| |
Considerations Regarding the Present and Future Roles of Colonoscopy in Colorectal Cancer Prevention
|
| |
| |
Clinical Gastroenterology and Hepatology
May 2008
Douglas K. Rex, Emely Eid
Department of Medicine, Division of Gastroenterology, Indiana University School of Medicine, Indianapolis, Indiana
Effective and safe colonoscopy is essential to colorectal cancer prevention, regardless of the method used for colorectal cancer screening. The level of colorectal cancer incidence reduction provided by colonoscopy and polypectomy varies widely in available studies. There are several mechanisms by which colonoscopy might fail to prevent colorectal cancer, and some of the mechanisms might be overcome by simple currently available measures. Further, advances in colonoscope technology could enhance the effectiveness of colonoscopy or render it less operator-dependent. The large market for colorectal cancer screening in the United States has spawned innovative noncolonoscopic technologies for colorectal cancer and polyp detection. Because these technologies are diagnosis only, their overall impact on outcomes ultimately may be determined by whether they successfully increase adherence to screening (which should reduce colorectal cancer incidence) versus displace patients from colonoscopy screening (which potentially could increase colorectal cancer incidence), as well as their cost effectiveness and the extent to which they reduce colonoscopy complications. As these strategies emerge, monitoring their effects on adherence, cancer prevention, and procedural complications will be needed to optimize their roles relative to primary colonoscopy screening.
Concluding Remarks
Ultimately, the role of colonoscopy in colorectal cancer prevention and in gastrointestinal practice will depend on developments in technology, physician and public perceptions of the risks and benefits of new screening tests, and insurers' judgments about the relative costs, cost effectiveness, and cost savings of different strategies. Effective and safe CT colonography without bowel preparation or development of an effective proteomic or serum DNA assay are examples of technologies that might dramatically impact the use of screening colonoscopy. Gastroenterologists currently are involved in many aspects of new technology development and will be uniquely suited to direct patients to appropriate strategies based on assessment of polyp and cancer risk and procedure risk. Even if the role of colonoscopy becomes confined everywhere to diagnostic examinations (Figure 1), highly effective colonoscopy will be critical to any screening approach to colorectal cancer.
The initial steps in enhancing colonoscopy's effectiveness seem straightforward (Table 3). Colonoscopists everywhere must become experts in bowel preparation science. Investigators in bowel preparation should continue to look for an optimal combination of efficacy, safety, and tolerability. Colonoscopy should be performed everywhere in a careful and meticulous fashion, and this approach should be documented by a record of adequate adenoma detection by individual colonoscopists.15, 16 Polypectomy should use effective methods of resection. When methods known to be associated with ineffective resection are used, particularly piecemeal resection of large sessile adenomas, there should be aggressive follow-up evaluation. With a commitment to quality by every colonoscopist, perhaps we may fulfill the promise of the National Polyp Study.
Article Outline
· Abstract
· How Well Does Colonoscopy Prevent Colorectal Cancer?
· Why Is Protection Against Colorectal Cancer by Colonoscopy Imperfect? What Corrective Actions Can Be Taken?
· Projecting the Future Role of Colonoscopy
· Plotting New Directions for Colonoscopy
· Improved Performance and Technology
· Cost Reductions
· Concluding Remarks
In the prevention of colorectal cancer, all roads lead to colonoscopy (Figure 1). In most of the world, patients with symptoms consistent with colorectal cancer are evaluated by colonoscopy with the goal of achieving early diagnosis. Surveillance after removal of colorectal cancers and precancerous polyps also typically is performed by colonoscopy, although the intervals at which colonoscopy is recommended often vary between countries.1, 2, 3, 4 The worldwide approach to colorectal cancer screening is much more variable. Several countries rely on the fecal occult blood test or sigmoidoscopy for screening,5, 6 but patients with positive tests undergo colonoscopy. In such settings, effective and safe colonoscopy remains critical to effective colorectal cancer screening (Figure 1), as has been appreciated in the United Kingdom7 and in Australia.8 In the United States and in emerging fashion in some European countries,9, 10 primary colorectal cancer screening is being dominated progressively by colonoscopy.
The emergence of screening colonoscopy in the United States has transformed the practice of gastroenterology and substantially has impacted general surgical practices. Demand for gastroenterologists is high, as is pursuit of gastroenterology fellowship positions by internal medicine residents. Arguably, the emergence of screening colonoscopy has impacted the practice of gastroenterology more than any event of the past quarter century, rivaling in its impact breakthroughs such as the identification of Helicobacter pylori and the discovery of chronic hepatitis viruses and their treatment.
