Training, financial considerations fuel debate about predict-resect-and-discard

Experts see ‘definite trend’ toward real-time optical biopsy of diminutive colorectal polyps.

Prompted by mounting research, advanced imaging technology and growing support from experts in the field, the gastroenterology community has been debating the pros and cons of a paradigm shift in the colonoscopic management of diminutive colorectal polyps over the better half of the last decade.

The replacement of routine postpolypectomy pathological analysis of diminutive polyps with real-time optical diagnosis using advanced imaging, also known as the predict-resect-and-discard strategy, is regarded by many gastroenterologists as a timely and sensible solution to the limitations of the current paradigm, primarily in terms of costs and efficiency.

Although studies have shown comparable accuracy between standard pathologic assessment and advanced imaging technologies when used by highly trained endoscopists, experts interviewed by HemOnc Today said several persisting obstacles must be overcome before widespread implementation of the predict-resect-and-discard strategy becomes possible in the United States.

Thresholds of care

Predict-resect-and-discard has already been endorsed by the European Society of Gastrointestinal Endoscopy in guidelines published in the May 2014 issue of Endoscopy. Although gastroenterological societies in America have yet to endorse it, Douglas K. Rex, MD, Indiana University School of Medicine, said the recent influx of research on the accuracy of real-time optical biopsy comes partly in response to the American Society of Gastrointestinal Endoscopy’s Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI) initiative, a summary of which was first published in the March 2011 issue of Gastrointestinal Endoscopy.

“The ASGE’s PIVI document is basically a proposal that establishes thresholds for performance,” said Rex, who chairs the PIVI committee. The document stipulates that for polyps 5 mm and smaller in size to be resected and discarded, the advanced endoscopic imaging technology used should provide greater than 90% agreement with pathology assessment in assignment of postpolypectomy surveillance intervals, at which point the ASGE would endorse that technology as an alternative paradigm.

Tonya Kaltenbach, MD, MS, from the Veterans Affairs Palo Alto, characterized the PIVI thresholds as reflective of the limitations of the current paradigm. “The thresholds were based on the current evidence that shows that pathology, although it is the reference standard, is not exactly a gold standard,” she said, adding that studies have shown that pathologists have variable interobserver agreement and 85% to 95% accuracy when diagnosing diminutive polyps. “Based on that alone, the thresholds determined by the ASGE group are very reasonable, and the technology has been shown to meet those thresholds when applied by users who are both experienced and motivated to use it.”

Data on the efficacy of optical biopsy has been accumulating since Rex’s first-of-its-kind study was published in Gastroenterology in 2009, which reported 91% and 95% accuracy in high-confidence predictions of adenoma and hyperplastic histology, respectively, of diminutive polyps using narrow-band imaging (NBI). Data confirming Rex’s findings has been consistently reported since; a study recently published in Gastrointestinal Endoscopy corroborated that high-confidence in vivo prediction of diminutive colorectal polyp histology by NBI appeared to meet the PIVI benchmarks.

Although the collective data of this nature supports the paradigm shift, according to Kaltenbach, what may be causing the American gastroenterological societies to be hesitant about endorsing it are the lack of long-term outcome studies, which would take 5 to 10 years, she said.

Dale R. Shepard, MD, PhD, FACP, department of hematology and oncology, Cleveland Clinic Taussig Cancer Institute, agrees that data from outcome studies are crucial, and that the “bottom-line” for the successful widespread adoption of predict-resect-and-discard as a clinical practice will be “partnering with oncologists to make some long-term assessments.”

“Long-term follow-up is exactly where the disconnect is going to be,” Shepard said. “It is difficult for procedural gastroenterologists to have meaningful follow-up — there is not the same continuity of care that you see in other disciplines, so what will be interesting to see is how many patients who have this procedure come in to see me with unexpected cancers. Unfortunately, I already see patients with missed cancers with the current colonoscopies, and I’m afraid that we might see similar problems.”

Many experts, including Kaltenbach, said the risk of undetected cancers is overestimated given the diminutive category of colorectal polyps that predict-resect-and-discard targets. “At the time of diagnosis,” she said, “when polyps are this small, rarely — exceedingly rarely — are they ever cancerous or have advanced pathology, so we have identified a target where we can apply the resect-and-discard strategy safely and effectively.”

