High Resolution Colonoscopy With Chromoscopy Versus Standard Colonoscopy for the Detection of Colonic Neoplasia: A Randomized Study

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2006;4: High Resolution Colonoscopy With Chromoscopy Versus Standard Colonoscopy for the Detection of Colonic Neoplasia: A Randomized Study MARC LE RHUN,* EMMANUEL CORON,* DAVID PARLIER, JEAN MICHEL NGUYEN, JEAN MARC CANARD, AFCHINE ALAMDARI,* DENIS SAUTEREAU, STANISLAS CHAUSSADE, and JEAN PAUL GALMICHE* *Digestive Diseases Institute, University Hospital Hôtel-Dieu, Nantes; Department of Gastroenterology, Cochin Hospital, Paris; Department of Biostatistics, University Hospital, Nantes; Service Department of Gastroenterology, Croix-Rouge, Paris; and Department of Gastroenterology, University Hospital, Limoges, France Background & Aims: High-resolution colonoscopy with chromoscopy (HRC) is a technique designed to improve the detection of colonic neoplasias. We prospectively compared standard colonoscopy (SC) and HRC in a randomized multicenter trial. Methods: Patients (n 203; age, years; sex ratio, 1) were recruited according to the following criteria: (1) a history of either familial or personal colonic neoplasia or (2) alarm symptoms after the age of 60 years. After randomization, an SC was performed in 100 patients (resolution, <410,000 pixels) and a HRC in 103 patients (Fujinon EC485ZW, 850,000 pixels). In the HRC group, each colonic segment was examined before and after spraying with indigo carmine 0.4%. Results: Two hundred seventy-six polyps were detected in 198 patients. One hundred sixty of them were hyperplastic polyps, 116 were adenomas, and 2 were carcinomas. The numbers of hyperplastic polyps and purely flat adenomas were significantly higher in the HRC group than in the SC group ( vs and vs , respectively; P.01 and P.04), but there was no significant difference in the total number of adenomas per patient (primary end point) detected between the HRC and the SC groups ( vs , respectively). Conclusions: Although HRC improves detection of purely flat adenomas and hyperplastic polyps, the overall detection of colonic adenomas in a population at increased risk of neoplasia is not significantly improved. These findings do not support the routine use of HRC in clinical practice. Colonoscopy is the only investigation that allows detection and resection of colorectal neoplastic lesions 1,2 and therefore can prevent the further development of colorectal cancer. 3 5 However, recent publications have suggested that 15% 27% of small adenomas might be missed during standard colonoscopy (SC). 6,7 Furthermore, in the National Polyp Study, 5 cancers were diagnosed in 1418 patients who had previously undergone a colonoscopy; among these cancers, 3 measured more than 15 mm and were found within 3 years of the first colonoscopy. 4 These results could be interpreted as a lack of sensitivity of SC to detect diminutive or flat adenomas because these lesions generally appear as subtle mucosal irregularities or minor changes in mucosal color or loss of vascular pattern. Moreover, recent studies in Europe and the US showed that flat lesions were found with a prevalence similar to that reported in Japan when investigated by Japanese endoscopists and/or Japanese techniques. For instance, in the United Kingdom, Rembacken et al 8 showed that flat lesions accounted for up to 36% of adenomas in a series of 1000 colonoscopies. These results have been confirmed recently by Hurlstone et al, 9 who reported that 45% of the 168 adenomas detected in 260 patients were flat, and also by various researchers in countries such as Sweden, 10 Germany, 11 and the USA. 12 Finally, malignant depressed lesions were associated with a higher risk of submucosal invasion. 13 Chromoscopy with various dyes has been developed for many years to enhance the detection of colorectal tumors. Recent uncontrolled studies have suggested that indigo carmine staining of the distal colon is able to increase the detection of adenomatous lesions that were otherwise overlooked by routine endoscopy. 11,14 In addition, highresolution colonoscopy coupled with chromoscopy (HRC) might further improve the detection of these non-polypoid or flat lesions. 15 To date, only 2 randomized controlled trials have compared chromoscopy with Abbreviations used in this paper: HRC, high-resolution colonoscopy with chromoscopy; SC, standard colonoscopy by the American Gastroenterological Association Institute /06/$32.00 doi: /j.cgh

