METHODS

Total colonoscopy was performed on 260 consecutive patients using the Olympus C240 colonoscope (Olympus Optical Co., Tokyo, Japan) from June 2002 to May 2003 by two experienced endoscopists at the Royal Hallamshire and Northern General Hospitals Sheffield, UK. Full ethics approval for the study was granted from the South Sheffield Research Ethics Committee. Signed informed consent was obtained prior to all procedures. Caecal intubation was verified by identification of the appendix orifice, triradiate caecal fold, ileocaecal valve, and terminal ileal intubation with small bowel biopsy. Patients with known familial adenomatous polyposis, inflammatory bowel disease, known colorectal cancer, or inadequate bowel preparation did not enter the study. Patients were randomly allocated to targeted (control group) or pan-colonic IC (0.5%) chromoscopy.

Bowel preparation comprised 2–4 litres of hypertonic polyethylene glycol solution (Kleanprep) 24 hours prior to the procedure. Quality of bowel preparation was graded by the endoscopist into one of three groups¹⁴:

1. good (near 100% mucosal visualisation following aspiration or any liquid residue);

2. fair (greater than 90% mucosal visualisation);

3. poor (less than 90% mucosal visualisation).

Patients graded as “poor” preparation on these criteria were excluded from the study. Patients requesting conscious sedation received intravenous midazolam (1–10 mg) prior to intubation.

Colonoscopic technique

For patients randomised to the pan-chromoscopy group, a chromoscopy diffusion catheter (Olympus PW-5L-1) was used to apply a maximum of 150 ml of 0.5% IC to the entire mucosa during withdrawal from the caecum. Chromoscopy was performed in 10 cm segments which were then reinspected following aspiration of any excess IC forming pools in the colon. Patients randomised to the control group received conventional colonoscopy with particular attention to subtle mucosal signs such as the presence of erythema, focal pallor, haemorrhagic spots, absence of vascular network pattern, discrete mucosal unevenness, or deformity.⁵ In this group, patients underwent segmental examination of the colon as in the pan-chromoscopy group but using a controlled volume of normal saline solution to a maximum of 150 ml applied through an identical diffusion catheter. Selective chromoscopy (2–5 ml of IC flushed down the side port of the colonoscope followed by a 20 ml air “push”) was permitted to further delineate suspected mucosal lesions which had been identified by observing the mucosal signs (described above) without the prior use of chromoscopy.

A minimum diagnostic extubation time from the caecum to the anus (excluding that required for interventional procedures) was set at eight minutes for all patients entering the study. Total diagnostic extubation times were recorded for all procedures.

Endoscopic classification and resection of lesions

Pedunculated lesions were defined as those with a distinct pedicle and sessile as raised lesions with no evidence of a stalk or pedicle where the diameter did not exceed twice the height.¹⁶ Flat lesions were defined as mucosal change with a flat or slightly rounded surface with a height of less than half the diameter of the lesion.¹⁷ The diameter of each lesion was estimated using a standard fully open biopsy forcep (4 mm) with the height estimated by placing the closed forcep tip adjacent to the lesion (2.1 mm) (Bard Inc., USA). Morphology, size, and anatomical location of all lesions were documented.

Sessile and flat lesions less than 20 mm in diameter were resected using endoscopic mucosal resection (EMR) with the saline snare assisted technique described by Karita and colleagues.¹⁸ A barbed snare (Olympus Keymed UK N110) was used for all resections with an “endo-coag mode” 25 W ERBE diathermy (ERBE Medical UK-ICC200). Lesions which failed to lift were cold biopsied only and a submucosal Indian ink tattoo applied adjacent to the lesion for future localisation. Lesions showing a depressed morphology larger than 10 mm diameter were also cold biopsied only as they demonstrate a high incidence of invasive neoplasia.^9,19 Lesions in excess of 20 mm diameter (without depression) were resected by piecemeal EMR or cold biopsies if anatomical problems made endoscopic resection impossible. Pedunculated polyps were removed by simple snare polypectomy. Lesions smaller than 4 mm diameter were treated by hot biopsy.

Histopathological analysis

Two experienced designated pathologists examined all retrieved tissue. Specimens were immediately fixed in 10% buffered formalin solution and stained with haematoxylin-eosin. Dysplasia was defined according to the Vienna criteria as either low grade (LGD) or high grade (HGD).²⁰ Neoplastic cellular proliferation extending into the deep submucosal layer 3 or beyond to the muscularis propria was classed as invasive cancer.²⁰ Lesions were classified as early CRC when intra- or submucosal carcinoma was present only, with no vertical extension into the muscularis propria.²⁰

Statistical analysis

A sample size calculation was based on the analysis by Brooker et al.¹⁴ We assumed from historical data that the mean number of adenomas per patient in the control group would be 0.36 (SD 0.3). A clinically significant increase in adenoma detection using pan-chromoscopy was set at 30%. Using these parameters, we calculated that a minimum of 118 patients in both the control and chromoscopy groups were required to gain significance (×level 80%). Statistical differences were analysed by χ² tests of independence, the Mann-Whitney U test, and Fisher’s exact test. A p value of less than 0.05 was considered statistically significant. All calculations were made using the SPSS statistical package for Windows 1998 (Microsoft Corporation USA). The University of Sheffield Statistics Unit provided statistical advice and support.

