Molecular Stool Testing Could Cut Colonoscopies by 15%-41%

Noninvasive surveillance with multitarget stool DNA testing or fecal immunochemical testing (FIT) could potentially match colonoscopy for reducing long-term colorectal cancer (CRC) incidence and mortality. It might also reduce colonoscopies by an estimated 15%-41%.
The greatest reduction would likely be achieved by annual FIT-based surveillance, especially with FIT FOB-Gold at a threshold of at least 32 µg/g feces, according to findings from the Dutch MOCCAS study published in Gastroenterology.
In this cross-sectional observational study, the multitarget DNA test outperformed FIT for detecting advanced precursor lesions, especially serrated polyps. According to long-term-impact mathematical modeling, however, DNA-based surveillance would be more costly than colonoscopy surveillance, whereas FIT would save costs.
“With the worldwide implementation of FIT-based screening programs, following a positive test, many more people enter surveillance programs after polypectomy. This results in an increased pressure on the colonoscopy capacity and healthcare budgets,” lead author Beatriz Carvalho, PhD, a molecular biologist in the Department of Pathology of the Netherlands Cancer Institute in Amsterdam, said in an interview.
A noninvasive strategy could ease the surveillance burden on healthcare resources and be more palatable to patients. Post-polypectomy guidelines have already been relaxed to allow less intensive surveillance.
“Our working hypothesis was that although the sensitivity of a singular molecular test to detect CRC or advanced adenomas is lower than that of colonoscopy, repeating molecular stool testing would yield similar detection rates as colonoscopy-based surveillance. And our hypothesis was confirmed,” Carvalho said.
The results of the MOCCAS study align with those of other studies that found that FIT could be safely applied as a triage test in post-polypectomy surveillance and could safely extend the interval of surveillance colonoscopy. “But these studies did not include a long-term impact analysis,” she said. “The next step is to run a prospective interventional study to validate the MOCCAS findings.”
Offering an outsider’s perspective on the findings, Uri Ladabaum, MD, director of the Gastrointestinal Cancer Prevention Program and a professor of medicine at Stanford University School of Medicine in Palo Alto, California, said the real-world results on lesion detection and the multi-year-horizon modeling performed are provocative and point to the potential to base post-polypectomy surveillance on stool tests.
He cautioned, however, that the proposed paradigm requires the ability to deploy FIT-based surveillance with broad flexibility in relation to hemoglobin-detection thresholds and testing interval, depending on the specific FIT that is chosen, with the possibility these may differ by setting based on the characteristics of the population and the relevant epidemiology.
“Such flexibility may or may not be technically feasible in all settings — for instance, in the current US regulatory context, it would be challenging to implement FIT-based testing at newly adjusted detection thresholds,” he said.
Nevertheless, the study provides a strong rationale for a real-world study of FIT-based surveillance, he added. “The choice of specific FIT and detection threshold will be critical. Multiple rounds of FIT-based surveillance, that is, years of prospective surveillance, will be needed to constitute a properly designed comparison with surveillance colonoscopy.”
Study Details
The cross-sectional observational study included individuals aged 50-75 years who provided stool samples for the DNA test and two FITs. Test accuracy was calculated for all surveillance indications.
For the post-polypectomy indication only, which is the most common and associated with a relatively low CRC risk, the long-term impact of stool-based surveillance was evaluated with the Adenoma and Serrated Pathway to Colorectal Cancer model. Stool-based strategies were simulated to tune each test’s positivity threshold to obtain strategies that are at least as effective as colonoscopy surveillance.
A total of 3453 individuals had results for all stool tests and colonoscopy; among them, 2226 had previously undergone polypectomy, 1003 had a history of CRC, and 224 had a familial risk.
Areas under the receiver operating characteristic curve for advanced neoplasia were as follows:
• 0.72 (95% CI, 0.69-0.75) for the multitarget stool DNA test
• 0.61 (95% CI, 0.58-0.64) for the FIT OC-SENSOR 
• 0.59 (95% CI, 0.56-0.61) for the FIT FOB-Gold 
Stool-based surveillance was estimated to be at least as effective as colonoscopy surveillance and required 5.6 to 9.5 stool tests over a person’s lifetime. DNA-based surveillance was more costly than colonoscopy surveillance, whereas FIT-based surveillance saved costs.
“These findings provide a basis to embark on a prospective intervention study to assess the clinical utility of FIT as an alternative to colonoscopy surveillance in a post-polypectomy CRC surveillance population,” the authors wrote.
In the United States, Ladabaum said, it would likely be possible to find FIT-based strategies that closely approximate or match surveillance colonoscopy — “if we could deploy FIT with the required flexibility, for example, by adjusting the threshold and if the reference surveillance standard were somewhat relaxed compared with current guidelines.”
He worries, however, that if FIT for screening and FIT for surveillance were optimized at different hemoglobin detection thresholds, “there could be confusion and room for error in real-world clinical implementation.”
The authors called for research to increase understanding of the mechanisms underlying progression from adenomas to malignancy over time, which may yield better biomarkers to improve stool test accuracy.
This study was funded by the Alpe d’HuZes charity and the Dutch Cancer Society. Exact Sciences provided test equipment and performed multitarget stool DNA test analysis. Sentinel Diagnostics provided equipment and reagents.
Carvalho and Veerle M. H. Coupé, PhD, disclosed several patents pending and/or issued. Other coauthors disclosed multiple financial relationships with private companies, including Exact Sciences and Sentinel, for research support, travel, board membership, advisory or speaker’s fees, consulting, employment, stock ownership, or patents.
Ladabaum disclosed no competing interests relevant to his comments.
 
Send comments and news tips to [email protected].