boston study conducted at the Harvard-affiliated hospitals reported results equally as impressive. TM In 49 selected patients with eight previously diagnosed cancers, all of the cancers were detected by virtual colonoscopy. The sensitivity/specificity for detection of large (>l.0cm), medium-sized (0.6-1.0cm), and small (->0.5 cm) polyps was 92%/100%, 72%/80%, and 27%/ 60%, respectively. The high false-positive rate (low specificity) for smaller polyps occurred mainly in patients with poorly prepped or poorly distended colonic segments. Dachman et al. from the University of Chicago compared virtual and conventional colonoscopy in a single-blinded study of 49 patients. 15 The sensitivity for detection of polyps >0.8 cm in diameter was 83%. The average time spent in interpretation was 28.5 min (range, 14-65 min). Three-dimensional endoluminal views were considered necessary in only 52% of cases. Others have not achieved results as good as those reported by Fenlon et al. Hara et al. from the Mayo Clinic reported a sensitivity of only 75% and a specificity of 90% for the detection of large adenomas (->1 cm). 16 Another recently published study by Rex et al. concluded that CT colography is not yet sufficiently
Table 1. Accuracy of virtual colonoscopy in three U.S. studies Polyp size (cm) >-1.0 0.6-0.9 -<0.5
Sensitivity/specificity(%) Fenlon et al.~3 Morrin et al.14 91/96 82/86 55/85
93/100 65/92 32/94
Hara et al.16 75/90 66/63 45/80
accurate to be promoted for screening. 17 In 46 patients at average risk for colorectal neoplasia, the sensitivity of virtual colonoscopy for the detection of adenomas 2 cm in diameter or larger was only 25% (three of four of these large polyps were missed), and for those 1-1.9cm in diameter it was 60%. In both the Hara and the Rex series, several large, flat, right-sided adenomatous polyps were missed by virtual colonoscopy. A London study reported this year that "CT pneumocolon" accurately detects invasive cancers but is not useful for detecting colorectal polyps, is In 201 selected consecutive patients, the sensitivity for detection of polyps >1 cm in diameter was only 50%. The difference between these results and those of Fenlon et al. can partly be explained by differences in study design, patient selection, and the state of this rapidly evolving technique at the time of examination of the patients.
Predicted improvements in virtual colonoscopy Although virtual colonoscopy has great potential as a screening test for polyps and cancers, several issues or questions about this method need to be addressed and resolved before population-based screening trials are conducted. It is not known how well asymptomatic, average-risk people would accept a screening CT procedure that requires a full cleansing bowel preparation followed by uncomfortable gas insufflation of the colon. Currently, because of the time it takes to process and interpret a virtual study, the onerous bowel preparation often would have to be repeated if conventional colonoscopy was required to evaluate an abnormality or resect detected polyps. Most people probably would prefer instead to have a conventional colonoscopic screening evaluation that would be both diagnostic and therapeutic at a single visit, with a single bowel preparation. In a series reported from San Francisco by Akerkar et al., 295 patients graded conventional colonoscopy as better tolerated than virtual colonoscopy59 Although some investigators believe that a way may be found to allow computer separation of tissue from bowel contents, obviating the need for a cleansing preparation, many others believe that such a "virtual prep" may be difficult to achieve. Current research also is being directed toward designing more effective and better-tolerated bowel-cleansing regimens. Bowel distention achieved with hand-held insufflation devices commonly causes appreciable patient discomfort. Automated systems capable of avoiding excessive luminal pressure that infuse carbon dioxide, a rapidly absorbed gas, are being tested and may eventually improve patient acceptance of virtual colonoscopy. Virtual colonoscopy currently is available at relatively few centers in the U.S., and most American radi-
J.H. Bond: Virtual colonoscopy screening ologists have had limited or no experience with the new technique. Published reports emphasize that considerable experience is required before virtual colonoscopy studies can be efficiently and accurately accomplished. It probably will be some time before others are able to achieve the level of accuracy for detecting colorectal neoplasia recently reported by both Fenlon et al. and some of the other centers.
Cost and cost-effectiveness of virtual colonoscopy The current cost of virtual colonoscopy probably renders it unsuitable as a screening test. The equipment needed to perform these studies is now available in most major medical centers. However, a substantial portion of the total cost of an examination is the time required for a radiologist to assemble and interpret each examination. Since virtual colonoscopy is only a diagnostic test, the cost would need to drop substantially below that of colonoscopy, which is capable of both diagnosis and required therapy in most cases. A substantial fraction of those undergoing screening virtual colonoscopy would need a subsequent conventional colonoscopy for the evaluation of abnormalities and performance of polypectomy. The relatively low specificity of virtual colonoscopy (many false positives) greatly adds to its cost, because expensive conventional colonoscopies are required in patients with no disease. Sonnenberg et al. employed the Markov modeling method to compare the cost-effectiveness of virtual colonoscopy with that of conventional colonoscopy in screening for colorectal cancer. 2~This analysis showed that virtual colonoscopy was more costly than colonoscopy (US $24586 vs $20930 per year of life saved). Even when the authors assumed that the sensitivity and specificity of virtual colonoscopy were 100%, conventional colonoscopy remained more cost-effective. Only when the cost of virtual colonoscopy was assumed to be <55% of that of conventional colonoscopy, or the compliance rate for virtual colonoscopy was 15%-20% higher than that for colonoscopy, did virtual colonoscopy become the more cost-effective option.
