[Skip to Navigation]
Sign In
June 21, 2016

Blood-Based Screening for Colon Cancer: A Disruptive Innovation or Simply a Disruption?

Author Affiliations
  • 1Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
  • 2Knight Cancer Institute, Division of Hematology and Medical Oncology, Oregon Health and Science University, Portland
  • 3Department of Public Health and Preventive Medicine, Oregon Health and Science University, Portland
JAMA. 2016;315(23):2519-2520. doi:10.1001/jama.2016.7914

On April 13, 2016, the US Food and Drug Administration (FDA) approved the first blood-based screening test for colon cancer. The assay (Epi proColon) relies on qualitative detection of the methylated septin 9 gene (SEPT9) and has been considered an innovation in screening. Despite multiple accepted options for colorectal cancer screening—including stool-based tests, such as the fecal occult blood test (FOBT) and lower endoscopy—handling, storing, and returning stool tests or prepping and undergoing an invasive procedure have limited adherence. Approximately one-fourth of eligible individuals aged 50 to 75 years have never been screened for colon cancer, and half are inadequately screened. Blood-based cancer screening has the potential to address this gap. Yet physicians must question how this newly approved assay might be incorporated in routine clinical practice, and, most important, whether use of the test will reduce colorectal cancer mortality.

The new blood-based screening test was approved on the basis of clinical data regarding the sensitivity and specificity of the test for detecting colon cancer. Although case-control studies reported an estimated sensitivity and specificity for cancer detection approaching 90%,1 in a clinical trial the assay’s sensitivity and specificity (68% and 79%, respectively) did not meet 1 of 2 prespecified goals (65% and 85%).2 A subsequent clinical trial found that, compared with fecal immunohistochemistry testing (FIT), SEPT9 testing significantly improved sensitivity (68% vs 73%) but markedly decreased specificity (97% vs 81%).3 The area under the receiver operator characteristic curve was greater for FIT than for SEPT9 testing (0.86 vs 0.82), suggesting worse overall test performance for the blood test.2 Poorer performance would increase false-positive results, with all the attendant anxiety, downstream procedures, and cost.

Importantly, no evidence has shown that this new assay improves disease-specific or overall mortality compared against no screening or against accepted screening methods. Sigmoidoscopy and FOBT have both demonstrated these benefits in randomized trials, while colonoscopy and FIT have not. Yet practically, colonoscopy includes sigmoidoscopy, and the rationale for FIT is more closely analogous to FOBT. Thus, in both cases, extrapolation may be more reasonable than blood-based screening, an entirely novel method.

The advantage of blood testing is the opportunity to expand the number of individuals who undergo colon cancer screening. Even among existing noninvasive tests for colon cancer, adherence is low: for annual FOBT, on average only half of participants return stool cards appropriately.4 Evidence suggests that some patients who are reluctant to undergo screening would be receptive to a blood test. In one survey, 97% (106/109) who refused colonoscopy accepted a noninvasive screening test, and 83% (90/109) of those preferred a blood test.5

Nevertheless, 4 concerns may limit the potential benefits of the blood-based screening test and thereby limit its clinical use. First, the adequacy of the end point for approval should be questioned. In cancer screening, proof a test can detect cancer is not the same as proof that the test can reduce disease-specific mortality. For instance, ovarian cancer screening with transvaginal ultrasound and a CA-125 measurement clearly increases cancer detection; however, there is no good evidence that acting on these findings improves disease-specific mortality, with at least 2 randomized trials failing to find such a benefit in the primary analysis. In contrast with imaging-based screening, the blood-based screening test for colon cancer was accepted without demonstrating an improvement in survival for colorectal cancer.

Why does disease-specific mortality matter? Not all colon cancer is biologically similar and amenable to mortality reduction through early detection. For instance, there is persistent debate as to whether colonoscopy improves disease-specific mortality beyond the benefits from sigmoidoscopy, even though only colonoscopy is able to screen the right-sided colon. Multiple observational studies suggest that the benefit of colonoscopy is limited to a reduction in death from left-sided but not right-sided colon cancer.6 One putative biological explanation for this is that right-sided cancers have more aggressive early genetic events and are more difficult to detect and resect with endoscopy.

For this reason it is unclear if the cancers detected by blood-based screening are ones for which identification and removal will yield real mortality gains. Secondary data suggest that this blood-based screening is more sensitive at detecting advanced-stage cancer (sensitivity = 80%; 4/5) than stage I cancer (sensitivity = 36.4%; 8/22)–undermining its theoretical utility.7 Furthermore, in so far as gains in colon cancer mortality are related to removal of precancerous polyps, this newly approved screening test falls short because the assay appears to only have a 10% sensitivity to detect advanced adenomas,7 which are considered precancerous and require removal.

Second, although blood-based screening should be used for individuals eligible for but nonadherent with other screening methods, it is possible that the test will undergo indication drift—in both directions. Patients with more severe comorbidities, whose life expectancy precludes benefit from screening, have been shown to be more likely to undergo blood-based screening than invasive screening tests. Availability of a blood test may increase colorectal cancer screening among those who should not be screened. In the other direction, given the convenience of blood testing, patients who are appropriate candidates for endoscopy may wish to first be tested with this assay in lieu or in advance of endoscopy. Given the limited sensitivity of this test for cancer and polyps, blood-based screening may erode, not enhance, potential benefits of colon cancer screening.

