|
Background
Virtual endoscopy combines the features of endoscopic viewing and computerized tomography (CT) data to create a virtual image that is artificially generated by a computer. In the evaluation of gastrointestinal cancers, virtual endoscopy has been most commonly used in colorectal carcinomas and to a much lesser extent in gastric carcinomas. The clinical application of virtual endoscopic techniques is also being used with other procedures such as bronchoscopy, gastroscopy, cystoscopy, sinus imaging, virtual angioscopy, and cerebral ventriculography. (Oto, 2002; Dykes, 2001). Virtual Colonoscopy Three-dimensional CT colography, or "virtual colonoscopy" combines the use of rapid helical CT with computer software capable of rendering images of the whole colon. Using a conventional workstation and a dynamic display of images, a radiologist conducts virtual examinations of the bowel, simulating the way endoscopists view the colon. Virtual colonoscopy provides a method for processing data that can display computer images of the colon in a more anatomic life-like format and is being promoted by some as a noninvasive screening test for colorectal neoplasia. Pickhardt, et al. (2003) reported that CT virtual colonoscopy with the use of a 3-D approach is an accurate screening method for the detection of colorectal neoplasia in asymptomatic average-risk adults and compares favorably with optical colonoscopy in terms of the detection of clinically relevant lesions. However, in an editorial that accompanied the study by Pickhardt et al, Morrin and LaMont (2003) stated that “if the results of this well-designed study are reproducible on a wider scale, and if the important questions regarding the appropriate size threshold and surveillance of smaller polyps can be resolved, then screening virtual colonoscopy is ready for prime time”. A study by Cotton, et al. (2004) reported that the accuracy of CT colonography (virtual colonoscopy) for the detection of colorectal cancer is less reliable than previously thought. CT colonography involves the examination of computer-generated images of the colon constructed from data obtained from an abdominal computed tomographic examination. Several studies have suggested a high degree of sensitivity for CT colonography; however, those results were obtained at single, specialized centers. Cotton reported on a new study that was designed to evaluate the accuracy of CT colonography in routine practice at nine major hospital centers. In this study, researchers assessed the accuracy of CT colonography in 615 patients aged 50 years or older who were referred for routine, clinically indicated colonoscopy (Cotton, et al., 2004). Colonoscopy was performed within 2 hours of the colonography and results were compared. The sensitivity of CT colonography for detecting one or more lesions sized at least 6 mm was 39% and for lesions sized at least 10 mm, it was 55%. These results were significantly lower than those for conventional colonoscopy, with sensitivities of 99% and 100%, respectively. CT colonography missed two of eight cancers. The accuracy of CT colonography varied considerably between centers. At the one center that had "substantial" prior experience with CT colonography, the sensitivity was 82% for lesions of 6 mm or more. Sensitivity at all of the other centers combined was 24%, with no improvement in accuracy as the number of cases at each center was increased. Preference questionnaires after both procedures were performed showed that 46% of the patients preferred CT colonography versus 41% who preferred conventional colonoscopy. The authors stated that "even if the results of CT colonography continue to be good in the hands of experts, it has yet to be proven that this expertise can be taught and disseminated reliably into daily practice". The authors concluded that CT colonography is not yet ready for widespread clinical application; techniques and training need to be improved. This is in agreement with the update of the clinical guidelines on colorectal cancer screening and surveillance that were prepared by a panel convened by the U.S. Agency for Health Care Policy and Research and published in 1997 under the sponsorship of a consortium of gastroenterology societies (Winawer, et al., 2003). It stated that promising new screening tests (virtual colonoscopy and tests for altered DNA in stool) are in development but are not yet ready for use outside of research studies. In addition, the American College of Gastroenterology does not recommend virtual colonoscopy for screening colorectal cancers. It states that more research is needed to verify the validity and generalizeability of the 3-D approach to polyp detection. It will also be necessary to develop recommendations for training in CT colonography as well as requirements of hardware and software systems and specification of methods for technical performance. Systems that allow same-day polypectomy on patients with positive CT colonography studies are not yet widely available (Rex, 2004) A technology assessment by the BlueCross BlueShield Association Technology Evaluation Center (2004) concluded that CT colonography does not meet the TEC criteria because the available evidence is insufficient to reach conclusions about the effect CT colonography on health outcomes. An assessment