Gastroenterology
Volume 133, Issue 4 , Pages 1364-1367, October 2007

Toward Risk Stratification for Screening and Surveillance of Colorectal Neoplasia: One Small Step for the Colonoscopist

  • Thomas F. Imperiale

      Affiliations

    • Corresponding Author InformationAddress requests for reprints to: Thomas F. Imperiale, MD, The Regenstrief Institute, Inc, 1050 Wishard Blvd, RG 6509, Indianapolis, Indiana 46202. Fax: 317-630-6611.

Division of Gastroenterology, Department of Medicine, Indiana University School of Medicine, and Regenstrief Institute, Inc, Indianapolis, Indiana

Article Outline

 

See “Five-year colon surveillance after screening colonoscopy” by Lieberman DA, Weiss DG, Harford WV, et al, on page 1077 and “Colonoscopic screening of first-degree relatives of patients with large adenomas: increased risk of colorectal tumors” by Cottet V, Pariente A, Nalet B, et al, on page 1086.

In 2002, more than 14 million colonoscopies were performed in the United States,1 of which an estimated 38%–45% were for screening and 22% for surveillance of persons with previous neoplasia.1, 2 Although several screening strategies are supported by professional organizations, colonoscopy is currently very popular. The general public seems to equate colorectal cancer (CRC) screening with colonoscopy, and appears to be less aware of other screening tests and strategies despite the fact that careful analyses by the U.S. Preventive Services Task Force and Institute of Medicine did not find colonoscopy every 10 years to be a superior screening strategy.3, 4 In addition to its use for screening, colonoscopy is the only modality used for surveillance of persons with previous neoplastic polyps. Although it is a powerful tool for screening, diagnosis, and management of colorectal neoplasia, colonoscopy is a limited resource; suboptimal allocation of it or lack of preference for it may prevent some persons from having any CRC screening. If we knew more about risk factors for advanced neoplasia and could effectively stratify risk, then screening and surveillance could be tailored—that is, targeted to persons with high risk and away from those at low risk, who could either be examined with less invasive methods or perhaps not at all. Risk stratification and tailoring would improve the overall efficiency and effectiveness of colonoscopic screening and surveillance.

One group generally thought to be at increased risk for CRC is patients with previous colorectal neoplasia. Periodic colonoscopic surveillance is commonly done in this group. Another group considered to be at high risk is patients with a family history of neoplastic polyps. Screening recommendations for this group are less consistent regarding when to begin screening and how to screen. Because both groups represent substantial proportions of the population, risk stratification within these groups would help to conserve resources and eliminate risk among persons who stand to gain little or nothing from colonoscopy. By providing a basis for risk stratification—and in particular, by identifying low-risk groups that do not need intense evaluation—2 studies in this issue have direct relevance for screening and surveillance. One study supports the current practice of colonoscopic surveillance for various types of advanced neoplasia, and suggests that persons with 1–2 small tubular adenomas may be treated nearly the same as those with no colorectal neoplasia.5 The other study identifies factors that may be useful for deciding when to screen persons with a family history of large adenomatous polyps.6

The study with relevance to surveillance is a follow-up study5 of patients who participated in VA Cooperative Study #380,7 which described the most advanced histologic findings in 3121 veterans who had colonoscopic screening.5 The objective was to quantify the 5.5-year incidence of advanced colorectal neoplasia as a function of baseline colorectal findings. Nearly 1200 patients with baseline neoplasia were assigned to colonoscopic surveillance, of whom 895 (76%) complied. A unique aspect of the study was the inclusion of 501 persons with no adenomas, 298 (59.5%) of whom complied with surveillance, to allow a comparison with the groups with different degrees of neoplasia. Relative risks of any and advanced neoplasia were generated based on initial findings, and were expressed relative to the “no neoplasia” group.

Among persons with no baseline neoplasia, 24.8% and 2.3% had any and advanced neoplasia, respectively, on follow-up colonoscopy, whereas corresponding proportions were 45.9% and 9.6% among those with baseline neoplasia. Importantly, a significant difference in the incidence of advanced neoplasia was not found between the subgroups with no baseline neoplasia and 1–2 small (<1 cm) tubular adenomas (relative risk, 1.92; 95% confidence interval [CI], 0.83–4.42). However, there was an increased risk for advanced neoplasia in all other subgroups. For the outcome of high-grade dysplasia or cancer, the incidence rates increased in a nearly linear fashion as baseline findings progressed from >2 small adenomas to large tubular or villous adenomas to high-grade dysplasia to cancer. Among persons with villous polyps, rates of advanced neoplasia were 13.7% for the first surveillance examination and remained high (11.9%) in those who had a second examination during the study period. Among the 1193 persons who had at least 1 follow-up colonoscopy, 21 (1.8%) developed “interval” lesions containing high-grade dysplasia or cancer, 11 of whom had baseline high-grade dysplasia or cancer; 6 of 11 lesions were found in a similar portion of the colon as the initial lesion.

