AGA Future Trends Committee Report: Colorectal Cancer: A Qualitative Review of Emerging Screening and Diagnostic Technologies

GASTROENTEROLOGY 2005;129:1083–1103

AMERICAN GASTROENTEROLOGICAL ASSOCIATION AGA Future Trends Committee Report: Colorectal Cancer: A Qualitative Review of Emerging Screening and Diagnostic Technologies CAROL R. REGUEIRO Pittsburgh, Pennsylvania

Part 1. The AGA Future Trends Commitee–Charge and Process The AGA Future Trends Committee (FTC) was created in 2004 to further the AGA Strategic Plan by identifying and characterizing important trends in clinical practice and scientific-technological developments in the world in general, and medicine and gastroenterology in particular, that potentially will impact the AGA and/or its members in the coming 3–5 years or beyond and to make strategic recommendations to the Governing Board on how AGA should deal with those trends and developments. These trends and developments may be economic, demographic, practice-based, scientific/technological, or political in nature. pecifically, the committee is charged with preparing a report (or reports) for the AGA Governing Board that describes the trends or developments it has identified, postulates their impact on gastroenterology practice and/or research as appropriate, and presents specific recommendations for action by the AGA in terms of policy and programs. The committee is also asked to monitor these trends and technologies as they play out over time. In July 2004, the AGA Leadership Cabinet suggested several topics that the FTC should address. Realizing that the FTC could not realistically consider all of them, criteria were developed to prioritize the topics and others that might be added in the future. These criteria were:

S

● Time variable, ie, “When will gastroenterology be affected?” ● Scale and magnitude ● Does the trend or development represent a threat and/or opportunity to gastroenterology? ● Effect on patient care quality and safety ● Effect on AGA members and the AGA per se ● Implications to reimbursement ● Impact on gastroenterologists’ training and education

In October, a crude Delphi process was used to determine the trends and developments that should be the focus of the committee’s work. Committee members were asked to assign priority scores to the items in the following list that was based on the Leadership Cabinet’s suggestions and supplemented by AGA staff and others. This process was done via the mail. ● The application of genomic and proteomic technologies to digestive disease diagnosis and treatment ● Major changes in the US healthcare system and reimbursement ● Increased median age of the population ● Changes in the ethnic and racial makeup of the US population ● Patients’ involvement in their own care ● New colorectal cancer screening and diagnostic technologies ● Biomedical research funding changes ● Changes in academic health centers ● Changes in physician education and training ● Obesity-related disease incidence and prevalence ● Computerization and digitization of gastroenterology practice Committee members were asked to score each item against each of the priority criteria noted above using a scale of 1 ⫽ large effect to 3 ⫽ small effect (on gastroenterology practice and research). The total scores of each topic were then summed and ranked. The four highest priority scores that resulted from this ranking were: 1. New colorectal cancer screening and diagnostic technologies 2. Obesity-related disease © 2005 by the American Gastroenterological Association

0016-5085/05/$30.00 doi:10.1053/j.gastro.2005.06.012

1084

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

3. Aging of the population 4. Genomic and proteomic technologies Because the AGA was already investigating the ramifications of the obesity epidemic, the FTC decided to concentrate on the other 3 topics. The committee determined that preparing the 3 reports on its own was not feasible. Hence, it decided that it would solicit proposals from potential qualified authors to draft the reports and would modify and supplement the drafts as necessary. A request-for-proposal (RFP) was prepared and disseminated in December 2004. The authors, who were paid for their work, were chosen by the committee from among the responses to the RFP. The manuscripts submitted by the authors were reviewed by the committee in February 2005. Among the changes to the draft reports were recommendations for action by the AGA–these were developed primarily by the committee and forwarded to the Governing Board. At its review meeting, the committee also developed a uniform format for the three reports. Revised manuscripts based on the committee’s critiques were completed in March 2005. The committee also had each report evaluated by an outside expert reviewer for completeness and to ensure that the authors had not made any egregious error that may have been overlooked. The Future Trends Committee’s recommendations for action by the AGA regarding new and potential CRC screening and diagnostic test methods were forwarded to the Governing Board. This report provides the background against which its recommendations were developed. It does not represent the Committee’s final word on the subject. Given its importance and relevance to gastroenterology practice, and the dynamic state of the technology, this topic will be reviewed periodically by the Committee.

Part 2. Executive Summary Colorectal cancer is the second leading cause of cancer death in the United States. It can be prevented in many cases, however, with early detection and removal of polyps. The effectiveness of this strategy is increasingly evident, and colorectal cancer screening is endorsed by physician associations and public health advocates, as both effective and cost-effective. There is less consensus regarding optimal screening strategies, as sensitivity, specificity, and patient acceptance limit current options. Investigators have evaluated new tools in attempts to overcome these barriers. A range of approaches, including proteomics-based testing, stool genetic testing, ra-

GASTROENTEROLOGY Vol. 129, No. 3

diologic imaging, and enhanced endoscopy has been the focus of intense research. Each of these promises potential patient benefits. Most, however, remain years away from routine clinical application. Proteomics, or the analysis of broad protein patterns, holds the greatest promise to radically alter medical diagnostics in general, including colorectal cancer screening. With such technology, it is theoretically possible to assess small amounts of protein for the presence of identified cancer markers using new, high-throughput protein assessment tools, and computerized artificial intelligence analysis. The benefits of such an approach are numerous, and include its rapid, noninvasive nature, and potential low cost. Practical challenges, such as marker identification and confirmation, and optimization of the testing platform, will likely require nearly a decade of work to overcome. When available, however, such tests will bring significant change to the practice of gastroenterology. New endoscopic tools, such as high magnification chromoscopic endoscopy, spectroscopy, and optical coherence tomography, are rapidly evolving adjuncts to traditional colonoscopy. High magnification chromoscopic colonoscopy has demonstrated improved sensitivity and yield in the surveillance of high-risk patients with ulcerative colitis. Spectroscopy may allow future endoscopists to more readily identify suspicious lesions, and tools such as optical coherence tomography may afford real-time evaluation of their histologic significance. Enhanced colonoscopy will likely enter the practice of specialized gastroenterologists within the coming years. Its broader adoption, however, will rely on convincing data demonstrating significant incremental patient benefits, especially in light of the training and additional procedure time such tools will require. Stool DNA tests have held great promise as a more sensitive, noninvasive screening tool. Despite early positive results, data from a recent large screening trial indicate that among asymptomatic patients, its sensitivity remains relatively poor. This coupled with its significant cost renders it unlikely to replace fecal occult blood testing in the near-term. Among new colorectal cancer screening tools, new radiologic techniques, termed virtual colonography, have received the majority of attention from the academic and lay press. Magnetic resonance colonography remains in its infancy. Computed tomographic colonography, however, is clinically available and offers potential near-term utility. Computed tomographic colonography offers high sensitivity for large (⬎10 mm) lesions and has rapidly improving ability to detect medium-sized polyps. Issues regarding optimal technique, patient preparation, and

September 2005

radiologist training are the subjects of on-going research. Cost remains the most substantial barrier to wider adoption, especially as most patients with positive results will need subsequent conventional colonoscopy. It is likely in the near term that computed tomographic colonography will play a screening role in patients unwilling to undergo existing screening or who are unlikely to require subsequent colonoscopy. Unanswered questions limit the use of computed tomographic colonography in general screening programs. These include: (1) What is the relevance of small or medium polyps? (2) What is the relevance of flat lesions? (3) What are the appropriate evaluation and attendant costs of extracolonic findings? (4) Will patients truly prefer this screening approach? and (5) What will the ultimate costs of such tests be? The range of new approaches to colorectal cancer screening is broad and promising. None appears poised to radically alter the lives of patients or their doctors over the next 2 to 5 years. Some strategies, such as computed tomographic colonography and enhanced endoscopy, will gradually enter practice over the coming years. These tests will likely be part of daily practice for a minority of gastroenterologists in the short term. All gastroenterologists, however, will need a broad understanding of them and other emerging tests to effectively counsel patients and other medical colleagues on these growing options. The number of effective strategies for colorectal cancer screening will continue to grow and may encourage greater numbers of patients to undergo screening. As the population increases and ages, colorectal cancer screening will remain a large part of the daily practice of gastroenterology.

Part 3. Literature Review I. Overview A. Introduction. Colorectal cancer (CRC) is the second leading cause of cancer death in the United States. In 2004, 106,370 new cases of colon cancer were diagnosed.1 Furthermore, its prevalence increases with age; thus, as the U.S. population ages, the number of colorectal cancer cases can be expected to rise. Research into reduction and prevention of CRC has focused on lifestyle modification and identification of earlier, potentially curable cancers and precancerous lesions. While better understanding of the role of lifestyle in CRC may offer significant long term benefits, more tangible gains have been made through CRC screening programs. Colorectal cancer screening is effective and the incidence of CRC has dropped in recent years, possibly attributable to screening programs.2 United States Pre-

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1085

ventive Services Task Force strongly recommends CRC screening and places it among a small number of cancer screening strategies for which good data support its use.3 Current colonoscopic techniques rely on thorough bowel preparation and inspection to function optimally. The efficacy of current CRC screening is predicated on identification and removal of early stage malignancies, or ideally, precancerous polyps. Polypectomy and subsequent surveillance are currently the cornerstones of CRC prevention. Issues of sensitivity, specificity, and patient acceptance limit existing CRC screening methods. Compared with screening for other common malignancies, such as breast and prostate cancer, CRC lags behind. Despite randomized trials demonstrating its ability to reduce CRCrelated mortality,4,5 the use of fecal occult blood testing among screening-eligible patients remains low. In recent data, only 21.8% of screening-eligible survey respondents reported fecal occult blood testing within the prior year.6 Even if widely adopted, Helm et al note, due to its poor sensitivity, fecal occult blood testing would save the lives of ⱕ15% of those who would otherwise succumb to colon cancer within the first decade of a screening program.7 Endoscopic screening is increasingly used in the U.S. due to superior sensitivity and specificity. Despite its gold standard status, lesions can be missed at colonoscopy. In one report, researchers estimated that as many as 4% of right-sided lesions were undetected at colonoscopy.8 Missed lesions are particularly concerning for patients at high-risk, such as those with inflammatory bowel disease, in whom visual surveillance can be hampered by extensive inflammation and the presence of pseudopolyps. Furthermore, colonoscopy is associated with uncommon but significant complications even in experienced hands. Patient acceptance remains a barrier to endoscopy; in recent data, only 40.5% of eligible patients reported either flexible sigmoidoscopic or colonoscopic screening within the previous 5 years.6 Poor acceptance stems from many factors, including dietary restrictions or burdensome cathartic preparation, the invasiveness, perceived discomfort, and risks of the test, and the anesthesia and recovery time associated with colonoscopy.9 Attempts to devise more sensitive, specific, and acceptable screening methods have increased as new technologies have emerged to exploit a growing understanding of CRC pathogenesis. New screening strategies can be broadly divided into (1) noninvasive tests performed on stool or serum, (2) less invasive radiologic tests, and (3) improved endoscopic techniques with potentially better sensitivity and specificity.