prepared for the Ontario Ministry of Health and Long-Term Care (2003) found that, "[a]lthough [CT colonography] offers the potential advantage of being less invasive than colonoscopy and has the ability to image the entire colon, it lacks the necessary sensitivity required for screening." The assessment noted that, in addition, unlike standard colonoscopy, CT colonography (CTC) "offers no therapy that can be applied once an abnormality is detected"; standard colonoscopy is necessary to resect lesions detected by CT colonography. The assessment concluded that "[w]ith the limited sensitivity and specificity of CTC relative to colonoscopy, together with the lack of therapeutic intervention, this method of screening may result in inconvenience, cost, and complications of both tests" and that "[b]ased on the current evidence, CTC cannot be proposed for population-based colorectal cancer screening." The assessment explained that "[p]atients with colonic symptoms or a personal/family history of polyps will benefit more in several ways if they undergo colonoscopy including excision of premalignant polyps." The assessment concluded, however, that CT colonography can be considered for diagnostic purposes in patients in whom performing colonoscopy is clinically contraindicated or for those patients who had incomplete colonoscopy because of stenosis or obstruction of the colon (Ontario Ministry of Health and Long-Term Care, 2003). In support of this conclusion, the assessment reasoned that that CT colonography is able to visualize the entire colon in most patients with occlusive tumors or stenosing lesions, and that CT colonography may be preferable to barium enema in terms of the extent of the proximal colon that can be visualized and in terms of detecting extracolonic lesions. An American Gastroenterological Association (AGA) Task Force Report (Van Dam, et al., 2004) concluded that, although virtual colonoscopy has significant promise, the technology is still evolving and results of virtual colonoscopy for screening are highly variable. According to the University of Michigan Health System’s guidelines on adult preventive health care (2004), recommended methods for colon cancer screening include fecal occult-blood testing, flexible sigmoidoscopy or colonoscopy. Winawer (2005) noted that “several options are now available for screening of colorectal cancer, and the emerging technology of stool DNA testing and virtual colonoscopy shows promise …There are quality-control issues at every step.” van Gelder, et al. (2005) stated that “despite a growing body of evidence, it remains uncertain to what extent patient acceptance, radiation issues, flat lesions, and extracolonic findings will be a stumbling block to using CT colonography for colorectal cancer screening.” An assessment by the Danish Centre for Health Technology Assessment (DACEHTA, 2005) concluded that "[d]espite the potential economic benefits, CT colonography should not replace colonoscopy as the primary diagnostic method in a Danish outpatient colonoscopy population. Such a strategy requires further research at a few centres before widespread use in routine clinical practice." Some reports of virtual colonoscopy demonstrate sensitivity similar to that of conventional colonoscopy for polyps 5 mm or larger (Pineau, et al., 2003; Paonessa, et al., 2005). However, in a meta-analysis on CT colonography, Mulhall, et al. (2005) concluded that “computed tomographic colonography is highly specific, but the range of reported sensitivities is wide. Patient or scanner characteristics do not fully account for this variability, but collimation, type of scanner, and mode of imaging explain some of the discrepancy. This heterogeneity raises concerns about consistency of performance and about technical variability. These issues must be resolved before CT colonography can be advocated for generalized screening for colorectal cancer”. In an editorial that accompanied the article by Mulhall, et al., Imperiale (2005) stated that “until we understand more about the factors – both within and among institutions – that are responsible for the varying sensitivity of CT colonography, we should not recommend it as a screening test.” There are also concerns that virtual colonoscopy cannot reliably detect flat lesions, and that flat lesions are more common than previously thought (Soetikno, et al., 2008). An assessment prepared for the Swedish Council on Health Technology Assessment (SBU, 2004) summarized current evidence for CT colonography: “Most studies have shown that CT colonography offers high diagnostic reliability for malignant tumors and polyps 10 mm or larger, inconsistent diagnostic reliability for changes of 5 to 9 mm, and insufficient diagnostic reliability for changes smaller than 5 mm. Some studies, however, have shown unacceptable diagnostic reliability even for malignant tumors and polyps larger than 10 mm.” The assessment stated that CT colonography may compare favorably to barium enema. The assessment explained: “The extent to which CT colonography can replace double-contrast barium enema in patients with disease symptoms has not been completely studied since CT colonography almost exclusively has been compared to colonoscopy. However, the diagnostic reliability of CT colonography compared to colonoscopy appears to be at least equal