Bypassing user error in optical diagnosis of diminutive colorectal polyps – at what cost?

Colonoscopy and polypectomy is the dominant screening modality in the United States for colorectal cancer [1] and in large part accounts for the progressive decline in colorectal cancer incidence and mortality over the last several decades [2]. Although colonoscopy is cost-effective when taking into account the high cost of colon cancer treatment [3], the overall cost will continue to rise as colonoscopy uptake improves and more patients enter polyp surveillance protocols. Diminutive (≤ 5 mm) colorectal polyps are the most commonly found polyps during colonoscopy, comprising 62.7 % of all polyps found in a screening cohort [4]. They rarely harbor advanced histology (1.2 % – 3.4 %) [5] [6] or malignancy (0 % – 0.08 %) [4] [7] [8]. Despite this low neoplastic potential, diminutive polyps are routinely resected and submitted for pathologic assessment in order to guide surveillance intervals. There is potential for enormous cost savings [9] [10] if the histology of diminutive polyps could be determined in real time and the costs of resecting nonadenomatous polyps and of pathologic examination of all diminutive polyps could be avoided.

The American Society for Gastrointestinal Endoscopy (ASGE) has accordingly published the Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI) [11] statement for the application of narrow band imaging (NBI) and other endoscopic imaging techniques in the real-time assessment of diminutive colorectal polyps. The PIVI statement suggests that for diminutive polyps characterized with “high confidence,” endoscopists should achieve a ≥ 90 % negative predictive value (NPV) for adenomas in the rectosigmoid colon and a ≥ 90 % agreement in surveillance intervals compared with those dictated by histology. There are now data showing that both experts and those without experience in NBI can achieve these thresholds with standardized training and real-time feedback [12] [13]. In a recent multicenter study including 26 endoscopists, we showed that with standardized training and real-time performance feedback, endoscopists without prior expertise in NBI can exceed the thresholds set forth by the ASGE for characterization of diminutive colorectal polyps [13].

Although the PIVI statement was specifically formulated for the application of NBI, the suggested performance thresholds can be applied to any optical diagnostic technology. Rath and colleagues [14] report results from a pilot study evaluating the diagnostic performance of a probe-based laser-induced autofluorescence spectroscopy system (WavSTAT4; SpectraScience, San Diego, USA) in determining diminutive colorectal polyp histology in real time. The probe is integrated into a single-use biopsy forceps and takes 1 second to analyze tissue fluorescence signals to determine whether a polyp is “suspect” or “not suspect.” The study included 27 patients referred for screening or surveillance colonoscopy in whom only diminutive polyps were found. A total of 137 diminutive colorectal polyps were analyzed, 115 (83.9 %) of which were nonadenomatous by histology. The WavSTAT technology achieved an overall accuracy of 84.7 %, sensitivity of 81.8 %, specificity of 85.2 %, and NPV of 96.1 %. When restricted to distal diminutive polyps, the NPV improved to 100 %. Using Multi-Society Task Force (MSTF) polyp surveillance guidelines [15], the agreement for surveillance intervals was 88.9 % with all disagreement bringing patients back sooner than recommended by histology.

The currently widely available optical diagnostic technologies such as narrow band imaging (Olympus) are operator-dependent. Although recent studies [12] [13] have demonstrated that experts and nonexperts in advanced imaging can as a group achieve the ASGE thresholds using NBI, there is significant variability in individual performance/learning curves and there are some individuals who just do not achieve competence [13]. Widespread adoption of NBI optical diagnosis will require intensive and standardized training and would likely require periodic audit of performance in real time to ensure thresholds are continually met. It is clear that the overall cost – effectiveness of this endeavor is dependent on two critical variables: the overall accuracy and the proportion of “high confidence” diagnosis in real-time prediction of diminutive colon polyp histology. These measures will require institutional resources and individual commitment to the technology which may partially offset the cost/resource effectiveness. Furthermore, cost savings may also be offset by earlier surveillance examinations and higher proportions of “low confidence” diagnoses which would require resection and pathologic analysis.

There are conceptual advantages to a technology such as WavSTAT in real-time characterization of diminutive polyp histology. WavSTAT adopts a standardized computer algorithm to analyze induced-fluorescence tissue signatures and completely circumvents user variability. Unfortunately, there are several unanswered questions, based on the data presented in this study. Furthermore, the ultimate purpose of a “characterize, resect, and discard” strategy is to minimize cost while preserving quality. Thus, the global cost of using this technology must be justified. With any new diagnostic test or technology, it is important to demonstrate reliability and reproducibility of results in addition to performance when compared with the current gold standard (histopathology) as reported in this study. It would be important to know the rate at which the same polyp would be identified in the same way by the same probe (for instance after analyzing a different polyp) and the rate at which different probes would analyze the same polyp with the same outcome.