1086

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

New, noninvasive strategies could potentially improve patient acceptance; these include new and refined stool screening tests and efforts to develop serum-based tests. Radiologic colonic imaging, such as computed tomographic and magnetic resonance imaging are the focus of intense research efforts. Enhancements of traditional endoscopy include magnification endoscopy, optical coherence tomography, and spectroscopy, which may offer improved sensitivity and specificity. This paper qualitatively reviews emerging new technologies and their potential application to CRC screening. Detailed information and supporting data are discussed for those technologies currently available or anticipated to be clinically available within the next 2–5 years. In conclusion, the implications of such tests on gastroenterology training and practice are reviewed. B. Methods. A comprehensive literature search was conducted using OVID and PubMed to identify English language articles published between 2000 and 2005 with relevance to diagnosis of colorectal cancer. This initial search was augmented through review of cited publications, abstracts, and presentations from recent gastroenterology meetings. C. New diagnostic technologies and targets: applicability to CRC screening. Detecting malignancy early

in its course requires detection of cancer-specific markers. Potential markers can reflect any stage of the disease process including genetic alterations, and abnormalities in transcription, translation, and protein expression. Identification of markers has been challenging. Potential markers are frequently present in small amounts compared with substances expressed by normal tissues. Additionally, techniques to identify specific proteins have been time and labor intensive. Enzyme-linked immunosorbent assay systems, the most widely available, reliable, and accurate systems to date, are limited to a single target.10 Other techniques, such as 2-dimensional polyacrylamide gel electrophoresis (2-D PAGE) coupled with mass spectrometry require large amounts of protein substrates. A refinement of this technique, 2-dimensional difference gel electrophoresis (2D-DIGE), incorporates fluorescent labeling to improve reproducibility, sensitivity, and quantitative aspects of gel analyses. Significant advancements, however, in the fields of proteomics, epigenetics, and nuclear matrix proteins are poised to improve the identification of biomarkers, and to facilitate their use in novel diagnostic strategies. 1. New techniques to identify proteins. ● Proteomics Proteomics has been defined as the “large-scale analysis of the expressed protein complement of the ge-

GASTROENTEROLOGY Vol. 129, No. 3

nome.”11 Proteomic techniques allow simultaneous assessment of many proteins. Through the use of bioinformatics systems, this information can be rapidly assessed leading to more rapid identification of cancerassociated biomarkers. The same technology can be used to detect such markers and distinguish cancerous from noncancerous tissues. While these techniques can be applied to protein obtained from any source, much of their application for CRC screening has explored serumbased tests. ● Surface-enhanced laser desorption ionization timeof-flight The use of surface-enhanced laser desorption ionization time-of-flight (SELDI-TOF) mass spectroscopy has greatly advanced proteomics. In SELDI-TOF, proteins from a patient sample are bound to a chip, which, after processing, is subjected to laser irradiation in a vacuum.10 The irradiated proteins discharge from the chip as charged ions. Proteins are separated by their time of flight, a property determined by mass-to-charge ratios. Thus, multiple data points are generated for each sample, information that has been likened to a unique “fingerprint.”10 Automated, artificial intelligence systems can rapidly detect abnormal proteins. SELDI-TOF affords many theoretical advantages: it can be performed on minute amounts of serum; it can accommodate high through-put, and may be very inexpensive. Some estimates suggest it may cost as little as $10 per sample compared to $200 –$1000 for serum-based genetic tests.12 Preliminary clinical application of SELDI-TOF has been promising. In a study reported by Petricoin et al, this technique correctly identified 100% of ovarian cancer patients, including those with stage I disease, and correctly identified 95% of healthy controls.13 Other researchers reported that SELDI-TOF enabled highly accurate discrimination between patients with benign prostatic disease and prostate cancer.14 Ornstein et al studied 154 men with prostate specific antigen (PSA) levels between 2.5 and 15.0 ng/mL and/or abnormal digital rectal exam before transrectal ultrasound guided biopsy. Approximately half of their data was used to create a diagnostic algorithm; the remainder was used in a blinded manner to test its function. They report a sensitivity of 100%, and note that such a test could obviate unnecessary biopsies in such patients. SELDI-TOF has been applied to CRC research as well.15–17 Drake and colleagues reported impressive results from a small study, in which SELDI-TOF distinguished cancers from adenomas with sensitivities of 91%–95%, and specificities of 80%–95%.16 More re-

September 2005

cently, Chinese researchers reported that, using a panel of 4 potential biomarkers, they could discriminate patients with colorectal cancer from those with adenomas with a specificity of 92%, sensitivity of 89%, and a positive predictive value of 86%.17 Despite this great promise, specific concerns regarding SELDI-TOF, including reproducibility and validity, remain. Additionally, new approaches to high-throughput protein analysis have recently been explored, including matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry and tandem hybrid mass spectrometry (Q-Star). Significant barriers remain between the promise and clinical applicability of these proteomic techniques. The optimal approaches to both protein and data analysis are subjects of much debate. These issues, as well as identification and confirmation of the most relevant CRC protein patterns are likely to delay a clinically relevant proteomic-based CRC screening test for a decade or more. 2. Potential new CRC diagnostic targets. ● Nuclear matrix proteins Nuclear matrix proteins are another area of intense research. Nuclear matrix proteins provide the structural framework for the cell nucleus and are the sites of messenger RNA transcription. Most nuclear matrix proteins are common to all cell types; some, however, are tissue and cell-line specific. Aberrant nuclear matrix proteins have been associated with many cancers, including esophageal, bladder, and CRC.18 –20 Nuclear matrix proteins are found in the urine and serum of cancer patients. A nuclear matrix protein-based urine assay for bladder cancer had reported sensitivity and specificity of over 96% and 100%, respectively.21 In 2004, researchers identified CRC-associated nuclear matrix proteins that were notably absent from adjacent tissue and normal colon tissue.22 Subsequent analysis identified specific nuclear matrix proteins that were unique to frank cancer and adenomatous polyps and could potentially discriminate between the two. Another NMP appeared to be expressed at the point of cancer invasion, indicating that nuclear matrix proteins may illuminate specific stages along the adenoma-carcinoma pathway. These markers may provide a basis for serumbased CRC screening and surveillance. Further analysis of these nuclear matrix proteins and development of antibodies towards them are currently underway. ● Minichromosome maintenance proteins In eukaryotic cells, minichromosome maintenance proteins are crucial to the initiation of DNA synthesis. As cells differentiate, the expression of these proteins is

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1087

greatly down-regulated.23 Minichromosome maintenance protein expression has been targeted as a potential marker of unregulated proliferation and dysplasia. In gastroenterology, minichromosome maintenance proteins have been primarily examined in relation to Barrett’s esophagus (BE).24,25 Sirieix et al report successful use of minichromosome maintenance proteins to detect dysplasia and adenocarcinoma in BE.25 Using archived tissue specimens and prospectively obtained endoscopic cytological brushings, the presence of minichromosome maintenance proteins was assessed and correlated with histopathologic diagnosis. Minichromosome maintenance protein expression was unique to dysplastic and cancerous BE cells. Specimens from the gastric antrum, normal esophagus, and duodenum had no detectable minichromosome maintenance protein. Of note, in patients who eventually developed adenocarcinoma (AC), biopsies taken of Barrett’s esophagus prior to detectable dysplasia demonstrated elevated minichromosome maintenance proteins. Minichromosome maintenance protein expression correlated strongly with histopathologic diagnosis of AC or dysplasia. In CRC, the use of minichromosome maintenance protein markers faced an additional hurdle: extraction of protein from stool. Davies and colleagues retrieved minichromosome maintenance-positive colonocytes in 37/40 samples obtained from patients with CRC.23 No minichromosome maintenance-positive cells were found in 25 healthy controls. Of note, however, timing and sample storage were crucial to cell retrieval. Colonocyte retrieval was adequate up to 8 hours after test application. Samples kept at 0°C yielded significantly more cells than those kept at either 21°C or 37°C. These technical aspects currently limit its clinical applicability. If overcome, minichromosome maintenance protein expression may have future utility in noninvasive CRC screening strategies. ● Messenger RNA-based tests RNA targets have also been the focus of promising research, and cancer associated RNA has been detected in serum, urine, and stool samples.26 –29 One target, telomerase, a ribonucleoprotein enzyme associated with cell division, is highly activated in a variety of malignancies.28 Melissourgos et al exploited this fact using a messenger RNA-based assay for human telomerase reverse transcriptase (hTERT) to detect transitional cell carcinoma (TCC) and other urothelial neoplasms.27 Voided urine samples from 146 patients diagnosed with TCC and urothelial neoplasms and 128 controls with either benign or no urologic disease, were tested with

1088

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

both an mRNA based assay and conventional cytology. The mRNA assay performed better than cytology in discriminating between the two groups, with sensitivity of 92% and specificity of 96%. The assay was bettersuited to differentiating between benign and low-grade TCC than cytology, with sensitivities of 93% and 28% respectively. Niiyama et al demonstrated that hTERT was significantly elevated in CRC and adenomas.28 Davidson et al reported the ability of a stool-based messenger RNA (mRNA) assay to detect putative CRC biomarkers extracted from exfoliated colonocytes in a human pilot study.29 In a subsequent study of 29 patients with and 22 without CRC, an mRNA COX-2 assay applied to processed stool samples had a sensitivity of 90% (95% confidence interval [CI] 73%–98%) and specificity of 100% (95% CI, 85%–100%).30 Detection of cyclooxygenase-2 (COX-2) mRNA has been an important target of these assays. Overexpression of the COX-2 gene has been found in a number of aerodigestive tumors and it is overexpressed in ⬎80% of CRC compared with normal controls.30,31 Its presence in cancers besides CRC has raised broader screening possibilities and concerns. Such a marker would allow broad screening for cancers above the level of the colon; however, specificity would suffer. Furthermore, the appropriate follow up of positive results for which no colorectal dysplasia can be found could be problematic.32 In addition, RNA is much more susceptible to degradation compared to DNA. Careful studies of reproducibility and repeatability will be needed to validate an RNA-based biomarker. ● Epigenetics Epigenetic processes are those that involve generelated phenomena but do not involve changes in the DNA sequence. These include “gene silencing” (the inactivation of 1 or 2 copies of a gene), and loss of imprinting (LOI), whereby a normally inactivated copy of a gene is activated. Somatic inactivation of a DNA mismatch repair gene can lead to genetic instability as replication errors go unchecked. This “microsatellite instability” plays a primary role in hereditary nonpolyposis colorectal cancer (HNPCC) and accounts for an estimated 15% of sporadic CRC.33 Thus, these epigenetic changes are increasingly the target of proteomic approaches to CRC screening, particularly among those at high risk. Hyper- and hypo-methylation of DNA is another promising epigenetic cancer marker. Methylation can directly influence transcription of the affected gene and

GASTROENTEROLOGY Vol. 129, No. 3

has been linked to tumorigenesis.34 In addition, methylation patterns have been likened to “epigenetic signatures” for subgroups of colorectal tumors35 and hold future promise as targets for noninvasive, potentially simple, and inexpensive tests for CRC.36,37 LOI of the maternal gene for insulin-like growth factor (IGF) has been described in a number of cancers, including colon cancer; however it has not yet proved sensitive or specific enough to serve as the basis of a CRC screening test. In a study of 172 people undergoing colonoscopy, Cruz-Correa et al found LOI of IGF-2 was increased more than 5-fold among those with colorectal neoplasia.38 Furthermore, LOI of IGF-2 was not associated with environmental exposures such as alcohol, tobacco, or nutrients. LOI of IGF-2, however, did not discriminate accurately enough between those with and those without CRC. LOI of IGF-2 was found in 47/62 patients with and in 7/107 without CRC. It appears that this marker may be poised to join the ranks of others under investigation as risk-stratification, not screening, tools. While the technologies described previously are poised to rapidly advance medical diagnostics in general, and CRC screening in particular, none are yet ready for clinical application. The number of new clinically available diagnostic technologies is decidedly smaller, and includes enhanced endoscopy, particularly suited to screening high risk patients, and new stool and radiologic tests with potential broader screening uses. II. New Endoscopic Tools–Potential for Use in High-Risk Screening Endoscopy both directly and as a follow up to abnormal stool testing has become an integral part of CRC screening. A limitation, however, is the need for visual inspection of all surface areas. Often early dysplasia is not readily apparent, leading to random biopsies especially for those with increased cancer risk. Conversely, many abnormalities are benign, and are likely to remain so. It is often difficult to distinguish between these and precancerous lesions. New endoscopic techniques appear to greatly increase the sensitivity, specificity, and yield of colonoscopy. A. High-magnification chromoscopic colonoscopy. High magnification chromoscopic colonoscopy in-

volves the use of a highly magnified image to reveal detailed structure of colonic epithelium. Chromoendoscopy refers to the addition of colorizing agents to the gastrointestinal tract to enhance visualization of surface features.39 Magnification, which can range up to 1125-