By every account, millions of colonoscopies are performed per year in the United States11 at a cost of billions of dollars, and with a nonnegligible risk of significant and occasionally deadly complications.12, 13, 14 Any health care leader should ask whether this investment and risk are associated with substantial and optimized benefits, particularly in a system that exerts limited control over which physicians perform colonoscopy, how well any physicians perform colonoscopy,15, 16 or how often physicians choose to repeat procedures in the same patients.17, 18 Colonoscopists should ask: are we providing for our patients the best possible chance of colorectal cancer prevention at the lowest risk? Are we serving our patients and country well by being effective and cost effective in preventing colorectal cancer? How can we improve?
This review discusses issues pertaining to the effectiveness of colonoscopy in colorectal cancer prevention, and discusses how resolution of these issues and the development of new alternative screening and imaging technologies may alter the role of colonoscopy in the United States.
How Well Does Colonoscopy Prevent Colorectal Cancer?
Recent evidence on cancer risk in the United States supports a general decline in cancer incidence, to which reductions in colorectal cancer incidence are an important contributor.19 In addition, the distribution of colorectal cancer stages has shifted toward earlier stages, coinciding with the emergence of screening colonoscopy.20 It is unknown to what extent the magnitude of these effects could be increased by improvements in the effectiveness of colonoscopy or, conversely, the extent to which the magnitude has been diminished by ineffective colonoscopy.
Evidence regarding the impact of colonoscopy on colorectal cancer prevention is mixed. The National Polyp Study found that patients with adenomas who had undergone clearing colonoscopy experienced a 76% to 90% reduction in the incidence of colorectal cancer compared with the expected incidence based on reference populations.21 Long-term follow-up evaluation identified a sustained impact on colorectal cancer incidence and mortality.22 A modeling analysis found that reduction in incidence and mortality during the first 10 years of follow-up evaluation was almost entirely the result of the baseline colonoscopy, and even at 20 years most of the mortality reduction resulted from the baseline procedure.23 The impact of the National Polyp Study cannot be underestimated because it shaped the public's and physicians' views of what is achievable through colonoscopy and polypectomy. The National Polyp Study was the foundation of the lobbying that led to legislation providing payment for screening colonoscopy for Medicare patients. The simple demonstration that screening colonoscopy could be performed safely and had a substantial yield24, 25, 26 was less important than evidence that colonoscopy and polypectomy prevent colorectal cancer. At least 4 other studies have reported 80% protection against colorectal cancer after colonoscopy.27, 28, 29, 30
However, other studies examining adenoma cohorts have documented greater rates of incident cancer after clearing colonoscopy than were observed in the National Polyp Study. Figure 2 displays incident cancer rates per thousand patient years of observation in the National Polyp Study, compared with the Wheat Bran Fiber Trial,31 the Polyp Prevention Trial,32 the Funen Adenoma Follow-up Study,33 and 3 chemoprevention trials performed in the United States.34 In the chemoprevention trials, colonoscopy and polypectomy had no measured protective effect against incident colorectal cancers compared with those expected based on Surveillance, End, Epidemiology Results.34 These results may not negate a protective effect because an adenoma cohort could have a higher incidence rate of cancer than the general population. Nevertheless, the protective level of colonoscopy and polypectomy in these adenoma cohorts was substantially less than that described in the National Polyp Study. The reasons for these differences are unclear, although one possibility is superior baseline clearing of adenomas in the National Polyp Study. For example, 13% of patients entering the National Polyp Study underwent 2 or more clearing baseline examinations before they were randomized. Further, patients with adenomas greater than 3 cm were excluded from the National Polyp Study, whereas all adenoma patients were included in some trials such as the Funen Adenoma Follow-up Study.33 Also, expected incidences of cancer in the various adenoma cohorts are difficult to predict given that the size and type of adenomas may vary between studies. Similarly, distinguishing prevalent from incident cancers in the control groups is problematic, and prevalent cancers were excluded from the adenoma cohorts. It is impossible to determine whether unmeasured differences in the quality of colonoscopy performance contribute to the observed differences in protection.