to the diagnostic reliability of double contrast barium enema compared to colonoscopy.” The assessment explained that one advantage of CT colonography is that, in findings of colon tumors, CT colonography can, in the same examination, also provide information on changes in adjacent tissues and metastases in the lymph nodes and liver. The assessment stated that another advantage of CT colonography is that patients usually experience less discomfort and pain with CT colonography than with conventional colonoscopy and double-contrast barium enema. The assessment noted, however, that it generally is not the examination per se that causes the most discomfort for the patient, but the pre-examination procedure, i.e., the use of laxatives, that is similar in all of the methods of examining the colon. On the other hand, the National Institute for Health and Clinical Excellence (2005) stated that "Current evidence on the safety and efficacy of computed tomographic colonoscopy (virtual colonoscopy) appears adequate to support the use of this procedure provided that the normal arrangements are in place for consent, audit and clinical governance." Guidelines on evaluation of patients with lower gastrointestinal bleeding from the American Society for Gastrointestinal Endoscopy (2005) state that "[v]irtual colonoscopy or computed tomographic (CT) colonography also can be used to rule out a proximal colonic lesion in patients who have had an incomplete colonoscopy." Virtual colonoscopies should only be performed at centers with an appropriate generation of CT scan – a minimum 4 detector CT scanner; collimation of 3 mm or less, overlapping sections at an interval that is two-thirds or less of the collimation, and scan times should be 30 seconds or less in order to minimize respiratory motion. Scholmerich (2003) stated that virtual colonoscopy using CT or MRI does not appear to offer much help in the diagnosis of inflammatory bowel disease (Crohn’s disease). The American Cancer Society guidelines on colorectal cancer screening recommend several methods of screening, including virtual colonoscopy, based in part upon the presumption that the availability of multiple methods of screening will improve compliance (Levin, et al., 2008). Colorectal cancer screening guidelines from the American Cancer Society recommend CT colonography (virtual colonoscopy) performed every 5 years as an acceptable alternative to optical colonoscopy performed every 10 years for screening of average-risk persons. However, there are no studies demonstrating that virtual colonoscopy does, in fact, increase compliance. Virtual colonoscopy is similar to optical colonoscoppy in that it requires completion of a preprocedure cathartic regimen. If a lesion in found on virtual colonoscopy, the patient must return another day and complete another cathartic regimen for an optical colonoscopy to remove the lesion. By contrast, optical colonoscopy allows for identification and removal of a lesion in one procedure. An assessment of CT colonography prepared by the Institute for Clinical and Economic Review (ICER) for the Washington State Health Care Authority (Scherer, et al., 2008) found that, in direct comparison to optical colonoscopy, CT colonography every ten years is substantially more expensive and marginally less effective in preventing cases of cancer (47 vs. 52 in a lifetime cohort of 1,000 individuals) and cancer deaths (24 vs. 26). The investigators reported that only one CT colonography screening strategy is as effective as optical colonoscopy every ten years, and that strategy is to perform CT colonography every five years with colonoscopy referral for polyps greater than 6mm. For this strategy, the cost per life-year gained for CT colonography versus optical colonoscopy was $630,700. Rex and colleagues (2009) updated the American College of Gastroenterology (ACG)'s recommendation on colorectal cancer (CRC) screening. The CRC screening tests are now grouped into cancer prevention tests and cancer detection tests. Colonoscopy every 10 years, beginning at age 50, remains the preferred CRC screening strategy. It is recognized that colonoscopy is not available in every clinical setting because of economic limitations. It is also realized that not all eligible persons are willing to undergo colonoscopy for screening purposes. In these cases, patients should be offered an alternative CRC prevention test (flexible sigmoidoscopy every 5 to10 years, or a CT colonography every 5 years) or a cancer detection test (fecal immunochemical test for blood, FIT). The U.S. Preventive Services Task Force (2008) provided the following statements on screening for CRC: (i) it recommends screening for CRC using fecal occult blood testing, sigmoidoscopy, or colonoscopy in adults, beginning at age 50 years and continuing until age 75 years. The risks and benefits of these screening methods vary (A recommendation), (ii) it recommends against routine screening for CRC in adults 76 to 85 years of age. There may be considerations that support CRC screening in an individual patient (C recommendation), (iii) it recommends against screening for CRC in adults older than age 85 years (D recommendation); and (iv) it concludes that the evidence is insufficient to assess the benefits and harms of CT colonography and fecal DNA testing as screening