September 2005

fold, can reveal dysplastic features not seen by the unaided eye.40 Magnification is not needed to detect masses or polypoid lesions, but greatly improves detection of flat lesions, and abnormal surface pit patterns, including the presence of aberrant crypt foci (ACF). ACF have gained increasing attention as an important cancer precursor. ACF, first described in the mid-1980s, are mucosal crypts with a larger, thicker, and darker microscopic appearance than normal.41 These abnormalities were initially identified in animals exposed to carcinogens, and subsequently associated with a range of bowel diseases, including CRC.42,43 Identification and classification of pit topography in vivo, using a magnifying endoscope, was pioneered in Japan. High magnification chromoscopic colonoscopy and pit classification, formalized by the Kudo classification system includes 5 groupings that range from benign to invasive neoplastic lesions.44 Further, ACF may be found at sites distant from cancer and dysplasia, and are thought to reflect “field effects” whereby the effects of carcinogenic insults can be detected concurrently throughout the colon.45 Besides revealing surface irregularities such as depression or minor elevation, magnification allows detection of more subtle changes such as mucosal pallor or erythema, and changes in vascular pattern.46 In high magnification chromoscopic colonoscopy, such changes prompt application of enhancing dyes to better clarify architecture and guide biopsy. A number of stains have been used and fall into 2 groups: contrast dyes and staining dyes. Contrast dyes, which include indigo carmine and cresyl violet, pool in grooves and highlight surface irregularities.42 These stains can be sprayed onto the mucosa with a special catheter or may be ingested by the patient. Indigo carmine is a blue dye that is applied topically to the mucosal surface in an aqueous solution of 0.2%–2.0% concentration (volume/volume). After a potential lesion is identified through the magnified view, the mucosal surface is flushed with water to remove secretions and dye is applied through the endoscope’s biopsy channel in volumes of 3–5 mL followed by 15 mL or air. Indigo carmine is nontoxic and not systemically absorbed.41 This technique has been primarily used in the identification of flat adenomas or cancer.39 Staining dyes color the convex or protuberant portion of a lesion; contrast dyes pool in spaces between protuberances. This leads to significantly different appearances depending upon the type of dye used.46 Additionally, staining dyes generally are not easily removed, and do not allow for subsequent use of contrast agents. The most commonly used vital or reactive stains include Lugol’s solution, methylene blue, and toluidine blue. These stains are applied after a mucosal surface wash of pro-

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1089

teinase solution. Small amounts of dye are applied directly via a specialized catheter and excessive dye is aspirated. After dye fixation of 2–3 minutes, the magnified image is viewed.41 Lugol’s solution and toluidine blue have been used to detect dysplasia in Barrett’s esophagus. In the colon, methylene blue has been used to detect flat adenoma and carcinoma, and to distinguish between hyperplastic and adenomatous polyps.39 Flat and depressed adenomas have been the focus of much chromoscopic colonoscopy. They are likely more prevalent than initially estimated and studies suggest that they may have greater malignant potential than polypoid lesions.47 In an English study of 1000 patients referred for routine colonoscopy, Rembacken et al reported that over one third had flat lesions detected by chromoscopic colonoscopy.48 Subsequent histological examination demonstrated that lesions were more likely to be dysplastic or frankly cancerous. Using chromoscopic techniques in an American population, Saitoh et al noted similar findings.49 In this prospective study of 211 patients, 22.7% had flat or depressed lesions; 80% of these were adenomatous. The authors conclude that chromoscopic techniques could aid in triage of lesions into high and low risk. They add, “American endoscopists should consider incorporating the use of a simple dye-spray method” into their practice. While high magnification chromoscopic colonoscopy is widely used in Japan for diagnosis of non-colitic dysplastic lesions,50 its use remains uncommon in the United States. Barriers to clinical adoption may include the lack of familiarity with the procedure, additional procedure time, and the lack of proven incremental benefit when added to traditional colonoscopic techniques. The technology is readily available, however, and could be adopted by interested clinicians. Training in the use of magnification endoscopy would be critical. Even slight unintended movement of the magnified field can significantly blur the images obtained.39 It is plausible that high magnification chromoscopic colonoscopy will gain wider use in surveillance of patients at particularly high risk, such as those with ulcerative colitis. Encouraging results have been obtained in this population. In a large, randomized controlled trial of chromoendoscopy, Kiesslich et al demonstrated that chromoendoscopy was more sensitive (93%) and specific (93%) in discriminating between nonneoplastic and neoplastic lesions in patients with chronic ulcerative colitis than conventional surveillance.51 In a subsequent prospective study of 162 patients with chronic ulcerative pancolitis, Hurlstone et al reported improved detection of neoplasia and dysplasia using high-magnification chromoscopic colonoscopy and indigo carmine dye.52

1090

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

High-magnification chromoscopic colonoscopy successfully differentiated between nonneoplastic groups, defined as having mucosal crypt patterns types I/II and neoplastic groups, with mucosal crypt patterns IIIs/IIIL and V. Sensitivity was 97%; specificity was 93%. Importantly, diagnostic yield in this high-risk group was 4 times greater than that of conventional colonoscopy and biopsy techniques. Another refinement to high-magnification chromoscopic colonoscopy is narrow-band imaging. Narrow band imaging entails the replacement of conventional broadband, red/green/blue (RGB) light filters by specific, narrow band imaging filters that result in illumination in a narrow wavelength range. This increases the contrast value of vascular patterns.53 In recently reported preliminary results, narrow-band imaging aided in the diagnosis of dysplasia in patients with chronic ulcerative pancolitis.54 In comparison to conventional colonoscopy and biopsy, narrow-band imaging resulted in fewer biopsies and yielded better sensitivity. Its application as an adjunct to other endoscopic techniques is undergoing further investigation for many gastrointestinal diseases. B. Spectroscopy. Many new endoscopically based diagnostic technologies exploit the unique interaction between light and target tissues. Limitations of traditional white-light endoscopy have been overcome by new optical technologies that employ fluorescence and different light sources to improve sensitivity. 1. Light-scattering spectroscopy. Light scattering spectroscopy is based on the idea that light absorption and scattering is related to the composition of the tissue being examined. The intensity of light that is elastically scattered, or scattered without a change in wavelength, reflects the properties of molecules in the tissue that absorb and scatter light.55 Early iterations of light scattering spectroscopy were not detailed enough to detect subtle changes associated with carcinogenesis. In a refinement called 4-dimensional elastic light scattering fingerprinting (4D-ELF), Roy et al describe the use of a “new generation” of this technique.55 The investigators note that 4-D ELF allows measurement of 4 characteristics of light: “wavelength, scattering angle (ie, the angle between the backward direction and the direction of the propagation of scattered light), azimuthal angle of scattering (ie, the angle between the incident light polarization and the projection of the direction of the scattered light propagated onto the place in which the incident electric oscillates), and polarization of scattered light.”55 The data obtained through this technique allow resolution of structures measured in nanometers. Tissues can be probed to greater depth than with conventional microscopy, and detail 20 –50 times greater than that of light

GASTROENTEROLOGY Vol. 129, No. 3

microscopy, can be obtained. The authors note that such resolution would allow detection of histological evidence of preneoplastic changes, or field effect. The investigators tested this hypothesis by using 4-D ELF to assess for the presence of ACF in a validated rat model of colorectal carcinogenesis. In this study, rats subjected to the carcinogen azoxythmethane were compared to saline-treated controls. The capacity for 4D-ELF to detect earliest stages of carcinogenesis, ACF, was compared with that of a blinded observer. The investigators noted that as early as 2 weeks after carcinogen exposure, 90% of probed areas in the distal colon were correctly classified as having exposure. This number increased to 100% at weeks 12 and beyond. Specificity was 100%. The authors also noted that ACF were detected as early as week 4 after carcinogen exposure and that numbers of ACF increased progressively over time. An important finding was that ACF occurred earlier and with greater frequency in the distal colon, lending credence to the presence of ACF as reliable markers of field effects. The authors caution that the azoxythmethane-treated rat model of carcinogenesis may not accurately reflect human colon cancer development. Right-sided field changes may not be reflected in the left colon. As the investigation of appropriate CRC markers proceeds, it is possible that other markers may be the focus of this technology. 2. Light-induced fluorescent spectroscopy. Light-induced fluorescence (LIFS) has been extensively used in cancer diagnosis, though its application to screening has been more recent. With this technique, low-power laser light is directed towards a tissue surface, and induces autofluorescence of endogenous chemicals.39,56 Changes between normal, precancerous, and cancerous tissues are reflected in alterations of the emitted spectra. Many studies in the early 1990s indicated great potential for light-induced fluorescent spectroscopy.57,58 In a blinded, human study, light-induced fluorescent spectroscopy demonstrated 90% sensitivity and 95% specificity in identifying colonic dysplasia.54 In 1996, Wang et al demonstrated that real-time, fluorescence imaging in vitro could reliably identify dysplastic tissue.59 The majority of initial studies used monochromatic laser energy, an expensive and often complicated technology. In a recent study, German investigators reported good in vivo results using a more practical, incoherent violet blue light as excitation energy obtained by a standard xenon-arc lamp.60 Twenty-three patients with high risk for CRC or dysplasia underwent light-induced fluorescent spectroscopy. Lesions were classified using a mathematical algorithm applied to the light-induced

September 2005

fluorescent spectra and compared to the biopsy results, considered the “gold standard.” Light-induced fluorescent spectroscopy identified CRC with sensitivity of 96% and specificity of 93%. It also performed well for the detection of dysplastic adenomas, with sensitivity and specificity of 98% and 89% respectively. General concerns remains about using light-induced fluorescent spectroscopy for cancer screening. Inflammation can cause a red fluorescence that is difficult to differentiate from dysplasia or cancer.39 Additionally, the relationship between the source of excitation light and the tissue needs to be fixed and constant to function optimally. As Pfau and Sivak note, “the human colon is never this accommodating.”39 Researchers have attempted to further hone the use of fluorescent molecules through the development of socalled “molecular beacons.”61 Researchers have developed an optically based protease activatable fluorescent sensor with the goal of using this sensor to highlight dysplastic tissue. Using a mouse model of familial adenomatous polyposis (aged ApcMin/⫹), investigators demonstrated that a novel, activatable, capthepsin-B sensing nearinfrared fluorescence (NIRF) imaging probe helped identify adenomatous polyps. This agent, given as an intravenous injection, fluoresced when exposed to confocal laser light, allowing identification of lesions as small as 50 ␮m in diameter. The authors note that such technology would allow the development of a range of probes that could “type” lesions through simultaneous probe imaging. Additionally, such technology could be adapted for detection in both conventional endoscopy or through external noninvasive radiologic techniques. 3. Confocal laser spectroscopy. Confocal colonoscopy is a significant technical advance that marries traditional endoscopy with real-time microscopy.62 In confocal laser colonoscopy, a confocal laser microscope is added to the distal tip of a traditional video colonoscope.63 Patients receive a contrast agent, and an endoscopically directed laser excites target tissue. Confocal images are collected at the microscopic tip and generated simultaneously with endoscopic images. This allows imaging of the mucosal subsurface and generation of real-time, in vivo histologic images. Kiesslich et al studied the utility of this technique in patients presenting for screening colonoscopy.63 Additionally, they evaluated the use of 2 varying contrast regimens: topical acriflavine hydrochloride or intravenously administered fluorescein sodium. Fluorescein enabled better resolution of structures within the lamina propia and was used in a prospective classification of lesions prior to biopsy. Confocal classification was compared to those obtained by conventional histological