Money talks

According to Rex, the essential goal of colonoscopists who support the predict-resect-and-discard paradigm is “to prevent people from getting colon cancer in a way that is both very effective and optimized from the standpoint of cost-effectiveness.”

As colonoscopy is the most widely performed procedure by gastroenterologists, with millions performed every year, “the major advantage of this algorithm would be humongous cost savings to the health care system, with estimates ranging from $33 million to a billion US dollars annually,” said Amit Rastogi, MD, associate professor of gastroenterology at University of Kansas Medical Center, and PIVI committee member.

Kaltenbach agreed that the predict-resect-and-discard proposal is “attractive and timely” primarily because the costs, time and demand for colonoscopy are increasing. In addition to savings on the direct costs of pathology processing, she said, there are a lot of potential savings in terms of time associated with polyp retrieval and pathology processing, which could make colonoscopy more efficient overall.

These time savings can benefit colonoscopy patients as well, Rastogi said. “If we can make the diagnosis during the colonoscopy, then right at the end of the procedure we can inform the patient as to what kind of polyps he had, and when he needs to have this procedure again.” Not only would this reduce patient anxiety associated with waiting for results, he said, but it also would make the process more efficient from the point of view of the endoscopist.

“We are looking at value-based medicine,” John Vargo, MD, MPH, department of gastroenterology and hepatology, Cleveland Clinic, said. “We need to maintain or improve our performance parameters while at the same time decreasing the cost to our patients, and to society. Since the fairly sizeable portion of polyps that we encounter during colonoscopy are diminutive, there are cost-effectiveness analyses that show that we can save $25 a patient, and if we look at the screening population in the US we can save tens of millions of dollars by employing resect and discard.”

While cost savings is the paradigm’s main advantage, short-term financial disincentives unique to the US health care system are one of its greatest impediments, according to Rastogi. “A lot of practices own their own pathology practice,” he said, “and in that situation sending fewer polyps [to pathology] brings you less revenue.”

One approach proposed by the American Gastroenterological Association that could drive the paradigm shift is the bundled payment model, Rastogi said. “As a gastroenterologist, or as a practice or hospital, you would get only a certain amount of fixed dollars for doing a colonoscopy, as well as sedation, the pathology, and the postcolonoscopy follow-up.” With a fixed rate, reducing expenditures is an obvious financial incentive, and would provide a stimulus to practice predict-resect-and-discard strategy, he said.

Bundled payment, said Vargo, “is basically an agreement with a payer or a company to cover the colonoscopy procedure for the workforce or the population that is under surveillance.” This could be a key opportunity if a practice can employ resect and discard and actually show savings in terms of decreased pathology costs, he said. “Driving the price point down while maintaining the same or perhaps even enhanced quality will obviously give any endoscopy group an advantage in the marketplace.”

“The real driver, as it is for so many things in medicine, is going to be when there are financial incentives that are aligned with the practice,” Rex said.

Closing the experience gap

One caveat to the data on optical diagnosis with advanced imaging, Rastogi said, is that the majority of studies with good results have been from centers with experienced endoscopists who have an interest in novel imaging technologies, whereas the few studies that have been performed in community practices have not shown similar results. His study published in the March 2014 issue of Gastrointestinal Endoscopy showed that academic and community gastroenterologists can learn how to characterize diminutive colon polyps with NBI using a computer-based teaching module, but the performance of experts was still significantly better.

“The results that have been shown by experts have not translated over to the community gastroenterologists,” he said, “so we have to develop appropriate training tools to teach community doctors to make them competent in this paradigm. Not only do we have to train, but we also have to document their competence in some fashion. Furthermore, there also has to be some method of auditing periodically to see if they are maintaining their performance.”

The experience gap is a major disadvantage of the reality of implementation, according to Shepard. “The PIVI thresholds are reasonable guidelines in the right hands,” he said, but questions whether anyone but the experienced expert “can really tell a 5-mm polyp from a 7-mm polyp in a setting where you may have a marginal prep. I appreciate that the guidelines are very clear, but I think where they fall apart is being able to put them into practical use, and being put into practical use by everybody.”