2 350 LE RHUN ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 3 SC in terms of the detection of colorectal adenomas. 16,17 A significant improvement in diagnostic yield was reported in 1 of these 2 studies. 16,17 It is of note that these studies involved a small number of very experienced endoscopists and might therefore not reflect the true diagnostic yield of colonic chromoscopy when performed in conditions of routine practice. Therefore, the aim of our randomized controlled study was to compare highresolution colonoscopy coupled with pancolonic indigo carmine chromoscopy (ie, HRC to SC) in a French population at increased risk of colonic neoplasia. In contrast to previous studies, the trial was conducted in a clinical setting more representative of real life. Methods General Study Design Between February 2002 and September 2004, patients referred to 4 French centers (Nantes, Paris Cochin, Paris Croix Rouge, Limoges) for colonoscopy were considered for enrollment into the study. The inclusion criteria were as follows: (1) surveillance colonoscopy of known colorectal adenomas or cancer or screening colonoscopy in patients with a first-degree relative with colon cancer or adenomatous polyps diagnosed at age 50 years and (2) diagnostic colonoscopy in patients older than 60 years of age presenting with symptoms (iron deficiency anemia, change in bowel habit, abdominal pain). In addition, patients with 1 or more of the following criteria did not enter the study: symptoms occurring before 60 years of age, digestive bleeding, symptoms suggestive of digestive obstruction, known inflammatory bowel disease, severe weight loss or ongoing organic disease, pregnancy or breastfeeding, recent inclusion in another randomized study, and refusal to give informed consent. Before randomization, patients were informed that if they were allocated the SC procedure, they could also choose to undergo a complementary HRC in addition to the previous SC (during the same anesthesia). All the patients randomized were allocated either to group A (HRC) or to group B (SC). No stratification was done according to the indication of colonoscopy (screening or surveillance or alarm symptoms). According to the International Conference on Harmonisation guideline on Statistical Principles for Clinical Trials, the randomization procedure was organized centrally, and the center was a stratification factor. The random allocation sequence was generated by using computer-generated random numbers. Blocking with randomly varying groups of 6 8 was used to restrict randomization within the strata (center). Each center received a package of sealed envelopes sequentially numbered, containing the intervention assigned. Envelopes were opened immediately before the endoscopic procedure. At the end of the trial, all envelopes not used were returned to the principal investigator for control. The protocol was approved by an Ethical Committee (CCPPRB no. 2 Pays de Loire). All patients agreed to give written informed consent. Colonoscopic Technique All patients underwent a 3-day dietary restriction and a bowel preparation consisting of at least 4 L of polyethylene glycol solution during the 24 hours before the procedure. The quality of the preparation was graded by the endoscopist as follows: (1) excellent (no solid or liquid residue), (2) good (complete mucosal examination after aspiration), (3) fair (persistence of residue despite aspiration, thus preventing correct examination of 5% 20% of the mucosa), and (4) poor (persistence of residue despite aspiration, thus preventing correct examination of more than 20% of the mucosa). Patients with a poor preparation were excluded from the study. The colonoscopies were performed with the patient under general anesthesia by a total of 12 endoscopists. Seven of the endoscopists had previous experience of more than 3000 colonoscopies, and 5 had more limited practice (between 300 and 3000 previous colonoscopies). The success of cecal catheterization was verified by identification of the usual landmarks (ileocecal valve, triradiate cecal folds, and appendix orifice). In the SC group, the colonoscopy was performed by using a standard resolution videoendoscope ( 410,000 pixels) without magnification or chromoscopy. During extubation, each segment was carefully examined by using a standardized procedure, at first with maximal insufflation to detect color changes and then with minimal insufflation to detect small mucosal surface abnormalities. In the HRC group, colonoscopy was performed with a high-resolution videoendoscope (Fujinon EC485ZW, 850,000 pixels; Fujinon, Saitama, Japan). To differentiate between the properties of high-resolution (HR) and chromoscopy, each segment was analyzed by 2 consecutive examinations, first with high-resolution alone and second with high-resolution at a 1.5-fold magnification coupled with 0.4% indigo carmine dye spray. The polyps detected were removed during each examination. The total extubation time was noted for both SC and HRC, excluding the time required for polyp resection. The morphology, size, and location of each polyp were recorded. Different types of polyps were prospectively defined. The polyp size was estimated by placing full open biopsy forceps next to the lesion before resection. All visible lesions were resected according to the following techniques; lesions 5 mm were resected by snare or hot biopsy. Pedunculated polyps were resected by a simple polypectomy snare. Protrusive nonpedunculated polyps and flat lesions 5 mm were resected by mucosectomy. In both groups of polyps ( or 5 mm in diameter) we defined purely flat polyps as mucosal elevations with a flat or slightly rounded surface and a height of less than half the diameter of the lesion. 17 The immediate outcome after colonoscopy was assessed by a questionnaire completed by the patient between the 7th and the 15th day after endoscopy. This questionnaire was based on a visual analog scale graded from 0 (poor tolerance or side effects) to 10 (excellent tolerance).