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1091

analysis. Sensitivity for detection of neoplastic changes was 97.4% and specificity was 99.4%. All patients who received intravenous fluorescein noted temporary yellow skin discoloration which lasted less than 1 hour, and no serious adverse events occurred. The authors note this technique may improve identification of flat and polypoid colonic lesions, and may help avoid repeat colonoscopies by empowering the endoscopist to evaluate the need for resection at time of initial endoscopy. The Pentax Corporation is poised to produce confocal endomicroscopes for commercial sale (personal communication, R. E. Schoen, February 7, 2005). Clinical impact, however, is unlikely to be realized until the instrument’s usefulness in validated in clinical studies. C. Optical coherence tomography. Optical coherence tomography provides a cross-sectional, subsurface image of the gastrointestinal tract.64 In theory, it is similar to endoscopic ultrasound. Light waves, instead of sound waves probe the tissues. Light waves afford optical coherence tomography substantially better image resolution, and allow optical coherence tomography to approach light microscopy in its detail. Optical coherence tomography is conducted using 2 separate light beams. They are simultaneously directed in 2 paths: one is directed towards the target while the second is directed towards a mirror at a known distance. A detector and interferometer process the light which returns from the sources to construct detail about the examined area.39 Optical coherence tomography has been likened to “virtual histology” and allows visualization of the gut’s layers.65 Placing the probe parallel and close to the mucosal surface (approximately 1 mm) enables resolution of colonic crypts, gastric pits, and duodenal villi.39 Technological advances have yielded real-time optical coherence tomography probes that are small enough to fit in the accessory channel of an endoscope– endoscopic optical coherence tomography.39 A number of studies indicate that optical coherence tomography can be performed safely in humans, and can be readily adapted to current endoscopic technology. The use of endoscopic optical coherence tomography in screening remains unexplored. These and other optical technologies such as Raman spectroscopy continue to evolve. It is likely in the future that 2 or more complementary endoscopic techniques will be combined and molecular targeting will be used to improve detection of dysplastic colon lesions.66 A prototype combination optical coherence tomography-light induced fluorescence miniature endoscope has been designed to investigate mouse colon cancer in vivo.67

1092

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

GASTROENTEROLOGY Vol. 129, No. 3

Table 1. Clinical Studies: Stool DNA Assays Study Dong82 Alqhuist80 Alqhuist80 Brand83 Tagore84 Syngal85 Calistri86 Imperiale87

DNA targets p53, P53, P53, P53, P53, P53, P53, P53,

K-ras, APC K-ras, APC, BAT-26, long APC, BAT-26, long DNA K-ras, APC, BAT-26 K-ras, APC, BAT-26, long K-ras, APC, BAT-26, long K-ras, APC, BAT-26, long K-ras, APC, BAT-26, long

DNA

DNA DNA DNA DNA

Sensitivity for cancer

Sensitivity for adenoma

Specificity

71% 91% 91% 69% 64% 68% 74% 52%

NA 82% 73% NA 57% 30% NA 15%

NA 93% 100% NA 96% NA 97% 94%

NA, not applicable.

III. New Clinically Available Screening Technologies for CRC A. Laboratory-based diagnostics. 1. Serum-based tests. The search for quick, noninvasive CRC methods has been under way for years. Serum screening appears to have broad patient acceptance, as evidenced by the high screening rates achieved with the prostate specific antigen (PSA) test.6 Because of the wide range of CRCassociated mutations, however, population screening for mutation carriers is currently not feasible.68 Serum-based tests exist for patients at risk for FAP, HNPCC, Peutz– Jeghers, Cowden’s disease, and juvenile polyposis.69 Tissue-based immunohistochemical studies can detect microsatellite instability and mismatch repair genes and may aid in identification of patients who warrant testing for germline mutations.70 Such tests, while inapplicable to general screening, are playing an increasing role in the evaluation and management of patients with these syndromes. Thus, it is increasingly important that clinical gastroenterologists remain knowledgeable about these tests and refer when appropriate. While the vast majority of practicing gastroenterologists obtain a family history, the numbers who are aware of genetic tests for FAP and HNPCC are much lower, 52% and 34%, respectively.71 As the number of specialized screening tests continues to grow, it will be important for clinicians to keep abreast of these developments. 2. Stool and rectal-mucus based testing. Stool-based screening, while somewhat less acceptable to patients than serum tests, remains noninvasive and may be better accepted by some than invasive alternatives. Assays to detect a range of substances, such as fecal calprotectin, fecal lactoferrin, lysozyme, albumin, and alpha-1 antitrypsin have all been evaluated. None of these has demonstrated better sensitivity and specificity than fecal occult blood testing.72–77 Two screening approaches using stool DNA and rectal mucus, however, have shown significant potential.

● Stool DNA tests

The feasibility of fecal DNA (fDNA) screening was first demonstrated over a decade ago when a successful assay for K-ras mutation was reported. Concordance between mutations found in stool and those in tumors has been high.78 Still, this approach to screening was hampered by the insensitivity inherent in testing for a specific single mutation. This, and difficulties attendant to extracting and isolating human DNA from the DNA of colonic bacteria, has been addressed by newer approaches using multitargeted assays. A number of advantages have been ascribed to stool DNA screening for CRC.74 It is noninvasive and requires neither cathartic preparation nor restrictions of diet or medication. Testing can be performed on a single specimen, and these can be shipped by patients, thus obviating an office visit. In theory, the test can be highly accurate and can reflect the entire gastrointestinal tract. In a pilot study, Traverso et al reported that success identifying proximal CRC through the use of stool-based, DNA MSI testing.79 Laboratory through-put can be high in automated systems. The sensitivity of DNA testing varies, as expected, by the target markers that are incorporated. In 2000, Ahlquist et al reported sensitivities of 91% and 82% in the detection of cancer and adenoma, respectively, using a panel incorporating APC, K-RAS, p53, BAT26, and long DNA.80 With this panel, specificity was 91%. This improved to 100% when K-ras was removed from the panel, but detection of adenomas worsened (Table 1). Sensitivity reported in other studies has been lower, between 71%– 88% for cancer, and lower still for adenomas, 51%.81– 87 In 2003, a commercially available stool DNA test was marketed by EXACT Sciences and large scale screening studies were initiated. Imperiale et al compared fecal occult blood testing, and fDNA testing using a 21marker panel to results obtained at screening colonoscopy among average risk, asymptomatic persons aged 50 years and older.87 Specificity was high for both techniques. fDNA had 94.4% specificity and fecal occult

September 2005

blood testing had 95.2%. Sensitivity, however, was poor. fDNA detected 51.6% of invasive cancers compared to 12.9% detected by Hemoccult II. When high grade dysplasia was included as a positive result, fDNA detected only 40.8%. While this result was superior to a conventional single application of fecal occult blood testing (which detected 14.1% of such cases), neither technique detected the majority of neoplastic lesions found at colonoscopy. The sensitivity of fDNA is an improvement on that currently available by noninvasive means. Its cost, however, remains a significant impediment to broader use, especially as those with positive tests will require colonoscopy. The current retail price of PreGenPlus is $795.88 In a recent decision analysis, fDNA was not as cost-effective as strategies using fecal occult blood testing or conventional colonoscopy.89 These calculations were predicated on assumptions regarding sensitivity, costs, and participation that could change. In late 2004, Whitney et al reported a new method for DNA recovery from stool.90 This technique increased DNA yield over 5 times; an amount that they note corresponds in an increased sensitivity of up to 70% (95% confidence interval, 59%–79%). The use of additional markers would also improve sensitivity but would likely add to test costs. More sensitive stool DNA assays combined with lower costs would change the relative cost-effectiveness of fDNA. Additionally, if fDNA screening strategies resulted in significantly improved participation, they might ultimately be superior to fecal occult blood testing programs. It remains to be seen whether patients will embrace this strategy in clinical practice, as it requires that an entire bowel movement be packaged and mailed to a testing center.88 Even among compensated and presumably motivated participants, over 10% of participants in the study of Imperiale et al did not return adequate specimens.87 Currently, it appears that screening fDNA will be an option only for those reluctant to participate in more conventional approaches.89 ● Rectal mucus tests The normal colonic lining secretes a number of mucins, the expression and structure of which become dysregulated during carcinogenesis.91 Disaccharide D-Galactose-N-Acetyl-D-Galactosamine (GalNAc) is a mucin expressed by precancerous and cancerous colon lesions.92 Furthermore, in rat models of CRC, it has been detected at sites distant from dysplastic tissue.93,94 This field effect offers the promising possibility that detection of this abnormal mucin through simple rectal swab could reli-

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1093

Table 2. Recent Clinical Studies: Rectal Mucus Galactose Oxidase Schiff Assays Sensitivity for cancer

Study Vucenik102 Horsewood105 Marcon106 aSensitivity

100%a 80.1% 49%

Specificity 97% 75% 85%

for Neoplasia (Cancer or Adenoma).

ably indicate whether CRC is present anywhere within the large intestine. The disaccharide GalNAc can be detected in a number of ways. The simplest and least expensive is through the galactose oxidase Schiff (GOS).95 After digital rectal swab, samples undergo oxidation (with D-galactose oxidase) and colorization with Schiff’s reagent. When GalNAc is present, the sample changes color. Studies performed in the late 1980s and early 1990s showed a range of sensitivity from 67%–100% (Table 2).96 –101 Specificity ranged from less than 50% to 92%. More recent reports are equally variable. Vucenik et al reported 100% sensitivity and 96.8% specificity using GOS techniques to distinguish between those with colorectal malignancy and polyps from normal controls.102 Abnormal mucins may be present in patients with inflammatory bowel disease and other benign diseases, raising concerns that mucins may not be specific enough for general screening.103 Similarly, little is known about the stability of these markers and how this might affect the performance of mucin-based screening.104 Despite these issues, a commercially available, mucinbased CRC screening test (ColoRectAlert) has been developed by International Medical Innovations. In a comparative study, over 600 patients scheduled for colonoscopy underwent ColoRectAlert and a fecal occult blood based assay, Hemoccult II SENSA.105 Both tests had similar sensitivity of 81%, but the mucin-based test had better specificity, 75% versus 57%. In a much larger study of over 1700 patients, ColoRectAlert’s results were less impressive, with sensitivity of 49%.106 Rectal mucus testing has advantages to fecal occult blood testing, including the lack of dietary and medication restrictions, and the fact that the physician in the office performs it, thus allowing clinicians more potential control over screening performance. Issues of sensitivity, specificity, and patient acceptability require more exploration. B. Radiologic testing. New radiologic approaches to the early detection of CRC have been the focus of much research. “Virtual colonography,” or the ability to obtain 3-dimensional images of the colon noninvasively, has gained much attention both in popular and scientific

1094

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

GASTROENTEROLOGY Vol. 129, No. 3

Table 3. Reports from 2004: CT colonography

Study (n ⫽ 625) (113) Cohnen (n ⫽ 137) (114) Hoppe (n ⫽ 100) (115) Iannacconec,d (n ⫽ 203) (116)

Cottona,b

van Gelderd (n ⫽ 249) (117)

Per-polyp sensitivity: medium (6–9 mm)

Per-polyp Sensitivity: large (ⱖ10 mm)

23%

52%

85.7%

78.6%

50%

71%

Per-patient sensitivity: medium (6–9 mm)

Per-patient sensitivity (%): large (ⱖ10 mm)

Per-patient specificity (%): medium (6–9 mm)

Per-patient specificity (%): large (ⱖ10 mm)

30%

55%

93%

96%

76% (ⱖ6 mm)

86.3% (ⱖ5 mm lesions) 95%

88% (ⱖ6 mm)

93.7% for ⱖ5 mm lesions 98%

100% all readers

R1 - 91.6%

100% all readers

R1 - 93.5%

100% all readers

R2 - 51.8% R3 - 56% (ⱖ6 mm) R1 - 64%

R1 - 75%

R2 - 93.7% R3 - 89.6% (ⱖ6 mm) R1 - 76%

R1 - 84%

R2 - 94.2% R3 - 91.3% (ⱖ6 mm) R1 - 71%

R1 - 92%

R2 - 75%

R2 - 77%

R2 - 80%

R2 - 84%

R2 - 69%

R2 - 92%

R1 - 86%

aHigher

sensitivity with fly-throughs. study. cStudy performed without colonic preparation. dMultiple reviewers. bMulticenter

publication. Many general advantages are ascribed to VC techniques, including fewer complications, no cathartic preparation, and better patient acceptance. Colonic imaging has been explored through both computed tomography (CT) and magnetic resonance imaging (MRI). VC CRC screening strategies rely, as conventional endoscopy does, upon detection of existing precancerous polyps. Thus, questions regarding the significance of polyp size and morphology apply. While total polyp number is an acknowledged predictor of future risk for advanced adenomas, it is less clear what constitutes a significant polyp. Small polyps (ⱕ5 cm) may be less ominous and their detection and removal may be less important. As discussed in the previous section, the prevalence and significance of flat adenomas are unclear; nonetheless, their detection in VC protocols has generated great interest. 1. Computed tomographic colonography. CT colonography is the most widely studied new radiologic CRC screening tool. Van Dam et al have recently published an excellent, detailed review of this technology and its potential place in CRC screening.107 The U.S. Food and Drug Administration (FDA) has approved CT colonography on any CT scanner able to obtain helical thin slice images. Generally, helical CT scanning is performed during a single breath hold. Multidetector (or multislice) CT (MCDT) scanners allow faster image acquisition with greater anatomic details.108 A 2-dimensional view can be seen on a work station or a 3-dimensional view can be

created using specialized software, such as Viatronix 3-D Colon (Viatronix, Stonybrook, NY). A wide range of sensitivities and specificities have been reported,109 –117 though in general, the operating characteristics of CT colonography have improved over time. Studies that are more recent have reported sensitivities and specificities of 100%; however, considerable variability persists between observers and centers (Table 3). In a recent report of multidetector CT among symptomatic patients, overall per-patient sensitivity and specificity of 70.3% and 80.8%, respectively, was reported.114 In another study, Hoppe et al report sensitivity of 76% for polyps larger than 6 mm in a group of high risk patients presenting with positive fecal occult blood testing or iron deficiency anemia.115 In both of these studies bowel cleansing was used, and in the study be Hoppe et al, an intravenous contrast agent was also administered. In 2004, Iannaccone et al evaluated the use of CT colonography without cathartic preparation to detect colorectal polyps in a group of 203 patients scheduled to undergo conventional colonoscopy.116 In their study, in lieu of cathartic preparation, fecal tagging was performed by adding diatrizaoate meglumine and diatrizaoate sodium to regular meals 3–7 days before CT colonography. They noted an average sensitivity of 95.5% and specificity of 92.2% for polyps at least as large as 8 mm. Also, inter-observer variability was low for polyps in this range. For smaller polyps, however, sensitivity decreased and inter-observer variability increased.