Other cohort and case-control studies have documented reductions in risk of colorectal cancer after colonoscopy, but the reductions were lower than observed in the National Polyp Study. For example, a case-control study found a 50% reduction in mortality after colonoscopy in US veterans.35 A large cohort of symptomatic patients had a less than 50% reduction in cancer incidence during the first 5 years after the colonoscopy.36 A case-control study in the Medicaid population in California involving 4458 cancer cases and 43,815 controls identified a 45% reduction in colorectal cancer incidence after colonoscopy.37 In both men and women, the reduction in left-sided cancers was 84%, whereas the reduction in right-sided colon cancer was 62% in men and only 18% in women.37 These studies19, 20 confirm that colonoscopy is associated with reductions in colorectal cancer risk, but the reductions are less than described in the National Polyp Study, although as noted previously, comparisons between the studies are complicated.
In summary, evidence on the protective level of colonoscopy against colorectal cancer is mixed. Why is colonoscopy not always as protective against colorectal cancer as we hoped for, and as many endoscopists and much of the public still believe? Several plausible contributors deserve investigation (Table 1). Whether these contributors account for all of colonoscopy's failures is uncertain, as is their relative quantitative importance. Nevertheless, pending definitive studies and explanations, they deserve careful consideration by colonoscopists.
Why Is Protection Against Colorectal Cancer by Colonoscopy Imperfect? What Corrective Actions Can Be Taken?
Table 1 lists a priori reasons why a patient might undergo colonoscopy and then present in the next few years with colorectal cancer. Of the 5 reasons listed, at least 4 appear to be correctable to some extent by improvements either in preparation, technique, or colonoscope technology. The relative quantitative importance of each of the explanations is uncertain and the cause often is hard to identify in individual cases of interval cancers. The least correctable etiology for interval cancers would appear to be biologic causes, in which a colorectal cancer develops rapidly from either endoscopically normal-appearing mucosa, or from a tiny precancerous lesion. The best understood mechanism for rapid growth is inactivation of a mismatch repair gene. Mismatch repair gene inactivation accounts for essentially all cases of colorectal cancer arising in hereditary nonpolyposis colorectal cancer (HNPCC) and contributes to about 15% of sporadic colorectal cancers.38, 39, 40 Sporadic cancers with microsatellite instability typically show inactivation of the hMLH1 gene by promoter hypermethylation. The appearance of colorectal cancers within a few years of a negative colonoscopy has been described repeatedly in HNPCC gene carriers41, 42 and serves as the basis for current recommendations to perform colonoscopy in HNPCC patients at short intervals.43 Hypermethylation of hMLH1, accompanied by microsatellite instability and a high prevalence of BRAF mutations, has been associated with a hyperplastic polyp to serrated adenoma to cancer sequence that appears to occur primarily in the proximal colon.44 This altered polyp to cancer sequence might be a contributor to the recent observation of lower levels of protection against proximal colorectal cancer in females because elderly women have the highest incidence of these tumors.44 However, it is possible that even interval cancers attributable to biologic mechanisms could be prevented by improvements in colonoscopy. For example, even small adenomas may be important in HNPCC because the adenoma-carcinoma sequence in HNPCC can be more rapid. Techniques such as chromoendoscopy45 and high-definition endoscopy46 appear capable of detecting tiny adenomas, and although the value of these technologies in the average-risk population is uncertain, they might have a useful impact in HNPCC. There is currently no understanding of the detection rate of colonoscopy for proximal colon hyperplastic polyps or serrated adenomas, or the variation between endoscopists in the detection of these lesions. That they are endoscopically a distinct group of lesions seems certain,47 but the general level of recognition of the lesions by endoscopists, the level of awareness of their potential importance, and the willingness to remove them effectively given that they often are flat and located in the proximal colon is entirely unknown at present. These questions are important areas of investigation for colonoscopists. Only further investigation can determine the extent to which biologic explanations for interval cancers48 are correctable by improvements in colonoscopy or meticulous inspection.
