Stool DNA testing has evolved into a highly accurate and well-validated test for the screen-detection of colorectal neoplasia. An optimized multi-target stool DNA test (MT-sDNA) achieves the same high point-sensitivities as reported with colonoscopy for early-stage colorectal cancer (CRC) and significantly higher sensitivity than by fecal immunochemical blood testing (FIT) for detection of both cancer and advanced precancer. Thus, MT-sDNA sets a new high criterion standard for the noninvasive screen-detection of colorectal neoplasia. With clinical application, MT-sDNA has potential to meaningfully address current gaps in our approach to CRC screening through benefits of its high accuracy on screening effectiveness, user-friendly features on patient compliance, and easy mail-out distribution on test access. After an unprecedented parallel review, the U.S. Food & Drug Administration (FDA) and Center for Medicaid and Medicare Services (CMS) simultaneously endorsed MT-sDNA for general CRC screening in August 2014, and the test is now available for clinical use. This overview briefly summarizes the clinical validation, addresses central clinical questions at implementation, and touches on future applications of stool DNA testing.

Despite efforts to date, CRC remains the 2nd leading cause of cancer mortality in the US and screening rates are suboptimal. Conventional screening tools are variably hampered by invasiveness, unpleasant bowel preparation, access limitations, interruption of work routines, and a bias toward left-sided lesion detection which compromise efficacy and participation. There is an imperative to improve screening effectiveness through more accurate, patient-friendly, and accessible approaches.

Multi-Target Stool DNA Test (MT-sDNA) Stool DNA testing is grounded in the rational biology of exfoliation, inherent discrimination afforded by tumor-associated genetic and epigenetic alterations, and exquisite analytical sensitivity of new assay methods (Pathology 2012;44:8-0). Key technical innovations leading to the current high test performance include use of a stool buffer to prevent DNA degradation, discriminant new marker panels, and advanced analytical platforms that substantially augment sensitivity. MT-sDNA, an optimized and automated new generation test and the product of an intensive research and development effort, has been described in detail (Clin Gastroenterol Hepatol 2013; 11:1313). It targets two highly discriminant methylated genes (BMP3 and NDRG4), the informative KRAS mutations, a marker for total human DNA (β-actin), and fecal hemoglobin. Quantitative outputs analyzed in a software system yield a positive or negative result based on an established cutoff score. With tightly reproducible results via automation, MT-sDNA avoids the operator variation that hinders quality with endoscopic and subjective in vitro screening methods. This noninvasive test requires no bowel preparation, diet or medication restrictions, or missed work time. Stools can be collected at home and mailed to a central laboratory for processing.

Clinical Validation Case-Control Studies. In the first large multicenter study using next generation technology (Gastroenterology 2012;142:248), a prototype MT-sDNA showed high neoplasm detection unaffected by tumor stage or site. In the second study (Clin Gastroenterol Hepatol 2013;11:1313), an optimized MT-sDNA detected 98% of CRC at specificity of 91% and again was unaffected by stage or site. Adenoma detection increased with size and degree of dysplasia: sensitivity for adenomas >1cm was 57%, >2 cm was 73%, and >3 cm was 83%; sensitivity was 83% with high-grade dysplasia (HGD), 94% of which occurred in lesions >2 cm. In a third study (PloS One 2014;9:e85659), MT-sDNA detected 55% of sessile serrated polyps (SSP) >1cm compared to 10% by fecal immune-testing for blood (FIT) at matched specificities.

Pivotal Screen-Setting Study. In a multicenter cross-sectional study comprising >10,000 average risk persons from across North America (NEJM 2014;370:1287), MT-sDNA detected 92% of CRC (94% of stage I-II), 69% of HGD, 66% of adenomas >2cm, and 42% of advanced adenomas (>1cm or with villous features) + SSP >1cm. Specificity was 90% based on normal colonoscopy. MT-sDNA superiority over FIT was statistically significant for all lesion categories.

