ORIGINAL CONTRIBUTIONS
nature publishing group
855
Detection of Colorectal Adenoma by Narrow Band Imaging (HQ190) vs. High-Definition White Light Colonoscopy: A Randomized Controlled Trial Wai K. Leung, MD, FACG1, Oswens S.H. Lo, MBBS, FRCS2, Kevin S.H. Liu, MBBS, MRCP1, Teresa Tong1, David Y.K. But, MBBS, MRCP1, Frank Y.F. Lam, MBBS, MRCP1, Axel S.J. Hsu, MBBS, MRCP1, S.Y. Wong, MBBS, FRCP1, W.K. Walter Seto, MD, MRCP1, Ivan F.N. Hung, MD, FRCP1 and W.L. Law, MS, FACS2 OBJECTIVES:
The benefits of narrow band imaging (NBI) on enhancing colorectal adenoma detection remain questionable. We tested whether the new generation of NBI (190-NBI), which is twice as bright as the previous version, would improve adenoma detection when compared with high-definition white light (HD-WL) colonoscopy.
METHODS:
It was a randomized controlled trial with tandem colonoscopy. We recruited patients who underwent colonoscopy for symptoms, screening, or surveillance. Patients were randomized for the use of either 190-NBI or HD-WL on withdrawal. Tandem colonoscopy was performed by using the same assigned colonoscope and withdrawal method. Lesions detected on first-pass and second-pass examination were used for adenoma detection and miss rates, respectively. The primary outcomes were adenoma and polyp detection rates.
RESULTS:
A total of 360 patients were randomized to undergo either 190-NBI or HD-WL colonoscopy. Both the adenoma and polyp detection rates were significantly higher in the 190-NBI group compared with the HD-WL group (adenoma: 48.3% vs. 34.4%, P = 0.01; polyps: 61.1% vs. 48.3%, P = 0.02). The mean number of polyps detected per patient was higher in the 190-NBI group (1.49% vs. 1.13, P = 0.07). There was no significant difference in the adenoma miss rates between the two groups (21.8% vs. 21.2%). Multivariate analysis showed that the use of 190-NBI (odds ratio (OR) 1.85; 95% confidence interval (CI) 1.10–3.12), withdrawal time (OR 1.29; CI 1.19–1.38), patient’s age (OR 1.04; CI 1.01–1.06), and male gender (OR 2.38; CI 1.42–3.99) were associated with adenoma detection.
CONCLUSIONS: 190-NBI colonoscopy was superior to the conventional HD-WL in detecting colorectal adenomas or
polyps, but there was no significant difference in adenoma miss rates. Am J Gastroenterol 2014; 109:855–863; doi:10.1038/ajg.2014.83; published online 22 April 2014
INTRODUCTION Colorectal cancer is the third most common cancer in the world and the incidence of colorectal cancer is rising rapidly in many Asian countries. In Hong Kong, colorectal cancer is now the second most common cancer with an age-standardized incidence of 38.1 per 100,00 persons (1). We have previously shown that the prevalence of both advanced colonic neoplasms and serrated lesions in our populations were very comparable with those in Western countries (2,3). Colonoscopy is the usual preferred method for diagnosis and surveillance of colorectal neoplasms. There are also emerg-
ing data to support that screening by colonoscopy reduces both the incidence and mortality of colorectal cancer (4). However, it is well known that colonoscopy could still miss colorectal adenomas and even cancer (5). The miss rate for colonic adenomas was reported to be ranging from 15 to 32% in tandem colonoscopy studies (6). The risk of interval cancer development after a negative colonoscopy was up to 1.7 per 1,000 person-years of follow-up (7), or 1 in 13 of all diagnosed cancers (8). In view of the limitations of colonoscopy examinations, different image-enhanced modules were developed with an aim to improve colorectal polyp or adenoma detection. One of the
1
Department of Medicine, Queen Mary Hospital, University of Hong Kong, Hong Kong, Hong Kong; 2Department of Surgery, Queen Mary Hospital, University of Hong Kong, Hong Kong, Hong Kong. Correspondence: Wai K. Leung, MD, FACG, Department of Medicine, Queen Mary Hospital, University of Hong Kong, Room 413, Professorial Block, 102 Pokfulam Road, Hong Kong, Hong Kong. E-mail:
[email protected] Received 2 December 2013; accepted 18 February 2014 © 2014 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
ENDOSCOPY
see related editorial on page 864
ENDOSCOPY
856
Leung et al.
