A Smarter Strategy? — Reflections on Fecal DNA Screening for Colorectal Cancer
http://www.100md.com
《新英格兰医药杂志》
Colorectal cancer, the second leading cause of death from cancer in the United States, will claim approximately 56,730 lives in 2004.1 Screening for colorectal cancer lowers both the mortality and the incidence of the disease and is widely recommended for persons 50 years of age or older. Interest in screening has increased in recent years but remains low. Only 53 percent of Americans who are 50 years of age or older have undergone screening within the recommended intervals.2 The needless deaths from and complications of colorectal cancer that occur as a consequence of this low rate of screening make correction of the problem an urgent societal priority.
Each existing screening option — fecal occult-blood testing, sigmoidoscopy, colonoscopy, and barium enema — has advantages and limitations, and no one option is ideal. For example, colonoscopy, the most accurate test, is invasive and requires an unpleasant preparation of the bowel. The risk of colonic bleeding and perforation from colonoscopy is small but large enough to raise questions about whether these risks outweigh the incremental health benefits that colonoscopy offers over other tests.3
Fecal occult-blood testing, the only test for which clinical trials have shown evidence of a decreased risk of death, is a noninvasive option that limits the need for follow-up colonoscopy to patients with evidence of bleeding. Neoplasms bleed intermittently, however, allowing many to escape detection with fecal occult-blood testing. Annual retesting is therefore necessary but is still insufficient, detecting only 25 to 50 percent of colorectal cancers and 10 percent of adenomas.3,4,5,6 The specificity of fecal occult-blood testing is also limited by frequent false positive reactions to dietary compounds, medications, and gastrointestinal bleeding from causes other than colorectal cancer.
Testing stool for DNA is a potentially smarter strategy. Oncogene mutations that characterize colorectal neoplasia are detectable in exfoliated epithelial cells in the stool. Whereas neoplastic bleeding is intermittent, epithelial shedding is continual, potentially making fecal DNA testing more sensitive. In 2000, Ahlquist et al. reported on a study showing that a fecal DNA test had a sensitivity of 91 percent for the detection of colorectal cancer and 82 percent for the identification of adenomas.7 The news attracted media8 and commercial9 attention. This study, like other encouraging studies that followed,10,11 drew its data from patients with known colorectal cancer. The accuracy of the test in asymptomatic persons — the target population for screening — was unclear.
This issue of the Journal presents the results of the first large study of fecal DNA testing in asymptomatic persons, reported by Imperiale et al.12 A cohort of 4404 average-risk adults who were at least 50 years of age underwent fecal occult-blood testing, fecal DNA testing, and colonoscopy; colonoscopy was used as the reference standard. Imperiale et al. compared test results in a random subgroup of 2507 subjects and found that the fecal DNA panel was more sensitive than fecal occult-blood testing in detecting advanced neoplasia in general and colorectal cancer in particular (sensitivity, 52 percent vs. 13 percent). The fecal DNA panel was not significantly more sensitive than fecal occult-blood testing for the detection of adenomas (sensitivity, 15 percent vs. 11 percent), except for adenomas with high-grade dysplasia. The fecal DNA panel had a specificity of 94 percent, and fecal occult-blood testing had a specificity of 95 percent.
What should we make of these findings? Should we begin using the fecal DNA panel as a screening test for colorectal cancer, perhaps replacing fecal occult-blood testing? The short answer is no, because at least six questions remain.
First, are the results of the study generalizable? Their generalizability is enhanced by the fact that the study included a large number of subjects from 81 sites, the study cohort was the type of population for which screening is recommended, and the results were carefully validated against a reference standard. There are constraints, however. For example, the analysis excluded 20 percent of subjects because they did not provide stool samples or complete colonoscopy. The investigators included a large number of persons who were 65 years of age or older. The results could be misleading if the epidemiologic characteristics of colorectal neoplasia or test performance differ substantially in underrepresented populations.
