Portal de Periódicos da CAPES

Screening by CT colonography: too early to pass judgment on a nascent technology

1999; Elsevier BV; Volume: 50; Issue: 3 Linguagem: Inglês

10.1053/ge.1999.v50.99617

1097-6779

David A. Ahlquist, C. Daniel Johnson,

Gastric Cancer Management and Outcomes

Most would agree that better screening tools for colorectal neoplasia are needed in terms of performance characteristics, acceptance by patients, credibility among physicians, and affordability. Opening the door to an intriguing minimally invasive approach, Vining et al.1Vining DJ Gelfand DW Bechtold RE Scharling ES Grishaw EK Shifrmn RY. Technical feasibility of colon imaging with helical CT and virtual reality.Am J Roentgenol. 1994; 162 ([abstract]): 104Google Scholar in 1994 introduced “virtual colonoscopy,” which involved the creation of a virtual colorectal image using data input from a brief helical CT. Rigorous diagnostic interrogation could then be performed on this virtual image rather than on the patient directly. Now generically referred to as CT colonography, this methodology yields an array of reformatted two- and three-dimensional views that can be evaluated in a highly interactive manner with multiple static and dynamic display options using commercially available workstations and software packages. Instrumentation, image processing techniques, and operator efficiencies continue to evolve, as reviewed.2Vining DJ. Virtual endoscopy: Is it reality?.Radiology. 1996; 200: 30-31PubMed Google Scholar Since the initial clinical feasibility studies were published in 1996,6Hara AK Johnson CD Reed JE Ahlquist DA Nelson H Ehman RL et al.Detection of colorectal polyps by computed tomographic colography: feasibility of a novel technique.Gastroenterology. 1996; 110: 284-290Abstract Full Text PDF PubMed Scopus (203) Google Scholar, 7Hara AK Johnson CD Reed JE Ehman RL Ilstrup DM. Colorectal polyp detection using computed tomographic colography: two- versus three-dimensional techniques.Radiology. 1996; 200: 49-54PubMed Google Scholar dozens of groups have presented their early clinical experience with CT colonography. Most of these series, largely in abstract form, have been descriptive in nature, comprised small numbers of selected patients, and shown favorable neoplasm detection rates with colonoscopy as the gold standard. In this issue of Gastrointestinal Endoscopy, Rex et al.8Rex DK Vining D Kopecky KK. An initial experience with screening for colon polyps using spiral CT with and without CT colography (virtual colonoscopy).Gastrointest Endosc. 1999; 50: 309-313Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar report the results of a blinded comparison between CT colonography and colonoscopy on 43 patients studied during 1995 to 1996. Their conclusion that CT colonography as performed is not adequate as a colorectal cancer screening test might be challenged by some. Several questions can be raised in an effort to reach for perspective at this juncture. The answer would seem to depend on the target population. If intended for colorectal neoplasia screening in those at average risk, CT colonography should be at least as sensitive and specific in detecting large adenomas and early-stage cancers as are the currently approved methods of fecal blood testing, flexible sigmoidoscopy, and barium enema x-ray. It should also be competitive with respect to patient comfort and safety, cost, and potential for wide distribution. If planned as a surveillance tool in higher-than-average risk populations, then CT colonography would have to approach the detection accuracy of colonoscopy for clinically important neoplasms whatever its other advantages may be. Although rigorous comparative data are yet to be produced, it is instructive to consider how the emerging technology of CT colonography may offer improvements over the other screening modalities. Fecal occult blood testing (FOBT), the most widely-used and extensively evaluated screening approach, modestly reduces colorectal cancer mortality when applied every 1 to 2 years over a decade or more but does not affect cancer incidence.9Mandel JS Bond JH Church TR Snover DC Bradley GM Schuman LM et al.Reducing mortality from colorectal cancer by screening for fecal occult blood.N Engl J Med. 1993; 328: 1365-1371Crossref PubMed Scopus (2902) Google Scholar, 10Kronborg 0 Fenger C Olsen J Jorgensen OD Sondergaard 0. Randomized study of screening for colorectal cancer with faecal-occult blood test.Lancet. 1996; 348: 1467-1471Abstract Full Text Full Text PDF PubMed Scopus (2179) Google Scholar, 11Hardcastle JD Chamberlain JO Robinson MHE Moss SM Amar SS Balfour TW et al.Randomized controlled trial of faecal-occult blood screening for colorectal cancer.Lancet. 