RESULTS

A total of 260 patients were recruited to the study. Patient demographics and indications for colonoscopy are shown in tables 1 and 2, respectively. The pan-chromoscopy group included 128 patients, of whom 58 (45%) were male (median age 58 years (range 22–92)). The control group included 132 patients (61 males (46%); median age 53 years (range 23–89)). There were no statistically significant differences between the pan-chromoscopy and control groups with respect to sex, age, indications for colonoscopy, or quality of bowel preparation. Twenty three cases were excluded prior to randomisation due to inadequate bowel preparation (n = 14), incomplete colonoscopy (n = 5), obstructing tumour mass (n = 2), and procedural diagnosis of severe colitis (n = 2). There were no immediate (<24 hours) or delayed (>24 hours) complications.

Lesion demographics

A total of 289 lesions were diagnosed in 138 patients. The median number of lesions detected was 0 (range 0–24). Tissue was retrieved for histological analysis from 288 (99%) lesions. One hundred and three lesions (36%) were diagnosed in the control group compared with 185 (64%) lesions in the pan-chromoscopy group. Hyperplastic histology was reported in 117 (41%) lesions with 45 (38%) and 72 (62%) being present in the control and pan-chromoscopy groups, respectively. Fifty nine hyperplastic lesions (51%) were flat with 58 (49%) being protuberant. Neoplastic lesions accounted for the remaining 170 (59%) lesions where 57 (33%) and 112 (66%) were present in the control and pan-chromoscopy groups, respectively. Of the neoplastic lesions identified, 168 (98%) were adenomas. Seventy six (45%) of all adenomas were flat with 92 (55%) being protuberant. One flat and one protuberant lesion had evidence of carcinoma in situ. No invasive carcinomas were identified in this cohort. Anatomical location, morphology, and histological analysis of the lesions are summarised in table 3.

The number of hyperplastic polyps detected in the pan-chromoscopy group was significantly higher than that in the control group (p<0.001). There was also a significant difference between the groups regarding the total number of adenomas detected (p<0.05). A significantly higher number of diminutive (<4 mm) and flat adenomas were detected in the pan-chromoscopy group compared with controls (p<0.01). Importantly, eight flat lesions (mean diameter 6.2 mm; range 5–8 mm) in the pan-chromoscopy group had HGD histologically. Three diminutive adenomas in this group also had HGD with one being predominantly villous. The remaining diminutive lesions all had LGD appearances histologically. These results are summarised in table 4.

Analysis of lesions according to anatomical site showed that significantly more hyperplastic polyps were detected in the left colon (86% within the rectosigmoid) using pan-chromoscopy compared with controls (p<0.04). Pan-chromoscopy also diagnosed more right hemicolonic diminutive and flat lesions compared with the control group (p<0.05). More patients with multiple adenomas (>3) were detected using pan-chromoscopy (n = 13) compared with controls (n = 4) (p<0.01).

Extubation times did not differ significantly between the control and pan-chromoscopy groups (median 15 minutes (range 8–41) v 17 minutes (8–39), respectively) (p>0.1). Median volume of IC used in the pan-chromoscopy group was 68 ml (range 65–90). Within the control group, a median of 69 ml of normal saline was used (range 60–93). No significant difference was observed in the volume of IC and normal saline used between the pan-chromoscopy and control groups.

DISCUSSION

Following the initial description of colonic chromoscopy by Tada and colleagues,²¹ the technique has been applied to many aspects of colonoscopy. Many authors report that chromoscopy is helpful for the detection and detailed morphological assessment of flat and depressed colorectal lesions.^{9,10,12,19,22,23} Kudo et al refined the technique further, combining colonic chromoscopy with high magnification colonoscopy for the in vivo prediction of histology (optical biopsy technique).²⁴ Saitoh and other groups have used both colonic chromoscopy and high magnification colonoscopy as an in vivo staging tool for the assessment of flat and depressed colorectal neoplasia.^25–27 Kiesslich et al showed that pan-colonic chromoscopy using methylene blue also increased the detection rates of intraepithelial neoplasia in colitis surveillance threefold,²⁸ data validated by our group.²⁹ However, only one randomised controlled trial has assessed detection rates of colorectal lesions using pan-colonic IC chromoscopy.¹⁴ This study did not control for diagnostic extubation time between the pan-chromoscopy and targeted chromoscopy control group and also did not control for the lavage effect encountered when using chromoscopy.¹⁴ Although showing a significant increase in diminutive adenoma detection,¹⁴ these data were unable to clarify the potential benefits of chromoscopy in routine clinical practice.