The authors reported low sensitivity for both fecal occult-blood testing and the fecal DNA panel. The fecal DNA panel detected only 52 percent of colorectal cancers and 15 percent of adenomas — rates that are far lower than earlier claims of respective rates of 64 to 91 percent and 57 to 82 percent for multitarget testing.7,10,11 At a detection rate of 52 percent, the fecal DNA panel is certainly less sensitive than colonoscopy and is probably less sensitive than flexible sigmoidoscopy, especially when sigmoidoscopy is combined with annual fecal occult-blood testing.3 The fecal DNA panel outperformed fecal occult-blood testing in this study, but the sensitivity of fecal occult-blood testing (13 percent) was unusually low. Imperiale et al. did not rehydrate the cards used to test for fecal occult blood, which lowers the sensitivity of the test, but in other studies that used the same approach, the sensitivity ranged from 30 to 40 percent.3,4 Perhaps, as the authors suggest, their results better reflect the accuracy of this approach in asymptomatic populations under normal practice conditions, but only future studies of such populations will clarify whether this explanation or something about the conduct of the study itself accounts for the low sensitivity. The results of another large trial, funded by the National Cancer Institute,13 should provide further context.
Second, although the study by Imperiale et al. increases the evidence that the fecal DNA panel detects more cancers than fecal occult-blood testing, to what degree does this superiority improve health? Some would contend that the answer requires a prospective trial with mortality as an end point, but the compelling evidence from the trials of fecal occult-blood testing3 makes it safe to infer that any measure that enhances the early detection of adenomas and colorectal cancer will benefit patients. What is unclear is the magnitude of those benefits. The magnitude is estimable if the true sensitivity of the fecal DNA panel is known, but wide confidence intervals surround the values reported by Imperiale et al. They note that, within a 95 percent probability, the sensitivity of the fecal DNA panel could lie anywhere between 35 percent and 68 percent. The accuracy of the fecal DNA panel must be clarified before its health benefits can be quantified.
Third, do the health benefits of fecal DNA testing, whatever their magnitude, outweigh the harms? Public demand for the test — and anxiety over abnormal results — may be intensified by the perception that the results of DNA testing cannot be wrong. The reported specificity of the fecal DNA panel (94 percent) suggests as much. In fact, owing to the low prevalence of colorectal cancer in the asymptomatic population (240 cases per 100,000 people 50 to 59 years of age14), colorectal cancer will be diagnosed in only 2 percent of adults in this age group who have a positive fecal DNA test. The remaining 98 percent can be told, after undergoing colonoscopy, that the screening test was in error, but how many will remain unconvinced and live with the fear that the DNA test was correct?
Fourth, is the fecal DNA test acceptable and available to patients? Its acceptability is enhanced by the fact that it is more convenient than fecal occult-blood testing because it requires no dietary or medication restrictions, only one sample (home fecal occult-blood testing requires two smears from three consecutive stool samples), and less frequent retesting. Its acceptability may suffer, however, from the requirement that patients must collect and refrigerate an entire bowel movement.15,16 Access to the test may be constrained by its limited availability at medical laboratories and coverage by health plans.
Fifth, is the fecal DNA test affordable? It costs $400 to $800,15,17 as compared with a cost of $3 to $40 for fecal occult-blood testing.6 As a result of other costs (e.g., the costs of colonoscopies triggered by test results), fecal occult-blood testing costs $5,700 to $17,800 per year of life gained. A cost-effectiveness ratio below $25,000 per year of life gained is considered a good value, however, and the ratio for existing tests (including sigmoidoscopy and colonoscopy) falls in this range.6 The fecal DNA panel is another matter, however. With a unit cost potentially 200 times that of fecal occult-blood testing, its cost-effectiveness ratio must be higher. According to one optimistic estimate, the fecal DNA panel costs $47,700 per year of life gained.18 If so, fecal DNA testing would cost more to save a life than would the use of more accurate tests, such as colonoscopy.