The accuracy and NPV reported in this study may be an overestimation due to a potential clustering effect, as acknowledged by the investigators. There were only 27 patients included in this study and a total of 137 polyps; thus each patient had over five diminutive polyps on average. The majority of patients in prior studies have had one to two diminutive polyps per patient [12] [13] [16]. The high number of polyps per patient, and significantly higher proportion of hyperplastic (84 %) or nonadenomatous polyps compared with prior studies (38 % – 47 %) [12] [13] [16], suggest possible observation of clusters of hyperplastic polyps, thus overestimating diagnostic accuracy compared with the value that might have been obtained by reviewing polyps across more patients.

It may be premature to declare that this technology accurately characterizes sessile serrated polyps even though the two histologically identified sessile serrated polyps were characterized as “suspect.” Though the authors do not present histology by segment of the colon, we presume that there was a substantial portion of ‘hyperplastic’ polyps in the ascending colon (since there were 49 total polyps in the ascending colon and only 22 adenomas in total). The authors specifically state that the nonadenomas were hyperplastic (as opposed to normal tissue, lymphoid aggregates, or a variety of other nonadenoma findings) raising the issue of whether the diminutive polyps on the right side of the colon were in fact sessile serrated polyps.

The authors report that the agreement in surveillance intervals predicted by using WavSTAT compared with histology was 88.9 % using the MSTF guidelines. The MSTF guidelines state that for 1 – 2 diminutive adenomas, surveillance can take place between 5 to 10 years, based on the discretion of the endoscopists. Adoption of an optical diagnosis technology should not alter the current standard of care, acknowledging the variability in standard of care among endoscopists. For instance, some endoscopists routinely recommend 5 years if 1 – 2 diminutive adenomas are found, whereas others may extend surveillance to 10 years with the same findings, depending on the case. Prior studies have therefore more stringently reported intraendoscopist agreement [12] [13] [16] instead of agreement with a range. Thus, agreement regarding surveillance intervals may be overestimated in this study.

There are important issues when considering the potential cost of implementing WavSTAT. In addition to the upfront cost of investing in the diagnostic unit, the probes are integrated into single-use biopsy forceps. Thus, every case/patient would require a new probe. Furthermore, for patients found to have “not suspect” polyps, the forceps would be discarded without being used. Finally, there is now ample evidence that even diminutive polyps have a higher incomplete resection rate (compared with snare removal) when removed by biopsy forceps [17] [18] [19]. Thus, practitioners might use the biopsy probe for diagnosis, but then use a separate snare for resection. This approach would add time and cost to the procedure. If an endoscopist settles on using a forceps when he/she would have otherwise used a snare, quality of resection may be compromised. Finally, the authors indicate that the technology is increasingly recognized by insurance carriers, and applications for integration of this technology into procedural costs and the reimbursement system are underway in parts of Europe. However even if insurance covers these costs (relieving the endoscopist from bearing the cost), the global cost to the healthcare system still rises.

Rath and colleagues present very interesting preliminary results on an emerging technology that potentially comes close to meeting the thresholds set forth by the ASGE for optical diagnosis of diminutive colorectal polyps. As the authors concede, there is need for large multicenter studies to validate these findings, and to demonstrate consistent achievement of the PIVI thresholds and comparability to current endoscopic imaging techniques for real-time prediction of polyp histology. Although it is highly appealing to embrace a technology that obviates user error, the entire purpose of a “characterize, resect, and discard” strategy is to decrease the overall cost of colonoscopy to healthcare systems without compromising quality of care. Even if this technology is validated and meets the ASGE thresholds, there is an important responsibility to demonstrate that the strategy is cost-effective compared with the current standard of practice, and that quality is not compromised by settling for a resection technique that increases rates of incomplete resection. Employing a single-use probe for the most commonly found polyps in colonoscopy, in addition to a snare to actually resect the polyp if needed, seems to pose a significant cost burden. This is especially the case when the same outcome can be achieved with training in a technology (such as NBI) that is already provided on widely used endoscopes.