This paper contains important messages. First, the findings confirm previous studies on subsequent risk for colorectal neoplasia and support the rationale for surveillance and the current guidelines for tailoring of surveillance intervals based on initial endoscopic and histologic findings (Table 1). The Multisociety Task Force (MSTF) recently revised the recommended surveillance interval for persons with 1 or 2 small tubular adenomas from 5 years to 5–10 years.8 With the most recent revision of its guidelines, the MSTF’s recommendations are now more in line with earlier statements from its predecessor, the GI Consortium. In 1997, the GI Consortium stated that persons with 1 or 2 small adenomas had a risk for colon cancer that was “no greater than the general population.”9 In fact, an earlier surveillance guideline stated “… patients undergoing resection of a single, small tubular adenoma (<1 cm) may not have an increased subsequent risk for cancer, follow-up surveillance may not be indicated.”10 These statements were based largely on a cohort study of 1618 persons with rectosigmoid adenomas who underwent rigid sigmoidoscopy and polypectomy, but who did not undergo subsequent colonoscopy to look for proximal neoplasia.11 Follow-up of persons with small rectosigmoid tubular adenomas showed a risk for colon cancer that was no greater than expected, with a standardized incidence ratio of 0.5 (95% CI, 0.1–1.3). Despite this evidence, and the statement by the Consortium that the risk in persons with 1–2 small tubular adenomas was no different than average risk, the recommendations of the consortium at that time were more aggressive than the evidence supported. With its most recent statement, the MSTF’s recommendation for this subgroup has evolved to more clearly reflect the evidence. The findings by Lieberman et al provide some support for this recent revision, as persons with 1–2 small tubular adenomas were not found to have a greater risk for subsequent advanced neoplasia than those with no baseline neoplasia. However, given the point estimate for the relative risk (1.92) and the width of the confidence interval (0.83–4.42), the flexibility of the 5- to 10-year interval seems appropriate for now.

Table 1. Summary of the Multi-society Task Force’s Recommendations for Colonoscopic Surveillance8
Polyp CategoryTime to First Surveillance
Small rectal hyperplastic polyps10 years (same as normal colonoscopy)
1–2 small (<1 cm) tubular adenoma5–10 years
3–10 adenomas, any adenoma ≥1 cm3 years
Adenoma with villous features or high-grade dysplasia3 years
>10 adenomas<3 years (per endoscopist judgment)
Sessile adenoma removed piecemeal2–6 months, then per endoscopist judgment
Post cancer resection1 year

Second, performing surveillance at shorter intervals is supported by the finding of higher relative risks for advanced neoplasia in the other subgroups of persons with baseline neoplasia. Third, and perhaps most importantly, the study raises concerns and important questions about those patients who developed interval high-grade dysplasia or cancer. Are the reasons procedure related (preparation quality), endoscopist related (adenoma detection rate, withdrawal time, polypectomy technique), or patient related (phenotypic factors such as cigarette smoking, body mass index vs genotypic factors), or some combination of all 3? Which factors are most important in understanding risk for serious interval lesions? Understanding which factors are related to high risk (as well as to low risk) for these lesions is necessary to further tailor surveillance, a goal that is worth pursuing.

The study with relevance to screening is a case-control study, the purpose of which was to determine whether first-degree relatives (FDRs) of patients with large (≥10 mm) adenomas are at increased risk for colorectal neoplasia.6 Cases were FDRs of persons with large adenomas who agreed to undergo colonoscopy; controls were randomly selected from patients who underwent colonoscopy for “minor” symptoms. Although there was no difference in the prevalence of any neoplasia (22.6% in cases vs 16.3% in controls; odds ratio [OR], 1.56; CI, 0.96–2.53), the combination of cancer and large adenomas was greater among cases (8.4%) than controls (4.2%; OR, 2.27; CI, 1.01–5.09). Neither cancer nor large or advanced adenomas alone was more prevalent among cases, however, likely owing to the small sample size. The relative risk of finding advanced neoplasia in cases was higher when the index case was <60 years old (OR, 3.82; CI, 0.92–15.87), male (OR, 4.01; CI, 1.45–11.09), or had large distal adenomas (OR, 3.14; CI, 1.27–7.73).

