Segmental re-examination of the proximal colon and its impact on adenoma detection: Does seeing twice make nice?

Colonoscopy is the accepted gold standard for detecting colorectal adenomas and neoplasms [1]. Yet studies have consistently shown that a significant percentage of adenomas, especially small lesions, can be missed during colonoscopy [2]. This is problematic as missed adenomas can lead to interval colorectal cancers, especially in the right colon [3] [4] [5]. Thus, there has been an array of emerging technologies attempting to improve adenoma detection (or conversely reduce adenoma miss rates) [6]. Others have attempted to develop practical colonoscopic techniques or maneuvers that may be utilized by the colonoscopist to improve adenoma detection [7]. These techniques include prolonged colonoscope retroflexion in the right colon and extended withdrawal times. However, these maneuvers may not be consistently performed by the colonoscopist or may be done using a less-than-optimal technique as they require additional time, technical skill, and may confer a limited patient risk.

“Do our brains synthesize or process this visual information differently on separate passes of the colonoscope?”

In this issue of Endoscopy, Guo et al. report the results of a single-center, parallel group, randomized, blinded study whereby adult patients (18 – 80 years of age) at intermediate or high risk for colorectal neoplasia and who were scheduled to undergo colonoscopy were randomly assigned to receive segmental re-examination or extended withdrawal time (EWT group) of the proximal colon, with the proximal colon defined as anywhere proximal to and including the splenic flexure [8] [9]. Segmental re-examination was a priori defined as follows. After initial cecal intubation, the colonoscope was withdrawn to the hepatic flexure with careful inspection of the colonic mucosa. The colonoscope was then re-advanced into the cecum and a second withdrawal was performed with re-examination to the hepatic flexure. This segmental re-examination technique was similarly repeated in the transverse colon to the splenic flexure. The withdrawal time of the segmental re-examination group was not restricted. Extended withdrawal time was defined as the colonoscope being withdrawn from the cecum directly to the splenic flexure in a slow meticulous fashion (minimum 3 – 4 minutes). Patients and pathologists were blinded to randomization group. Four endoscopsists performed all colonoscopies using high-definition colonoscopes. In a per patient analysis, the primary end point of the study was the proximal colon adenoma detection rate (ADR).

In the per patient analysis, the authors report that the proximal colon ADR was significantly higher in the segmental re-examination group compared with the EWT group (33.1 % vs. 23.6 %, respectively; P = 0.045). However, there was no statistically significant difference between groups in advanced ADR of the proximal colon, which might be expected as the study was not powered to detect such a difference. In addition, in a per lesion analysis, a significantly higher proportion of proximal colon adenomas were identified using segmental re-examination compared with EWT (72.1 % vs. 56.0 %, respectively; P = 0.008). Moreover, of the adenomas identified in the segmental re-examination group, 38 /98 were identified during the second pass, translating into a proximal colon adenoma miss rate of 38.8 %. All missed adenomas were diminutive tubular adenomas, with most located in the ascending or transverse colon. It should also be pointed out that in the segmental re-examination group, 9/10 (90 %) of the adenomas found in the hepatic and splenic flexures were found during the second examination, and that 19 patients (10 %) had adenomas detected only on the second re-examination of the proximal colon. Although defined in the Methods section of the article, there was no reported histological evaluation of low grade vs. high grade dysplasia within the adenomas, so this additional marker of histological risk cannot be further commented upon. Overall, there was no difference between randomization groups in mean withdrawal times of the proximal colon (approximately 4.5 minutes each), nor of total withdrawal time (just under 8 minutes in each group).

So how can the results of this prospective randomized blinded study be interpreted? What might explain the observed differences between groups? And how can this study potentially benefit us in our approach to colonoscope withdrawal and reducing missed adenomas, especially in the proximal colon? Overall, this appears to be a well-conducted study and the authors are to be commended. The authors have added to the growing literature on techniques and practical maneuvers using our existing colonsocope that may increase adenoma detection (or reduce adenoma miss rates – however you wish to look at it) without needing to use chromoendoscopy (e. g. vital dyes or electronic), add-on devices (e. g. caps, cuffs, rings), additional imagers (e. g. Third-Eye Retroscope, Third-Eye Panoramic, Fuse), or retrofitted colonoscopes (NaviAid G-EYE). These emerging technologies all have associated learning curves for the colonoscopist, can be cumbersome and/or inefficient to use, and have associated financial costs.