widely available modules is the narrow band imaging (NBI) developed by the Olympus Company. The NBI system uses an optical filter of blue (440–460 nm) and green (540–560 nm) wavelength, which enhances the visualization of blood vessels and mucosal pit patterns. It has also been applied in the virtual diagnosis of polyp histology to possibly save the cost and burden related to histological interpretation of colorectal polyps (9). However, it remains controversial whether the NBI system actually helps to improve colonic polyp detection. Three recently published meta-analyses failed to demonstrate any superiority of the current NBI system over white light (WL) colonoscopy in detecting colonic polyps. In particular, there may not be any additional benefits in using the NBI colonoscope over the high-definition WL (HD-WL) colonoscopy for colorectal polyp detection (10–12). The previous version of NBI provides dimmer images of the colonic mucosa, which may limit its performance on polyp or adenoma detection (13). The newly available second generation of NBI using the 190 system (190-NBI) provides an at least twofold brighter image compared with the previous version and in full HD. This may possibly increase polyp detection by enhancing visibility of the colonic mucosa with brighter NBI images. In this study, we hypothesize that the 190-NBI system would enhance the detection rate of colorectal adenomas as compared with conventional HD-WL colonoscopy. The primary aim of the current study was to compare the adenoma and polyp detection rates of the 190-NBI colonoscopy with conventional HD-WL colonoscopy in a prospective randomized trial with tandem colonoscopy. We also compared the adenoma and polyp miss rates of the 190-NBI with HD-WL colonoscopy during tandem colonoscopy as secondary outcomes.
METHODS Study design and setting
This was a prospective randomized trial with tandem colonoscopy comparing the 190-NBI with the conventional HD-WL colonoscopy. It was a single-center study conducted in the Integrated Endoscopy Center of the Queen Mary Hospital of Hong Kong, which is a regional hospital serving the Hong Kong West Cluster and a university teaching hospital. Patients
We enrolled consecutive patients who were referred for colonoscopy to our center for diagnostic work up of colonic symptoms, surveillance of colorectal polyps, and colorectal cancer screening. All patients were aged 40 or above. We excluded patients with familial colorectal cancer syndrome including familial adenomatous polyposis and hereditary non-polyposis colorectal cancer syndrome, personal history of colorectal cancer or inflammatory bowel disease and those who had previous colonic resection. Patients who were considered to be unsafe for polypectomy, including patients with bleeding tendency and those with severe comorbid illnesses, were excluded. A written informed consent was obtained from all patients for participation into this The American Journal of GASTROENTEROLOGY
trial. The study protocol was approved by the Hospital Authority Hong Kong West Cluster & Hong Kong University Institutional Review Board. The study was registered with Clinictrials.org (NCT01725321). Randomization
Patients were randomized in a 1:1 ratio in blocks of 10 to undergo either the 190-NBI or the standard HD-WL colonoscopy. Randomization was carried out by computer-generated random sequences and stratified according to the endoscopist’s experiences (experienced vs. fellows) and indications of colonoscopy (symptomatic vs. screening/surveillance). Individual random sequence was placed in an opaque envelope and kept by an independent research assistant who was not involved in this study. Once informed consent was obtained, the research assistant would disclose the assigned colonoscopic imaging technique (190-NBI or HD-WL) to the responsible endoscopist immediately before the procedure. Colonoscopy procedure
All patients were given dietary instructions before colonoscopy and were given three liters of polyethylene glycol in a split dose for bowel cleansing 1 day before colonoscopy. Colonoscopy was performed under conscious sedation with intravenous midazolam and pethidine. Tandem, or back-to-back, colonoscopy was performed in all patients by the same endoscopist using the same assigned method. After the first complete colonic examination with colonoscope withdrawn from the anus, the colonoscope was reinserted to the cecum for the second colonic examination (Figure 1). In the 190-NBI group, the Olympus CF-HQ190 colonoscope (Olympus Optical, Tokyo, Japan) and the EVIS-EXERA III CLV-190 video system (Olympus) were used for both the firstpass and second-pass examinations. The system was attached to a 26-inch full HD monitor (OEV261H; Olympus). Magnification