September 2005

CT colonography, like conventional colonoscopy, has generally been less sensitive in the detection of flat adenomas versus polypoid lesions.117 In a recently published study, Pickhardt et al have attempted to address concerns regarding the prevalence and significance of such lesions.118 In a prospective study of 1233 asymptomatic Western adults undergoing same-day CT colonography and conventional colonoscopy, they compared overall polyp detection and detection of flat lesions. They reported a rate of flat adenomas similar to that reported in many studies, approximately 10%. CT colonography sensitivity for flat adenomas was 82.8% (24/29), and 80% for all flat lesions (47/59) 6 mm or larger. These rates were similar to those of CT colonography in detection of polypoid lesions 6 mm or larger, of 81%. An important finding was that none of 148 flat lesions ⬍6 mm that were detected by conventional colonoscopy was histologically advanced. The authors concluded that no significant differences in the detection of flat or polypoid lesions were found in the use of CT colonography. Of note, this study employed a cathartic regimen of 90 mL of oral sodium phosphate (PhosphSoda, Fleet) 1 day before CT colonography and conventional colonoscopy examination day. In addition, patients consumed 500 mL of dilute barium and 120 mL of water-soluble iodinated contrast material for the purposes of fecal tagging and electronic fluid subtraction. The authors identified adherent stool as a major source of false-positive identification of flat lesions at CT colonography, and considered barium stool tagging “vital” in order to enhance specificity. The impressive sensitivity and specificity attained in this CT colonography study must be considered in light of this preparation. Some may find this difficult and certainly not appreciably more appealing that the preparation required for colonoscopy, especially since a positive CT colonography will generally require a subsequent colonoscopy. Few direct side effects have been reported with CT colonography; however, case reports of perforation have emerged.119,120 The effects of radiation exposure, a consideration in serial screening or surveillance, have been cited by some as a potential public health concern if CT colonography were widely adopted, and as a reason that MR colonography might be a preferable approach.121 Additionally, the significance and appropriate handling of incidental findings have not been fully characterized. Such findings are common. Rajapaska et al reported extracolonic findings, defined by the reviewing radiologist in over 33% of 250 patients undergoing CT colonography, including solitary lung nodules, lymphadenopathy, adrenal masses, and evidence of metastases.122 Of the “highly significant” findings (n ⫽ 17), the

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1095

vast majority were not previously known, and 11 required further diagnostic evaluation. Hellstrom et al report similar results, noting that 23% of patients undergoing CT colonography had potentially important extracolonic findings, including aortic aneurysms, hepatic and renal masses, and gallstones.123 The authors note that some findings were clinically important while others were not and led to “unnecessary work-ups.” Extracolonic findings have been called a “double-edged sword.”124 While early detection of important remediable conditions may be beneficial, the detection of ultimately unimportant lesions will increase costs, raise patient anxiety, and may cause harm. Such indirect consequences of CT colonography must be better characterized and accounted for when the acceptability and cost-effectiveness of CT colonography is considered. Patient preferences, often assumed to be more favorable for VC, have proved variable. In one study, patients perceived VC more favorably than conventional colonoscopy before examination, but this preference decreased significantly after undergoing both tests.125 In a study of patients undergoing both CT colonography and conventional colonoscopy, 88/104 patients surveyed reported no difference in embarrassment associated with either examination.126 Among these respondents, only 68/104 favored one exam over another, though among those with a preference, CT colonography was preferred. Akerkar et al explored the preferences of patients who had undergone both CT colonography and conventional colonoscopy both immediately after the tests, and 24 hours later.127 These 378 patients tolerated both procedures well, but reported more pain, discomfort and less respect with virtual colonoscopy. In Iannaccone and coworkers recent study, despite cathartic preparation, over one third of patients noted that they would prefer conventional colonoscopy to CT colonography.116 The authors note that, while a formal explanation of such preferences was not performed, it is possible that the inherent therapeutic capabilities of colonoscopy could account for this preference. In an attempt to assess the impact of choices on CRC screening participation, Scott et al invited community dwelling, asymptomatic adults to participate in CRC using CT colonography, conventional colonoscopy, or their choice of either modality.128 In this 3-arm study, those offered a choice of screening by either CT colonography or conventional colonoscopy did not participate at greater rates than those offered screening by CT colonography or conventional colonoscopy. Both tests were associated with high levels of acceptability, and the majority of subjects found both tests less unpleasant than they had anticipated. Patient acceptance and participa-

1096

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

Table 4. Clinical Studies MR Colonography Study

Subjects

Pappalardo135 Luboldtb 136

70 132

Lama 138

36

Per-patient sensitivity

Per-patient specificity

96% (all polyps) 75% (ⱖ7 mm) 93% (ⱖ10 mm) 75% (⬎5 mm)

93% (all polyps) 92% (ⱖ7 mm) 99% (ⱖ10 mm) 93.3% (⬎5 mm)

aAir-inflated bStudy

technique. performed without colonic preparation using fecal tagging.

tion in CT colonography-based CRC screening programs need further exploration. The impact of CT colonography on gastroenterology has been explored in a number of mathematical models.129,130 Ladabaum and Song report that CT colonography could be a cost-effective approach, compared with no screening.130 However, for CT colonography to be more cost-effective than conventional colonoscopy, it would need to cost 25%– 40% less than conventional colonoscopy. In their cost-effectiveness model, Heitman et al considered factors such as deaths related to adenomas missed by CT colonography and deaths associated with perforations due to conventional colonoscopy.131 They concluded that CT colonography was associated with 6.6 deaths per 100,000; conventional colonoscopy was associated with 6 deaths/100,000 and cost $8.6 million less over a 3-year period. In a recently published article, Hur et al estimated that the use of CT colonography could decrease the overall conventional colonoscopy rate by 19% and decrease the use of screening and surveillance colonoscopy by as much as 49%.132 2. Magnetic resonance colonography. The clinical use of MR colonography began in 1997.133 In theory, magnetic resonance imaging (MRI) technology affords better soft-tissue contrast capabilities and lacks exposure to ionizing radiation. Drawbacks include contraindications to MRI in general (pacemakers, claustrophobia), metal and motion related artifact, and that MR colonography protocols often use fluid contrast, which may add expense and compliance issues. A variety of MR colonographic approaches have been reported, including those using 1.5 and 1.0 Tesla technologies, methods with and without cathartic preparation, and with and without fluid contrast media. (Table 4).134 –140 In general, MR colonography can be performed on any magnetic resonance imaging unit capable of performing angiography.134 In a study of 132 patients, MR colonography using a bowel cleansing regimen followed by rectal tube placement and colonic filling with a combination of water and gadopentetate dimeglumine was evaluated.136 MR colonography detected large lesions (⬎10

GASTROENTEROLOGY Vol. 129, No. 3

mm) with a per-patient sensitivity of 93% and a specificity of 99.7%. More problematic was the detection of intermediate lesions. The per-patient sensitivity of MR colonography for lesions with a cut off of 7 mm was 75%; specificity remained high, at 98.6%. Specificity was still high, 96.5%, using a cut-off of 5 mm. Per-patient sensitivity, however, dropped to 47%. In another study, using MR colonography, sensitivity for lesions larger than 5 mm was 98% and was 91% in general. The feasibility of performing air-inflated MR colonography was demonstrated in a number of small studies.137,138 Reported sensitivities, however, were disappointingly low. In a larger study (n ⫽ 36), Lam et al performed air-inflated MR colonography in a population of 36 patients, noting an overall sensitivity of 38%.138 Air-inflated MR colonography performed significantly better for lesions ⬎5 mm, with sensitivity of 75% and specificity of greater than 93%. The substitution of room air for traditional gadolinium-based contrast could generate significant cost savings. Air-inflated MR colonography, however, requires a significant, cathartic bowel preparation. Attempts to develop MR colonography without bowel preparation have included fecal tagging and the use of intravenous contrast.139,140 A number of groups have reported the feasibility of fecal tagging in MR colonography. Lauenstein et al reported successful fecal tagging through ingestion of barium sulfate containing contrast with 4 low-fiber meals within 36 hours of examination.139 Subjects were instructed to avoid foods high in manganese such as fruits and chocolates before study. In addition to fecal tagging, patients were administered an enema of water and barium and an intravenous injection of gadobenate dimeglumine at time of the procedure. Using this regimen, 2 colonic carcinomas and 2 polyps were correctly identified among 6 study subjects. Few data about the relative cost of MR colonography exist. Similar to CT colonography, the cost-effectiveness of MR colonography will depend on the direct procedural costs, indirect costs associated with incidental findings, and the rate at which subsequent conventional colonoscopy is required.

Part 4. Implications for Gastroenterology Clinical Practice A. General factors influencing the impact of new technologies The impact of new technologies on CRC screening and on the practice of gastroenterology is challenging

September 2005

to predict. This is reflected in the very divergent estimates of future conventional colonoscopy use posited by recent authors. As cited, Hur et al anticipate that the use of conventional colonoscopy may decrease by as much as 19% as CT colonography gains broader use.132 In a different view, Kahi and Rex predicted that the role of conventional colonoscopy in CRC screening would increase.141 This latter view was based on the authors’ assessment that CT colonography and MR colonography in their current forms are not ready for screening use, and the fact that the relative use of fecal occult blood testing, flexible sigmoidoscopy, and a combination of the 2 are decreasing as screening tools.142 A shift from flexible sigmoidoscopy to conventional colonoscopy would increase demand for gastroenterologists as they are the predominant providers of colonoscopy; in contrast, flexible sigmoidoscopy is performed by both gastroenterologists and other providers. Seeff et al recently analyzed the capacity in the United States for endoscopic CRC screening.143 This assessment, though limited by its nonstandardized approach to capacity assessment and reliance on physician self-report, concluded that a decade or more would be required to screen the approximately 42 million unscreened, asymptomatic population aged 50 and older using flexible sigmoidoscopy and conventional colonoscopy as the primary approaches. They note fecal occult blood testing screening followed by conventional colonoscopy for positive results could be accomplished more readily. Vijan et al note that in order to offer screening conventional colonoscopy to all patients aged 50 years and older every 10 years, as many as 32,700 additional gastroenterologists would need to be trained.144 These varying outlooks reflect the variety of assumptions that must be made to attempt to quantitatively model the impacts of new diagnostic technologies. These assumptions, thoroughly outlined in the previously cited CT colonography review of Van Dam et al, are based on factors that remain unknown.107 These include: (1) overall screening participation rates and the rates at which specific tests are used, (2) the frequency of general screening (ie, every 5 years, every 10 years?), (3) the prevalence of specific lesions and their need for subsequent polypectomy and surveillance (ie, 5-mm polyps, flat lesions?), and (4) the specific characteristics and costs of each competing screening strategy. The goal of many new CRC screening methods is improved overall participation. It is hoped that less invasive approaches will result in higher screening rates. While intuitive, this has yet to be proved. As noted previously, in a community-based study, participation rates did not increase when a choice of screening tests was