3 March 2006 HIGH-RESOLUTION CHROMOSCOPY AND COLONIC NEOPLASIA 351 Table 1. Patient Demographics Table 2. Indications for Colonoscopy Standard colonoscopy High-resolution chromoscopy P value Standard colonoscopy High-resolution chromoscopy P value Number of patients Male sex 53 (53%) 49 (48%).48 Age (y) (median, range) 59 (30 78) 59 (31 80).15 Preparation.24 Excellent 48 (48%) 49 (48%) Good 45 (45%) 38 (37%) Fair 6 (6%) 12 (11%) Poor 1 (1%) 4 (4%) Histopathologic Analysis All resected tissues were examined by the pathologists of each center after fixation in buffered formalin solution and hematoxylin-eosin staining. The Vienna criteria were used to classify dysplasia as either low grade or high grade. Early cancer was defined as intramucosal or submucosal carcinoma without vertical extension into the muscularis propria. Invasive cancer was defined as neoplastic proliferation beyond the muscularis propria. Statistical Analysis To set power calculations, we made a preliminary analysis of a cohort of similar patients recruited in the Nantes center. In this cohort, the mean number of adenomas detected was per SC. The controlled study was powered to establish a 30% significant increase in the rate of adenoma detection with HRC. With these parameters, we calculated that a total of 310 patients in each group were required to achieve significance ( 0.05, 0.2). Because of uncertainty concerning the exact diagnostic yield of HRC and variability concerning the number of polyps detected, an interim analysis was planned after inclusion of 200 patients (100 in each group) according to a procedure by O Brien and Fleming. 18 All calculations were made using the S-PLUS 6.2 software (Insightful Corporation, Seattle, WA). Results The interim analysis was performed after inclusion of 203 consecutive patients recruited from February 2002 September 2004; 103 were randomized to the HRC group (49 men; median age, 59 years; range, years) and 100 to the SC group (53 men; median age, 57 years; range, years). The mean number of colonoscopies per endoscopist was 17 (range, 2 38). Among the 100 patients randomized to the SC group, a subgroup of 36 patients decided to undergo a HRC immediately after the SC. There were no differences between groups with respect to sex, age, quality of bowel preparation (Table 1), or indications for colonoscopy (Table 2). Five cases were excluded because of poor bowel Number of patients Screening or surveillance 88 (88%) 86 (83%) NS Symptoms after 60 y of age 12 (12%) 18 (17%) NS Iron deficiency anemia 2 (2%) 4 (4%) Change in bowel habit 5 (5%) 8 (7%) Abdominal pain 5 (5%) 6 (5%) preparation (4 in the HRC group and 1 in the SC group). The extubation time was significantly longer in the HRC group when compared with the SC group (26 11 vs 8 4 minutes; P.0001). The median volume of indigo carmine solution used was 62.5 ml (range, ml). Tolerance of the colonoscopy was good in both the HRC and the SC groups (visual analog scale, and in the HRC and SC groups, respectively; P.24). There were no immediate ( 24 hours) or delayed ( 24 hours) complications. Lesion Demographics The overall findings are summarized in Table 3.A total of 276 polyps were seen in 198 patients. One hundred seventy two (62%) polyps were found in the HRC group and 104 (38%) in the SC group (P.01). One hundred sixteen polyps were found to be adenomas, Table 3. Summary of Lesions and Patient Numbers in the Standard Colonoscopy and High-Resolution Chromoscopy Groups Number of polyps per patient Number of adenomas per patient Number of adenomas 5 mm per patient Number of purely flat adenomas per patient Number of hyperplastic polyps per patient Patients with at least 1 polyp (%) Patients with at least 1 adenoma (%) Patients with at least 3 adenomas (%) Patients with at least 1 hyperplastic polyp (%) Standard colonoscopy (n 99) High-resolution chromoscopy (n 99) P value NS NS (45) 65 (66) (31) 40 (39) NS 6 (6) 7 (7) NS 26 (26) 48 (48).0001

4 352 LE RHUN ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 3 and 2 were carcinomas (1 in the rectum in the HRC group and 1 in the left colon in the SC group). Discounting histology, HRC found more polyps than SC ( vs polyps per patient, respectively; P.01). However, before chromoscopy, the mean number of polyps detected per patient by HR alone and by SC was nonsignificantly different ( vs ; P.75). The mean size of polyps was also nonsignificantly different between both groups ( vs mm for HRC and SC, respectively; P.07). Histopathologic analysis revealed that HRC detected significantly more hyperplastic polyps per patient than SC ( vs , respectively; P.01). In contrast, there was no difference between the number of adenomas per patient found with HRC and SC ( vs , respectively; P.38). However, the