function was not used in this study. Insertion to cecum was performed under WL and once the cecum was reached, the NBI mode was switched on during withdrawal of endoscope for complete colonic examination. Second colonoscopic examination was performed in a similar manner after the first complete withdrawal of the colonoscope. WL was used on insertion and NBI was used on withdrawal. For the HD-WL arm, the Olympus HD-colonoscope (CF-H260) and the EVIS-LUCERA CV-260SL system were used. The same HD monitor was used for the HD-WL colonoscopy examination. In this arm, WL only was used for both insertion and withdrawal of the colonoscope during the tandem examinations. Withdrawal time was defined as the time of the initiation of cecal inspection to the time when the colonoscope was completely removed from anus. The time for polypectomy was not included. A dedicated research assistant measured the withdrawal time by using a stopwatch. The withdrawal time was set to a minimum of 6 min for both the first-pass and the second-pass examinations even in patients in whom no polyps were found (14). Bowel preparation of the whole colon during the first examination was VOLUME 109 | JUNE 2014 www.amjgastro.com
190-NBI for colorectal adenoma
857
569 Patients screened
Age 90% mucosal visualization), and poor ( < 90% mucosal visualization) as described previously (15). The sizes and locations of all colonic polyps were marked contemporaneously. The locations of colonic lesions were identified by anatomical landmarks or by transillumination. The size of colonic lesions was measured against the span of an opened biopsy forceps. All lesions found on first examination were immediately removed and labeled for histological examination before the second examination. The proportion of cases in which polyps were detected during insertion was not recorded. Similarly, polyps © 2014 by the American College of Gastroenterology
found on second examination were characterized and removed. Polyps found on second examination were labeled separately for histological assessment. Procedures were performed by either experienced endoscopists who had performed at least 2,000 colonoscopies or by endoscopy fellows. All fellows who participated in this study had completed their training in endoscopy and had performed more than 500 colonoscopies. All endoscopists had prior experiences with the previous version of NBI. To familiarize them with the 190-NBI, all endoscopists were asked to carry out at least five examinations with the new NBI system before performing study cases. The American Journal of GASTROENTEROLOGY
ENDOSCOPY
209 Excluded due to:
ENDOSCOPY
858
Leung et al.
Definition
RESULTS
Complete colonoscopy was defined as successful cecal intubation. All colonic polyps removed during each examination were sent for histological examination with clear labeling of location and sequences of colonoscopy. Histological interpretation of all colonic lesions followed the World Health Organization system (16). Advanced adenoma was defined as adenoma ≥ 10 mm in diameter, with any villous histology, high-grade dysplasia, or invasive carcinoma (3,17). Adenoma detection rate and polyp detection rate were defined as the proportion of patients with at least one adenoma and one polyp found on first colonoscopic examination, respectively. Adenoma miss rate or polyp miss rate were defined as the proportion of patients with adenomas or polyps found on secondpass colonoscopy.
Patient characteristics
Outcomes
The primary outcomes of this study were to compare the adenoma and polyp detection rates of the 190-NBI with the HD-WL colonoscopy. Secondary outcomes included adenoma miss rate and polyp miss rate between the two groups.
During the 12-month study period, 569 patients were screened. Two hundred and nine patients were excluded for reasons listed in Figure 1. A total of 360 patients were enrolled and randomized to receive either the 190-NBI or HD-WL colonoscopy, with 180 patients in each group. The mean age of the patients was 62.2 (s.d. 11.8) years and 48.3% of them were men. There was no significant difference in the baseline characteristics between the two groups (Table 1). The most common indication for colonoscopy was for diagnostic work up of colonic symptoms (59.4%). Forty-two (11.7%) and 104 (28.9%) patients underwent colonoscopy for screening and surveillance, respectively. Eight endoscopists, including four experienced endoscopists and four endoscopy fellows, performed all the study examinations. No major complications were noted during tandem colonoscopy. There was also no significant difference in the quality of bowel preparation (P = 0.86) and withdrawal time (10.4±4.1 min in new NBI group vs. 9.9±.6.9 min in the HD-WL; P = 0.47) between the two groups (Table 1). Adenoma and polyp detection rates