The remaining explanations for interval cancers (Table 1) each appear to be at least partly correctable. Bowel preparation has been associated with impaired detection of small49, 50 and large adenomas50 and recently has been associated with impaired polyp detection in prospective clinical trials.51, 52 Clinical Outcomes Research Initiative investigators reported that 24% of colonoscopies in the United States were designated as having ginadequateh bowel preparation.49 The economic impact of inadequate bowel preparation is enormous.53
The colonoscopy community thus far has failed to identify a bowel preparation that optimizes efficacy, tolerability, and safety. Bowel preparation frequently is worse in the right colon compared with the left colon, and many clinical trials now report separate cleansing scores for the ascending colon.54 We believe the most consistent error in bowel preparation instructions contributing to ineffective preparation is failure to use split dosing. Split dosing refers to administration of half of the preparation on the evening before the colonoscopy and the other half on the day of the colonoscopy, preferably a few hours before the scheduled time for the examination. Direct testing has now established that split dosing is superior to dosing all of the preparation on the day before colonoscopy for both sodium phosphate51, 55 and polyethylene glycol.51, 56 Anecdotally, the most commonly forwarded arguments against split dosing include the inconvenience of arising during the night to ingest preparation when patients are scheduled for early morning colonoscopy and objections by some anesthesiologists who prefer that patients ingest nothing by mouth after midnight the night before colonoscopy. In our experience, patients who understand the importance of split dosing almost invariably are willing to arise during the night to take preparation doses at the preferred times, although the issue deserves formal study. Insistence by anesthesiologists that patients be nil per os after midnight the night before colonoscopy is inconsistent with American Society of Anesthesiology guidelines indicating that patients may take clear liquids until 2 hours before being sedated.57 Indeed, there is no evidence that split dosing increases the risk of aspiration during colonoscopy.
A second factor that is correctable and a potential cause of interval cancers is ineffective inspection technique. Older evidence that adenoma detection varies substantially between endoscopists suggested the need for quality indicators and targets for colonoscopy that measure the quality of mucosal inspection. In 2002, the US Multisociety Task Force recommended that endoscopists should know their individual adenoma detection rates and that they should be at least 25% in men and 15% in women, age 50 and older, undergoing colonoscopy.15 These recommendations were reiterated by a joint task force of the American Society for Gastrointestinal Endoscopy and the American College of Gastroenterology in 2006.16 As a secondary end point, individual endoscopists are recommended to have average withdrawal times of at least 6 minutes in normal colonoscopies in which no biopsies or polypectomies are performed.15, 16 Despite this, few colonoscopists in the United States have yet measured their adenoma detection rates.
Recent evidence has extended the importance of measuring these indicators by establishing that highly variable detection extends to large adenomas58, 59 and by validating the 6-minute withdrawal recommendation.58, 60 In a prospective study of screening colonoscopy, adenoma detection rates were found to vary by more than 10-fold among 12 experienced gastroenterologists in a private practice in Rockford, Illinois.58 There was a strong correlation between improved overall adenoma detection, as well as large adenoma detection, and longer withdrawal time in patients who had normal colonoscopies. Among those gastroenterologists with average withdrawal times greater than 6 minutes, the detection of adenomas 1 cm in size or greater was more than twice that (6.4% vs 2.6%) of those with average withdrawal times of less than 6 minutes.58 Similarly, in a study of 9 endoscopists performing more than 10,000 colonoscopies at Indiana University Hospital, there was a 4-fold difference among endoscopists in the range of adenoma detection, and the highest detector found 11 times more patients with 3 or more adenomas and 3.7 times more patients with adenomas 1 cm or greater compared with the lowest detector.59
The potential impact of this magnitude of variable detection on the effectiveness of postpolypectomy surveillance guidelines is staggering. By necessity, postpolypectomy surveillance guidelines assume that the performance of colonoscopy is comparable between operators,1, 2, 3, 4 although it is increasingly clear that this is far from the case. Highly operator-dependent detection of adenomas creates a scenario in which the most effective colonoscopists bring back more patients at short intervals, although these patients have had their colons cleared more effectively and are in less need of short-term follow-up evaluation. At the same time, the least effective adenoma detectors bring back fewer patients at short intervals, and they tell more patients who have had both large and small adenomas go undetected to return at 10-year intervals (although the possibility that low adenoma detectors are the same individuals that overuse colonoscopy for surveillance deserves investigation). The need for processes, standards, and incentives to reduce the operator dependency of adenoma detection, particularly large adenoma detection, is clear.
Although measurement of adenoma detection rates could be facilitated by electronic medical records that link pathology reports to procedure reports, the process is relatively easy and inexpensive even when it requires chart review. There has been no suggestion that adenoma detection rates have to be measured in a continuous or ongoing fashion, or even in a repeated fashion when the rates exceed recommended thresholds. Overall adenoma detection as an end point is certainly a surrogate for the detection of advanced adenomas and cancer. However, the feasibility of the review process is facilitated greatly by using overall adenoma detection (fewer charts need review to establish detection rates with narrow confidence intervals), problems with interobserver variation in polyp size measurement are averted, and overall adenoma detection and resection correlates with large adenoma detection and resection.58, 59 Given that adenoma detection is the purpose of most colonoscopies, it seems reasonable to establish adenoma detection rates as a priority for colonoscopists.