number of purely flat adenomas was significantly higher in the HRC group than in the SC group ( vs , respectively; P.04). HRC revealed more flat adenomas 5 mm in diameter than SC ( vs ; P.09), whereas there was no significant difference between the numbers of flat adenomas 5 mm detected by HRC and SC ( vs , respectively; P.17). There were 2 adenomas with high-grade dysplasia in the SC group versus 0 in the HRC group (not significant). The number of patients with at least 1 adenoma and the number of individuals with multiple ( 3) adenomas were nonsignificantly different between the HRC and the SC groups. Similarly, a significant difference was not noted between the HRC and the SC groups according to the anatomic location of adenomas (ie, proximal versus distal colon; data not shown). The total number of adenomas detected was not significantly affected by the experience of endoscopists (data not shown). In the subgroup of 36 patients who had HRC immediately after SC, the number of adenomas per patient found with SC and HRC were and , respectively. In the 5 patients (14%) with negative SC, HRC found 5 adenomas. Discussion In our study, HRC did not detect significantly more adenomas than SC. In addition to the lack of a statistical difference in the overall detection of colorectal adenomas, no statistically significant difference was detected in patients with multiple ( 3) adenomas, who represent a subgroup of patients at a higher risk of developing colorectal neoplasia, 4,19 or in the detection of diminutive ( 5 mm) adenomas between HRC and SC. In contrast, the number of purely flat adenomas was significantly increased. These results provide a contrast to 2 other recently published trials. 16,17 Brooker et al 16 showed that chromoscopy significantly increased the proportion of patients found to have multiple adenomas and also the detection of diminutive adenomas in the proximal to the sigmoid-descending junction. However, the overall detection rate of adenomas did not differ significantly between the 2 techniques. Hurlstone et al 17 showed a significant increase in the detection of colonic adenomas including diminutive polyps. In comparison, we detected fewer adenomas per patient in our chromoscopy group than Brooker et al and Hurlstone et al. Technical shortcomings are unlikely to explain these differences. In fact, the quality of the bowel preparation was excellent or good in 85% of cases. Both the volume and the concentration of the dye were consistent with other studies; indigo carmine was generally noted to have been used at a concentration of 0.2% 0.5% with volumes up to 150 ml. 8,9,14 Furthermore, the significant increase in the number of purely flat adenomas and hyperplastic polyps detected in the colon with HRC was consistent with other reports, 11 and our number of hyperplastic polyps was even higher than that reported by Brooker et al and Hurlstone et al. In addition, the issue of endoscopists expertise and their experience with chromoendoscopy in particular deserve further consideration. Indeed, a learning curve effect has been reported to occur in colonoscopic training especially for chromoscopic recognition of the pit pattern of intestinal lesions. 9,19 However in a recent study, no statistical difference was found in the mean number of polyps (including flat lesions) detected by using chromoscopy by a Western endoscopist with no previous experience of chromoscopy and an expert Japanese endoscopist ( vs ; P NS). 20 In our series, the number of procedures previously performed by endoscopists did not seem to influence the diagnostic performance. However, the learning curve influence should be addressed more specifically in further studies of chromoscopy. Another hypothesis that might account for the absence of a statistical difference between HRC and SC in the detection of colorectal adenomas could be, at least in part, related to the excellent diagnostic performance of SC in our study. In fact, the numbers of adenomas detected per patient were higher than those reported by Brooker et al 16 and Hurlstone et al 17 in their control groups, in addition to Rembacken et al, 8 who found 321 adenomas in 1000 colonoscopies performed by a single experienced endoscopist. Moreover, the mean size of