Sample size estimation
The sample size estimation was based on the assumption that the 190-NBI colonoscopy was superior to the HD-WL colonoscopy for colorectal adenoma and polyp detection. As this study included symptomatic subjects as well as subjects who underwent colonoscopy for surveillance and screening, we anticipated the prevalence of colorectal adenomas or polyps to be higher than that in previous reports (10–12). We estimated the overall prevalence of colorectal adenoma in the HD-WL colonoscopy to be 35%. In order to show a clinically important improvement of adenoma detection by the 190-NBI, we assumed that the new system should increase the adenoma detection rate by ~15%. With a statistical power of 80% and a two-sided significance level of 0.05, 182 patients were needed in each study arm. Statistical analysis
Adenoma detection rate or polyp detection rate were calculated on the actual number of patients who were randomized to each group. Adenoma miss rate or polyp miss rate was based on the percentage of patients who had completed the second colonoscopy. Bowel preparation was regrouped into satisfactory (excellent to good) and unsatisfactory (fair to poor) for statistical analysis. Categorical data were compared by the χ2-test or Fisher Exact test where appropriate. Numerical data were analyzed by the Student’s t-test. Statistical significance was taken as a twosided P value < 0.05. Factors associated with adenoma detection on first colonoscopy were first identified by univariate analysis. Factors with a P value < 0.1 on univariate analysis were further entered into forward stepwise logistic regression analysis. The adjusted odds ratio with 95% confidence interval (CI) was used to describe the influence of various factors on adenoma detection rate. All statistical analysis was performed by SPSS statistics software (version 19.0, SPSS, Chicago, IL). The American Journal of GASTROENTEROLOGY
First-pass colonoscopy was completed in 355 (98.6%) patients. Two (1.1%) patients in the 190-NBI group and three (1.7%) patients in the HD-WL group failed to achieve cecal intubation during the first-pass examination due to obstructing tumors and looping of the endoscope (Figure 1). As shown in Table 2, both the adenoma and polyp detection rates of the 190-NBI colonoscopy were significantly higher than those of the HD-WL colonoscopy. The percentage of patients who were found to have at least one adenoma in the first examination in the 190-NBI and HD-WL groups was 48.3% (95% CI 41.1–55.6%) and 34.4% (95% CI 27.9–41.7%; difference: 13.9%, 95% CI 3.7–23.7%, P = 0.01), respectively. The corresponding proportion of patients who were found to have at least one colorectal polyp was 61.1% (95% CI: 53.8–67.9%) and 48.3% (95% CI: 41.1–55.6%; difference: 12.8%, 95% CI 2.5–22.7%, P = 0.02). However, there was no significant difference in the rate of advanced colorectal neoplasm detection between the two groups (7.8% in the new NBI vs. 8.3% in HD-WL, P = 1.0). The percentage of patients with serrated polyps also tended to be higher in the 190-NBI group than in the HD-WL group (19.4% vs. 12.8%, P = 0.11). The mean number of adenomas and polyps that were detected on first colonoscopy in the 190-NBI group tended to be higher than in the HD-WL group. The mean number of adenomas in the new NBI group was 0.94 (s.d. = 1.37) and in the HD-WL group was 0.76 (s.d. = 1.58, P = 0.23). The corresponding mean numbers of polyps detected on first colonoscopy by the new NBI and HD-WL were 1.49 (s.d. = 1.84) and 1.13 (s.d. = 1.86, P = 0.07), respectively. We further characterized the distribution of adenomas detected by first colonoscopy. As shown in the left panel of Figure 2, the new NBI detects more adenomas than the HD-WL colonoscopy in all segments of the colon. Although the differences VOLUME 109 | JUNE 2014 www.amjgastro.com
190-NBI for colorectal adenoma
Table 2. Colonoscopy findings of 190-NBI colonoscopy and HD-WL colonoscopy
190-NBI colonoscopy (n=180)
HD-WL colonoscopy (n=180)
P
Mean age in years (s.d.)
61.9 (10.8)
62.6 (10.6)
0.53
Male
84 (46.7%)
90 (50%)
0.60
27 (15%)
15 (8.3%)
Bowel symptoms
101 (56.1%)
113 (62.8%)
Polyps surveillance
52 (28.9%)
52 (28.9%)
Incomplete first colonoscopy
2 (1.1%)
3 (1.7%)
Incomplete second colonoscopy
8 (4.4%)
12 (6.7%)
190-NBI colonoscopy
HD-WL colonoscopy
(n=180)
(n=180)
Patients with polyps
110 (61.1%)
87 (48.3%)
0.02
Patients with adenoma
87 (48.3%)
62 (34.4%)
0.01
Patients with advanced neoplasm
14 (7.8%)
15 (8.3%)
1.0
Patients with serrated polyps
35 (19.4%)
23 (12.8%)
0.11
1.0
Mean number of polyps per patient (s.d.)
1.49 (1.84)
1.13 (1.86)
0.07
0.49
Mean number of adenoma per patient (s.d.)
