by Charles Bankhead
Staff Writer, MedPage Today
March 19, 2014

A stool DNA test detected significantly more colon cancers in average-risk patients as compared with a conventional fecal immunochemical test (FIT), a large prospective clinical study showed.

The DNA test detected 92.3% of colorectal cancers versus 73.8% for FIT in a study involving almost 10,000 patients scheduled for colonoscopy. The DNA test also had an 20%+ absolute advantage for detecting advanced precancerous lesions.

The FIT was associated with significantly fewer false-positive results, but the DNA test nonetheless had a much lower number needed to detect (166 versus 208), Thomas F. Imperiale, MD, of Indiana University in Indianapolis, and co-authors reported online in the New England Journal of Medicine.

“Sensitivity is the most important characteristic for screening tests because the primary role of such testing is to rule out diseases such as cancer,” the authors said of their findings. “The sensitivity of the DNA test for the detection of advanced precancerous lesions was approximately half that for the detection of colorectal cancer.

“[The DNA test] exceeded the performance of the FIT overall and in important subsets of lesions, including adenomas measuring 2 cm or more … and large, sessile serrated polyps (which may account for up to one-third of colorectal cancers.”

The DNA test evolved from recognition that colorectal cancer arises from an accumulation of genetic and epigenetic changes, providing the basis for a noninvasive stool test to identify tumor-specific changes. Early versions of DNA tests lacked the necessary sensitivity for colorectal cancer and had poor sensitivity for advanced precancerous lesions.

Advances in DNA-analysis technology have resulted in stool DNA tests that have higher sensitivity for both cancerous and precancerous lesions, the authors noted in their introduction. Preliminary studies confirmed the improved performance characteristics in archived stool samples, leading to a clinical trial to compare performance characteristics of stool DNA testing and FIT.

Investigators at 90 sites in North America enrolled average-risk patients 50 to 84 who were scheduled to undergo screening colonoscopy. The study population was enriched with patients ≥65 in an attempt to increase cancer prevalence.

Each test was used to analyze specimens obtained before routine bowel preparation for colonoscopy. The primary endpoint was the DNA test’s performance for detecting cancerous lesions identified during colonoscopy, and the secondary endpoint was the test’s ability to detect precancerous lesions.

The final analysis included data for 9,989 patients, 65 of whom had colorectal cancer on colonoscopy and 757 with advanced precancerous lesions.

The DNA test identified 60 of 65 malignancies, including 56 of 60 considered to be screening relevant (stages I-III). Sensitivity was 92.3% for all malignancies and 93.3% for the screening-relevant lesions. Corresponding values for FIT were 73.8% and 77.3% (P=0.002). The DNA test was more accurate for detection of proximal versus distal cancers.

The DNA test detected 321 advanced precancerous lesions identified by colonoscopy, resulting in a specificity of 42.4%. The FIT had a specificity of 23.8% (P<0.001). A subgroup analysis found the DNA test superior for detection of high-grade dysplasia (69.2% versus 46.2%, P=0.004) and sessile polyps ≥1 cm (42.4% versus 5.1%, P<0.001), as well as distal and proximal advanced precancerous lesions (P<0.001).

The DNA test had lower specificity for nonadvanced precancerous lesions (86.6% versus 94.9%) and for negative colonoscopy results (89.8% versus 96.4%).

Analysis of the number of persons screened to detect one cancer (NND) yielded values of 154 for colonoscopy, 166 for the DNA test, and 208 for FIT. Corresponding NND values for precancerous lesions were 13, 31, and 55.

Authors of an accompanying editorial raised several issues with the DNA test’s performance. The DNA test was associated with a much higher rate of test exclusions related to sample collection or the assay’s application (6.3% versus 0.3%).

“Given that colorectal cancer was detected in nearly one of 154 participants on colonoscopy, it is possible that four cancers would have been missed simply because of the complexity of the test,” said Douglas J. Robertson, MD, of White River Junction VA Medical Center in New Hampshire, and Jason A. Dominitz, MD, of the VA Puget Sound Health Care System in Seattle.

Robertson and Dominitz also were bothered by the proportion of DNA tests that were positive when colonoscopy was negative (10.2%). Finally, they questioned the applicability of the results (complete data for both stool tests and colonoscopy) to real-world clinical practice.

“The new multitarget stool DNA test is clearly an improvement over its predecessors, and the results of this study will help to inform the current effort by the U.S. Preventive Services Task Force to re-evaluate screening tests,” Robertson and Dominitz concluded. “Comparative-effectiveness studies are now needed to clarify the role of stool DNA testing with respect to programmatic screening with other test options.”

LAST UPDATED 03.19.2014