Although evidence that variable adenoma detection is a significant problem among gastroenterologists is overwhelming, there is also substantial evidence that on average detection by gastroenterologists during colonoscopy is superior to that of primary care physicians. At least 3 large studies have documented this effect with regard to cancer detection.61, 62, 63 The differences presumably reflect differences in training. There is less evidence about the relative performance of gastroenterologists compared with general surgeons, although at least one study found that cancer detection by general surgeons was inferior (on average) to gastroenterologists.61 There is no evidence regarding the comparative performance of gastroenterologists and colorectal surgeons. Overuse of postpolypectomy surveillance colonoscopy relative to gastroenterologists has been described for general surgeons17 and primary care physicians.18 These differences underscore the importance of incorporating quality guidelines and measurements into the practices of nongastroenterologist colonoscopists.
A third potentially correctable cause of interval cancers is ineffective polypectomy (Table 1). Two studies have suggested that ineffective polypectomy underlies more than one quarter of interval cancers, although the possibility of missed synchronous neoplasms cannot be excluded.64, 65 How to correct ineffective polypectomy is uncertain. Current postpolypectomy surveillance guidelines recommend aggressive follow-up evaluation of patients undergoing piecemeal resection of large sessile adenomas1 because a piecemeal technique is associated with a 14% to 55% chance of incomplete polypectomy at the first attempt.66 The extent to which such patients are followed up closely in clinical practice is uncertain. One small randomized trial found that treatment of the edges of the polypectomy site with the argon plasma coagulator effectively lowered the recurrence rate after piecemeal polypectomy.67 Beyond this, we have no data from randomized controlled trials to guide us in any specific aspects of polypectomy resection technique.66 As a result, polypectomy technique in clinical practice is remarkably variable,68 and based on anecdotally developed techniques that have not been tested scientifically to compare their safety and efficacy.
Finally, it is possible that advances in colonoscope or accessory technology could improve adenoma detection or reduce interobserver variability in adenoma detection (Table 2). These technologies have been discussed elsewhere.69 To summarize, although several hold considerable promise, none of them has been established clearly as both effective and practical in Western populations.69 Certainly, there has been no suggestion that any of these technologies can substitute for an examination that is performed carefully using an adequate interval of time in a well-cleansed colon.
Projecting the Future Role of Colonoscopy
Will colonoscopy continue to dominate colorectal cancer screening in the United States, or will its role become confined to diagnostic and/or surveillance colonoscopy? Whether improvements in colonoscopy performance and technology occur, as already discussed, could impact whether the role of colonoscopy in the United States changes. More effective, safer, and less costly colonoscopy could prolong the domination of colonoscopy in colorectal cancer screening. However, the expansion of colorectal cancer screening in the United States has created an economic market that has attracted technology innovators, typically seeking to find a niche that fulfills a deficiency of colonoscopy screening. Indeed, colonoscopy has major deficiencies other than its failure to detect and remove all precancerous lesions and early cancers from the colon. These include a substantial risk of perforation,12, 13, 14 the need for bowel preparation, significant numbers of patients who are unwilling to undergo the procedure for a variety of reasons,70 and the need for sedation (which is arguably both an advantage and a disadvantage for colonoscopy). The use of current alternatives to colonoscopy such as guaiac-based fecal occult blood testing, flexible sigmoidoscopy, and barium enema has decreased in the United States whereas colonoscopy use has skyrocketed.71 New innovations need to overcome some or all of the problems with current alternative tests to substantially alter the role of colonoscopy.
New tests to screen for colorectal cancer can be roughly grouped into 2 categories. One category includes new stool-based assays and potentially blood tests that compete primarily with the guaiac-based fecal occult blood tests for detection of colorectal cancer. The other category is alternative imaging strategies that aim, such as colonoscopy, to detect precancerous polyps and thus prevent colorectal cancer.