modalities for CRC. The Canadian Agency for Drugs and Technologies in Health's technology assesssment on CT colonography for CRC screening in an average risk screening population (Ho, et al., 2008) concluded that (i) CT colonography and colonoscopy have comparable sensitivity and specificity for detecting polyps that are 10 mm or greater and for detecting CRC, and CT colonography has lower sensitivity and specificity than colonoscopy for detecting polyps that are smaller than 10 mm, and (ii) colonoscopy is more cost-effective than CT colonography. Colonoscopy leads to reduced disease burden at less cost than CT colonography. Compared with no screening, colonoscopy is likely to reduce disease burden at additional cost and may be perceived as cost-effective. However, increased screening with colonoscopy requires additional infrastructure. If colonoscopy is not available, CT colonography is the next most cost-effective strategy for detecting CRC. The California Technology Assessment Forum (CTAF) concluded that virtual colonoscopy for colorectal cancer screening does not meet CTAF criteria (Walsh, 2009). The CTAF assessment found that the recent results of the ACRIN [American College of Radiology Imaging Network] trial [citing Johnson, et al., 2008] have shown that virtual colonoscopy has diagnostic accuracy comparable to that of optical colonoscopy and importantly that this level of diagnostic accuracy can be achieved in the “real world.” The CTAF assessment explained that the ACRIN study was done in multiple institutions, using multiple bowel preparations and colonoscopy procedures in place at the participating institutions, thereby increasing its generalizability. The CTAF assessment noted, however, that the radiologists were highly trained in how to interpret virtual colonoscopy results. The CTAF assessment stated that, despite the exciting results of the ACRIN trial, several important questions remain before virtual colonoscopy screening can be recommended for widespread use. One unanswered question is how well would virtual colonoscopy screening perform in a setting where the radiologists were not so highly trained. A second unanswered question is what is the clinical impact of the possible harms of the procedure, including radiation risk (especially with virtual colonoscopy repeated periodically) and the high incidence of extra-colonic findings. The CTAF assessment concluded that, despite its diagnostic accuracy, because the impact of the potential harms is not currently known, virtual colonoscopy "is not currently recommended for screening asymptomatic individuals for CRC." On May 12, 2009, the Centers for Medicare & Medicaid Services (CMS) issued a final coverage determination that refused coverage of CT colonography for colorectal screening. It stated that the evidence is inadequate to conclude that CT colonography is an appropriate colorectal cancer screening test. In addition, the BlueCross BlueShield Association Technology Evaluation Center is currently completing an assessment and cost-effectivenss analysis of virtual colonoscopy for colorectal cancer screening. Virtual Upper Endoscopy Virtual upper gastrointestinal (GI) endoscopy is a minimally invasive test that utilizes three-dimensional (3-D) computed tomography (CT) to simulate conventional upper endoscopy images. Potential clinical applications include the evaluation of early gastric carcinoma, advanced gastric carcinoma, leiomyoma, lymphoma, and benign ulcer. For dedicated imaging of the stomach, an oral contrast agent (e.g., water) is administered to opacify and distend the stomach and GI tract and an intravenous contrast agent is used (e.g., Omnipaque 350) for complete evaluation. Virtual upper GI endoscopy has not been studied as extensively as virtual colonoscopy. A limited number of studies have been published and most of these studies have been conducted outside the United States involving small numbers of patients. Early reports of 3-D imaging of the stomach by spiral CT were limited to shaded-surface display (Ogata, et al., 1999). However, the development of multidetector row scanners has improved the visualization of subtle tumors by allowing thinner collimation. The detection rate of gastric lesions using virtual GI endoscopy has been reported to be between 73% - 96.7% in early gastric cancer and between 90% - 100% in advanced gastric cancer (Kim, et al., 2001; Bhandari, et al., 2004). The overall accuracy, sensitivity, and specificity for endoscopic ultrasound and 3-D multidetector row CT in the pre-operative determination of depth of invasion of gastric cancer (T stage) have been reported to be 87.5%, 82.4%, and 96%; and 83.3%, 69.1%, and 94.4%, respectively. The accuracy, sensitivity, and specificity of endoscopic ultrasound and 3-D multidetector row CT for lymph node staging were reported to be 79.1%, 57%, and 89.5%; and 75%, 57.4%, and 89.3%, respectively (Bhandari, et al., 2004). In a prospective study, Kim, et al. (2005) evaluated the accuracy of multi-detector row CT gastrography for the preoperative staging of gastric cancer, with pathologic and surgical results as the reference standard. One hundred six patients with endoscopically proved gastric cancer underwent unenhanced and contrast material-enhanced multidetector row CT gastrography, with effervescent granules used as