Program Performance MT-sDNA could compare favorably with conventional methods in a longitudinal screening program. MT-sDNA achieves high point sensitivity for CRC, similar to that of colonoscopy. However, because it is practical to perform more frequently, MT-sDNA (followed by colonoscopy if positive) could improve program sensitivity by catching aggressive metachronous lesions missed by colonoscopy alone done at typical 10 year intervals. Point sensitivities of MT-sDNA for precancers could translate into very high programmatic sensitivity with repeated screening rounds, as we have estimated (Pathology 2012;44:80), and, therefore, into effective CRC prevention. Finally, MT-sDNA's point specificity of 90%, while lower than the average 95% reported for FIT, bodes well within a screening program as cumulative false-positives are most relevant. For example, MT-sDNA applied every 3 years (frequency supported by early modeling and approved by CMS) would yield roughly 3-4% false-positives/year, fewer than the 5% rate by FIT applied annually.

Implementation With the clinical entry of MT-sDNA come both opportunities to improve screening outcomes and a collective responsibility to apply rational principles and evidence-based decision-making to guide and refine its use.

Clinical Questions. Multiple questions will be briefly addressed in the oral presentation, including: Can FIT achieve same sensitivity by lowering its specificity?, How does MT-sDNA compare to plasma-based DNA tests?, Is MT-sDNA cost-effective?, What does the clinical use algorithm look like?, and Who would benefit from its use? With respect to the last question, while the FDA approved MT-sDNA for average-risk screening, other indications may evolve. A sound case can be made for interval testing between screening colonoscopies. As currently practiced, screening colonoscopy has a much lower benefit on incidence and mortality reduction from right-sided than left-sided CRC, interval cancers are disproportionately right-sided, and quality metrics of this operator-dependent procedure vary widely. Because of its very high sensitivity for CRC that is unaffected by tumor site and its superior sensitivity over FIT for detection of sessile serrated polyps and those adenomas at greatest risk of progression, MT-sDNA seems well-suited for consideration as an interval test. In instances of inadequately prepared or incomplete colonoscopy, which occur in 15 - 25% of colonoscopies, MT-sDNA testing may offer a convenient follow-up alternative to repeat colonoscopy or other imaging methods, particularly among those reticent or unwilling to undergo repeat bowel preparation and invasive testing. In low CRC risk surveillance populations, such as patients with a single small adenoma, use of MT-sDNA might be explored as a resource-savings approach. Those at high CRC risk represent another potential target group for surveillance. A premise for the recommendation of colonoscopy as the choice for high risk surveillance has been its unparalleled high sensitivity. However, in light of the recent MT-sDNA validation studies showing similarly high point-sensitivity for early-stage CRC and estimates suggesting comparable program-sensitivity for advanced precancers, it is intriguing to re-consider noninvasive surveillance in this setting. Further studies are clearly needed. Meanwhile, in high risk individuals refusing or unable to access colonoscopy and in those with relative contraindications to colonoscopy, surveillance by MT-sDNA might be considered.

The Future: Expanded Applications Inflammatory Bowel Disease (IBD). IBD patients are at increased CRC risk, and stool DNA testing may help to improve surveillance detection and compliance rates. In a feasibility study (Aliment Pharmacol Ther 2013;37:546), highly accurate detection of both CRC and HGD in IBD patients was observed by stool assay of methylated DNA markers alone.

Supra-colonic Cancers: There remains an enormous unmet need for early detection of cancers above the colon, which collectively exact a death toll 2-times higher than that of CRC in the US and 4-times higher globally. Based on our early observations, stool DNA testing represents a potential strategy to fill this void. If combined with CRC screening, it would be ideal to differentiate upper from lower GI lesions; and preliminary data suggest that site-specific DNA tumor markers can be identified (Gastroenterology 2013;144:S-84). Clearly, more studies are needed and warranted.

As stool DNA testing moves from historical promise into primetime, efficient implementation systems and essential linkages with endoscopy will need to be established to achieve highest impact. Logical extensions of this powerful molecular technology open intriguing opportunities to expand its value.

Citation Format: David A. Ahlquist. Stool DNA detection of colorectal neoplasia: A new high bar for noninvasive screening. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr PL02-01. doi:10.1158/1538-7445.AM2015-PL02-01

©2015 American Association for Cancer Research.
2015