Third, there is no evidence to guide how positive results will be handled in clinical practice. Given the baseline prevalence of colorectal cancer in the United States, the positive predictive value of blood-based screening is just 5.2%,7 compared with more than 30% for FOBT. No matter which noninvasive test is performed, positive results cannot improve outcomes without further intervention. However, it is possible that patients undergoing blood-based testing may be less amenable to endoscopy than those who complete stool testing. Completing stool testing serves as a marker of compliance by requiring collection and returning of the sample. In contrast, giving blood may foster less of a psychological commitment. Blood-based screening may merely shift loss to follow-up to after the test is completed, as opposed to before it is returned.

Fourth, cancer screening remains a widely contested topic, prompting vigorous debates. A compromise that has emerged is, as in all cases in which the full risks and benefits are not certain, for physicians to engage patients in shared decision making that requires a commitment to full disclosure of the limits of the current evidence.8 The uncertainty regarding the true clinical utility of this newly approved test makes shared decision making difficult. Physicians must inform patients of the unknown effects of the test on mortality, the lower rate of detection of precancerous polyps, and the presence of better alternatives. Patients should be told that positive test results commit them to further investigations, likely including endoscopy. The prospect of such a lengthy discussion becoming routine seems unlikely. Moreover, SEPT9 blood testing raises concerns that like prostate-specific antigen testing, it may be ordered without specific patient consent, as part of routine blood work.

Questions about comparative efficacy may explain the test’s approval history. In March 2014, the FDA’s Molecular and Clinical Genetics Panel evaluating Epi proColon voted 9-0 (1 abstention) in favor of adequate evidence of safety, but voted 4-5 against adequate evidence of efficacy (1 abstention).9 The panel recommended further testing showing benefit. The manufacturer of this test proposed including a warning to address the low sensitivity, specifying that a negative result “does not guarantee absence of cancer” and that patients should still pursue other screening methods.2 This warning along with further prospective evidence that SEPT9 blood testing improved detection of colon cancer led to the FDA’s eventual approval in 2016.

As more molecular ways to detect cancer are discovered, the medical community must consider the standard for integrating these markers as screening tests. For imaging, the bar has been randomized trials demonstrating improved disease-specific mortality, though some in the research community favor even higher standards.8 The standard for blood screening tests should be no lower. While there may be value to enhanced population-wide cancer screening, physicians and policy makers must be objective about the comparative efficacy of novel blood-based screening tests. As the experience with SEPT9 blood testing shows, blood-based testing must balance convenience against the risk of inadequate screening, underdiagnosis, and reluctance to pursue further testing. Allowing blood-based screening tests for colon cancer to have a lower standard than that of other screening tests risks prioritizing convenience over patient safety and health care value.

Back to top
Article Information

Corresponding Author: Vinay Prasad, MD, MPH, Department of Public Health and Preventive Medicine, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Portland, OR 97239 (prasad@ohsu.edu).

Published Online: June 15, 2016. doi:10.1001/jama.2016.7914.

Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Warren  JD, Xiong  W, Bunker  AM,  et al.  Septin 9 methylated DNA is a sensitive and specific blood test for colorectal cancer.  BMC Med. 2011;9(1):133.PubMedGoogle ScholarCrossref
Food and Drug Administration Molecular and Clinical Genetics Panel. FDA Executive Summary: Epigenomics AG Epi proColon. Silver Spring, MD: Food and Drug Administration; 2014. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/MolecularandClinicalGeneticsPanel/UCM390221.pdf. Accessed May 8, 2016.
Johnson  DA, Barclay  RL, Mergener  K,  et al.  Plasma Septin9 versus fecal immunochemical testing for colorectal cancer screening: a prospective multicenter study.  PLoS One. 2014;9(6):e98238.PubMedGoogle ScholarCrossref
Lee  JK, Reis  V, Liu  S,  et al.  Improving fecal occult blood testing compliance using a mailed educational reminder.  J Gen Intern Med. 2009;24(11):1192-1197.PubMedGoogle ScholarCrossref
Adler  A, Geiger  S, Keil  A,  et al.  Improving compliance to colorectal cancer screening using blood and stool based tests in patients refusing screening colonoscopy in Germany.  BMC Gastroenterol. 2014;14:183.PubMedGoogle ScholarCrossref
Neugut  AI, Lebwohl  B.  Colonoscopy vs sigmoidoscopy screening: getting it right.  JAMA. 2010;304(4):461-462.PubMedGoogle ScholarCrossref
Church  TR, Wandell  M, Lofton-Day  C,  et al; PRESEPT Clinical Study Steering Committee, Investigators and Study Team.  Prospective evaluation of methylated SEPT9 in plasma for detection of asymptomatic colorectal cancer.  Gut. 2014;63(2):317-325.PubMedGoogle ScholarCrossref
Prasad  V, Lenzer  J, Newman  DH.  Why cancer screening has never been shown to “save lives”—and what we can do about it.  BMJ. 2016;352:h6080.PubMedGoogle ScholarCrossref
FDA Center for Devices and Radiological Health Medical Devices Advisory Committee, Molecular and Clinical Genetics Panel. March 26, 2014, transcript. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/MolecularandClinicalGeneticsPanel/UCM395028.pdf. Accessed June 1, 2016.