oral contrast material. Gastric cancer was detected in 92 (87%) of 106 patients with transverse CT imaging and in 104 (98%) with volumetric CT imaging. The overall accuracy of the tumor staging was 77% with transverse CT imaging and 84% with volumetric CT imaging (p < 0.001). The overall accuracy for lymph node staging was 62% with transverse CT imaging and 64% with volumetric CT imaging (p = 0.057). For staging of metastases, there was no difference between transverse and volumetric CT imaging (86% for both) (p > 0.99). The authors concluded that multidetector row CT gastrography with multiplanar reformation and virtual endoscopy, compared with transverse CT imaging, can improve the accuracy of preoperative staging of gastric cancer. This difference was significant for tumor staging but not for the staging of lymph nodes and metastases. A prospective study of the preoperative assessment of gastric cancer tumors using 32-multidetector row CT was carried out by Kikuchi, et al. (2006) on patients (n = 74) with adenocarcinoma of the stomach (T1 tumors, n = 38; T2 and T3 tumors, n = 36). In 35 (47%) out of the 74 patients, the primary lesions could be detected on 2-D images obtained by CT. In these patients, virtual endoscopic images of these tumors could be created. Twenty-seven advanced cancer tumors (75%) were assessed based on 2-D CT images and 27 larger tumors (greater than 40 mm) (69%) were assessed based on 2-D CT images. Significant differences were found with respect to depth of tumor (p < 0.0001) and tumor size (p < 0.0001) between tumors that could or could not be assessed on multidetector CT. The authors concluded that future studies are required to fully explore the ability of multidetector CT to assess tumor volume in advanced gastric cancer cases and to determine the optimum application of this approach. In a prospective study, Mazzeo, et al. (2004) assessed the diagnostic capabilities of multidetector CT in various esophageal pathologic conditions. Thirty-three patients underwent a multidetector CT study after esophageal distention by means of effervescent powder administered after induction of pharmacologic esophageal hypotonia. All acquired images were post-processed with 2-D and 3-D software tools. The CT data were compared with the results of conventional radiology (n = 33), endoscopy (n = 28), endoscopy ultrasonography (n = 14), or surgery (n = 14). Follow-up ranged between 4 and 15 months. Final diagnoses were leiomyoma (n = 6), squamous cell carcinoma (n = 6), adenocarcinoma (n = 4), esophageal infiltration by thyroid cancer (n = 1), benign polyposis (n = 2), chronic esophagitis (n = 5), post-sclerotherapy stenosis (n = 1), and no abnormalities (n = 7). Pathologic wall thickening was observed in 25 of 33 cases (76%), with values ranging between 3.6 and 36 mm (mean, 9.6 mm). Spiral CT demonstrated 21 true positive cases, and seven true negative cases. There were four false negative cases and one false positive case. Sensitivity was 84%, specificity was 87%, diagnostic accuracy was 85%, positive predictive value was 95%, and negative predictive value was 64%. The authors concluded that evaluation of the esophagus with multidetector CT is a promising technique and easy to use, allowing panoramic exploration, virtual endoluminal visualization, accurate longitudinal and axial evaluations, and simultaneous evaluation of T (tumor penetration) and N (lymph node involvement) parameters. The first virtual gastroscopy study in North American patients assessed the feasibility of performing virtual gastroscopy on 10 patients with no reported GI abnormalities. The authors stated that the anatomy of the stomach including the lumen, the cardia, the pylorus, gastric folds and the incisura angularis were well-visualized using 3-D spiral CT. It was not possible to visualize the esophagus by virtual endoscopy because of difficulties keeping the lumen patent long enough to provide accurate imaging. The authors concluded that further development is needed before virtual gastroscopy can be considered for clinical application (Ezzeddine, et al., 2006). Virtual gastroscopy is also being evaluated to assess the gastric mucosa in patients who have undergone laparoscopic Roux-en-Y gastric bypass. One small case series reported promising results (Alva, et al., 2008). Conventional upper GI endoscopy provides direct visualization of the mucosa, permits evaluation of color changes that may be indicative of pathology, and suspicious lesions can be biopsied and the tissue sample evaluated histologically. While virtual upper GI endoscopy using CT is a promising method for the detection and evaluation of upper GI lesions, randomized controlled studies comparing it to conventional upper GI endoscopy are needed to determine its clinical value. Appendix Virtual colonoscopies should only be performed at centers with an appropriate generation of multi-detector CT scan – a minimum 4 detector CT scanner; collimation of 3 mm or less, overlapping sections at an interval that is two-thirds or less of the collimation, and scan times should be 30 seconds or less in order to minimize respiratory motion (Taskar, et al., 1995; Scherer, et al., 2008). In addition, scans should be read by trained readers by virtue of having read least 30 CT scans (Halligan, et al., 2005; ACR, 2006).
|
Close Window