0.94 (1.37)
0.76 (1.58)
0.23
(n=170)
(n=165)
Patients with polyps
54 (31.8%)
52 (31.5%)
1.0
Patients with adenoma
37 (21.8%)
35 (21.2%)
0.90
Patients with serrated polyps
12 (7.1%)
17 (10.3%)
0.33
Mean number of polyps per patient (s.d.)
0.50 (0.92)
0.46 (0.82)
0.66
Mean number of adenoma per patient (s.d.)
0.29 (0.64)
0.27 (0.58)
0.73
Indications Screening
Second colonoscopy
Bowel preparation Excellent Good
15 (8.3%)
18 (10%)
90 (50%)
94 (52.2%)
Fair
58 (32.2%)
53 (29.4%)
Poor
17 (9.4%)
15 (8.3%)
10.4 (4.1)
9.9 (6.9)
Experienced endoscopists
89 (49.4%)
75 (41.7%)
Fellows
91 (50.6%)
105 (58.3%)
Mean withdrawal time in minutes (s.d.)
First colonoscopy
0.86
0.47
Colonoscopy by
P
NBI, narrow band imaging; HD-WL, high-definition white light.
0.17
NBI, narrow band imaging; HD-WL, high-definition white light.
did not reach statistical significance, the differences were more prominent in the cecum and ascending colon (22.8% vs. 14.4%, P = 0.06), particularly for patients with small ( < 5 mm) adenomas (17.2% vs. 10%, P = 0.06). In the recto-sigmoid region, the new NBI also tended to detect more adenomas compared with the HD-WL group (20.6% vs. 13.3%, P = 0.09). Adenoma and polyp miss rates
Second-pass colonoscopy was completed in 170 (94.4%) and 165 (91.7%) patients in the NBI and HD-WL groups, respectively (P = 0.41). The reasons for incomplete second-pass examinations were listed in Figure 1 and were largely due to poor patient tolerance or failure to reach cecum again by the endoscopist. Twenty-seven (6.5%) patients who had normal first-pass examinations were found to have colorectal adenomas on tandem colonoscopy only: 11 (6.5%) in the new NBI group and 16 (9.7%) in the HD WL group (P = 0.32). Overall, 72 (21.4%) patients had adenomas missed by first colonoscopy: 37 (21.8%) patients in the new NBI group and 35 (21.2%) patients in the HD-WL colonoscopy (P = 0.90). The corresponding polyp miss rate of the two groups was 31.8% and 31.5%, respectively (P = 1.0). © 2014 by the American College of Gastroenterology
A total of 94 adenomas were missed by the first colonoscopic examination and detected by second colonoscopy: 49 in the new NBI group and 45 in the HD-WL group (P = 0.73). The characteristics of missed adenomas were shown in the right panel of Figure 2. The majority of the missed adenomas were small ( < 5 mm) adenomas. The 190-NBI tended to pick up more patients with small ( < 5 mm) recto-sigmoid adenomas than the HD-WL on tandem colonoscopy (7.1% vs. 2.4%; P = 0.07). No advanced colorectal neoplasm was missed by first colonoscopy. When the findings of the two colonoscopies were combined, the overall adenoma detection rate of the 190-NBI group was still significantly higher than that of the HD-WL group (54.4% vs. 43.3%; P = 0.045). The corresponding polyp detection rate was 68.9% vs. 58.9% (P = 0.06). The total number of adenomas and polyps detected by the two colonoscopies combined also tended to be higher in the new NBI group (mean number of adenomas, 1.22±1.66; and mean number of polyps, 1.96±2.30) than in the HD-WL group (mean number of adenomas, 1.00±1.74, P = 0.08; and mean number of polyps, 1.55±2.15, P = 0.23). Adenoma detection and miss rates of different endoscopists
The overall adenoma detection rate tended to be higher in experienced endoscopists (45.1%) compared with fellows (38.3%, P = 0.20). The adenoma detection rates were higher in all eight endoscopists with the use of 190-NBI compared with HD-WL. Notably, 190-NBI significantly increased the adenoma detection The American Journal of GASTROENTEROLOGY
ENDOSCOPY
Table 1. Patient’s characteristics
859
860
Leung et al.
Second colonoscopy
ENDOSCOPY
First colonoscopy
190-NBI
190-NBI
HD-WL
190-NBI
HD-WL
All adenoma
22.8%
18.3%
All adenoma
8.2%
9.1%