In contrast to several previous studies that have assembled study groups and data retrospectively or that have used interview responses alone to categorize the neoplasia status of cases and/or controls,12 this study prospectively evaluated with colonoscopy the FDRs of patients with large adenomas. As pointed out by the authors, a more desirable study design would be to perform colonoscopy on FDRs of persons with large adenomas and on FDRs of persons with no adenomas to ensure accurate ascertainment of family history and colonoscopic findings for both groups. However, such a study would have logistical and ethical challenges and is probably not feasible. Despite the prospective evaluation of cases, the study included some FDRs with a family history not just of large adenomas, but of CRC. Although exclusion of these cases and their matched controls had a clinically insignificant effect on the magnitude of relative risk (2.27 with vs 2.09 without), there was a loss of statistical significance. Other study limitations are well-described by the authors, although it is unclear what the cumulative effect of these limitations may have had on the study’s relative risk estimates. In addition, there is a practical limitation to the clinical application of this study: Information about the size, number, and type of polyps of an FDR is often unavailable when deciding about whether and how to screen the patient under consideration.

Despite its limitations, the study by Cottet et al contains some potentially useful information. The results suggest that FDRs of patients with large adenomas have a roughly 2-fold increased risk (and absolute risk of about 8%) for the combined finding of CRC and large adenoma. Of particular interest is the finding of increased risk when the index case had the features of age <60 years, male gender, or distal adenoma location, a finding consistent with at least one other study of adenomas and FDR risk (for adenoma location)13 and with this same group’s work on familial risk of CRC.14 On the flip side, the relative risk was not significantly elevated (and in some cases, numerically lower) among FDRs of index cases aged ≥60, female, and proximal adenoma location. Thus, the authors may have identified features that help to stratify risk for the subgroup of persons who have FDRs with large adenomas. Although these exploratory findings require validation, how should the current results affect clinical practice, if at all?

First, we need further high-quality studies with adequate statistical power to validate the current findings. A similar study should be done in FDRs of patients with 1–2 small tubular adenomas. Some guidelines recommend screening persons with an FDR with an adenomatous polyp of any size as aggressively as those with an FDR with CRC.15, 16 The study by Cottet et al implies that this practice may be overly aggressive: If low-risk subgroups risk were identifiable among persons who have an FDR with a large adenoma, then it is very likely that a low-risk subgroup can be identified among persons who have an FDR with a small adenoma. Second, we need to think about solutions to the logistical limitations of applying this study to clinical practice, which would involve taking a good family history and verifying the findings (ie, histology, size, and location) of all polyps of FDRs. Currently, we are often challenged by establishing information about previous polyps for patients in the endoscopy suite who are prepped for surveillance colonoscopy. Finally, as raised by the authors, we need to consider how best (ie, effectiveness, cost-effectiveness, acceptability) to screen persons at increased risk for colorectal neoplasia based on an FDR with adenomatous polyps. Periodic colonoscopy will likely be used for this indication in the United States, although other recommended strategies may be as good or better. Cost-effectiveness models, of which there are several,4, 17, 18, 19, 20 may be able to inform this issue.

In summary, these 2 articles have important clinical implications for how screening and surveillance are performed. The article by Lieberman et al supports current practice and less aggressive surveillance of persons with 1–2 small tubular adenomas. At the same time, it raises questions about the early recurrence of advanced neoplasia, the answers to which may enhance our ability to further stratify risk and tailor surveillance. With regard to screening, the article by Cottet et al suggests that FDRs of persons with a large adenoma are a high-risk group, and identifies factors that may allow risk stratification within this group. Both articles advance our knowledge about tailoring colonoscopic screening and surveillance; both articles suggest directions for future research. Risk stratification of screening and surveillance based on phenotypic (and perhaps proteomic and genotypic) factors is the giant leap; these studies are a small step in the right direction.