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1097

provided.128 As this study did not offer a wide range of screening choices, anticipated improvements with broader choices may yet be realized. Data indicate that screening participation and choices are also influenced by practical factors. Out-of-pocket expenses may drive screening choices more than the characteristics of the tests themselves145 and patients with a usual source of care are more likely to undergo screening.146 Patients’ perceptions of screening tests may affect their choices. As options become increasingly available, information about these tests from reliable sources are likely to be important. Also, alternative strategies that screen for many diseases simultaneously might emerge. Within the past year, an MR-based full-body preventative cardiovascular and tumor imaging strategy was explored by German investigators.147 They note that within the framework of an hour-long test, it was feasible to examine the brain, lung, colon, and arterial system, including the heart, aorta, and renal arteries. While such a “pan-scan” may not be indicated in all patients, for those with high risk of arterial disease and who need CRC screening, such a strategy might be more appealing and/or most cost-effective. Certainly, in its preliminary stages, such considerations are speculative. Increasingly, direct-to-consumer marketing of diagnostic tests has raised concerns of inappropriate screening test use.148,149 It will be important to insist that information reflect medical science and that providers remain informed with data from well-conducted research. B. Specific factors influencing the impact of new technologies The adoption of specific screening strategies and their impact on gastroenterology practice will rely on the complex interplay of physician, patient, and insurer acceptance and feasibility. As the number and nature of new screening technologies rapidly change their effects are difficult to estimate in all but a qualitative manner (Table 5). Technical challenges represent the major limitations to proteomic approaches; thus, a decade may pass before they are clinically applicable. Enhanced endoscopy, including high magnification chromoscopic colonoscopy, spectroscopy, and optical coherence tomography, remain unproven in general CRC screening. Broader use will be limited unless substantial clinical data support their benefit, particularly in light of the added procedure time and required training. It is likely, however, that such techniques will gradually enter the practice of a minority of gastroenterologists specializing in the care of patients with high-risk conditions, such as

1098

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

GASTROENTEROLOGY Vol. 129, No. 3

Table 5. Estimated Clinical Impact of New CRC Screening Technologies Technology

Clinical application within 5 years?

Clinical impact ⱕ5 years

Clinical impact ⱖ 10 years

Proteomics Stool DNA Rectal mucus CT colonography-General screening CT colonography - Low-risk and surveillance MR colonography Enhanced endoscopy - General screening Enhanced endoscopy - High-risk

no low low no Selected populations no no Selected populations

none 3 3 3 2 3 3 2

1 2 3 2 2 2 2 2

1, great impact on gastroenterology practice; 2, moderate impact on gastroenterology practice; 3, minimal impact on gastroenterology practice.

ulcerative colitis and Barrett’s esophagus, in whom these techniques may provide important incremental benefits. Stool DNA has demonstrated limited sensitivity, significantly higher costs, and unknown patient acceptance compared with standard fecal occult blood tests. Thus, it is not likely to gain broader use until these issues are resolved. CT colonography remains less cost-effective than other existing options at its current costs and sensitivity. As these aspects change and as patient acceptance is clarified, CT colonography may become cost-neutral or costeffective. In the near term, CT colonography may find a role in screening of patient reluctant to consider other techniques, interval surveillance between colonoscopies in high-risk groups or those unlikely to need subsequent colonoscopy.

Part 5. Implications for Gastroenterologists’ Education and Training While conventional colonoscopy is likely to remain a significant part of the practice of most gastroenterologists, it is likely that more emphasis in the future will be on specialized diagnostic and therapeutic procedures. To function in the endoscopy suite, a greater number of technical skills and broader knowledge base will be needed. Emerging endoscopic techniques will require familiarity with new equipment. Certain techniques, such as high magnification chromoscopic colonoscopy, require specific skills, as the degree of magnification will greatly increase movement related artifact.39 It is possible that these specialized techniques may become the purview of more specialized gastroenterologists. The level of training and practice required to attain competency in those technologies that ultimately enter practice will need to be defined and disseminated.

Technologies such as optical coherence tomography and confocal laser spectroscopy will allow real-time decisions about the need for biopsy; however, these decisions will be based on a strong working knowledge of histology.62 A broader knowledge base will be required in the office as well. A working understanding of cancer genetics and associated diagnostic testing will be paramount. Proteomics will continue its rapid expansion and become increasingly clinically relevant. Currently, many practicing physicians understand little of this field and its potential uses. The armamentarium to screen for CRC is rapidly evolving. Insights into colorectal carcinogenesis, and protein characterization and detection, are positioned to radically change diagnosis in the future. Technological advances, particularly in the area of optical endoscopy are already making inroads in the surveillance of patients with high-risk conditions and they are likely to continue to significantly alter CRC screening in both incremental and revolutionary ways. A preliminary report of the successful use of self-propelling, self-navigating miniaturized colonoscopes in pigs holds the promise of one such revolutionary approach.150 The number of new clinically available tests indicated for mass screening, however, is modest, and includes stool DNA tests, rectal mucus tests, and CT colonography. None of these tests, however, in their current forms is superior to existing screening methods and is unlikely to substantially change patient care or physician practice over the next 2 to 5 years. Computed tomographic colonography and enhanced endoscopy will gradually enter the practice of some gastroenterologists over the coming years. As our knowledge of new screening tools grows, quantitative modeling of relative costs, benefits, and risks will enable more informed decision-making. All gastroenterologists, their medical colleagues, and patients will benefit from a broad understanding of these evolving options. It is hoped that increasing screening options will

September 2005

improve participation. It is certain that as the population grows and ages, colorectal cancer screening will remain a large part of the daily practice of gastroenterology.

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

8.

9.

Part 6. Suggested Key Research Questions The degree to which specific new technologies enter future clinical practice depends on the answers to questions that are currently unanswered. Regarding CT colonography: ● What is the natural history of small polyps? ● Do small polyps invariably enlarge? ● Do small polyps impart increased risk of malignancy, and, if so, to what degree? ● What is the relevance of extracolonic findings? ● Does the identification of extracolonic findings improve health outcomes? ● How does the identification and evaluation of extracolonic findings affect costs, after direct and indirect economic, societal (ie, time off from work), and personal costs (ie, anxiety and/or relief) are considered? Regarding enhanced endoscopy: ● What is the ultimate sensitivity and specificity of such tools? ● For whom are these tools best suited? In general, does a greater array of screening choices improve participation rates, and, if not, why not? Somewhat related to this, the causes of disparities in CRC screening rates and practices among different populations need to be investigated.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

References 1. American Cancer Society Cancer Facts and Figures 2004; Estimated New Cancer Cases and Deaths by Sex for All Sites, US, 2004. Available: http://www.cancer.org, accessed January 27, 2005. 2. Surveillance, Epidemiology and End Results (SEER) data, available: http://seer.cancer.gov, accessed January 27, 2005. 3. www.ahrq.gov/clinic/uspstf/uspscolo.htm 4. Mandel JS, Church TR, Bond JH, et al. The effect of fecal occult blood screening on the incidence of colorectal cancer. JAMA 2000;343:1603–1607. 5. Scholefield JH, Moss S, Sufi F, Mangham CM, Hardcastle JD. Effect of faecal occult blood screening on mortality from colorectal cancer: results from a randomised controlled trial. Gut 2002;50:840 – 844. 6. American Cancer Society Cancer Facts and Figures 2004; Cancer Prevention & Early Detection Facts & Figures 2004, available: http://www.cancer.org, accessed January 27, 2005. 7. Helm JF, Russo MW, Biddle AK, Simpson KN, Ransohoff DF, Sandler RS. Effectiveness and economic impact of screening for

23.

24.

25.

26.

27.

1099

colorectal cancer by mass fecal occult blood testing. Am J Gastroenterol 2000;95:3250 –3258. Bressler B, Paszat LF, Vinden C, Li C, He J, Rabeneck L. Colonoscopic miss rates for right-sided colon cancer: a populationbased analysis. Gastroenterology 2004;127:452– 456. Vernon SW. Participation in colorectal cancer screening: a review. J Natl Cancer Inst 1997;89:1406 –1422. Wulfkuhle JD, Liotta LLA, Petricoin EF. Proteomic applications for the early detection of cancer. Nature Rev Cancer 2003;3: 267–275. Feldman AL, Espina V, Petricoin EF, Liotta LA, Rosenblatt KP. Use of proteomic patterns to screen for gastrointestinal malignancies. Surgery 2004;135:243–247. Wilson JF. The promise of disease proteomics: faster detection, diagnosis and drug development. Ann Intern Med 2004;140: 317–319. Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002;359: 772–777. Ornstein DK, Rayford W, Fusaro VA, et al. Serum proteomic profiling can discriminate prostate cancer from benign prostates in men with total prostate-specific antigen levels between 2.5 and 15.0 ng/ml. J Urol 2004;172:1302–1305. Lawrie LC, Dundas SR, Curran S, Murray GI. Liver fatty acid binding protein in colorectal neoplasia. Br J Cancer 2004;90: 1955–1960. Drake RR, Upender M, Bao-Ling A, et al. SELDI-TOF-MS profiling of serum for early detection of colorectal cancer. Gastroenterology 2003;124(Suppl 1):A650. Yu JK, Chen YD, Zheng S. An integrated approach to the detection of colorectal cancer utilizing proteomics and bioinformatics. World J Gastroenterol 2004;10:3127–3131. Xiong XD, Li EM, Xu LY, et al. Separation and identification of differentially expressed nuclear matrix proteins between human esophageal immortalized and carcinomatous cell lines. World J Gastroenterol 2003;9:2143–2148. Getzenberg RH, Konety BR, Oeler TA, et al. Bladder cancerassociated nuclear matrix proteins. Cancer Res 1996;56: 1690 –1694. Brunagel G, Shah U, Scheon RE, Gentzenberg RH. Identification of calreticulin as a nuclear matrix protein associated with human colon cancer. J Cell Biochem 2003;89:238 –243. Konety BR, Nguyen TST, Dhir R, et al. Detection of bladder cancer using a novel nuclear matrix protein, BLCA-41. Clin Cancer Res 2000;6:2618 –2625. Brunagel G, Schoen RE, Getzenberg RH. Colon cancer specific nuclear matrix protein alterations in human colonic adenomatous polyps. J Cell Biochem 2004;91:365–374. Davies JR, Freeman A, Morris LS, et al. Analysis of minichromosome maintenance proteins as a novel method for detection of colorectal cancer in the stool. Lancet 2002;359:1917–1919. Going JJ, Keith WN, Neilson L, Stoeber K, Stuart RC, Williams GH. Aberrant expression of minichromosome maintenance proteins 2 and 5, and Ki-67 in dysplastic squamous oesophageal epithelium and Barrett’s mucosa. Gut 2002;50:373–377. Sirieix PS, O’Donovan M, Brown J, Save V, Coleman N, Fitzgerald RC. Surface expression of minichromosome maintenance proteins provides a novel method for detecting patients at risk for developing adenocarcinoma in Barrett’s esophagus. Clin Cancer Res 2003;9:2560 –2566. Funaki NO, Tanaka J, Itami A, et al. Detection of colorectal carcinoma in circulating peripheral blood by reverse transcription-polymerase chain reaction targeting cytokeratin-20 mRNA. Life Sci 1997;60:643– 652. Melissourgos N, Kastrinakis NG, Davilas I, Foukas P, Farmakis A, Lykourinas M. Detection of human telomerase reverse tran-

1100

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41.

42.

43. 44. 45.

46.

47.