5 March 2006 HIGH-RESOLUTION CHROMOSCOPY AND COLONIC NEOPLASIA 353 adenomas detected in SC and HRC groups did not differ (about 4 mm). This size was generally smaller than that reported by Kiesslich et al, 11 who found a mean size of adenomas of 14 mm before indigo carmine spraying of the 30 cm of the distal colon. The quality of bowel preparation as well as the careful examination during extubation in conditions of both insufflation and exsufflation with a mean examination time of 8 minutes might have increased the quality of the SC procedure and consequently the overall detection of adenomas in our control group. Extubation time in the HRC group was almost 3 times as long as in the SC group, correlating with the data of Brooker et al. Some authors have tried to control for extubation time by spraying a saline solution during SC, thus prolonging the duration of SC. In our study our choice was rather to reproduce the conditions of routine use of SC and chromoscopy, therefore allowing for a true head-to-head comparison. Moreover, the failure to detect a higher rate of adenomas with chromoscopy, even though there were longer extubation times, argues in favor that we are not missing a lack of effect of chromoscopy. Regarding the description of polyp morphology, flat adenomas were defined according to previous literature, and we did not take photographs of all lesions, allowing independent interpretation. Therefore, we cannot completely exclude potential bias in interpreting the morphology of adenomas. However, it is not really surprising that the number of non-flat adenomas was similar in both groups, because indigo carmine is a contrast dye that accumulates between pits and valleys and enhances mucosal architecture. Therefore, chromoscopy is not expected to improve the detection of protrusive or large lesions (which can be easily identified without indigo carmine) but rather the diagnosis of small tiny lesions. This assumption is further supported by the fact that HRC revealed more flat adenomas 5 mm in diameter than SC, whereas there was no significant difference between the numbers of flat adenomas 5 mm detected by HRC and SC. We were unable to apply the Paris Endoscopic Classification of Superficial Neoplastic Lesions 21 because this system was published after our study had commenced. Therefore, differentiation between small protruded sessile polyps and flat sessile lesions was a particular difficulty. However, our proportion of purely flat adenomas was similar to that of other recent Western studies, ranging from 36% 45%. 8,10,12 It is important to note that almost 50% of the lesions diagnosed in our HRC group were not detected by high-resolution alone and only after chromoscopy. This aspect was not evaluated in the trials of Brooker et al 16 and Hurlstone et al 17 ; however, our data are consistent with other studies. 11,14 Indeed, Kiesslich et al 11 noted that 165 of 210 hyperplastic polyps were seen after staining of the distal 30 cm. In addition, Lee et al 14 reported that 158 of 174 hyperplastic or inflammatory lesions were seen only after chromoscopy. Finally, statistical considerations regarding previous studies should also be discussed. Brooker et al 16 assumed from historical data that the mean number of adenomas per patient was 0.36 with a standard deviation of 0.3. Therefore, they calculated that 117 patients were required in each group for a 30% increase in adenoma detection with dye spray. It is somewhat surprising that they did not reach the statistical difference (P.06) because they finally reported 49 adenomas in 135 patients in the control group, compared with 125 adenomas in 124 patients in the chromoscopy group, corresponding to 0.36 and 1 adenoma per patient, respectively. These data resulted in a 56% increase instead of the presumed 30%. We calculated that these results should have shown a statistical difference (P.0001) if the initial hypothesis of a standard deviation equal to 0.3 was verified. These negative results could only be explained by a higher than expected standard deviation; however, this information was not available in the publication. Hurlstone et al 17 based their sample calculation on the data of Brooker et al. They detected 57 adenomas in 132 patients in the control group and 111 adenomas in 128 patients in the chromoscopy group, corresponding to 0.43 and 0.87 adenoma per patient, respectively. However, their results differed from those of Brooker et al in that they noted a statistically significant increase in detection with chromoscopy; the standard deviation and therefore the variability were not mentioned in the article. At the time of our trial design, neither of these 2 studies had been published. Therefore, we assumed from a pilot study that the mean number of adenomas would be per patient in the SC group. This number was in fact slightly greater than the one actually observed in our study. Moreover, the interim analysis showed that the increase in adenoma detection in the HRC group was 20%, slightly lower than the expected 30% but with considerable intersubject variability. Although our study might not have been sufficiently powered to ascertain a statistical significance regarding the primary end point (ie, the number of adenomas detected), to establish a 20% advantage with the results of the interim analysis would require a recalculation of the sample with considerably higher recruitment (ie, 1127 patients per group). These data led us to discontinue the study for a number of reasons, namely (1) the