In countries (and certain regions and health care systems in the United States) with limited resources to perform screening colonoscopy or where the wisdom of screening colonoscopy is not accepted, the role of colonoscopy is unlikely to change given the emergence of new tests. Colonoscopy will continue to be used in the screening process primarily to evaluate other positive screening tests (Figure 1). In these settings, fecal immunochemical tests and Hemoccult SENSA (Beckman Coulter, Fullerton, CA) appear promising as substitutes for older guaiac-based fecal occult blood tests because they are more effective and still quite affordable.72 At current performance levels fecal DNA73 testing offers no clear advantage over improved fecal occult blood tests. The alternative imaging strategies could have an impact in these settings if the manpower to perform colonoscopy is limited and there is interest in imaging the colon for the purpose of adenoma detection. Thus, both capsule colonoscopy74, 75 and the Aer-O-Scope76 (GI View Ltd, Ramat Gan, Israel) are operator-independent procedures with regard to their technical performance (perhaps not their interpretation), and could be performed by nonphysician personnel, and thereby could increase the availability of polyp imaging. Similarly, computerized tomography (CT) colonography could be used to increase the availability of imaging technology and could be interpreted by nonradiologist personnel.77, 78, 79, 80 Because of limited availability, colonoscopy would be used in a therapeutic role when other imaging tests have detected polyps, particularly large polyps. In these settings the roles of colonoscopy for diagnosis in symptomatic patients and for surveillance also might be supplanted by alternative imaging techniques.
The projected role of colonoscopy in the United States and any other country that has adopted screening colonoscopy is more complex. For patients unwilling to undergo colonoscopy,70 the situation is analogous to that in countries that do not use primary screening colonoscopy. The role of alternative imaging tests (Figure 3) is particularly important to consider because these tests have polyp-detection capabilities, and thus could serve as an alternative to colonoscopy for cancer prevention. These tests might be used to increase adherence or to displace colonoscopy from a primary screening role in the segment of the population that is currently willing to undergo primary screening colonoscopy. However, each of the new imaging tests is expensive and to be competitive from a cost perspective, each test must significantly limit the number of patients with detected polyps who undergo colonoscopy. Otherwise, the strategy cost of using these tests will be too high. American College of Radiology reporting recommendations for CT colonography effectively reduce the proportion of patients undergoing polypectomy because they recommend that polyps 5 mm or smaller should not be mentioned on CT colonography reports, and patients with 1 or 2 polyps 6 to 9 mm in size may be offered a gsurveillanceh CT colonography in 3 years, rather than polypectomy.81 The medical rationale for this approach is that the risk of conversion of large adenomas to cancer over several years appears small.82 Although the natural history of small polyps has received only limited study,83 the risk of conversion to cancer in the short term is certainly lower than for large polyps. Therefore, the cancer risk associated with CTC surveillance of 6- to 9-mm polyps might be minimal. In addition, these strategies could reduce the complication burden associated with colonoscopic resection of small polyps. However, in a recent decision analysis, delaying polypectomy in 6- to 9-mm polyps for even 3 years resulted in a 20-fold greater rate of colorectal cancers and a 10-fold greater rate of colorectal cancer mortality compared with immediate polypectomy.84 Thus, if a new imaging test is used primarily to displace patients from colonoscopy (Figure 3), the effect potentially could be to increase colorectal cancer compared with continued use of colonoscopy screening. Alternatively, if new imaging tests are used primarily to increase adherence to colorectal cancer screening, then the benefits would be clear. Increased adherence results in improved cost effectiveness85 and identification of medium- and large-sized polyps in previously nonadherent patients would increase polypectomy rates and reduce colorectal cancer deaths (Figure 3). Unfortunately, there is no clear evidence that any of the tests actually will significantly improve adherence in their current forms.86 Investigation of the impact on adherence of these tests should be a top priority for colorectal cancer prevention investigators.
Overall, the impact of alternative colorectal imaging strategies on outcomes such as colorectal cancer prevention and colonoscopy complications is uncertain and hard to predict, and will be important to monitor as these strategies emerge.
