The ability of a novel blue laser imaging system for the diagnosis of invasion depth of colorectal neoplasms

Background

Fujifilm has developed a novel endoscope system with two kinds of lasers that enables us to allow narrow-band light observation with blue laser imaging (BLI). The aim of this study was to evaluate BLI magnification in comparison with narrow-band imaging (NBI) magnification for the diagnosis of colorectal neoplasms.

Methods

This was a multicenter open study. A total of 104 colorectal neoplasms were examined with BLI and NBI magnifications in Kyoto Prefectural University of Medicine and Fukuoka University Chikushi Hospital. Vascular and surface patterns of tumors under BLI magnification were compared with those under NBI magnification, using a published NBI classification. The main outcome was the correlation between the NBI classification diagnosed by BLI or NBI magnification and the histopathological analyses.

Results

Sixty-two cases of adenoma, 34 cases of intramucosal cancer and shallowly invaded submucosal cancer, and eight cases of deeply invaded submucosal cancer were diagnosed. The diagnostic accuracy of BLI magnification in the NBI classification was 74.0 % (77/104), similar to that of NBI magnification (77.8 %). The consistency rate between BLI and NBI magnification in the NBI classification was 74.0 %. Concerning image evaluation, the interobserver variability of two expert endoscopists (N.Y. and T.H.) in BLI magnification was κ = 0.863. On the other hand, the intraobserver variability of the two endoscopists was κ = 0.893 (N.Y.) and 0.851 (T.H.).

Conclusions

BLI magnification by laser source could predict histopathological diagnosis and invasion depth of colorectal neoplasms. The diagnostic effectiveness of this method was similar to that of NBI magnification.     

The detection of surface patterns by flexible spectral imaging color enhancement without magnification for diagnosis of colorectal polyps

Purpose

Flexible spectral imaging color enhancement (FICE), or image-enhanced endoscopy, can enhance visualization of surface and vascular patterns of colorectal polyps. Resolution of FICE has recently been improved. We evaluated diagnostic accuracy for neoplastic and non-neoplastic colorectal polyp differentiation with detection of surface patterns by FICE without magnification.

Methods

Retrospective analysis of 151 colorectal polyps evaluated by FICE without magnification was performed. Neoplastic surface patterns were defined as tubular and oval pit. We aimed to determine sensitivity, specificity, positive and negative predictive values (PPV and NPV), and accuracy in correlating diagnosis by FICE without magnification with histology. Moreover, findings were compared to those of white-light endoscopy (WL) and chromoendoscopy (CHR).

Results

Of the 151 colorectal polyps, 95 were identified as neoplastic and 56 were identified as non-neoplastic. FICE without magnification had a sensitivity of 89.4%, specificity of 89.2%, PPV of 93.4%, NPV of 83.3%, and accuracy of 89.4%. The accuracy of FICE value was higher than that of WL (sensitivity of 74.7%, specificity of 73.2%, PPV of 82.5%, NPV of 63.0%, and accuracy of 74.1%) and was worse than that of CHR (sensitivity of 96.8%, specificity of 89.2%, PPV of 93.9%, NPV of 96.1%, and accuracy of 94.7%). Imaging evaluation was validated by inter-/intra-observer measurements, demonstrating consistent results.

Conclusions

The detection of surface patterns by FICE without magnification is useful for differential diagnosis of colorectal polyps. We believe that FICE without magnification is more convenient and easier method than CHR.     

A randomized crossover open trial of the adenoma miss rate for narrow band imaging (NBI) versus flexible spectral imaging color enhancement (FICE)

Purpose

Narrow band imaging (NBI) and flexible spectral imaging color enhancement (FICE) allow improved contrasted evaluation of the mucosal surface. However, no study has compared the utility of these two modalities. Therefore, the aim of this study was to compare the adenoma miss rate (AMR) between NBI and FICE.

Methods

A total of 55 patients (38 men, 17 women) were enrolled in this study. Patients were randomly assigned to the NBI–FICE group (NBI followed by FICE) or the FICE–NBI group (FICE followed by NBI). NBI and FICE total colonic observations were tandemly performed for each patient during the scope withdrawal with white light following cecal intubation. All detected polyps with the NBI or FICE observation were categorized into three groups according to the size and number of polyps missed.