6 354 LE RHUN ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 3 recruitment of 2254 patients seemed difficult to achieve within a reasonable time period, (2) a better classification of lesions is now available (Paris classification), and (3) a period of chromoscopic training should be included in the protocol to assess more accurately the influence of the learning curve. Consideration of these different issues would result in a new trial rather than continuation of the present study design. In conclusion, HRC is a time-consuming procedure that did not significantly increase the overall detection of colorectal adenomas in a moderately high-risk population. Although the number of purely flat adenomas detected was increased by HRC, we do not believe that this finding, in isolation, justifies the recommendation of HRC in routine practice, but rather it highlights the importance of performing SC of excellent technical quality. References 1. Sonnenberg A, Delco F, Inadomi JM. Cost-effectiveness of colonoscopy in screening for colorectal cancer. Ann Intern Med 2000;133: Rex DK, Lieberman DA. Feasibility of colonoscopy screening: discussion of issues and recommendations regarding implementation. Gastrointest Endosc 2001;54: Winawer SJ, Zauber AG. Colonoscopic polypectomy and the incidence of colorectal cancer. Gut 2001;48: Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993;329: Citarda F, Tomaselli G, Capocaccia R, et al. Efficacy in standard clinical practice of colonoscopic polypectomy in reducing colorectal cancer incidence. Gut 2001;48: Hixson LJ, Fennerty MB, Sampliner RE, et al. Prospective study of the frequency and size distribution of polyps missed by colonoscopy. J Natl Cancer Inst 1990;82: Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies. Gastroenterology 1997;112: Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neoplasms: a prospective study of 1000 colonoscopies in the UK. Lancet 2000;355: Hurlstone DP, Cross SS, Adam I, et al. Efficacy of high magnification chromoscopic colonoscopy for the diagnosis of neoplasia in flat and depressed lesions of the colorectum: a prospective analysis. Gut 2004;53: Tsuda S, Veress B, Toth E, et al. Flat and depressed colorectal tumours in a southern Swedish population: a prospective chromoendoscopic and histopathological study. Gut 2002; 51: Kiesslich R, von Bergh M, Hahn M, et al. Chromoendoscopy with indigocarmine improves the detection of adenomatous and nonadenomatous lesions in the colon. Endoscopy 2001; 33: Saitoh Y, Waxman I, West AB, et al. Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population. Gastroenterology 2001;120: Kudo S, Kashida H, Tamura T, et al. Colonoscopic diagnosis and management of nonpolypoid early colorectal cancer. World J Surg 2000;24: Lee JH, Kim JW, Cho YK, et al. Detection of colorectal adenomas by routine chromoendoscopy with indigocarmine. Am J Gastroenterol 2003;98: Bruno MJ. Magnification endoscopy, high resolution endoscopy, and chromoscopy; towards a better optical diagnosis. Gut 2003; 52: Brooker JC, Saunders BP, Shah SG, et al. Total colonic dye-spray increases the detection of diminutive adenomas during routine colonoscopy: a randomized controlled trial. Gastrointest Endosc 2002;56: Hurlstone DP, Cross SS, Slater R, et al. Detecting diminutive colorectal lesions at colonoscopy: a randomised controlled trial of pan-colonic versus targeted chromoscopy. Gut 2004; 53: O Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics 1979;35: Togashi K, Konishi F, Ishizuka T, et al. Efficacy of magnifying endoscopy in the differential diagnosis of neoplastic and nonneoplastic polyps of the large bowel. Dis Colon Rectum 1999; 42: Togashi K, Radford-Smith G, Hewett D, et al. The use of indigo carmine increases the colonoscopic detection rate of flat adenomas and large sessile hyperplastic polyps. Gastrointest Endosc 2004;59:AB The Paris endoscopic classification of superficial neoplastic lesions: esophagus, stomach and colon. Gastrointest Endosc 2003;58(Suppl 6). Address requests for reprints to: Jean-Paul Galmiche, MD, Digestive Diseases Institute, University Hospital Hôtel-Dieu, Nantes, France. jeanpaul.galmiche@chu-nantes.fr; fax: 33 (0) Supported by the DRC Nantes and CIC-INSERM CHU. Marc Le Rhun and Emmanuel Coron contributed equally to this work. The authors thank the anesthesiologists and nurses from the various GI and endoscopy departments for their valuable help during all the trials, and their pathologist colleagues for characterization of neoplastic polyps.