Endoscopy With NBI in Detecting Colorectal Adenomas
Endoscopy With NBI in Detecting Colorectal Adenomas
Eight elgible studies of 3049 participants with both Exera II and Lucera types of the Olympus (Olympus Medical Systems Corp, Tokyo, Japan) NBI system were retrieved using the search strategy, and they were published as full articles and abstracts, except for one study that was written in Italian. Another study which presented as an abstract was replaced by a full article. Three studies from the same group had no overlap and were included in the meta-analysis. The principal characteristics and quality of included studies are reported in Table 1 and Table 2.
The pooled data from eight trials (fix effects analysis) showed that the rates of adenoma detection with NBI in patients with at least one adenoma, regardless of the tumor size, were marginally higher than that with white light, but the difference was not significant (RR: 1.09, 95%CI: 1.00–1.19, P = 0.05 Z = 1.95). There was no significant heterogeneity among the studies (P = 0.32, I = 15.4). The funnel plots of publication bias for patients with at least one adenoma are shown in Figure 1, indicating no evidence of publication bias, as determined by Egger's asymmetry regression test (two tailed, P = 0.185).
(Enlarge Image)
Figure 1.
Funnel plot from clinical trials of colonoscopy with narrow band imaging for the detection of adenomas. Data are expressed as the values from individual studies. There is no evidence of publication bias (P = 0.185), as determined by Egger's asymmetry regression test.
Stratification analysis by excluding the study with high-definition television (HDTV) modality revealed that there was a significant difference in the frequency of patients with at least one adenoma between the NBI and conventional colonoscopy groups (RR 1.36, 95%CI 1.07–1.71, P = 0.02, Z = 2.43). There was no significant heterogeneity among these studies (P = 0.66, I = 0). When the abstracts were excluded, the pooled results from the remaining six studies showed that there was no significant difference in the percentages of patients with adenoma as detected by these two modalities (RR 1.07, 95%CI 0.97–1.17, P = 0.17, Z = 1.39). There was also no significant heterogeneity among these studies (P = 0.45, I = 0).
The mean number of adenomas detected in one patient from five available publications was various among these studies, but was not significantly different between these two modalities (weighted mean difference [WMD]: −0.04, 95%CI: − 0.12–0.04, P = 0.36, Z = 0.92). However, there was a marginally significant heterogeneity among these studies (P = 0.06 I = 56.6).
There were five studies with the data of the detection of flat adenomas. The rates of flat adenoma detection with NBI were significantly higher than that with white light (pooled RR 1.96, 95%CI 1.09–3.52, P = 0.02, Z = 2.26), as determined by the random effects model because of significant heterogeneity (P < 0.01, I = 84.4). However, there was no publication bias (P = 0.33). We further stratified the data by factors that could potentially influence the pooled results, including study publication types (full articles vs abstracts), different NBI systems (Lucera spectrum vs Exera II), and the type of imaging modalities (HDTV vs conventional imaging). We found that there were significant results pooled from abstract versus article publications (P = 0.0001), with the Lucera spectrum system versus Exera system (P = 0.0001) and HDTV versus non-HDTV modalities (P < 0.001).
Eight studies measured the colonoscopy withdrawal time in patients. There was significant heterogeneity among these studies (P < 0.001, I = 89.4). Using the random effects model, there was a statistically significant difference between these two groups (WMD: 0.90, 95%CI: 0.38–1.42, P = 0.0006, Z = 3.42). Because of significant heterogeneity, we preformed sensitivity analyses. By repeating the meta-analysis with exclusion of each study per analysis, we found no significant difference between the initial pooled results and further pooled data. Subsequently, we performed sensitivity analyses by stratifying studies, according to the factors that could potentially influence the pooled results, including study publication types (full articles vs abstracts), the number of experienced colonoscopists (performing examination by more than six colonoscopists vs less than six), type of imaging modalities (HDTV imaging vs conventional imaging), and type of NBI system (Lucera spectrum vs EXERAII). We found that the results pooled only from full articles showed less heterogeneity, and the pooled WMD was 0.64 (95%CI: 0.28–0.99; test of heterogeneity, P = 0.03). Furthermore, we found that the results pooled from studies only with HDTV modality also displayed less heterogeneity, and the pooled WMD was 0.71 (95%CI: 0.35–1.06; test of heterogeneity, P = 0.02). When the withdrawal time was defined as mean extubation time, the pooled WMD was 0.69 (95%CI: 0.01–1.37; test of heterogeneity, P = 0.07). In addition, analysis of the studies performed by more than six experienced colonoscopists indicated that the pooled WMD was 0.73 (95%CI: 0.26–1.19; test of heterogeneity, P = 0.04). Finally, we stratified these studies with EXERA II vs Lucera spectrum, and we found that there was a significant difference (WMD: 0.98, 95%CI: 0.30–1.66, P = 0.005) ( Table 3 ).
