AGA: Gastrointestinal Innovation at the Forefront of Clinicians’ Minds

AGA Center for GI Innovation and Technology hosted “Colorectal Cancer Screening and Surveillance: Role of Emerging Technology and Innovation to Improve Outcomes.”

In 2018, the American Gastroenterological Association’s (AGA) Center for GI Innovation and Technology hosted a program titled “Colorectal Cancer Screening and Surveillance: Role of Emerging Technology and Innovation to Improve Outcomes.” Attendees were invited to disregard currently accepted state-of-the-art practices, reimagine the present as an “imperfect stepping stone,” and envision a future defined by optimized colorectal cancer screening and surveillance.

According to Neelendu Dey, MD, of the Clinical Research Division at the Fred Hutchinson Cancer Research Center and the Division of Gastroenterology at the University of Washington in Seattle, Washington, the overarching objectives of the conference were to identify current practice gaps, identify promising new innovations, understand where these innovations are needed, and define the requisite clinical studies required to validate new colorectal cancer screening paradigms.

To execute this objective, the hosts inverted the traditional conference format, prioritizing hour-long group discussions. A meeting summary of these discussions was published in Gastroenterology.1

Current Gaps in Practice

According to both conference attendees and the World Health Organization’s International Agency for Research on Cancer, an organized, population-based screening approach for colorectal cancer — including defined target populations, timely access, and quality assurance — is preferred to an opportunistic approach. “Screening is a process, not a single test, that is comprised of stages,” Dr Rabeneck emphasized.1

One major gap she identified is the need for an effective information technology platform in order to best organize and deploy population-based screening.

In addition to the overall need for targeted screening, attendees, led by AGA president David Lieberman, MD, of Oregon Health and Science University, discussed guidelines issued in 2016 by the US Preventive Services Task Force, in 2017 by the Multi-Society Task Force, and in 2018 by the American Cancer Society.1 Attendees considered whether the presence of multiple guidelines is representative of “a process problem” in gastroenterology. Multiple attendees shared their viewpoints, noting that the guidelines are based on imperfect science and that the implementation of shared decision making, often touted in guidelines, is unclear.

Because the presentations primarily focused on practices based in the United States and Canada, colonoscopy represented a large part of the attendee discussions.1 One attendee likened colonoscopy to “no other [screening test] in medicine,” citing the preparation that patients must take to ensure surveillance success.

Due to the commitment required in addition to the demographic risk factors predictive of colonoscopy uptake and adherence (socioeconomic disparities, lifestyle factors, race, ethnicity, education, and language barriers), physicians must recognize and address potential barriers such as embarrassment, fear, and personal or cultural factors that prevent patients from undergoing colonoscopy screening.1

Despite the challenges the method is associated with, Steven Itzkowitz, MD, of the Icahn School of Medicine at Mount Sinai, estimated that if colonoscopy screening began at 45 years of age, each 1000 individuals undergoing a colonoscopy would result in a gain of 25 additional life-years and the prevention of 3 cases of colorectal cancer.1

However, it is essential to acknowledge the implicit bias that gastroenterologists have when evaluating the limitations of colonoscopy or assessing competing procedures, the authors noted. 

Another attendee, Jeffrey Lee, MD, of Kaiser Permanente San Francisco, discussed the quality metrics currently employed throughout the multiple stages of the colonoscopy procedure: quality of bowel prep, cecal intubation rate, withdrawal time, adenoma detection rate, polyp characterization, completeness of polypectomy, rate of post-polypectomy pleading, and colorectal cancer incidence rate. Despite the multitude of quality metrics, gastroenterologists currently lack a metric for the quality of the endoscopists inspection of colonic mucosa. Dr Lee notes that adenoma detection rate is imperfect due to its inability to distinguish between multiple adenomas per colonoscopy. Endoscopists, he added, “may be susceptible to gaming the system via a ‘one-and-done’ approach.”

Jason Dominitz, MD, of the University of Washington School of Medicine, outlined the strengths, caveats, and limitations of stool-based fecal immunochemical testing.1 His presentation reemphasized patient buy-in as a central meeting theme: A screening test with 100% sensitivity has a 0% so-called effective sensitivity if no patients are willing to use the test.1

Additionally, Dr Dominitz presented data illustrating that an organized colorectal screening program that integrated both fecal immunochemical testing and lower endoscopy was associated with a 25% and 50% reduction in incidence and mortality.2

However, stool-based testing can be rife with barriers, including incomplete follow-up after a positive test result, incomplete knowledge of test performance over multiple consecutive intervals, and, according to Dr Dey and colleagues, “the lingering question of what to do after a negative colonoscopy preceded by a positive [fecal immunochemical test].”1

