Does advanced endoscopic imaging increase the efficacy of surveillance colonoscopy?

Colonoscopy with adenoma resection is thought to reduce the incidence of colorectal cancer (CRC) by up to 80 % [1], and recent randomized population-based data confirm that flexible sigmoidoscopy can reduce left-sided CRC development and death from CRC [2]. However, recent reports demonstrate that some patients under colonoscopic surveillance still develop CRC, particularly in the right colon [3] [4]. This may be explained by a rapid progression of adenomas, by incomplete removal of adenomas, or by the fact that colonoscopy is not infallible for the detection of adenomas. Indeed, a systematic review of back-to-back colonoscopies demonstrated that 15 %–32 % of adenomas were overlooked [5]. On top of this, subtle flat or depressed lesions are more difficult to detect and may have been missed by both colonoscopies in study designs of this type.

Over the past few years increasing scientific data have demonstrated that the quality of colonoscopy can be improved by attention to basic colonoscopic technique. These quality indicators include adequate bowel preparation, confirmation of cecal intubation, an endoscope withdrawal time of at least 6 minutes, and an adenoma detection rate by the endoscopist of at least 20 % [6] [7] [8]. In response, professional societies have proposed the use of various quality assessment indicators and the implementation of continuous quality improvement programs in daily clinical practice. It remains unclear whether such programs lead to improvements in adenoma detection [9] and CRC-related death.

Over the past 40 years, colonoscopic equipment has progressed rapidly from fiberoptic colonoscopes, to video colonoscopes (standard definition), to high resolution colonoscopes (increased number of pixels in charged coupled device [CCD], with output to a standard resolution monitor), to the current series of high definition colonoscopes (substantially increased number of pixels in CCD with output to a high definition monitor). Each new generation of instruments has improved the quality of the image presented to the endoscopist. Theoretically, this should improve diagnostic accuracy and thus increase the effectiveness of colonoscopic surveillance and screening.

Five colonoscopic studies have been published comparing standard white light video endoscopy with high definition white light endoscopy involving more than 4000 patients, in both cohort and randomized designs [10] [11] [12] [13] [14]. Only one large cohort study has demonstrated a benefit from high definition white light endoscopy ([Table 1]). In the majority of cases the high definition colonoscopes were also wide angle instruments, potentially further increasing the potential advantage over the non-wide angle standard definition instruments used as comparators. Therefore, when considered in the round, the published studies to date do not show an increased adenoma detection rate for high definition white light colonoscopy when compared with standard colonoscopy, although a small benefit may exist.

Alongside the development of high definition endoscopy, additional advanced imaging techniques have been developed. Application of dye during colonoscopy with a white light endoscope, chromoendoscopy, is used to improve the detection of adenomas. Several alternative approaches to enhance the contrast of the mucosal surface without the use of dye have been introduced as “digital chromoendoscopy” – narrow band imaging (NBI, Olympus, Tokyo, Japan), Fujinon intelligent color enhancement (FICE, Fujinon, Saitama, Japan), and i-Scan (Pentax, Tokyo, Japan). These technologies are all integrated in the endoscopic system and activated at the push of a button on the endoscope. NBI makes use of optical filters creating narrowed wavelength bands of blue and green light for illumination of the mucosa. As blue light has only a superficial penetration depth into the mucosa and is the main color absorbed by hemoglobin, this setting allows for detailed mucosal imaging with enhancement of small superficial capillaries. Adenomas appear darker than their background due to a higher vascular pattern intensity, and it is also possible to predict histology with a high accuracy using the Kudo pit pattern. FICE and i-Scan also enhance contrast for superficial mucosal capillaries, but are based on a computed spectral estimation technology that arithmetically processes the reflected photons to reconstitute virtual images for a choice of different wavelengths.

To justify the acquisition of new and expensive advanced imaging techniques, thorough scientific evaluation should demonstrate its effectiveness when compared with standard techniques. Today, standard equipment is providing high quality white light colonoscopy. What is the additional value of the different additional techniques in adenoma detection?

Chromoendoscopy is the only technique that has been shown to improve the detection of adenomas during colonoscopy, both for patients with at least one adenoma detected, and for the detection of three or more adenomas, an important end point for risk stratification for surveillance [16] [17]. However, this is a time-consuming method and therefore impractical during routine practice. On the other hand, for this technique the endoscopist does not have to buy new and expensive endoscopic equipment, only dye (usually indigo carmine or methylene blue) and a spray catheter. The technique is cheap and an easy addition to regular colonoscopy that can be used in high risk patients such as those with Lynch syndrome [18] or, for example, when in doubt on the presence of a flat lesion.