1996; 348: 1472-1477Abstract Full Text Full Text PDF PubMed Scopus (2422) Google Scholar Such outcomes can be explained by the low sensitivity of FOBT for colorectal cancer, 20% to 50%, and the very low detection rates for premalignant adenomas, only 5% to 15%, based on comparison against structural tests.12Ahlquist DA Wieand HS Moertel CG McGill DB Loprinzi CL O'Connell MJ et al.Accuracy of fecal occult blood screening for colorectal neoplasia: a prospective study using HemoQuant and Hemoccult.JAMA. 1993; 269: 1262-1267Crossref PubMed Scopus (309) Google Scholar, 13Bang TM Tillett 5 Noar SK Blair A McDougall V. Sensitivity of fecal Hemoccult testing and flexible sigmoidoscopy for colorectal cancer screening.J Occup Med. 1986; 28: 709-713Crossref PubMed Scopus (39) Google Scholar, 14Demers RY Stawick LE Demers P. Relative sensitivity of the fecal occult blood test and flexible sigmoidoscopy in detecting polyps.Preventive Med. 1985; 14: 55-62Crossref PubMed Scopus (41) Google Scholar, 15Jahn H Joergensen OK Kronborg O Fenger C. Can Hemoccult II replace colonoscopy after radical surgery for colorectal cancer or after polypectomy?.Dis Colon Rectum. 1992; 35: 253-256Crossref PubMed Scopus (59) Google Scholar, 16Robinson MHE Kronborg O Williams CB Bostock K Rooney PS Hunt LM. Fecal occult blood testing and colonoscopy in the surveillance of subjects at high risk of colorectal neoplasia.Br J Surg. 1995; 82: 318-320Crossref PubMed Scopus (36) Google Scholar, 17Rozen P. Screening for colorectal neoplasia in the Tel Aviv area: cumulative data 1979-89 and initial conclusions.Isr J Med Sci. 1992; 28: 8-20PubMed Google Scholar CT colonography would appear to be substantially more sensitive for colorectal neoplasms, which would translate into greater efficacy. Because it would detect premalignant adenomas, CT colonography could be applied less frequently than FOBT to yield a benefit. Case-control studies suggest that proctosigmoidoscopy reduces both incidence and mortality for rectosigmoid cancers but has no affect on more proximally located colon cancers.18Selby JV Friedman GD Quesenberry CP Weiss NS. A case control study of screening sigmoidoscopy and mortality from colorectal cancer.N Engl J Med. 1992; 326: 653-657Crossref PubMed Scopus (1567) Google Scholar, 19Newcomb PA Norfleet RG Storer BE Surawicz TS Marcus PM. Screening sigmoidoscopy and colorectal cancer mortality.J Natl Cancer Inst. 1992; 84: 1572-1575Crossref PubMed Scopus (864) Google Scholar, 20Muller AD Sonnenberg A. Prevention of colorectal cancer by flexible endoscopy and polypectomy—a case control study of 32,702 veterans.Ann Int Med. 1995; 123: 904-910Crossref PubMed Scopus (561) Google Scholar, 21Muller AD Sonnenberg A. Protection by endoscopy against death from colorectal cancer. A case control study among veterans.Arch Int Med. 1995; 155: 1741-1748Crossref PubMed Scopus (376) Google Scholar Like proctosigmoidoscopy, CT colonography would detect adenomas and its use in screening should lower cancer incidence. However, the proximal colon is also examined with CT colonography which would potentially result in a more universal benefit. CT colonography appears to have several advantages over barium enema x-ray, a radiographic procedure that would seem to compete most directly. Although barium enema x-ray displays the entire colorectum, images are limited to a few planar views with distraction and obscuration caused by superimposed radiodense shadows. Polyp detection rates by barium enema x-ray have varied widely in published reports. Preliminary data from the National Polyp Study, in which more than 3000 patients underwent both air-contrast barium enema and colonoscopy, suggest that barium enema x-ray may detect only 44% of colorectal neoplasms 1 cm or more in diameter.22Ott DJ. Analysis of accuracy, complications, and cost of barium enema for CRC diagnosis.in: Presented at the International Workshop on Colorectal Cancer Screening National Institute of Health,, Bethesda, MarylandJune 1994Google Scholar Neoplasm detection rates by barium enema x-ray have been higher in earlier studies at referral centers.23Johnson CD Carlson HC Taylor WF Weiland LP. Barium enemas in carcinoma of the colon: sensitivity of double- and single-contrast studies.AJR. 1983; 140: 1143-1149Crossref PubMed Scopus (69) Google Scholar Nevertheless, CT colonography is theoretically a more powerful imaging method, as the endoluminal displays improve diagnostic accuracy by eliminating the confounding caused by superimposed structures. Preliminary results from early comparison studies suggest that CT colonography does indeed detect more polyps with fewer false-positive results than does barium enema x-ray,24Hara AK Johnson CD Reed JE Ahlquist DA Nelson H Ehman RL. Completed colography (virtual colonoscopy): early comparison with barium enema for polyp detection.Gastroenterology. 1997; 112 ([abstract]): A575Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 25Ogura T Koizumi K Kai S Maruyama M Terahira T. Three-dimensional CT colonoscopy: comparison with colonoscopy and barium enema examination.Radiology. 