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

scriptase mRNA in urine of patients with bladder cancer: evaluation of an emerging tumor marker. Urology 2003;62: 362–367. Niiyama H, Mizumoto K, Sato N, et al. Quantitative analysis of hTERT mRNA expression in colorectal cancer. Am J Gastroenterol 2001;96:1895–1900. Davidson LA, Lupton JR, Miskovsky E, Fields AP, Chapkin RS. Quantification of human intestinal gene expression profiles using exfoliated colonocytes: a pilot study. Biomarkers 2003;8: 51– 61. Kanaoka S, Yoshida KI, Miura N, Sugimura H, Kajimura M. Potential usefulness of detecting cyclooxygenase 2 messenger RNA in feces for colorectal cancer screening. Gastroenterology 2004;127:422– 427. Mohan S, Epstein JB. Carcinogenesis and cyclooxygenase: the potential role of COX-2 inhibition in upper aerodigestive tract cancer. Oral Oncol 2003;39:537–546. Ouyang DL, Chen JJ, Getzenberg RH, Schoen RE. Non-invasive testing for colorectal cancer: a review. Am J Gastroenterol 2005;100:1393–1403. Chung DC, Rustgi AK. The hereditary nonpolyposis colorectal cancer syndrome: genetics and clinical implications. Ann Intern Med 2003;138:560 –570. Frigola J, Sole X, Paz MF, et al. Differential DNA hypermethylation and hypomethylation signatures in colorectal cancer. Hum Mol Genet 2005;14:319 –326. Lenhard K, Bommer GT, Asutay S, et al. Analysis of promoter methylation in stool: a novel method for the detection of colorectal cancer. Clin Gastroenterol Hepatol 2005;3:142–149. Anacleto C, Rossi B, Lopes A, et al. Development and application of a multiplex PCR procedure for the detection of DNA methylation in colorectal cancer. Oncol Rep 2005;13:325–328. Schatz P, Dietrich D, Schuster M. Rapid analysis of CpG methylation patterns using RNase T1 cleavage and MALDI-TOF. Nucleic Acids Res 2004;32:e167. Cruz-Correa M, Cui H, Giardiello FM, et al. Loss of imprinting of insulin growth factor II gene: a potential heritable biomarker for colon neoplasia predisposition. Gastroenterology 2004;126: 964 –970. Pfau PR, Sivak MV. Endoscopic diagnostics. Gastroenterology 2001;120:763–781. Kumagai Y, Monma K, Kawada K. Magnifying chromoendoscopy of the esophagus: in-vivo pathological diagnosis using an endocystoscopy system. Endoscopy 2004;36:590 –594. Hurlstone DP, Fujii T, Lobo AJ. Early detection of colorectal cancer using high-magnification chromoscopic colonoscopy. Br J Surg 2002;89:272–282. Pretlow TP, Roukhadze EV, O’Riordan MA, Chan JC, Amini SB, Stellato TA. Carcinoembryonic antigen in human colonic aberrant crypt foci. Gastroenterology 1994;107:1719 –1725. Jass JR, Whitehall VLJ, Young J, Legget BA. Emerging concepts in colorectal neoplasia. Gastroenterology 2002;123:862– 876. Kudo S. Early colorectal cancer: detection of depressed types of colorectal carcinoma. Tokyo: Igaku-Shoin, 1996. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res 2003; 63:1727–1730. Hata K, Watanabe T, Shinozaki M, Kojima T, Nagawa H. To dye or not to dye? That is beyond question! Optimising surveillance colonoscopy is indispensable for detecting dysplasia in ulcerative colitis. Gut 2004;53:1722. Tsuda S, Veress B, Toth E, Fork FT. Flat and depressed colorectal tumours in southern Swedish population: a prospective chromoendoscopic and histopathological study. Gut 2002;51: 550 –555.

GASTROENTEROLOGY Vol. 129, No. 3

48. Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neoplasms: a prospective study of 1000 colonoscopies in the UK. Lancet 2000;355:1211–1214. 49. Saitoh Y, Waxman I, West BA, et al. Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population. Gastroenterology 2001;120:1657–1665. 50. Kudo S, Tamura S, Nakajima T, Yamano H, Kusaka H, Watanabe H. Diagnosis of colorectal tumorous lesions by magnifying endoscopy. Gastrointest Endosc 1996;44:8 –14. 51. Kiesslich R, Fritsch J, Holtmann M, et al. Methylene blue-aided chromoendoscopy for the detection of intraepithial neoplasia and colon cancer in ulcerative colitis. Gastroenterology 2003; 124:880 – 888. 52. Hurlstone DP, McAlindon ME, Sanders DS, Keogh R, Lobo AJ, Cross SS. Further validation of high-magnification chromoscopic colonoscopy for the detection of intraepithelial neoplasia and colon cancer in ulcerative colitis. Gastroenterology 2004; 126:376 –377. 53. Gono K, Obi T, Yamaguchi M, et al. Appearance of enhanced tissue features in narrow-band endoscopic imaging. J Biomed Optics 2004;9:568 –577. 54. Dekker E, van Deventer SJ, Hardwick JC, Offerhaus JH, Fockers P, Hommes DW. The value of narrow band imaging for the detection of dysplasia in longstanding ulcerative colitis. Gastroenterology 2004;126(Suppl 2):A77. 55. Roy HK, Liu Y, Wali RK, Kim YL, et al. Four-dimensional elastic light-scattering fingerprints as preneoplastic markers in the rat model of colon carcinogenesis. Gastroenterology 2004;126: 1071–1081. 56. Mallery S, van Dam J. Endoscopic practice at the start of the new millennium. Gastroenterology 2000;118(Suppl 1):S129 – S147. 57. Kapadia CR, Cutruzzola FW, O’Brien KM, Stetz M, Enriquez R, Deckelbaum L. Laser-induced fluorescence spectroscopy of colonic mucosa. Detection of adenomatous transformation. Gastroenterology 1990;99:150 –157. 58. Cothren RM, Sivak MV, van Dam J, Petras R, Fitzmaurice M, et al. Detection of dysplasia at colonscopy using laser-induced fluorescence: a blinded study. Gastrointest Endosc 1996;44: 168 –176. 59. Wang TD, van Dam J, Crawford JM, Presinger EA, Wang Y, Feld MS. Fluorescence endoscopic imaging of human colonic adenomas. Gastroenterology 1996;111:1182–1191. 60. Mayinger B, Jordan M, Horner P, et al. Endoscopic light-induced autofluorescence spectroscopy for the diagnosis of colorectal cancer and adenoma. J Photochem Photobiol 2003;70:13–20. 61. Marten K, Bremer C, Khazaie K, et al. Detection of dysplastic intestinal adenomas using enzyme-sensing molecular beacons in mice. Gastroenterology 2002;122:406 – 414. 62. Nathanson MH. Confocal colonscopy: more than skin deep. Gastroenterology 2004;127:987–989. 63. Kiesslich R, Burg J, Vieth M, et al. Confocal laser endoscopy for diagnosing intraepithelial neoplasia and colorectal cancer in vivo. Gastroenterology 2004;127:706 –713. 64. van Dam J. Novel methods of enhanced endoscopic imaging. Gut 2002;52(Suppl IV):iv12–iv16. 65. Inoue H, Cho JY, Satodate H, et al. Development of virtual histology and virtual biopsy using a laser-scanning confocal microscopy. Scand J Gastroenterol Suppl 2003;237:37–39. 66. Dacosta RS, Wilson BC, Marcon NE. Optical techniques for the endoscopic detection of dysplastic colonic lesions. Curr Opin Gastroenterol 2005;21:70 –79. 67. Tumlinson AR, Hariri LP, Utzinger U, Barton JK. Miniature endoscope for simultaneous optical coherence tomography and laser-induced fluorescence measurement. Applied Optics 2004; 43:113–121.

September 2005

68. Mitchell RJ, Farrington SM, Dunlop MG, Campbell H. Mismatch repair genes hMLH1 and hMSH2 and colorectal cancer: a HuGE review. Am J Epidemiol 2002;156:885–902. 69. Starosik P, Muller-Hermelink HK. Diagnosis of microsatellite instability-positive colorectal cancer. Exp Rev Molec Diag 2001; 1:71– 80. 70. Terdiman JP, Gum JR, Conrad PG, et al. Efficient detection of hereditary nonpolyposis colorectal cancer gene carriers by screening for tumor microsatellite instability before germline genetic testing. Gastroenterology 2001;120:21–30. 71. Batra S, Valdimarsdottir H, McGovern H, Itzkowitz S, Brown K. Awareness of genetic testing for colorectal cancer predisposition among specialists in gastroenterology. Am J Gastroenterol 2002;97:729 –733. 72. Limburg PJ, Devens ME, Harrington JJ, et al. Prospective evaluation of fecal calprotectin as a screening biomarker for colorectal neoplasia. Am J Gastroenterol 2003;98:2299 –2305. 73. Roseth AG, Kristinnson J, Nygaard K. Faecal lactoferrin, a novel marker of colorectal cancer in rats: a longitudinal in vivo study. Gastorenterology 1995;108:A530. 74. Dubrow R, Kim CS, Eldred AK. Fecal lysozyme: an unreliable marker for colorectal cancer. Am J Gastroenterol 1992;87:617– 621. 75. Nakayama T, Yasuoka H, Kishino T, et al. ELISA for occult faecal albumin. Lancet 1987;1:1368 –1369. 76. Dubrow R,Yannielli L. Fecal protein markers of colorectal cancer. Am J Gastroenterol 1992;87:854 – 858. 77. Kim Y, Lee S, Park S, et al. Gastrointestinal tract cancer screening using fecal carcinoembryonic antigen. Ann Clin Lab Sci 2003;33:32–38. 78. Ahlquist DA. Stool-based DNA tests for colorectal cancer: clinical potential and early results. Rev Gastroenterol Disord 2002; 2(Suppl 1):S20 –S26. 79. Traverso G, Shuber A, Olsson L, et al. Detection of proximal colorectal cancers through analysis of faecal DNA. Lancet 2002;359:403– 404. 80. Ahlquist DA, Skoletsky JE, Boynton KA, et al. Colorectal cancer screening by detection of altered DNA in stool: feasibility of a multitarget assay panel. Gastroenterology 2000;119:1219 1227. 81. Traverso G, Shuber A, Levin B, et al. Detection of APC mutations in fecal DNA from patients with colorectal tumors. N Engl J Med 2002;346:311–320. 82. Dong SM, Traverso G, Johnson C, et al. Detecting colorectal cancer in stool with the use of multiple genetic targets. J Nat Cancer Inst 2001;93:858 – 865. 83. Brand RE, Ross ME, Shuber AP. Reproducibility of a multitarget stool-based DNA assay for colorectal cancer detection. Am J Gastroenterol 2004;99:1338 –1341. 84. Tagore KS, Lawson MJ, Yucaitis JA, et al. Sensitivity and specificity of a stool DNA multitarget assay panel for the detection of advanced colorectal neoplasia. Clin Colorectal Cancer 2003;3: 47–53. 85. Syngal S, Chung D, Willet C, et al. Stool DNA analysis for the detection and follow-up of colorectal cancer and advanced adenomas: sensitivity in a prospective series. Am J Gastroenterol 2002;(Suppl)97:S109. 86. Calistri D, Rengucci C, Bocchini R, et al. Fecal multiple molecular tests to detect colorectal cancer in stool. Clin Gastroenterol Hepatol 2003;1:377–383. 87. Imperiale TF, Ransohoff DF, Itzkowitz SH, Turnbull BA, Ross ME. Fecal DNA versus Fecal Occult Blood for colorectal-cancer screening in an average risk population. N Engl J Med 2004; 351:2704 –2714. 88. available: http://www.directlabs.com/testtypes.php, accessed January 27, 2005.