In light of the limitations discussed above, before a technology such as WavSTAT can be widely implemented, there needs to be convincing evidence that the results are reliable, valid, and reproducible in multicenter trials, and that they consistently meet the ASGE thresholds. This technology should be compared with previously rigorously studied technologies such as NBI. It would be important to understand user practices in the use of a forceps versus a snare and how that may impact cost and incomplete resection rates. Finally, before this technology can be widely implemented, it is very important to carefully determine its cost – effectiveness, compared with the current gold standard of resecting all polyps and sending for histological examination, as well as with current optical diagnostic technologies such as NBI.

• References

• 1 Steele CB, Rim SH, Joseph DA et al. Colorectal cancer incidence and screening – United States, 2008 and 2010. MMWR Surveill Summ 2013; 62 (Suppl. 03) 53-60
• 2 Edwards BK, Ward E, Kohler BA et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010; 116: 544-573
• 3 Lansdorp-Vogelaar I, Knudsen AB, Brenner H. Cost-effectiveness of colorectal cancer screening. Epidemiol Rev 2011; 33: 88-100
• 4 Lieberman D, Moravec M, Holub J et al. Polyp size and advanced histology in patients undergoing colonoscopy screening: implications for CT colonography. Gastroenterology 2008; 135: 1100-1105
• 5 Aldridge AJ, Simson JN. Histological assessment of colorectal adenomas by size. Are polyps less than 10 mm in size clinically important?. Eur J Surg 2001; 167: 777-781
• 6 Gschwantler M, Kriwanek S, Langner E et al. High-grade dysplasia and invasive carcinoma in colorectal adenomas: a multivariate analysis of the impact of adenoma and patient characteristics. Eur J Gastroenterol Hepatol 2002; 14: 183-188
• 7 Pickhardt PJ, Choi JR, Hwang I et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 2003; 349: 2191-2200
• 8 Butterly LF, Chase MP, Pohl H et al. Prevalence of clinically important histology in small adenomas. Clin Gastroenterol Hepatol 2006; 4: 343-348
• 9 Hassan C, Pickhardt PJ, Rex DK. A resect and discard strategy would improve cost-effectiveness of colorectal cancer screening. Clin Gastroenterol Hepatol 2010; 8: 865-869 , e1 – e3
• 10 Kessler WR, Imperiale TF, Klein RW et al. A quantitative assessment of the risks and cost savings of forgoing histologic examination of diminutive polyps. Endoscopy 2011; 43: 683-691
• 11 Rex DK, Kahi C, O'Brien M et al. The American Society for Gastrointestinal Endoscopy PIVI (Preservation and Incorporation of Valuable Endoscopic Innovations) on real-time endoscopic assessment of the histology of diminutive colorectal polyps. Gastrointest Endosc 2011; 73: 419-422
• 12 Kaltenbach T, Rastogi A, Rouse RV et al. Real-time optical diagnosis for diminutive colorectal polyps using narrow-band imaging: the VALID randomised clinical trial. Gut 2015; 64: 1569-1577 DOI: 10.1136/gutjnl-2014-307742. Epub 2014 Nov 11
• 13 Patel SG, Schoenfeld P, Kim H et al. Real-time characterization of diminutive colorectal polyp histology using narrow-band imaging: implications for the resect and discard strategy. Gastroenterology 2016; 150: 406-418 DOI: 10.1053/j.gastro.2015.10.042. Epub 2015 Oct 30
• 14 Rath T, Tontini G, Vieth M et al. In vivo real-time assessment of colorectal polyp histology using an optical biopsy forceps system based on laser-induced spectroscopy. Endoscopy 2015; 48: 557-562
• 15 Lieberman DA, Rex DK, Winawer SJ et al. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2012; 143: 844-857
• 16 Ladabaum U, Fioritto A, Mitani A et al. Real-time optical biopsy of colon polyps with narrow band imaging in community practice does not yet meet key thresholds for clinical decisions. Gastroenterology 2013; 144: 81-91
• 17 Jung YS, Park JH, Kim HJ et al. Complete biopsy resection of diminutive polyps. Endoscopy 2013; 45: 1024-1029
• 18 Lee CK, Shim JJ, Jang JY. Cold snare polypectomy vs. Cold forceps polypectomy using double-biopsy technique for removal of diminutive colorectal polyps: a prospective randomized study. Am J Gastroenterol 2013; 108: 1593-1600
• 19 Kim JS, Lee BI, Choi H et al. Cold snare polypectomy versus cold forceps polypectomy for diminutive and small colorectal polyps: a randomized controlled trial. Gastrointest Endosc 2015; 81: 741-747