One possible approach to noncolonoscopic imaging tests that appears to optimize their benefits and cost effectiveness is to triage their use to populations with low anticipated polyp prevalence based on demographic factors such as age and sex.87 In this approach, diagnosis-only strategies are used for primary screening in low-polyp prevalence populations, and colonoscopy is used in high-polyp prevalence populations.87 This intermediate approach would mitigate potential negative effects on cancer prevention of reduced polypectomy, and at the same time reduce complications of diagnostic colonoscopy in low-prevalence populations. Further, the approach appears cost effective.87
In summary, the use of alternative imaging strategies for colon polyps in countries such as the United States that currently are using colonoscopy for screening and that are committed to polyp detection and resection as a means of cancer prevention, will require careful monitoring. The use of these tests to increase adherence to colorectal cancer screening or to overcome manpower deficiencies in providing colonoscopy would appear to have clear benefits for cancer prevention. The use of these tests on a widespread basis as an alternative to colonoscopy screening would reduce polypectomy substantially, which might increase colorectal cancer incidence. Research into the potential of alternative imaging tests to improve adherence is as important as establishing the performance characteristics of the techniques. The ultimate position of colonoscopy in primary screening will depend on evidence and perceptions regarding the issues of adherence, effectiveness, and complications.
Plotting New Directions for Colonoscopy
Improved Performance and Technology
For the foreseeable future, colonoscopy is likely to be the diagnostic test of choice in patients with positive screening tests and patients with symptoms strongly predictive of colorectal cancer, particularly the bleeding symptoms: hematochezia, iron-deficiency anemia, and melena with a negative upper endoscopy.88 Therefore, every possible advance with regard to correcting causes of interval cancers is worthy of pursuit (Table 3). As noted earlier, widespread use of measures that improve colonoscopy efficacy and safety and improved colonoscope technology are likely to improve the position of colonoscopy as a continued primary screening tool relative to emerging imaging technologies.
In addition to advances (Table 3) that are immediately achievable, colonoscopy is no longer a stagnant technology and new developments in imaging (Table 2) and possibly even in insertion platforms89, 90 could enhance the effectiveness or cost effectiveness of colonoscopy.
Cost Reductions
As imaging technology advances, the detection of tiny adenomas is likely to increase,46 which could negatively impact the cost effectiveness of colonoscopy by increasing the associated polypectomy and pathology charges, perhaps without substantially enhancing effectiveness. One potential approach to this development is to ignore tiny adenomatous lesions, but the strength of colonoscopy has always been its potential for simultaneous recognition and resection of precancerous lesions. An alternative approach would be to recognize and remove tiny adenomas, accompanied by expansion of the portion of the adenoma cohort considered at low risk, and for whom postpolypectomy surveillance intervals can be up to 10 years, and perhaps even longer if examination includes complete clearing of even the tiniest and flattest lesions. The cost of pathology evaluation of very small polyps is quite substantial and the main use of the information obtained from pathologic resection of polyps 5 mm or less in size is to guide postpolypectomy surveillance intervals. If reliable real-time methods of polyp histology could be developed, small polyps could be destroyed or resected and discarded, and postpolypectomy surveillance could be based on real-time endoscopic assessment of histology. The cost savings associated with this development are obvious and the relative cost effectiveness of colonoscopy and polypectomy compared with alternative imaging strategies would be enhanced by this development.
Whether involvement in colonoscopy sedation by anesthesia specialists increases or decreases will be another key determinant of the relative cost effectiveness of colonoscopy compared with alternative diagnosis-only imaging strategies.
Concluding Remarks
Ultimately, the role of colonoscopy in colorectal cancer prevention and in gastrointestinal practice will depend on developments in technology, physician and public perceptions of the risks and benefits of new screening tests, and insurers' judgments about the relative costs, cost effectiveness, and cost savings of different strategies. Effective and safe CT colonography without bowel preparation or development of an effective proteomic or serum DNA assay are examples of technologies that might dramatically impact the use of screening colonoscopy. Gastroenterologists currently are involved in many aspects of new technology development and will be uniquely suited to direct patients to appropriate strategies based on assessment of polyp and cancer risk and procedure risk. Even if the role of colonoscopy becomes confined everywhere to diagnostic examinations (Figure 1), highly effective colonoscopy will be critical to any screening approach to colorectal cancer.
The initial steps in enhancing colonoscopy's effectiveness seem straightforward (Table 3). Colonoscopists everywhere must become experts in bowel preparation science. Investigators in bowel preparation should continue to look for an optimal combination of efficacy, safety, and tolerability. Colonoscopy should be performed everywhere in a careful and meticulous fashion, and this approach should be documented by a record of adequate adenoma detection by individual colonoscopists.15, 16 Polypectomy should use effective methods of resection. When methods known to be associated with ineffective resection are used, particularly piecemeal resection of large sessile adenomas, there should be aggressive follow-up evaluation. With a commitment to quality by every colonoscopist, perhaps we may fulfill the promise of the National Polyp Study.
|
|
| |
| |
|
|
|