In our study, pan-colonic chromoscopy did significantly improve the total number of adenomatous lesions detected and significantly improved detection rates of diminutive and flat adenomas. Despite their small size, 3% of diminutive lesions in Brooker’s cohort had moderate dysplastic histology with 2% having villous architecture.¹⁴ Our data showed a similar prevalence, with three HGD diminutive adenomas diagnosed with one villous adenoma. All are therefore significant lesions at high risk of progression to carcinoma. The three diminutive HGD adenomas were present in the right colon where pan-chromoscopy in our series diagnosed more right hemicolonic diminutive and flat lesions compared with controls. Also, the number of hyperplastic lesions detected by pan-chromoscopy was significantly higher than the control group, with 86% being present within the rectosigmoid. These data confirm those observed by Brooker and colleagues.¹⁴

Importantly, pan-chromoscopy increased detection of patients with multiple adenomas (>3) who, according to the data of Atkin and colleagues³⁰ and Winawer and colleagues,³¹ are a group at high risk of developing advanced neoplasia at follow up. We additionally showed that significantly more flat and diminutive adenomas were present in the right colon compared with controls, which is consistent with Woolfson³² and Hofstad’s³³ reporting of a proximal shift in adenoma distribution. Brooker’s data also confirm this observation.¹⁴

There are limitations to our study. Two experienced endoscopists who have trained in Japanese colonoscopic diagnostic techniques performed all examinations. In this scenario, detection rates of lesions within the control group may have been increased. As we used endoscopists with additional training, we are unable to hypothesise if pan-chromoscopy would improve overall adenoma detection rates compared with “conventionally” trained UK endoscopists. Alternatively, this approach ensured that the detection rates within the control and chromoscopy groups were not a result of the endoscopist’s experience. However, we would suggest that although colonic chromoscopy is important as an adjunctive diagnostic tool, further training and education would be helpful for many practitioners. Indeed, should the clinical evidence for chromoscopy and magnification “zoom” endoscopy become apparent in long term follow up studies, issues of additional training will need to be addressed within the UK.

The increase in detection rate of lesions using pan-chromoscopy, although improving the yield of clinically relevant lesions, also results in an increased number of hyperplastic and LGD adenomatous lesions being diagnosed.¹⁴ The risk of progression to carcinoma in both of these groups requires clarification with long term follow up studies. High magnification chromoscopic colonoscopy has been used as an in vivo optical biopsy tool to distinguish non-neoplastic, neoplastic, and invasive characteristics in polyps but with variability in reported sensitivity and specificity.^24,34,35 However, further prospective studies using this technology are required as it may offer a useful tool to the endoscopist when deciding “on table” management policy such as biopsy, polypectomy, or EMR.³⁶ It may therefore be possible to lower the number of biopsies taken from non-neoplastic or low risk lesions while allowing selective resection of lesions considered clinically significant. Such a strategy may reduce the number of resection associated complications³⁷ and reduce the secondary burden to histopathology services.

In conclusion, disruption of the adenoma-carcinoma sequence by endoscopic polypectomy has been shown to reduce the incidence of CRC.² Despite reduction rates of 76–90% compared with reference populations in the National Polyp Study, five interval cancers were diagnosed.² The Veterans Affairs study of 32 702 patients demonstrated a lower cancer preventative rate of 50% although not all cases received total colonoscopy.³⁸ Evidence suggests that between 15% and 27% of lesions remain undiagnosed at colonoscopy, even by experienced practitioners.^39,40 Many reasons can account for these findings. Inadequate bowel preparation may result in suboptimal mucosal views, with flat and diminutive lesions being reported by many authors as difficult to detect with conventional colonoscopic techniques.^9,10,19 Traditional colonoscopic training within the UK has also focused on the diagnosis of exophytic or polypoid lesions with less emphasis given to flat and diminutive colorectal lesions. However, data from the UK now suggest that such lesions have a significant prevalence and malignant potential in certain morphological subgroups.^9,12 It is therefore possible that some cancers may develop from diminutive and flat colorectal lesions where failure to diagnose and treat ultimately leads to invasive neoplasia. Colonic chromoscopy may improve our detection of such lesions. In routine practice, chromoscopy may benefit patients, assuming a high risk of CRC, and may be a helpful tool used at index colonoscopy when applied to risk stratification and planning follow up colonoscopy intervals.