Summary and conclusions Virtual colonoscopy has been greatly improved since its introduction in 1994. However, before it can replace established methods of screening for colorectal cancer and polyps, the problems and limitations discussed in this review need to be satisfactorily addressed. In addition, several technical advances are required. These include the development of scanning methods with greater resolution for detecting small and flat polyps.
95 Refinements in computer software that allow fast, automated interpretation of images are needed to reduce expensive radiologist time. Accepted methods of achieving colonic distention and of either ensuring a well-tolerated bowel preparation or obviating the need for bowel cleansing probably will be a prerequisite for achieving satisfactory levels of screening compliance. After these technical improvements are in place and a sufficient fraction of centers have experience with virtual colonoscopy, larger multicenter, controlled trials or outcome studies should be conducted to see if virtual colonoscopy can replace current methods for population-based screening of people at average risk for colorectal neoplasia.
References 1. Guide to clinical preventive services. 2nd ed. Report of the U.S. Preventive Services Task Force. Washington, DC: Department of Health and Human Services, 1995. 2. Winawer SJ, Fletcher RH, Miller L, Godlee F, Stolar MH, Mulrow CD, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594-642. 3. Byers T, Levin B, Rothenberger D, Dodd GD, Smith RA. American Cancer Society guidelines for screening and surveillance for early detection of coiorectal polyps and cancer: update 1997. CA-A Cancer J Clin 1997;47:154-60. 4. Rex DK, Rahmani EY, Haseman JH, Lemmel GT, Kaster S, Buckley JS. Relative sensitivity of colonoscopy and barium enema for detection of colorectal cancer in clinical practice. Gastroenterology 1997;112:17~3. 5. Winawer SJ, Stewart ET, Zauber AG, Bond JH, Ansel H, Waye JD, et al. A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. N Engl J Med 2000;342:1766-72. 6. Hixson LJ, Fennerty MB, Sampliner RE, Garewal HS. Prospective blinded trial of the colonoscopic miss-rate of large colorectal polyps. Gastrointest Endosc 1991;37:125-7. 7. Rex DK, Cutler CS, Lemmel GT, Rahmani EY, Clark DW, Helper D J, et al. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies. Gastroenterology 1997; 112:24-8. 8. Ahlquist DA, Hara AK, Johnson CD. Computed tomographic colography and virtual colonoscopy. Gastrointest Endosc Clin No Am 1997;7:439-52. 9. Hara AK, Johnson CD, Reed JE, Ahlquist DA, Nelson H, Ehman RL, et al. Detection of colorectal adenomas by computed tomographic colography: feasibility of a novel technique. Gastroenterology 1996;110:284-90. 10. Bond JH. Virtual colonoscopy--promising, but not ready for widespread use. N Engl J Med 1999;341:1540-2. 11. Rankin GB. Indications, contraindications and complications of colonoscopy. In: Sivak MV, editor. Gastrointestinal endoscopy. Philadelphia: WB Saunders; 1987, p. 873-8. 12. Vining D J, Telgen EL, Stetts D, Vanderwerken B, Kopecky KK, Rex D. Virtual colonoscopy: a 60-second colon examination. Radiology 1995;197:281 (abstract). 13. Fenlon HM, Nunes DP, Schiroy PC 3rd, Barish MA, Clarke PD, et al. A comparison of virtual and conventional colonoscopy for the detection of colorectal polyps. N Engl J Med 1999;341:14961503. 14. Morrin M, Farrell RJ, Kruskal JB, McGee JB. Prospective comparison of virtual CT colonography in patients undergoing elective colonoscopy. Gastroenterology 1999;116:A461 (abstract).
96 15. Dachman AH, Kuniyoshi JK, Boyle CM, Samara Y, Hoffmann KR, Rubin DT, Hanan I. CT Colonography with three-dimensional problem solving for detection of colonic polyps. Am J Roentgenol 1998;171:989-95. 16. Hara AK, Johnson CD, Reed JE, Ahlquist DA, Nelson H, MacCarty RL, et al. Detection of colorectal polyps with CT colography. Initial assessment of sensitivity and specificity. Radiology 1997;205:59-65. 17. Rex DK, Vining D, Kopecky KK. An initial experience with screening for colon polyps using spiral CT with and without CT colography. Gastrointest Endosc 1999;50:309-13.
J.H. Bond: Virtual colonoscopy screening 18. Miao YM, Paul B, Amin Z, Healy J, Murugan N, Westaby D, et al. A prospective single centre study comparing CT pneumocolon against conventional colonoscopy for detecting colorectal neoplasms. Gastrointest Endosc 2000;51:AB92 (abstract). 19. Akerkar GA, Hung RK, Yee J, Terdiman JP, McQuaid KR. Virtual colonoscopy: real pain. Gastroenterology 1999;116:A44 (abstract). 20. Sonnenberg A, Delco F, Bauerfeind P. Is virtual colonoscopy a cost-effective option to screen for colon cancer? Am J Gastroenterol 1999;94:2268-74.