Other groups have similarly reported additional practical techniques or maneuvers during withdrawal of the colonoscope in an attempt to reduce the adenoma miss rate. These have included prolonged colonoscope retroflexion in the right colon and repeat “en face” re-examination. A forthcoming systematic review and meta-analysis including 3660 patients reports that colonoscopy performed with right-sided retroflexion demonstrated an adenoma miss rate of 16.9 %, technical success in 91.9 %, and rare adverse events in 0.03 % [7].

Although the Guo et al. study was well conducted, there are several questions and comments that need to be put forth. One limitation appears to be the lack of a true “control” group. A control group could include patients in whom colonoscopy withdrawal was performed as per the colonoscopists’ usual routine technique, without segmental re-examination or extended withdrawal time. This could have strengthened the methodology of the study by adding another comparator group and increasing the generalizability of the study findings. Moreover, during performance of segmental re-examination, each colonoscopist acted as their own control. In other words, the same colonoscopist performed both first and second passes of the colonoscope, thus possibly introducing bias in a subconscious attempt to demonstrate the added value of segmental re-examination [9]. In the reported exploratory analyses using logistic regression evaluating potential predictors of detecting at least one adenoma in the proximal colon during segmental re-examination, the only reported significant predictor was proximal colon withdrawal time. However, as there were only 32 such events, the regression model most likely was “overfitted” and its results may be less than robust. Finally, there was no reported significant difference in withdrawal times between randomization groups; thus, it appears that the technique of segmental re-examination is what improved adenoma detection in the proximal colon. Why is that? When re-examining or re-visualizing the colonic mucosa do we (the colonoscopist) see the mucosa differently on the second pass? Do we somehow focus on other areas of the colonic mucosa that we did not focus on during the initial pass? Do we do this consciously? Do we do this subconsciously? Do our brains synthesize or process this visual information differently on separate passes of the colonoscope? The answers are probably “Yes,” and this topic is well beyond the scope of this editorial. Suffice it to say, however, there is much work under way in an attempt to utilize computer-aided detection technologies to help colonoscopists reduce their polyp/adenoma miss rate [10]. I believe that in the not-too-distant future, we will routinely have embedded computer software within our endoscopy platforms that will detect and automatically alert the colonoscopist to possible adenomas. Until that day arrives, we will need to continue to strive to improve and maximize our colonoscopy technique to optimize adenoma detection, possibly by performing segmental re-examination, thereby continuing to reduce the incidence of colorectal cancer.

• References

• 1 Zauber AG, Winawer SJ, O’Brien MJ. et al. Colonoscopy polypectomy and long-term prevention of colorectal-cancer deaths. N Engl J Med 2012; 366: 687-696
• 2 van Rijn JC, Reitsma JB, Stoker J. et al. Polyp miss rate determined by tandem colonoscopy: a systematic review. Am J Gastroenterol 2006; 101: 343-350
• 3 Kaminski MF, Regula J, Kraszewska E. et al. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010; 362: 1795-1803
• 4 Cooper GS, Xu F, Barnholtz Sloan JS. et al. Prevalence and predictors of interval colorectal cancers in Medicare beneficiaries. Cancer 2012; 118: 3044-3052
• 5 Corley DA, Jensen CD, Marks AR. et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014; 370: 1298-1306
• 6 Gralnek IM. Emerging technological advancements in colonoscopy: Third Eye Retroscope and Third Eye Panoramic, Fuse Full Spectrum colonoscopy platform, Extra-Wide-Angle-View colonsocope, and NaviAid G-EYE balloon colonoscope. Dig Endosc 2015; 27: 223-231
• 7 Cohen J, Grunwald D, Grossberg LB. et al. The effect of right colon retroflexion on adenoam detectio: a systematic review and meta-analysis. J Clin Gastroenterol 2016; DOI: 10.1097/MCG.0000000000000695.
• 8 Guo CG, Zhang F, Ji R. et al. Efficacy of segmental re-examination of proximal colon on adenoma detection during colonoscopy: a randomized controlled study. Endoscopy 2017; 49: 243-250
• 9 Van den Broek FJC, Kuiper T, Dekker E. et al. Study designs to compare new colonoscopic techniques: clinical considerations, data analysis, and sample size calculations. Endoscopy 2013; 45: 922-927
• 10 Tajbakhsh N, Gurudu SR, Liang J. A comprehensive computer-aided polyp detection system for colonoscopy videos. Inf Process Med Imaging 2015; 24: 327-338