Sixth, beyond refining tests, could we do more to improve the outcomes of colorectal cancer by making the public's access to screening more systematic and of higher quality? Whether new or old, screening tests cannot reach patients and improve health outcomes in the absence of systems that recognize when screening is due, overcome access and attitudinal barriers, and ensure that the response to test results is appropriate. Improving screening rates is a less breathtaking accomplishment than are technological advances in testing and is less likely to make the evening news. But it may do greater good, especially for those minority populations at greatest risk of dying from colorectal cancer. Colorectal cancer is not the only disease whose treatment would benefit from system changes; these redesigns will improve the quality of screening for other conditions.
Efforts to improve both the delivery of tests and the tests themselves save lives and deserve priority. Both efforts are mutually beneficial. But the greater good that comes from the former should give it higher priority, and yet reverse priorities prevail. Society invests billions of dollars in developing new tests and drugs but only a fraction to modernize systems of care. Realigning priorities is not just a smarter strategy; it saves lives and thus becomes a matter of ethics.
From May to November 2004, Dr. Woolf was executive vice president of Partnership for Prevention. Partnership for Prevention receives less than 10 percent of its funding from contributions from its 75 member medical professional societies, voluntary health associations, state and local health departments, academic institutions, trade associations, and companies, of which Exact Sciences, sponsor of the study by Imperiale et al., is one.
Source Information
From the Departments of Family Medicine, Preventive Medicine, and Community Health, Virginia Commonwealth University, Richmond.
References
Cancer facts & figures 2004. Atlanta: American Cancer Society, 2004.
Colorectal cancer test use among persons aged > or = 50 years -- United States, 2001. MMWR Morb Mortal Wkly Rep 2003;52:193-196.
Pignone M, Rich M, Teutsch SM, Berg AO, Lohr KN. Screening for colorectal cancer in adults at average risk: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:132-141.
Ransohoff DF, Lang CA. Screening for colorectal cancer with the fecal occult blood test: a background paper. Ann Intern Med 1997;126:811-822.
Ahlquist DA, Shuber AP. Stool screening for colorectal cancer: evolution from occult blood to molecular markers. Clin Chim Acta 2002;315:157-168.
Pignone M, Saha S, Hoerger T, Mandelblatt J. 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.
Ahlquist DA, Skoletsky JE, Boynton KA, et al. Colorectal cancer screening by detection of altered human DNA in stool: feasibility of a multitarget assay panel. Gastroenterology 2000;119:1219-1227.
DNA testing may help detect colon cancers earlier. Atlanta: CNN, October 24, 2000. (Accessed December 2, 2004, at http://www.cnn.com/2000/HEALTH/cancer/10/24/colon.cancer/index.html.)
Pollack A. A positive culture for making profits: buoyed by mergers, medical labs await era of gene testing. The New York Times. June 14, 2001:C1.
Dong SM, Traverso G, Johnson C, et al. Detecting colorectal cancer in stool with the use of multiple genetic targets. J Natl Cancer Inst 2001;93:858-865.
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.
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.
Screening tests in detecting colorectal cancer (description of protocol ID MAYO-MC9944). Bethesda, Md.: National Cancer Institute, 2004. (Accessed December 2, 2004, at http://www.nci.nih.gov/search/ViewClinicalTrials.aspx?cdrid=68783&version=patient&protocolsearchid=1179145#TrialDescription_CDR0000068783.)
Colon and rectum cancer. In: Ries LAG, Eisner MP, Kosary CL, et al., eds. SEER cancer statistics review, 1975-2001. Bethesda, Md.: National Cancer Institute, 2004 (Table VI-18).
Levin B, Brooks D, Smith RA, Stone A. Emerging technologies in screening for colorectal cancer: CT colonography, immunochemical fecal occult blood tests, and stool screening using molecular markers. CA Cancer J Clin 2003;53:44-55.
PreGen-Plus colorectal cancer detection. Burlington, N.C.: Laboratory Corporation of America, 2003. (Accessed December 2, 2004, at http://www.labcorp.com/datasets/labcorp/html/chapter/mono/nf10002330.htm.)
PreGen-Plus: FAQs. Marlborough, Mass.: Exact Sciences Corporation, 2004. (Accessed December 2, 2004, at http://www.exactsciences.com/medical/faqs.html.)