1995; 197 ([abstract]): 444Google Scholar but more rigorous comparisons are clearly needed to settle this issue. Unlike barium x-rays, CT colonography provides information on lesion density, colon wall thickness, and pericolonic structures; and diagnostic manipulations are carried out on the virtual image rather than on the patient. The radiation exposure associated with CT colonography is comparable with that associated with barium x-ray.26Hara AK Johnson CD Reed JE Ahlquist DA Nelson H Ehman RL Harmsen WS. Reducing data size and radiation dose for CT colonography.AJR. 1997; 168: 1181-1184Crossref PubMed Scopus (148) Google Scholar CT colonography involves no liquid contrast so that subsequent endoscopic polypectomy could be accomplished the same day without the need for repeated bowel preparation. Finally, it may be feasible some day to perform digitalized subtraction of stool from the virtual image (“virtual preparation”), as has already been accomplished to a limited extent,4Ahlquist DA Johnson CD Hara AK. Computed tomographic colography and virtual colonoscopy.Gastrointest Endosc Clin N Am. 1997; 7: 439-452PubMed Google Scholar which would obviate the chore and disincentive of cathartic lavage. Although considered to be the most sensitive and specific diagnostic tool available, colonoscopy has historically not been used for colorectal cancer screening in the general population because of its expense, small risk of inflicting harm, and perceived discomfort. Prospective trials in higher-than-average risk groups have demonstrated that periodic colonoscopy markedly lowers incidence and mortality from cancers at all colorectal sites.27Jarvinen HJ Mecklin JP Sistonen P. Screening reduces colorectal cancer rates in families with hereditary nonpolyposis colorectal cancer.Gastroenterology. 1995; 108: 1405-1411Abstract Full Text PDF PubMed Scopus (457) Google Scholar, 28Winawer J Zauber AG Ho MN O'Brien MJ Gottlieb L Prevention of colorectal cancer by colonoscopic polypectomy.N Engl J Med. 1993; 329: 1977-1981Crossref PubMed Scopus (3874) Google Scholar CT colono-graphy offers possible advantages of greater safety, no need for sedation, less discomfort, and lower cost. Furthermore, assessment of lesion site is inaccurate in some instances by colonoscopy as a result of ambiguous landmarks, bowel stretching, or obstructing lesions; and this would be less likely to occur with CT colonography. Finally, incidental or associated health-threatening pathologic conditions in the extracolonic abdomen or pelvis may be discovered during CT colonography which would be an important added dimension. Are the data at hand sufficient to indict CT colonography as an inadequate screening tool? Rex et al.8Rex DK Vining D Kopecky KK. An initial experience with screening for colon polyps using spiral CT with and without CT colography (virtual colonoscopy).Gastrointest Endosc. 1999; 50: 309-313Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar report a sensitivity for CT colonography of 57% for detection of colorectal neoplasms 1 cm or more in diameter with a specificity of 89% in their initial clinical series of 46 patients compared against colonoscopy, which is below that generally observed in other studies to date, as reviewed.5Johnson CD Ahlquist DA. CT colography (virtual colonoscopy): a new method for colorectal screening.Gut. 1999; 44: 301-305Crossref PubMed Scopus (51) Google Scholar For example, our group found a sensitivity by CT colonography of 75% and specificity of 90% in an early series done in blinded fashion on 70 patients scanned in supine position alone.29Hara AK Johnson CD Reed JE Ahlquist DA Nelson H MacCarty FL et al.Detection of colorectal polyps with CT colography: initial assessment of sensitivity and specificity.Radiology. 1997; 205: 59-65PubMed Google Scholar In a more recent blinded analysis,30Fletcher JG Johnson CD MacCarty RL Welch TJ Reed JE Ahlquist DA. CT colonography in 180 patients: the benefits of prone imaging.Gastroenterology. 1999; 116 ([abstract]): A404Google Scholar our preliminary data on 180 patients scanned in both supine and prone positions (as was done in the study by Rex et al.) yield a sensitivity for CT colonography of 85% and specificity of 93% for neoplasms 1 cm or more in size. Shared shortcomings in statistical power and experimental design can be identified across all of these early studies. Because Rex et al. assert that CT colonography is inadequate as a screening tool, several weaknesses in their study (some of which the authors acknowledge in their Discussion) should be emphasized.Small sample size. The degree of uncertainty in their sensitivity estimate was not shown but might be considered too high to draw firm conclusions on the performance of CT colonography. Of the 14 polyps seen on colonoscopy that were 1 cm or more in diameter, 8 (57%) were detected by CT colonography in blinded fashion. However, based on this small number, the 95% confidence interval for this estimate is very wide, ranging from 