Results (Tables 1–3)
Description of Study Characteristics
Eight elgible studies of 3049 participants with both Exera II and Lucera types of the Olympus (Olympus Medical Systems Corp, Tokyo, Japan) NBI system were retrieved using the search strategy, and they were published as full articles and abstracts, except for one study that was written in Italian. Another study which presented as an abstract was replaced by a full article. Three studies from the same group had no overlap and were included in the meta-analysis. The principal characteristics and quality of included studies are reported in Table 1 and Table 2.
Detection of Adenomas
The pooled data from eight trials (fix effects analysis) showed that the rates of adenoma detection with NBI in patients with at least one adenoma, regardless of the tumor size, were marginally higher than that with white light, but the difference was not significant (RR: 1.09, 95%CI: 1.00–1.19, P = 0.05 Z = 1.95). There was no significant heterogeneity among the studies (P = 0.32, I = 15.4). The funnel plots of publication bias for patients with at least one adenoma are shown in Figure 1, indicating no evidence of publication bias, as determined by Egger's asymmetry regression test (two tailed, P = 0.185).
(Enlarge Image)
Figure 1.
Funnel plot from clinical trials of colonoscopy with narrow band imaging for the detection of adenomas. Data are expressed as the values from individual studies. There is no evidence of publication bias (P = 0.185), as determined by Egger's asymmetry regression test.
Stratification for Any Adenoma Detection
Stratification analysis by excluding the study with high-definition television (HDTV) modality revealed that there was a significant difference in the frequency of patients with at least one adenoma between the NBI and conventional colonoscopy groups (RR 1.36, 95%CI 1.07–1.71, P = 0.02, Z = 2.43). There was no significant heterogeneity among these studies (P = 0.66, I = 0). When the abstracts were excluded, the pooled results from the remaining six studies showed that there was no significant difference in the percentages of patients with adenoma as detected by these two modalities (RR 1.07, 95%CI 0.97–1.17, P = 0.17, Z = 1.39). There was also no significant heterogeneity among these studies (P = 0.45, I = 0).
Mean Number of Adenomas per Patient
The mean number of adenomas detected in one patient from five available publications was various among these studies, but was not significantly different between these two modalities (weighted mean difference [WMD]: −0.04, 95%CI: − 0.12–0.04, P = 0.36, Z = 0.92). However, there was a marginally significant heterogeneity among these studies (P = 0.06 I = 56.6).
Detection of Flat Adenoma
There were five studies with the data of the detection of flat adenomas. The rates of flat adenoma detection with NBI were significantly higher than that with white light (pooled RR 1.96, 95%CI 1.09–3.52, P = 0.02, Z = 2.26), as determined by the random effects model because of significant heterogeneity (P < 0.01, I = 84.4). However, there was no publication bias (P = 0.33). We further stratified the data by factors that could potentially influence the pooled results, including study publication types (full articles vs abstracts), different NBI systems (Lucera spectrum vs Exera II), and the type of imaging modalities (HDTV vs conventional imaging). We found that there were significant results pooled from abstract versus article publications (P = 0.0001), with the Lucera spectrum system versus Exera system (P = 0.0001) and HDTV versus non-HDTV modalities (P < 0.001).
Colonoscopy Withdrawal Time
Eight studies measured the colonoscopy withdrawal time in patients. There was significant heterogeneity among these studies (P < 0.001, I = 89.4). Using the random effects model, there was a statistically significant difference between these two groups (WMD: 0.90, 95%CI: 0.38–1.42, P = 0.0006, Z = 3.42). Because of significant heterogeneity, we preformed sensitivity analyses. By repeating the meta-analysis with exclusion of each study per analysis, we found no significant difference between the initial pooled results and further pooled data. Subsequently, we performed sensitivity analyses by stratifying studies, according to the factors that could potentially influence the pooled results, including study publication types (full articles vs abstracts), the number of experienced colonoscopists (performing examination by more than six colonoscopists vs less than six), type of imaging modalities (HDTV imaging vs conventional imaging), and type of NBI system (Lucera spectrum vs EXERAII). We found that the results pooled only from full articles showed less heterogeneity, and the pooled WMD was 0.64 (95%CI: 0.28–0.99; test of heterogeneity, P = 0.03). Furthermore, we found that the results pooled from studies only with HDTV modality also displayed less heterogeneity, and the pooled WMD was 0.71 (95%CI: 0.35–1.06; test of heterogeneity, P = 0.02). When the withdrawal time was defined as mean extubation time, the pooled WMD was 0.69 (95%CI: 0.01–1.37; test of heterogeneity, P = 0.07). In addition, analysis of the studies performed by more than six experienced colonoscopists indicated that the pooled WMD was 0.73 (95%CI: 0.26–1.19; test of heterogeneity, P = 0.04). Finally, we stratified these studies with EXERA II vs Lucera spectrum, and we found that there was a significant difference (WMD: 0.98, 95%CI: 0.30–1.66, P = 0.005) ( Table 3 ).
Source...