FDA Panel Unanimously Recommends New DNA Stool Test

David Ransohoff, MD

The FDA’s Molecular and Clinical Genetics advisory committee has unanimously supported the safety, efficacy, and positive risk-benefit profile of the noninvasive stool-based DNA test Cologuard in a 10-0 vote.

The Cologuard test is a DNA screening tool that involves quantitative molecular assays for KRAS mutations, abnormalities in methylated bone morphogenetic protein 3 (BMP3) and N-Myc downstream-regulated gene 4 (NDRG4), and β-actin. It also includes a hemoglobin immunoassay. The panel’s recommendation follows a review of data submitted to the FDA in an application for premarket approval suggesting that the Cologuard DNA test was more effective than currently available fecal immunochemical testing (FIT). These claims were based largely on findings from a large study that were published March 19 in The New England Journal of Medicine.

“We are pleased the Committee strongly supported Cologuard’s approval,” said Kevin T. Conroy, the chairman and chief executive of the company developing the test, Exact Sciences. “We look forward to continuing our work with the FDA to complete its review of Cologuard and remain committed to addressing the growing unmet needs in colorectal cancer screening.”

In the study, specimens from 9989 individuals at average risk of colorectal cancer scheduled to undergo colonoscopy screening within 90 days of sample collection were obtained between June 2011 and November 2012 at 90 sites throughout the United States and Canada. Study locations included both private practice and academic centers. To increase the likelihood of cancer, enrollment was weighted toward those aged ≥65 years. This study differs from previous evaluations of fecal testing options that relied largely on archived stool samples, the investigators noted.

Both the DNA and the FIT test comparator (OC FIT-CHEK, Polymedco) were evaluated based on their detection sensitivity—defined as the true positive rate or the proportion of individuals with disease who have a positive test—in four areas: colorectal cancer (ie, adenocarcinoma); advanced precancerous lesions; polyps with high-grade dysplasia, and serrated sessile polyps measuring ≥1 cm.

The DNA test was found to be more sensitive across all four measures. The sensitivity for detecting CRC with Cologuard was 92.3% versus 73.8% with FIT (P = .002). For Cologuard and FIT, respectively, the sensitivity for detecting advanced precancerous lesions was 42.4% versus 23.8% (P <.001); the sensitivity for detecting polyps with high-grade dysplasia was 69.2% and 46.2% (P = .004), and the sensitivity for the detection of serrated sessile polyps was 42.4% and 5.1% (P = .001).

The researchers noted, however, that although high sensitivity is the most important aspect of cancer screening, specificity (the proportion of individuals without disease who have a negative test result) also must be factored in, and here, results for FIT were better. Specificity with FIT was 94.9% to 96.4% versus 86.6% to 89.8% with Cologuard.

The majority of those with positive results, the researchers explained, will be false positives due to the low prevalence of cancer. The false positive rates for FIT were 3.6% to 5.1% versus 10.2% to 13.4% with Cologuard. Among those with a negative colonoscopy result, the DNA test’s specificity was high—nearly 90%&mdash;though that result was still inferior to FIT specificity for this group (>96%).

Cologuard identified 60 of 65 study participants who were found to have cancer. Sensitivity of the DNA test was not significantly affected by cancer stage or location in the colon. The number of persons who would need to be screened to detect one colorectal cancer using colonoscopy was 154 compared with 166 with Cologuard and 208 with FIT. For the detection of one advanced precancerous lesion the numbers were 13, 31, and 55, respectively.

For the hemoglobin assay component of both tests, performance was similar, both for sensitivity (72.3% and 73.8% for Cologuard and FIT, respectively) and specificity (94.8% and 94.9%, respectively).

"Detection of 92% of colon cancer is extremely high for a noninvasive test, so that a negative test result means that no further evaluation, like colonoscopy, is needed at that time," said one of the study’s authors, David Ransohoff, MD, professor of Medicine at the University of North Carolina School of Medicine and UNC Lineberger Comprehensive Cancer Center. "Having such a sensitive, noninvasive option could have an important effect on screening rates for colorectal cancer."

Despite the known benefits of screening for CRC and modest increases in uptake in the last 15 years, 41% of Americans aged ≥50 years are not current with their recommended screening, according to the American Cancer Society; for certain populations, such as African-Americans, screening rates are even lower.

Imperiale TF, Ransohoff DF, Itzkowitz SH, et al. Multitarget stool DNA testing for colorectal-cancer screening [published online ahead of print March 19, 2014]. N Engl J Med.