This aspect of the debate is where a key strength of the current paradigm becomes clear, according to an editorial published in Gastrointestinal Endoscopy. In it, Reed B. Hogan III, MD, and colleagues wrote that “although limitations exist (inadequate specimen, pathology error), the current histology-based paradigm provides extremely accurate and easily verifiable results. This allows all levels of endoscopists to diagnose and schedule follow-up intervals with great confidence.”

The data from non-academic centers is indeed mixed, according to Vargo, and they generally have two components. “There’s the training session which, usually utilizes optimized static images, and usually most of the physicians do well with that. It is the in vivo piece that seems to pose a challenge.” Vargo said the solution may lie in innovations in training, including stop-motion video or working more closely with experts.

“The other concern I have is the timing of utilizing the optical imaging following the computerized training,” Vargo said. “If there is a lag time between achieving competency on the training module and applying it into colonoscopy, I theorize that there could be a diminution in the skill set.”

According to Kaltenbach, the greatest barriers to predict-resect-and-discard are indeed “the teaching material, disseminating it and motivating people to use it and continue to learn over time. What we have observed both from our own experience and in published studies is this is not an innate type of knowledge,” rather it is something that needs to be taught and practiced, she said. Although the technology is able to meet the PIVI thresholds, “it is the user of the technology who needs to make sure they continue to meet those thresholds. If this becomes part of peoples’ curriculums, and becomes more of an expectation than an interest, it has been shown across the board that you can learn it.”

A credentialing system may need to be developed, Rex said, where the trainee’s performance can be measured in real-time and compared with pathology to demonstrate competence in the predict-resect-and-discard practice.

Risk and responsibility

According to Kaltenbach, “because the endoscopist or gastroenterologist is not directly paid to make this diagnosis or inherit this risk, they have reservations,” but because predict-resect-and-discard targets only diminutive polyps that rarely harbor cancer or advanced pathology, she said the risk is overestimated.

Rastogi agrees. “If the predict-resect-and-discard paradigm becomes standard of care, then obviously the responsibility then lies on the shoulders of the endoscopist rather than the pathologist,” he said, adding that one way to combat the endoscopist’s reservation toward taking the responsibility of accurately predicting the histology of these polyps is by saving high-definition images of polyps as supporting evidence for their real-time diagnosis. “The medicolegal liability can be a deterrent to this practice, but a good image would serve as proof that the polyp is either an adenoma or a hyperplastic polyp,” he said.

“We need to move into an era of very high-quality image storage because an endoscopist needs to be able to recreate the basis for their decision,” Rex said. “There must be a photographic record of the polyp in the same way that the pathologists have a glass slide and a photo-micrograph.”

Importance of data

According to Rex, recent meta-analyses of the literature on predict-resect-and-discard showed that several advanced imaging modalities have performed well enough to be used in clinical practice, particularly NBI as it has been tested in the most studies. However, the endorsements of medical professional societies are crucial, he said. “I would like to see all of them endorse resect and discard as an alternative approach. Endorsement will not mean everyone will have to start doing resect and discard immediately, but rather that those with the skill and inclination to do it will have some back-up to get started.”

“There is a definite trend for adoption and I see in the future multi-society approval of the PIVI,” said Vargo. “It is evidence based, the data is growing, and I think the challenge now to the gastroenterology community is to optimize training and ensure appropriate competency before it is employed.” – by Adam Leitenberger

References:

Hogan RB. Gastrointest Endosc. 2012;75(3):503-505.

Kaminski MF. Endoscopy. 2014;46(5):435-449.

Rastogi A. Gastrointest Endosc. 2014;79(3):390-398.

Repici A. Gastrointest Endosc. 2013;78(1):106-114.

Rex DK. Gastrointest Endosc. 2011;73(3):419-422.

Rex DK. Gastroenterology. 2009;136(4):1174-1181.

For more information:

Tonya Kaltenbach, MD, MS, can be reached at endoresection@me.com.

Amit Rastogi, MD, can be reached at arastogi@kumc.edu.

Douglas K. Rex, MD, can be reached at drex@iu.edu.

Dale R. Shepard, MD, PhD, FACP, can be reached at shepard@ccf.org.

John Vargo, MD, MPH, can be reached at vargoj@ccf.org.

Disclosure: Kaltenbach, Rastogi, Rex and Vargo are consultants to Olympus America Inc., and Kaltenbach, Rastogi and Rex have also received research funding from Olympus America Inc. Shepard reports no relevant financial disclosures.