Back to Article Outline

References 

  1. Seeff LC, Richards TB, Shapiro JA, et al. How many endoscopies are performed for colorectal cancer screening? (Results from CDC’s survey of endoscopic capacity). Gastroenterology. 2004;127:1670–1677
  2. Lieberman DA, Holub J, Eisen G, et al. Utilization of colonoscopy in the United States: results from a national consortium. Gastrointest Endosc. 2005;62:875–883
  3. Pignone M, Russell L, Wagner J. Economic models of colorectal cancer screening in average-risk adults: workshop summary. Washington, DC: Academies Press; 2005;
  4. Pignone M, Saha S, Hoerger T, et al. Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med. 2002;137:96–104
  5. Lieberman DA, Weiss DG, Harford WV, et al. Five-year colon surveillance after screening colonoscopy. Gastroenterology. 2007;133:1077–1085
  6. Cottet V, Pariente A, Nalet B, et al. Colonoscopic screening of first-degree relatives of patients with large adenomas: increased risk of colorectal tumors. Gastroenterology. 2007;133:1086–1092
  7. Lieberman DA, Weiss DG, Bond JH, et al. #380 VG (Use of colonoscopy to screen asymptomatic adults for colorectal cancer). N Engl J Med. 2000;343:162–168
  8. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US multi-society task force on colorectal cancer and the American Cancer Society. Gastroenterology. 2006;130:1872–1885
  9. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology. 1997;112:594–642
  10. Bond JH. Polyp guideline: diagnosis, treatment, and surveillance for patients with nonfamilial colorectal polyps. Ann Intern Med. 1993;119:836–843
  11. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med. 1992;326:658–662
  12. Johns LE, Houlston RS. A systematic review and meta-analysis of familial colorectal cancer risk. Am J Gastroenterol. 2001;96:2992–3003
  13. Lynch KL, Ahnen DJ, Byers T, et al. First-degree relatives of patients with advanced colorectal adenomas have an increased prevalence of colorectal cancer. Clin Gastroenterol Hepatol. 2003;1:96–102
  14. Pariente A, Milan C, Lafon J, et al. Colonoscopic screening in first-degree relatives of patients with “sporadic” colorectal cancer: a case-control study. The Association Nationale des Gastroenterologues des Hopitaux and Registre Bourguignon des Cancers Digestifs (INSERM CRI 9505). Gastroenterology. 1998;115:7–12
  15. Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the early detection of cancer. CA Cancer. 2005;55:31–44
  16. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance clinical guidelines and rationale—update based on new evidence. Gastroenterology. 2003;124:544–560
  17. Frazier AL, Colditz GA, Fuchs CS, et al. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA. 2000;284:1954–1961
  18. Lieberman DA. Cost-effectiveness model for colon cancer screening. Gastroenterology. 1995;109:1781–1790
  19. Maciosek MV, Solberg LI, Coffield AB, et al. Colorectal cancer screening: health impact and cost-effectiveness. Am J Prev Med. 2006;31:80–89
  20. Wagner JL, Tunis S, Brown M. Cost-effectiveness of colorectal cancer screening in average-risk adults. In:  Young G,  Levin B editor. Prevention and early detection of colorectal cancer. London: Saunders; 1996;

 Supported by grant R-01 CA 104459.

PII: S0016-5085(07)01592-2

doi:10.1053/j.gastro.2007.08.067

Refers to article:

  • Five-Year Colon Surveillance After Screening Colonoscopy , 12 July 2007

    David A. Lieberman, David G. Weiss, William V. Harford, Dennis J. Ahnen, Dawn Provenzale, Stephen J. Sontag, Thomas G. Schnell, Gregorio Chejfec, Donald R. Campbell, Jayashri Kidao, John H. Bond, Douglas B. Nelson, George Triadafilopoulos, Francisco C. Ramirez, Judith F. Collins, Tiina K. Johnston, Kenneth R. McQuaid, Harinder Garewal, Richard E. Sampliner, Romeo Esquivel, Douglas Robertson
    Gastroenterology October 2007 (Vol. 133, Issue 4, Pages 1077-1085)

  • Colonoscopic Screening of First-Degree Relatives of Patients With Large Adenomas: Increased Risk of Colorectal Tumors , 28 July 2007

    Vanessa Cottet, Alexandre Pariente, Bernard Nalet, Jacques Lafon, Chantal Milan, Sylviane Olschwang, Catherine Bonaiti–Pellié, Jean Faivre, Claire Bonithon–Kopp, ANGH Group
    Gastroenterology October 2007 (Vol. 133, Issue 4, Pages 1086-1092)

Gastroenterology
Volume 133, Issue 4 , Pages 1364-1367, October 2007

Article Tools