Promising Innovations and Studies for the Future of Colorectal Cancer

One promising so-called “great equalizer” among endoscopists is artificial intelligence.1 One 2018 study, published in Gastroenterology,3 identified a 96.4% rate of accuracy in polyp identification with the implementation of convolutional neural networks and deep learning. Such a system, according to study authors, could increase the adenoma detection rate and decrease interval colorectal cancers.3

While the data so far has been promising, more work is needed before artificial intelligence in this area is truly ready. One barrier Dr Bryne noted was that current verified videos are low quality and siloed; this can be combatted through the establishment of centralized or cloud-based colonoscopy video repositories, creating a training data program.1

Martin Golding, MD, of the US Food and Drug Administration, added that the use of artificial intelligence could potentially reduce colorectal cancer incidence, increase complications of colonoscopy, or both.1

Another promising screening option is computed tomography colonoscopy (CTC), a minimally invasive screening option. While it is potentially less costly, as it does not require patient sedation, CTC has not yet become a first-line, Tier 1 test. According to Judy Yee, MD, of the Albert Einstein College of Medicine, clear guidelines on the management of extra-colonic findings must be developed. If this criteria is met, Dr Yee believes that CTC may become a “better accepted” solution for screening younger patients.1

In addition to these screening options, future research to identify novel molecular and cellular strategies for targeting both host and microbial populations should result in the development of additional tools for colorectal cancer screening and surveillance. William Grady, MD, of the Fred Hutchinson Cancer Research Center, described multi-target stool DNA and plasma methylated Septin 9 testing, which are currently approved for use in screening.1 Serum protein panel testing is promising for the detection of adenoma, while emerging genetic variants, epigenetic variants, microRNA, and metabolome-based assays require validation studies.1

Related Articles

Robert Bresailer, MD, of the University of Texas MD Anderson Cancer Center, posited that “liquid biopsy” technology as a potentially useful noninvasive screening tool, attractive to patients who refuse lower endoscopy.1 This technology could also be used as an adjunct screening tool for the detection of proximal colonic lesions, as well as a reflex test following inadequate negative colonoscopy. Discussions of this development, however, pointed out that a blood-based cancer marker nonspecific to colorectal cancer could pose “special challenges.”1

Although epidemiologic studies of patients with colorectal cancer have “variably implicated different individual gut bacterial species,”1 a barrier to integrating the gut microbiome into clinical decision making is the inadequate data on the microbiome and metabolome, which is required to “yield insights into causal relationships of the microbiome in both health-to-adenoma and adenoma-to-carcinoma transitions,” Dr Dey argued.1

“With improved understanding of bacterial genomes and metabolism, as well as regulators of bacterial metabolism, microbiome engineering will ultimately permit clinicians to use microbe-based technologies to monitor and/or modulate the colon,” Dr Dey suggested.1

Molecular diagnostics could theoretically be incorporated into a screening program; however, key considerations must be factored in, including patient perspectives, test characteristics and reliability, interpretability, and benchmarks.1 One attendee proposed fecal immunochemical testing as an acceptable benchmark while other, newer tools are developed.

Two fundamental questions were debated over the course of the meeting: first, what is the precise goal of screening and surveillance, and second, what ought to be the measurable readouts or targets? Ultimately, attendees agreed that carcinogenesis — not cancer — should be the target.1

Finally, John Baron, MD, of the University of North Carolina at Chapel Hill, closed the meeting by reviewing the safety and efficacy data for chemoprevention.1 Dr Baron cautioned that “off-target effects are essential to consider,” adding that toxicity can negate effective agents. Robert Schoen, MD, of the University of Pittsburgh School of Medicine, posed goals for academic-industrial integration, arguing that physicians must identify precursor biomarkers.1 While >60% of colonoscopies are negative, the current generalized screening approach is inefficient.1 Dr Schoen argued that the identification of precursors to colorectal cancer — not just colorectal cancer itself — is essential to reduce incidence rates.

“This [Association’s Center for GI Innovation and Technology] meeting was motivated by the concept that efficient innovation directly aligns research and development with the most pressing clinical gaps and needs, and it serves as a model for organizing academic-industrial collaboration,” Dr Dey and colleagues concluded.

Disclosure: Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.


  1. Dey N, Kochman ML, Komanduri S, Melson JE, Muthusamy VR. Report from the AGA Center for GI Innovation and Technology’s Consensus Conference: Envisioning next-generation paradigms in colorectal cancer screening and surveillance [published online September 13, 2019]. Gastroenterology. doi: 10.1052/j.gastro.2019.05.071
  2. Levin TR, Corley DA, Jensen CD, et al. Effects of organized colorectal cancer screening on cancer incidence and mortality in a large community-based population. Gastroenterology. 2018;155(5):1383-1391.
  3. Urban G, Tripathi P, Alkayali T, et al. Deep learning localizes and identifies polyps in real time with 96% accuracy in screening colonoscopy. Gastroenterology. 2018;155(4):1069-1078.