Of the different digital chromoendoscopy techniques, NBI has been the most extensively studied for adenoma detection. A total of six randomized controlled trials have been published, of which five did not demonstrate an increased adenoma detection rate with NBI compared with white light endoscopy [19] [20] [21] [22] [23] [24]. Interestingly, NBI does seem to improve the learning curve of trainees and poor adenoma detectors. FICE was used for adenoma detection in four studies, three comparing FICE with standard definition white light and the other with targeted chromoendoscopy. All failed to show an improved adenoma detection rate [25] [26] [27] [28]. To date, no detection studies using i-Scan have been published.

In this issue of Endoscopy, Hoffman et al. present their study comparing the adenoma detection rate of high definition (described as HD+) colonoscopy in combination with i-Scan surface enhancement (SE-mode) vs. standard video colonoscopy [15]. As a secondary aim they assessed the accuracy of HD+ colonoscopy with i-Scan v-mode (vessel architecture) and p-mode (pattern architecture) enhancement, which are analogous to NBI or FICE, for the differentiation of colorectal lesions using Kudo pit patterns. In their study, at a single academic center with considerable experience of advanced imaging, 220 patients at intermediate and high risk for neoplasia were randomized to either an HD+ colonoscopy with i-Scan surface enhancement or standard video colonoscopy. Indications for colonoscopy were screening, postpolypectomy surveillance or a positive fecal occult blood test. In this study HD+ colonoscopy with i-Scan detected significantly more patients with at least one neoplasia compared with standard resolution endoscopy (38 % vs. 13 %; P < 0.001). Furthermore, polyp histology was predicted accurately with i-Scan with a sensitivity of 98 % and a specificity of 100 % (95 % confidence interval: sensitivity 90.6 – 99.6; specificity 92.8 – 100. Our calculation as per STARD guidelines [29]); however i-Scan v- and p-modes were not compared with standard or high definition white light characterization.

Whereas the polyp differentiation characteristics of i-Scan are within the expected range of other narrowed spectrum technologies [30] [31], the increased neoplasia detection rate is a remarkably positive result in the light of the previously mentioned published data on high definition colonoscopy and digital chromoendoscopy. A three-fold increase in adenoma detection rate (192 % relative increase, [Table 1]) for a new imaging technique is unprecedented and deserves further scrutiny.

The most important issue raised in the interpretation of this study is the surprisingly low neoplasia detection rate of 13 % in the reference group, who underwent standard colonoscopy, coupled with a relatively large mean adenoma size of 6.7 mm. Adenoma detection rates in populations undergoing colonoscopy at academic centers are generally higher than those reported here, with other published studies looking at high definition having a standard definition detection rate ranging from 20 % to 60 % ([Table 1]), and mean adenoma sizes smaller, with the majority of adenomas being less than 5 mm in size. What is the reason for this low detection rate in the reference group? To gain more insight we should focus on data on the baseline characteristics of patients and quality of colonoscopies in both groups.

Table 1 Studies comparing high definition with standard definition colonoscopy for adenoma detection. Study Design N Patients with ≥ 1 adenomas detected, n/N (%) P Absolute increase, % † Relative increase, % ‡       High definition Standard definition East et al. 2008 10 Cohort 130 41/58 (71) 43/72 (60)  0.20 11 18 Pellise et al. 2008 11* Randomized 620 82/310 (26) 79/310 (25)  0.85 1 4 Burke et al. 2010 12* Cohort 852 105/426 (25) 93/426 (22)  0.36 2.8 13 Tribonias et al. 2009 13* Randomized 390 111/193 (58) 99/197 (50)  0.16 8 16 Buchner et al. 2010 14* Cohort 2430 235/823(29) 278/1188 (23)  0.01 4.2 17 Hoffman et al. 2010 15 (current issue) Randomized 220 38/100 (38) 13/100 (13) < 0.001 25 192 * High definition scopes in these studies were also wide angle instruments (170° field of view). †Absolute increase: proportion of patients with at least one adenoma in high definition group – proportion of patients with at least one adenoma in standard definition group. ‡Relative increase: absolute increase/proportion of patients with at least one adenoma in standard definition group × 100.

The authors note correctly that there was a statistical trend (P = 0.08) towards imbalance between the baseline characteristics between the two groups, with the standard definition group having more screening (with the lowest risk for adenomas) cases. Patients with insufficient bowel preparation were excluded from the study, but it is unclear whether the quality of the preparation in the studied patients was equal. Variations in bowel preparation, even in what might be considered acceptable preparation, appear to effect adenoma detection rates [6]. Another important quality parameter is the withdrawal time of the colonoscope, with a linear increase in adenoma detection reported with increasing withdrawal time [7]. The average withdrawal time was much longer than the minimum of 6 minutes in both groups, but 2 minutes longer in the high definition group (17 % relative increase), perhaps suggesting that colonoscopists were unconsciously trying harder or looking more meticulously, the so called “new tool” effect. This is not entirely surprising given the much improved image quality presented, but this may lead to unintentional bias.