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1101

89. Song K, Fendrick AM, Ladabaum U. Fecal DNA testing compared with conventional colorectal cancer screening methods: a decision analysis. Gastroenterology 2004;126:1270 –1279. 90. Whitney D, Skoletsky J, Moore K, et al. Enhanced retrieval of DNA from human fecal samples results in improved performance of colorectal cancer screening test. J Molec Diag 2004; 6:385–395. 91. Boland CR, Montgomery CK, Kim YS. Alterations in human colonic mucin occurring with cellular differentiation and malignant transformation. Proc Natl Acad Sci U S A 1982;79:2051– 2055. 92. Kellokumpu IH, Andersson LC, Kellokumpu SJ. Detection of colorectal neoplasia with peanut-agglutinin (PNA) reactive carbohydrate structures in rectal mucus. Int J Cancer 1997;74: 648 – 653. 93. Filipe MI, Branfoot AC. Abnormal patterns of mucus secretion in apparently normal mucosa of large intestine with carcinoma. Cancer 1974;34:282–290. 94. Dawson PA, Filipe MI. An ultrastructural and histochemical study of the mucous membrane adjacent to and remote from carcinoma of the colon. Cancer 1976;37:2388 –2398. 95. Shamsuddin AM. A simple mucus test for cancer screening. Anticancer Res 1996;16:2193–2199. 96. Schreiber JB, Knodell RG, Shamsuddin AKM. Utilization of alterations in colonic mucus as a screening test for detecting colonic neoplasia. Am J Gastroenterol 1998;83:1058. 97. Knodell RG, Hale E, Antos M. A new screening test for the detection of colon cancer utilizing neoplasia associated alterations in rectal mucus. Gastroenterology 1988;94:A230. 98. Sakamoto K, Nakano G, Nagamachi Y. A pilot study on the usefulness of a new test for mass screening of colorectal cancer in Japan. Gastroenterol Jpn 1990;25:432– 436. 99. Sakamoto K, Muratani M, Ogawa T, et al. Evaluation of a new test for colorectal neoplasms: a prospective study of asymptomatic population. Cancer Biother 1993;8:49. 100. Macket C III, Wilson C, Brna T. Comparison of a galactose oxidase test to Hemoccult in colorectal cancer screening. Am J Gastroenterol 1989;84:1190. 101. Zhou DY, Feng FC, Zhang YL, et al. Comparison of Shams’ test for rectal mucus to an immunological test for fecal occult blood in large intestinal carcinoma screening. Analysis of a check-up of 6480 asymptomatic subjects. Chin Med J 1993;106:739 – 742. 102. Vucenik I, Gotovac J, Druzijanic N, Shamsuddin AM. Usefulness of galactose oxidase-Schiff test in rectal mucus for screening of colorectal malignancy. Anticancer Res 2001;21:1247–1255. 103. Vironen J, Kellokumpu S, Andersson LC, Kellokumpu I. Comparison of a peanut agglutinin test and an immunochemical faecal occult blood test in detecting colorectal neoplasia in symptomatic adults. Scand J Clin Lab Invest 2004;64:140 –145. 104. Hoskins LC. Mucin degradation in the human gastrointestinal tract and its significance to enteric microbial ecology. Eur J Gastroenterol Hepatol 1993;5:205–212. 105. Horsewood P, Evelegh MJ, Marcon NE, et al. Measurement of hue and chroma in assessment of biomarkers of colon and lung cancer: AACC Press, 2002. 106. Marcon N, Syngal S, Brenner D, et al. Quantitative galactose oxidase Schiff’s reactivity of rectal mucus correlates with colorectal cancer. In: American Association for Cancer Research Annual Meeting, 2003. 107. van Dam J, Cotton P, Johnson CD, et al. AGA future trends report: CT colonography. Gastroenterology 2004;127:970 –984. 108. Levin B, Brooks D, Smith RA, Stone A. Emerging technologies in screening for colorectal cancer. CA Cancer J Clin 2003;53:44 – 55.

1102

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

109. Fletcher JG, Johnson CD, Welch TJ, et al. Optimization of CT colonography technique: prospective trial in 180 patients. Radiology 2000;216:704 –711. 110. Pineau BC, Paskett ED, Chen GJ, et al. Virtual colonoscopy using oral contrast compared with colonoscopy for the detection of patients with colorectal polyps. Gastroenterology 2003;125: 304 –310. 111. Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evaluation of computed tomographic colonography for screen detection of colorectal polyps. Gastroenterology 2003; 125:311–319. 112. Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 2003;349:2191–2200. 113. Cotton PB, Durkalski VL, Pineau BC, et al. Computed tomographic colonography (virtual colonoscopy). JAMA 2004;291: 1713–1719. 114. Cohnen M, Vogt C, Beck A, et al. Feasibility of MDCT colonography in ultra-low-dose technique in the detection of colorectal lesions: comparison with high-resolution video colonoscopy. Am J Radiol 2004;183:1355–1359. 115. Hoppe H, Netzer P, Spreng A, Quattropani C, Mattich J, Dinkel HP. Prospective comparison of contrast enhanced CT colonography and conventional colonoscopy for detection of colorectal neoplasms in a single institutional study using second-look colonoscopy with discrepant results. Am J Gastroenterol 2004; 99:1924 –1935. 116. Iannaccone R, Laghi A, Catalano C, et al. Computed tomographic colonography without cathartic preparation for the detection of colorectal polyps. Gastroenterology 2004;127:1300 –1311. 117. van Gelder RE, Nio CY, Florie J, et al. Computed tomographic colonography compared with colonoscopy in patients at increased risk for colorectal cancer. Gastroenterology 2004;127: 41– 48. 118. Pickhardt PJ, Nugent PA, Choi JR, Schindler WR. Flat colorectal lesions in asymptomatic adults: implications for screening with CT virtual colonoscopy. Am J Radiol 2004;183:1343–1347. 119. Kamar M, Portnoy O, Bar-Dayan A, et al. Actual colonic perforartion in virtual colonscopy: report of a case. Dis Colon Rectum 2004;47:1242–1244. 120. Coady-Fariborzian L, Angel LP, Procaccino JA. Perforated colon secondary to virtual colonoscopy: report of a case. Dis Colon Rectum 2004;47:1247–1249. 121. Debatin JF, Lauenstein TC. Virtual magnetic resonance colonography. Gut 2003;52(Suppl IV):iv17–iv22. 122. Rajapaska RC, Macari M, Bini EJ. Prevalence and impact of extracolonic findings in patients undergoing CT colonography. J Clin Gastroenterol 2004;38:767–771. 123. Hellstrom M, Svensson MH, Lasson A. Extracolonic and incidental findings on CT colonography (virtual colonoscopy). Am J Roentgenol 2004;182:631– 638. 124. Tagore KS, Levin TR, Lawson MJ. Review article: the evolution to stool DNA testing for colorectal cancer. Aliment Pharmacol Ther 2004;19:1225–1233. 125. van Gelder RE, Birnie E, Florie J, et al. CT Colonography and colonoscopy: assessment of patient preference in a 5 week follow up study. Radiology 2004;233:328 –337. 126. Svensson MH, Svensson E, Lasson A, Hellstrom M. Patient acceptance of CT colonography and conventional colonoscopy: prospective comparative study in patients with or suspected of having colorectal disease. Radiology 2002;222:337–345. 127. Akerkar GA, Yee J, Hung R, McQuaid K. Patient experience and preferences toward colon cancer screening: a comparison of virtual colonoscopy and conventional colonscopy. Gastrointest Endosc 2001;54:310 –315. 128. Scott RG, Edwards JT, Fritschi L, Foster NM, Mendelson RM, Forbes GM. Community based screening by colonoscopy

GASTROENTEROLOGY Vol. 129, No. 3

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141. 142.

143.

144.

145. 146.

147.

or computed tomographic colonography in asymptomatic average-risk subjects. Am J Gastroenterol 2004;99:1145– 1151. Song K, Ladabaum U. Estimated national clinical and economic colorectal cancer burden and resource demand with different screening strategies. Gastroenterology 2004;126(Suppl 2):A14. Ladabaum U, Song K, Fendrick M. Colrectal neoplasia screening with virtual colonscopy: when, at what cost and with what national impact? Clin Gastroenterol Hepatol 2004;2:554 –563. Heitman S, Fong A, Dean S, Hilsden R, Manns B, Romagnuolo J. Cost-effectiveness study of CT colonography versus colonoscopy using decision analysis. Gastroenterology 2004; 126(Suppl 2):A202. Hur C, Gazelle GS, Zalis ME, Podolsky DK. An analysis of the potential impact of computed tomographic colonography (virtual colonoscopy) on colonoscopy demand. Gastroenterology 2004; 127:1312–1321. Saar B, Rosch T, Rummeny EJ. Colorectal cancer screening: a challenge for magnetic resonance colonography. Top Magn Reson Imaging 2002;13:427– 434. Saar B, Heverhagen JT, Obst T, et al. Magnetic resonance colonography and virtual magnetic resonance colonoscopy with the 1.0 T system. Invest Radiol 2000;35:521–526. Pappalardo G, Polettini E, Frattaroli FM, et al. Magnetic resonance colonography versus conventional colonoscopy for the detection of colonic endoluminal lesions. Gastroentereology 2000;119:300 –304. Luboldt W, Bauerfeind P, Wildermuth S, Marincek B, Fried M, Debatin JF. Colonic masses: detection with MR colonography. Radiology 2000;216:383–388. Morrin MM, Hochman MG, Farrell RJ, Marquesuzaa H, Rosenberg S, Edelman RR. MR Colonography using colonic distension with air as the contrast material: work in progress. Am J Radiol 2001;176:144 –146. Lam WW, Leung WK, Wu JK, So NM, Sung JJ. Screening of colonic tumors by air-inflated magnetic resonance colongraphy. J Magn Reson Imag 2004;19:447– 452. Lauenstein T, Holtman G, Schoenfelder D, Bosk S, Ruehm SG, Debatin JF. MR Colonography without colonic cleansing: a new strategy to improve patient acceptance. Am J Radiol 2001;177: 823– 827. Weishaupt D, Patak MA, Froelich J, Ruehm SG, Debatin JG. Faecal tagging to avoid colonic cleansing before MRI colonography. Lancet 1999;354:835– 836. Kahi CJ, Rex DK. Current and future trends in colorectal cancer screening. Cancer Met Rev 2004;23:137–144. Ko CW, Kreuter W, Baldwin LM. Utilization of colorectal cancer screening tests in Medicare beneficiaries in Year 2000. Gastroenterology 2004;126(Suppl 2):A57. Seeff LC, Manninen DL, Dong FB, et al. Is there endoscopic capacity to provide colorectal cancer screening to the unscreened population in the United States? Gastroenterology 2004;127:1661–1669. Vijan S, Inadomi J, Hayward RA, Hofer TP, Fendrick AM. Projections of demand and capacity for colonoscopy related to increasing rates of colorectal cancer screening in the United States. Aliment Pharmacol Ther 2004;20:507–515. Pignone M, Bucholtz D, Harris R. Patient preferences for colon cancer screening. J Gen Intern Med 1999;14:432– 437. Fisher DA, Martin C, Galanko J, Sandler RS, Noble MD, Provenzale D. Risk factors for advanced disease in colorectal cancer. Am J Gastroenterol 2004;99:2019 –2024. Goyen M, Goehde SC, Herborn CU, et al. MR-based full-body preventative cardiovascular and tumor imaging: technique and preliminary experience. Eur Radiol 2004;14:783–791.

September 2005

148. O’Malley PG, Taylor AJ. Unregulated direct-to-consumer marketing and self-referral for screening imaging services—a call to action. Arch Intern Med 2004;164:2406 –2408. 149. Fenton JJ, Deyo RA. Patient self-referral for radiologic screening tests: clinical and ethical concerns. J Am Board Fam Pract 2003;16:494 –501. 150. Pferrer N, Arber N, Halpern Z. Aer-O-Scope: Self-propelling, self-navigating, skill-independent colonoscopy; a new era in screening colonoscopy. [Abstract] Presented at Digestive Disease Week 2005.

Address requests for reprints to: Chair, Clinical Practice and Economics Committee, AGA National Office, c/o Membership Depart-

AMERICAN GASTROENTEROLOGICAL ASSOCIATION

1103

ment, 4930 Del Ray Avenue, Bethesda, Maryland 20814. Fax: (301) 654-5920. This report was prepared by Dr. Regueiro under the direction of the AGA Future Trends Committee. It was approved by the committee on May 15, 2005. Members of the committee: Nicholas F. LaRusso (chair), Juan R. Malagelada, Walter J. McDonald, Pankaj J. Pasricha, Suzanne Rose, Michael Lee Weinstein. The author gratefully acknowledges the expert advice and guidance of Dr. Robert E. Schoen, Professor of Medicine; Division of Gastroenterology, Hepatology and Nutrition; University of Pittsburgh, regarding content of this paper.