Song K, Fendrick AM, Ladabaum U. Fecal DNA testing compared with conventional colorectal cancer screening methods: a decision analysis. Gastroenterology 2004;126:1270-1279.(Steven H. Woolf, M.D., M.)
Each existing screening option — fecal occult-blood testing, sigmoidoscopy, colonoscopy, and barium enema — has advantages and limitations, and no one option is ideal. For example, colonoscopy, the most accurate test, is invasive and requires an unpleasant preparation of the bowel. The risk of colonic bleeding and perforation from colonoscopy is small but large enough to raise questions about whether these risks outweigh the incremental health benefits that colonoscopy offers over other tests.3
Fecal occult-blood testing, the only test for which clinical trials have shown evidence of a decreased risk of death, is a noninvasive option that limits the need for follow-up colonoscopy to patients with evidence of bleeding. Neoplasms bleed intermittently, however, allowing many to escape detection with fecal occult-blood testing. Annual retesting is therefore necessary but is still insufficient, detecting only 25 to 50 percent of colorectal cancers and 10 percent of adenomas.3,4,5,6 The specificity of fecal occult-blood testing is also limited by frequent false positive reactions to dietary compounds, medications, and gastrointestinal bleeding from causes other than colorectal cancer.
Testing stool for DNA is a potentially smarter strategy. Oncogene mutations that characterize colorectal neoplasia are detectable in exfoliated epithelial cells in the stool. Whereas neoplastic bleeding is intermittent, epithelial shedding is continual, potentially making fecal DNA testing more sensitive. In 2000, Ahlquist et al. reported on a study showing that a fecal DNA test had a sensitivity of 91 percent for the detection of colorectal cancer and 82 percent for the identification of adenomas.7 The news attracted media8 and commercial9 attention. This study, like other encouraging studies that followed,10,11 drew its data from patients with known colorectal cancer. The accuracy of the test in asymptomatic persons — the target population for screening — was unclear.
This issue of the Journal presents the results of the first large study of fecal DNA testing in asymptomatic persons, reported by Imperiale et al.12 A cohort of 4404 average-risk adults who were at least 50 years of age underwent fecal occult-blood testing, fecal DNA testing, and colonoscopy; colonoscopy was used as the reference standard. Imperiale et al. compared test results in a random subgroup of 2507 subjects and found that the fecal DNA panel was more sensitive than fecal occult-blood testing in detecting advanced neoplasia in general and colorectal cancer in particular (sensitivity, 52 percent vs. 13 percent). The fecal DNA panel was not significantly more sensitive than fecal occult-blood testing for the detection of adenomas (sensitivity, 15 percent vs. 11 percent), except for adenomas with high-grade dysplasia. The fecal DNA panel had a specificity of 94 percent, and fecal occult-blood testing had a specificity of 95 percent.
What should we make of these findings? Should we begin using the fecal DNA panel as a screening test for colorectal cancer, perhaps replacing fecal occult-blood testing? The short answer is no, because at least six questions remain.
First, are the results of the study generalizable? Their generalizability is enhanced by the fact that the study included a large number of subjects from 81 sites, the study cohort was the type of population for which screening is recommended, and the results were carefully validated against a reference standard. There are constraints, however. For example, the analysis excluded 20 percent of subjects because they did not provide stool samples or complete colonoscopy. The investigators included a large number of persons who were 65 years of age or older. The results could be misleading if the epidemiologic characteristics of colorectal neoplasia or test performance differ substantially in underrepresented populations.
