Efforts to curb colorectal cancer clearly have been successful, as colorectal cancer death rates have fallen 46% from their peak, according to the American Cancer Society. This good news has to be tempered, though, by the fact that about one-third of the recommended population has not been screened, despite colorectal cancer having very high treatment success when detected early.

There are many reasons behind this screening gap, but one no doubt has to do with what Victoria Pratt, PhD, called the “ick factor” associated with colonoscopy—the gold standard method—as well as flexible sigmoidoscopy, fecal occult blood testing (FOBT), and iFOBT, also known as fecal immunochemical test (FIT) (Clin Chem 2014;60:1141–42).

A concentrated research effort has blossomed around finding biomarkers that could be incorporated in new screening tests that would be better accepted by individuals currently reluctant to be screened, ultimately leading to higher screening rates. Among the many candidates, 2014 was a milestone year for two promising tests, Cologuard (Exact Sciences; Madison, Wisconsin), and Epi proColon­ (Epigenomics; Berlin, Germany and Germantown, Maryland). Important studies were published about both, and both also were the subject of significant regulatory actions.

A Multi-target Fecal DNA Test

In a study of nearly 10,000 patients, researchers found that in comparison to colonoscopy, Cologuard, a multi-target fecal DNA test, had a sensitivity of 92.3% for detecting colorectal cancer, and a specificity of 86.6% (N Engl J Med 2014;370:1287–97). This compared with FIT’s 73.8% sensitivity and 94.9% specificity. Cologuard, which includes quantitative molecular assays for KRAS mutations, NDRG4 and BMP3 methylation, β-actin, as well as a hemoglobin assay, performed particularly well in comparison to FIT in detecting advanced precancerous lesions (42.4% sensitivity versus 23.8%) and polyps with high-grade dysplasia (69.2% sensitivity versus 46.2%).

“Our study assessed the test performance in an asymptomatic average risk population, which would be the exact population for which the application would be recommended,” explained the study’s lead author, Thomas Imperiale, MD, a professor of medicine at the Indiana University School of Medicine in Indianapolis. “Everything was measured against colonoscopy to determine the multi-target assay’s sensitivity and specificity. The other purpose was to compare it against another popular, non-invasive test, FIT. The multi-target assay was found to be more sensitive but less specific than FIT.”

Based on the results of this study, the Food and Drug Administration (FDA) in August 2014 approved Cologuard, and in an unprecedented move, the Centers for Medicare and Medicaid Services on the same day issued a proposed national coverage determination for the test. The latter would support reimbursement for Cologuard every 3 years in Medicare beneficiaries age 50 to 84 with average colorectal cancer risk and no signs or symptoms of the disease.

A Blood Test for DNA Methylation

Epi proColon is an entirely different type of test. This blood test detects methylated SEPT9 in circulating cell-free DNA isolated from plasma. In a study assessing the analytical and clinical performance of Epi proColon¬≠, four laboratories tested 1,544 samples from subjects enrolled in the PRESEPT trial, a prospective study aimed at assessing the clinical utility of SEPT9 based on the outcome of colonoscopy, polypectomy, biopsy, and surgical tissue histopathology (Clin Chem 2014;60:1183–91). They found the test’s limit of detection for methylated SEPT9 DNA was 7.8 pg/mL, which corresponds to less than two genome copies of methylated SEPT9 per milliliter of plasma.

Epi proColon’s sensitivity for detecting colorectal cancer was 68%, and its specificity was 79%. Sensitivity was lower for advanced adenomas (22%) and small polyps (20%). Although this particular study did not involve a head-to-head comparison between Epi proColon and FIT, a separate analysis found it “noninferior” to FIT, with 72% sensitivity versus 68%, respectively.

“The primary goals of our study were to determine the analytical as well as clinical performance data for Epi proColon, which is a real-time PCR test targeting a differentially methylated sequence in the promoter of the septin 9 gene,” explained lead author Nicholas Potter, PhD, FACMG, executive vice president of clinical affairs at Molecular Pathology Laboratory Network in Maryville, Tennessee. “We also sought to provide data in the context of Epi proColon’s potential utilization for screening average risk individuals for colorectal cancer.”

Epi proColon has a CE mark in Europe and applied to FDA in 2013 for premarket approval in the United States. In March 2014, the Molecular and Clinical Genetics Panel of FDA’s Medical Devices Advisory Committee narrowly endorsed premarket approval for the test. However, FDA has asked Epigenomics for additional data on how Epi proColon might improve screening rates in patients not already compliant with screening recommendations who are offered a blood-based test versus FIT. In addition, Epigenomics intends to assess compliance with colonoscopy among patients who have positive Epi proColon or FIT tests, according to a press release issued after the company met with FDA.

The Twists and Turns of Development

If Epi proColon’s commercialization process hit a speed bump, it in some ways parallels challenges Cologuard faced in its development process. During an initial study, Cologuard detected just half of the invasive cancers and less than 20% of advanced adenomas.

Imperiale, who also was involved in that effort, recalled the earlier study and subsequent changes in Cologuard. “It was the first time that anybody had put a multi-target assay together and tested it in a screening setting. It didn’t do as well as it had done in preliminary studies though the preliminary studies were done in the same way,” he explained. “We realized the error that was made, with specimen collection. It didn’t do as well because of the failure to apply a buffer that preserved the human DNA and kept it from being broken down by bacterial enzymes.”

Cologuard’s subsequent overhaul included changing the markers—only KRAS made it in both versions—and adding an immunoassay for heme, according to Imperiale. Exact Sciences now expects to conduct post-approval testing to look at the validity of the proposed 3-year testing interval for Cologuard, he added.

On the Horizon

Even as Cologuard and Epi proColon continue to make their way into the marketplace, research also is ongoing for yet other candidate biomarkers for detecting colorectal cancer. CLN’s brief survey of the literature identified recent studies looking at methylated SFRP2, GATA4/5, NDRG4, and VIM, ALU in circulating cell-free DNA, and at serum microRNAs miR-21, let-7g, miR-31, miR-92a, miR-181b, miR-203, and miR-135b, among others. Potter’s analysis of the colorectal cancer diagnostic pipeline as of October 2014 revealed quite a range of candidate biomarkers in proposed blood, serum, plasma, and stool-based tests. In addition, emerging research is exploring the potential of emitted fecal volatile organic compounds and exhaled breath compounds.

Does the pathophysiology of colorectal cancer favor any particular molecular pathways, and hence markers? Imperiale has an open mind about the possibility. “The likelihood is that over time, better markers will be identified and will likely be incorporated into future multi-target assays that might help better optimize discrimination or improve cost-effectiveness,” he predicted. “It’s going to be a moving target, in a good way. Because the markers could change and as could the ways in which the markers are detected.”

Potter emphasized the need to put any candidate markers—regardless of type—through their paces. “There are a number of panel-based approaches in the preliminary stage of evaluation for use as screening targets for colorectal cancer. It’s a question of how the markers behave in preliminary trial investigations, and how those performance characteristics compare to potential markers that other research teams have been exploiting,” he explained. “It then becomes an exercise in exploratory due diligence with regard to setting a threshold of what you need the performance to be.” In the United States, large prospective clinical trials also would be required to provide data in support of any proposed screening test, he added.

If the markers and methods differ, researchers and clinicians agree the hunt should continue. As Pratt put it, “One thing is certain: early detection improves cancer survival rates. Because only approximately 65% of eligible people are getting the recommended screening for [colorectal cancer], any test that improves screening compliance will save lives.”

Disclosures: Dr. Potter serves on the medical advisory board of Epigenomics AG.