The experience, adenoma detection rate, and dedication of the endoscopist could also cause an important bias in this study. Adenoma detection rates between even experienced operators vary 10-fold [7]. Ideally, colonoscopists should be randomized and equally distributed between the groups to prevent unequal distribution of highly experienced and relatively inexperienced colonoscopists between the two study arms. Data on the distribution of the six endoscopists are not presented; however the authors do note a statistical trend towards a difference in detection rates of neoplastic and non-neoplastic lesions among endoscopists (P < 0.063). In this study it also remains unclear whether it was the high definition (HD+) or the i-Scan SE-mode that caused the improvement in neoplasia detection rates because inspection on withdrawal was always performed with the SE-mode of i-Scan activated, and this was not used with the standard definition examinations. This is in some ways analogous to the combination of high definition and wide angle colonoscopy vs. standard colonoscopy in other previously published high definition studies ([Table 1]), making it very difficult to dissect out the true additive value of high definition alone. It therefore is not possible to tell from the current study how much high definition alone contributed to the striking increase in adenoma detection reported, and how much can be explained by patient, endoscopist, and other endoscope factors (e. g. surface enhancement). It is also impossible to assess the impact of the fact that the current study compares an old generation of instruments, light sources, and processors with an entirely new generation system. There may be numerous other unmeasured differences between the two generations of systems, including scope handling, image brightness, and stability, before adding high definition and i-Scan; hence this trial might instead be considered a comparison between the ”old” and ”new” generation systems rather than high versus standard definition. This is an observation that could apply, to some degree, to most published high definition studies.

In conclusion, HD+ with i-Scan in SE-mode seems to lead to an adenoma detection rate comparable to other high definition studies, and this appears to exceed that of standard definition white light at this academic center. However important caveats remain, particularly regarding the magnitude of effect seen, which appears inconsistent with the current body of data on high definition endoscopy. We look forward to multicenter, randomized studies to provide improved clarity on this issue. Manufacturers are moving strongly towards high definition systems with digital chromoendoscopy and therefore endoscopists may lose the ability to choose to purchase standard definition. Nevertheless, we need to know whether or not the extra time needed to comprehend all the fine detail in a high resolution image improves clinical outcomes; in short, we still don’t know whether high definition is really worth it.

Competing interests: E. Dekker has received a research grant and equipment on loan from Olympus and equipment on loan from Pentax to conduct research. J. East has conducted research with endoscopic equipment on loan from Olympus within the past 3 years.