The authors reported low sensitivity for both fecal occult-blood testing and the fecal DNA panel. The fecal DNA panel detected only 52 percent of colorectal cancers and 15 percent of adenomas — rates that are far lower than earlier claims of respective rates of 64 to 91 percent and 57 to 82 percent for multitarget testing.7,10,11 At a detection rate of 52 percent, the fecal DNA panel is certainly less sensitive than colonoscopy and is probably less sensitive than flexible sigmoidoscopy, especially when sigmoidoscopy is combined with annual fecal occult-blood testing.3 The fecal DNA panel outperformed fecal occult-blood testing in this study, but the sensitivity of fecal occult-blood testing (13 percent) was unusually low. Imperiale et al. did not rehydrate the cards used to test for fecal occult blood, which lowers the sensitivity of the test, but in other studies that used the same approach, the sensitivity ranged from 30 to 40 percent.3,4 Perhaps, as the authors suggest, their results better reflect the accuracy of this approach in asymptomatic populations under normal practice conditions, but only future studies of such populations will clarify whether this explanation or something about the conduct of the study itself accounts for the low sensitivity. The results of another large trial, funded by the National Cancer Institute,13 should provide further context.
Second, although the study by Imperiale et al. increases the evidence that the fecal DNA panel detects more cancers than fecal occult-blood testing, to what degree does this superiority improve health? Some would contend that the answer requires a prospective trial with mortality as an end point, but the compelling evidence from the trials of fecal occult-blood testing3 makes it safe to infer that any measure that enhances the early detection of adenomas and colorectal cancer will benefit patients. What is unclear is the magnitude of those benefits. The magnitude is estimable if the true sensitivity of the fecal DNA panel is known, but wide confidence intervals surround the values reported by Imperiale et al. They note that, within a 95 percent probability, the sensitivity of the fecal DNA panel could lie anywhere between 35 percent and 68 percent. The accuracy of the fecal DNA panel must be clarified before its health benefits can be quantified.
Third, do the health benefits of fecal DNA testing, whatever their magnitude, outweigh the harms? Public demand for the test — and anxiety over abnormal results — may be intensified by the perception that the results of DNA testing cannot be wrong. The reported specificity of the fecal DNA panel (94 percent) suggests as much. In fact, owing to the low prevalence of colorectal cancer in the asymptomatic population (240 cases per 100,000 people 50 to 59 years of age14), colorectal cancer will be diagnosed in only 2 percent of adults in this age group who have a positive fecal DNA test. The remaining 98 percent can be told, after undergoing colonoscopy, that the screening test was in error, but how many will remain unconvinced and live with the fear that the DNA test was correct?
Fourth, is the fecal DNA test acceptable and available to patients? Its acceptability is enhanced by the fact that it is more convenient than fecal occult-blood testing because it requires no dietary or medication restrictions, only one sample (home fecal occult-blood testing requires two smears from three consecutive stool samples), and less frequent retesting. Its acceptability may suffer, however, from the requirement that patients must collect and refrigerate an entire bowel movement.15,16 Access to the test may be constrained by its limited availability at medical laboratories and coverage by health plans.
Fifth, is the fecal DNA test affordable? It costs $400 to $800,15,17 as compared with a cost of $3 to $40 for fecal occult-blood testing.6 As a result of other costs (e.g., the costs of colonoscopies triggered by test results), fecal occult-blood testing costs $5,700 to $17,800 per year of life gained. A cost-effectiveness ratio below $25,000 per year of life gained is considered a good value, however, and the ratio for existing tests (including sigmoidoscopy and colonoscopy) falls in this range.6 The fecal DNA panel is another matter, however. With a unit cost potentially 200 times that of fecal occult-blood testing, its cost-effectiveness ratio must be higher. According to one optimistic estimate, the fecal DNA panel costs $47,700 per year of life gained.18 If so, fecal DNA testing would cost more to save a life than would the use of more accurate tests, such as colonoscopy.
Sixth, beyond refining tests, could we do more to improve the outcomes of colorectal cancer by making the public's access to screening more systematic and of higher quality? Whether new or old, screening tests cannot reach patients and improve health outcomes in the absence of systems that recognize when screening is due, overcome access and attitudinal barriers, and ensure that the response to test results is appropriate. Improving screening rates is a less breathtaking accomplishment than are technological advances in testing and is less likely to make the evening news. But it may do greater good, especially for those minority populations at greatest risk of dying from colorectal cancer. Colorectal cancer is not the only disease whose treatment would benefit from system changes; these redesigns will improve the quality of screening for other conditions.