29% to 82%. They do not state how specificity was calculated, but this estimate is also likely to be uncertain as only 13 of their subjects had no polyps.Learning curve bias. As with other early studies, Rex et al. identified remediable technical and perceptive errors in their unblinding, and four of the six lesions 1 cm or more in size initially missed by CT colonography were apparent in retrospect. Such an internal reassessment is critical to improving the diagnostic process and underscores the importance of attention to detail with this technique. One could speculate that their results may have been better if repeated in blinded fashion to allow for these technical and interpre- tive adjustments. If the four misses could then have been avoided, the estimated detection rate by CT colonography would have risen to 86%: 95% CI [57%, 98%].Gold standard bias. It is axiomatic that if one of two tests being compared is declared the gold standard, then that one will necessarily outperform the other test. Such could be the case with the study by Rex et al. and most other early studies on CT colonography. As these authors appreciate from their previous studies and from others, colonoscopy is not perfectly golden and misses both cancers and large adenomas.31Rex DK Rahmani EY Haseman JH Lemmel GT Kaster S Buckely JS. Relative sensitivity of colonoscopy and barium enema for detection of colorectal cancer in clinical practice.Gastroenterology. 1997; 112: 17-23Abstract Full Text PDF PubMed Scopus (515) Google Scholar, 32Waye ID. What is the gold standard for colon polyps?.Gastroenterology. 1997; 112: 292-294Abstract Full Text PDF PubMed Scopus (20) Google Scholar, 33Rex DK Cutler CS Lemmel GT Rahmani EY Clark DW Helper DJ et al.Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies.Gastroenterology. 1997; 112: 24-28Abstract Full Text PDF PubMed Scopus (1377) Google Scholar, 34Haseman JH Lemmel GT Rahmani EY Rex DK. Failure of colonoscopy to detect colorectal cancer: evaluation of 47 cases in 20 hospitals.Gastrointest Endosc. 1997; 45: 451-455Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar Without follow-up imaging data on the four lesions noted by CT colonography but not by colonoscopy, it could be argued that these represent either false-positives on CT colonography or false-negatives on colonoscopy or a combination of both. In one small study of colorectal cancer patients destined for surgery in which the resected specimen was the final arbiter, CT colonography detected 14 of 15 associated polyps whereas colonoscopy detected 13 of the 15.35Royster AP Fenlon HM Clarke PD Nunes DP Ferrucci JT. CT colonoscopy of colorectal neoplasms: two-dimensional and three-dimensional virtual reality techniques with colonoscopic correlation.AJR. 1997; 169: 1237-1242Crossref PubMed Scopus (204) Google ScholarSubject selection bias. Patients studied by Rex et al. represented a small group enriched with polyps, a portion of which had already been detected on screening sigmoidoscopy. An unintended selection bias favoring endoscopic detection may have resulted. Also, three of the four lesions 2 cm or more in size were described as flat and initially missed by CT colonography. Although it is intuitive that flat lesions may prove to be more evasive targets for this radiographic technique, truly flush neoplasms without associated focal mural thickening or any other distinguishing features on CT colonography have been rare in our experience to date. Almost certainly yes. Given the rapid evolution that has already occurred with this technology, it seems reasonable to prognosticate on further changes. More rapid, higher-resolution CT scanners will further minimize any artifact caused by biomotions and enhance lesion discrimination. As mentioned, digitized subtraction of stool based on ingested contrast or native radiotexture differences could allow for a virtual preparation to improve patient acceptance or help to differentiate sessile polyps from adherent stool to increase specificity. Image processing may incorporate artificial intelligence methodology and more efficient display options to enhance the accuracy and speed of interpretation. Refinements in display of mucosal surface texture may even allow better detection of flat neoplasms. CT colonography has proven to be a powerful new approach to image the colorectum. Although instrumentation and operator skill will continue to evolve, CT colonography as it is currently configured possesses several real or potential advantages over conventional screening tools. Given the limited statistical power and design biases common in many of the early-series studies to date, lack of direct comparisons with other screening modalities, and uncertainties of costing, the verdict on the role for CT colonography is still out. Votes would be better cast after rigorously done prospective studies involving appropriate target populations and criterion standards are weighed in.