References

• 1 Winawer S J, Zauber A G, Ho M N. et al . Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup.  N Engl J Med. 1993;  329 1977-1981
• 2 Atkin W S, Edwards R, Kralj-Hans I. et al . Once-only flexible sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial.  Lancet. 2010;  375 1624-1633
• 3 Baxter N N, Goldwasser M A, Paszat L F. et al . Association of colonoscopy and death from colorectal cancer.  Ann Intern Med. 2009;  150 1-8
• 4 Brenner H, Hoffmeister M, Arndt V. et al . Protection from right- and left-sided colorectal neoplasms after colonoscopy: population-based study.  J Natl Cancer Inst. 2010;  102 89-95
• 5 van Rijn J C, Reitsma J B, Stoker J. et al . Polyp miss rate determined by tandem colonoscopy: a systematic review.  Am J Gastroenterol. 2006;  101 343-350
• 6 Lai E J, Calderwood A H, Doros G. et al . The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research.  Gastrointest Endosc. 2009;  69 620-625
• 7 Barclay R L, Vicari J J, Doughty A S. et al . Colonoscopic withdrawal times and adenoma detection during screening colonoscopy.  N Engl J Med. 2006;  355 2533-2541
• 8 Kaminski M F, Regula J, Kraszewska E. et al . Quality indicators for colonoscopy and the risk of interval cancer.  N Engl J Med. 2010;  362 1795-1803
• 9 Shaukat A, Oancea C, Bond J H. et al . Variation in detection of adenomas and polyps by colonoscopy and change over time with a performance improvement program.  Clin Gastroenterol Hepatol. 2009;  7 1335-1340
• 10 East J E, Stavrindis M, Thomas-Gibson S. et al . A comparative study of standard vs. high definition colonoscopy for adenoma and hyperplastic polyp detection with optimized withdrawal technique.  Aliment Pharmacol Ther. 2008;  28 768-776
• 11 Pellisé M, Fernández-Esparrach G, Cárdenas A. et al . Impact of wide-angle, high-definition endoscopy in the diagnosis of colorectal neoplasia: a randomized controlled trial.  Gastroenterology. 2008;  135 1062-1068
• 12 Burke C A, Choure A G, Sanaka M R, Lopez R. A comparison of high-definition versus conventional colonoscopes for polyp detection.  Dig Dis Sci. 2010;  55 1716-1720
• 13 Tribonias G, Theodoropoulou A, Konstantinidis K. et al . Comparison of standard versus high-definition, wide-angle colonoscopy for polyp detection: a randomized controlled trial.  Colorectal Dis. 2009;  ohne Seitenzahl DOI: 10.1111/j.1463-1318.2009.02145.x
• 14 Buchner A M, Shahid M W, Heckman M G. et al . High-definition colonoscopy detects colorectal polyps at a higher rate than standard white-light colonoscopy.  Clin Gastroenterol Hepatol. 2010;  8 364-370
• 15 Hoffman A, Sar F, Goetz M. et al . High definition colonoscopy combined with i-Scan is superior in the detection of colorectal neoplasias compared with standard video colonoscopy: a prospective randomized controlled trial.  Endoscopy. 2010;  10 827-833
• 16 Brown S R, Baraza W, Hurlstone P. Chromoscopy versus conventional endoscopy for the detection of polyps in the colon and rectum.  Cochrane Database Syst Rev. 2007;  (4) CD006439
• 17 Kahi C J, Anderson J C, Waxman I. et al . High-definition chromocolonoscopy vs. high-definition white light colonoscopy for average-risk colorectal cancer screening.  Am J Gastroenterol. 2010;  105 1301-1307
• 18 Lecomte T, Cellier C, Meatchi T. et al . Chromoendoscopic colonoscopy for detecting preneoplastic lesions in hereditary nonpolyposis colorectal cancer syndrome.  Clin Gastroenterol Hepatol. 2005;  3 897-902
• 19 Adler A, Pohl H, Papanikolaou I S. et al . A prospective randomised study on narrow-band imaging versus conventional colonoscopy for adenoma detection: does narrow-band imaging induce a learning effect?.  Gut. 2008;  57 59-64
• 20 Rex D K, Helbig C C. High yields of small and flat adenomas with high-definition colonoscopes using either white light or narrow band imaging.  Gastroenterology. 2007;  133 42-47
• 21 Inoue T, Murano M, Murano N. et al . Comparative study of conventional colonoscopy and pan-colonic narrow-band imaging system in the detection of neoplastic colonic polyps: a randomized, controlled trial.  J Gastroenterol. 2008;  43 45-50
• 22 Adler A, Aschenbeck J, Yenerim T. et al . Narrow-band versus white-light high definition television endoscopic imaging for screening colonoscopy: a prospective randomized trial.  Gastroenterology. 2009;  136 410-416
• 23 Kaltenbach T, Friedland S, Soetikno R. A randomised tandem colonoscopy trial of narrow band imaging versus white light examination to compare neoplasia miss rates.  Gut. 2008;  57 1406-1412
• 24 Paggi S, Radaelli F, Amato A. et al . The impact of narrow band imaging in screening colonoscopy: a randomized controlled trial.  Clin Gastroenterol Hepatol. 2009;  7 1049-1054
• 25 Pohl J, Lotterer E, Balzer C. et al . Computed virtual chromoendoscopy versus standard colonoscopy with targeted indigocarmine chromoscopy: a randomised multicentre trial.  Gut. 2009;  58 73-78
• 26 Cha J M, Lee J I, Joo K R. et al . A prospective randomized study on computed virtual chromoendoscopy versus conventional colonoscopy for the detection of small colorectal adenomas.  Dig Dis Sci. 2010;  55 2357-2364
• 27 Aminalai A, Rösch T, Aschenbeck J. et al . Live image processing does not increase adenoma detection rate during colonoscopy: a randomized comparison between FICE and conventional imaging (Berlin Colonoscopy Project 5, BECOP-5).  Am J Gastroenterol. 2010;  ohne Seitenangabe DOI: 10.1038/ajg.2010.273-0
• 28 Chung S J, Kim D, Song J H. et al . Efficacy of computed virtual chromoendoscopy on colorectal cancer screening: a prospective, randomized, back-to-back trial of Fuji Intelligent Color Enhancement versus conventional colonoscopy to compare adenoma miss rates.  Gastrointest Endosc. 2010;  72 136-142
• 29 Bossuyt P M, Reitsma J B, Bruns D E. et al . Standards for reporting of diagnostic accuracy. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative.  BMJ. 2003;  326 41-44
• 30 van den Broek F J, Reitsma J B, Curvers W L. et al . Systematic review of narrow-band imaging for the detection and differentiation of neoplastic and nonneoplastic lesions in the colon (with videos).  Gastrointest Endosc. 2009;  69 124-135
• 31 East J E, Tan E K, Bergman J J. et al . Meta-analysis: narrow band imaging for lesion characterization in the colon, oesophagus, duodenal ampulla and lung.  Aliment Pharmacol Ther. 2008;  28 854-867