Efforts to improve both the delivery of tests and the tests themselves save lives and deserve priority. Both efforts are mutually beneficial. But the greater good that comes from the former should give it higher priority, and yet reverse priorities prevail. Society invests billions of dollars in developing new tests and drugs but only a fraction to modernize systems of care. Realigning priorities is not just a smarter strategy; it saves lives and thus becomes a matter of ethics.
From May to November 2004, Dr. Woolf was executive vice president of Partnership for Prevention. Partnership for Prevention receives less than 10 percent of its funding from contributions from its 75 member medical professional societies, voluntary health associations, state and local health departments, academic institutions, trade associations, and companies, of which Exact Sciences, sponsor of the study by Imperiale et al., is one.
Source Information
From the Departments of Family Medicine, Preventive Medicine, and Community Health, Virginia Commonwealth University, Richmond.
References
Cancer facts & figures 2004. Atlanta: American Cancer Society, 2004.
Colorectal cancer test use among persons aged > or = 50 years -- United States, 2001. MMWR Morb Mortal Wkly Rep 2003;52:193-196.
Pignone M, Rich M, Teutsch SM, Berg AO, Lohr KN. Screening for colorectal cancer in adults at average risk: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:132-141.
Ransohoff DF, Lang CA. Screening for colorectal cancer with the fecal occult blood test: a background paper. Ann Intern Med 1997;126:811-822.
Ahlquist DA, Shuber AP. Stool screening for colorectal cancer: evolution from occult blood to molecular markers. Clin Chim Acta 2002;315:157-168.
Pignone M, Saha S, Hoerger T, Mandelblatt J. 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.
Ahlquist DA, Skoletsky JE, Boynton KA, et al. Colorectal cancer screening by detection of altered human DNA in stool: feasibility of a multitarget assay panel. Gastroenterology 2000;119:1219-1227.
DNA testing may help detect colon cancers earlier. Atlanta: CNN, October 24, 2000. (Accessed December 2, 2004, at http://www.cnn.com/2000/HEALTH/cancer/10/24/colon.cancer/index.html.)
Pollack A. A positive culture for making profits: buoyed by mergers, medical labs await era of gene testing. The New York Times. June 14, 2001:C1.
Dong SM, Traverso G, Johnson C, et al. Detecting colorectal cancer in stool with the use of multiple genetic targets. J Natl Cancer Inst 2001;93:858-865.
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.
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.
Screening tests in detecting colorectal cancer (description of protocol ID MAYO-MC9944). Bethesda, Md.: National Cancer Institute, 2004. (Accessed December 2, 2004, at http://www.nci.nih.gov/search/ViewClinicalTrials.aspx?cdrid=68783&version=patient&protocolsearchid=1179145#TrialDescription_CDR0000068783.)
Colon and rectum cancer. In: Ries LAG, Eisner MP, Kosary CL, et al., eds. SEER cancer statistics review, 1975-2001. Bethesda, Md.: National Cancer Institute, 2004 (Table VI-18).
Levin B, Brooks D, Smith RA, Stone A. Emerging technologies in screening for colorectal cancer: CT colonography, immunochemical fecal occult blood tests, and stool screening using molecular markers. CA Cancer J Clin 2003;53:44-55.
PreGen-Plus colorectal cancer detection. Burlington, N.C.: Laboratory Corporation of America, 2003. (Accessed December 2, 2004, at http://www.labcorp.com/datasets/labcorp/html/chapter/mono/nf10002330.htm.)
PreGen-Plus: FAQs. Marlborough, Mass.: Exact Sciences Corporation, 2004. (Accessed December 2, 2004, at http://www.exactsciences.com/medical/faqs.html.)
Song K, Fendrick AM, Ladabaum U. Fecal DNA testing compared with conventional colorectal cancer screening methods: a decision analysis. Gastroenterology 2004;126:1270-1279.(Steven H. Woolf, M.D., M.)