Results

Twenty-nine patients were assigned to the NBI–FICE group, and 26 patients were assigned to the FICE–NBI group. There was no significant difference in the overall AMR when comparing the image-enhanced endoscopy technologies (17.9 % for NBI, 26 % for FICE, p?=?0.159). AMR was lower for NBI than for FICE for adenomas <5 mm in diameter (5.7 % for NBI, 12.6 % for FICE, p?=?0.036). AMR was not significantly different when comparing NBI and FICE for lesions 5 to 10 mm (p?=?0.967) or for lesions ≥10 mm (p?=?0.269).

Conclusions

This study demonstrated that overall AMR was not different when comparing NBI and FICE.     

Characteristic magnifying narrow-band imaging features of colorectal tumors in each growth type

Purpose

We investigated the surface characteristics and vascular patterns of colorectal tumors according to growth type by means of magnifying narrow-band imaging (NBI).

Methods

Four hundred ninety-seven colorectal tumors larger than 10 mm (204 tubular adenomas [TAs], 199 frankly invasive intramucosal carcinomas to shallow invasive submucosal [M/SM-s] carcinomas, and 94 deeply invasive submucosal [SM-d] carcinomas) were analyzed. These colorectal tumors were classified according to growth type as follows: polypoid type, n?=?224; laterally spreading tumor-granular (LST-G) type, n?=?133; and LST-non-granular (LST-NG) type, n?=?140. Surface and vascular patterns were evaluated in relation to histology and growth type.

Results

The absent and irregular surface patterns were observed in approximately 40 % of the SM-d carcinomas of the polypoid and LST-G type. The unclear surface pattern was more frequent in tumors of the LST-NG type than in those of other growth types, regardless of histology. Among TAs and M/SM-s carcinomas, the dense vascular pattern was most frequent in polypoid type, the dense and corkscrew vascular patterns were most frequent in the LST-G type, and the honeycomb and avascular and/or fragmentary patterns were most frequent in the LST-NG type. The avascular and/or fragmentary vessel pattern was more frequent in SM-d carcinomas than in TA and M/SM-s carcinomas, regardless of growth type.

Conclusions

A part of LST-NG was difficult to identify the NBI magnifying surface pattern. Although NBI magnifying findings were almost same in each type lesion in SM-d lesion, those of LST-NG were different from those of LST-G and polypoid type in M/SM-s lesion.     

White Opaque Substance Visualized Using Magnifying Endoscopy with Narrow-Band Imaging in Colorectal Epithelial Neoplasms

Background

The presence of a white opaque substance (WOS) on magnifying endoscopy (ME) with narrow-band imaging (NBI) has been reported for gastric epithelial neoplasms, but the presence of WOS in colorectal epithelial neoplasms has not been investigated.

Aims

The purpose of this study was to determine whether WOS is present in colorectal epithelial neoplasms and to clarify its clinical significance.

Methods

A total of 590 colorectal epithelial neoplasms from 368 consecutive patients were retrospectively analyzed using prospectively collected data. Presence or absence of WOS in colorectal epithelial neoplasms was recorded based on the findings of ME with NBI.

Results

White opaque substance was present in 236 of the 590 (40 %) colorectal epithelial neoplasms. Compared with WOS-negative patients, WOS-positive patients showed significantly larger tumors (p < 0.0001) and significantly more tumors in the proximal colon (p = 0.0003). WOS was more frequently present in carcinomas (66.0 %) than in adenomas (31.8 %; p < 0.0001). WOS was also more frequent in submucosal carcinomas (75.9 %) than in intramucosal carcinomas (59.0 %; p = 0.0380).

Conclusions

This study confirmed the presence of WOS in colorectal epithelial neoplasms, and prevalence increased with the progression of cancer, from adenoma to carcinoma and from intramucosal carcinoma to submucosal carcinoma.     

Comparison of Clinical Outcomes Among Different Endoscopic Resection Methods for Treating Colorectal Neoplasia

Background/Aim

Endoscopic treatments of colorectal neoplasms have yet to be standardized. This study aimed to compare efficacy and tolerability of different endoscopic resection methods for colorectal epithelial tumors.

Methods

Patients with non-pedunculated colorectal tumors undergoing endoscopic treatments were consecutively enrolled, and their medical records were reviewed retrospectively. The resection methods were classified into three groups: endoscopic mucosal resection with circumferential precutting (EMR-P), endoscopic submucosal dissection with snaring (ESD-S), and endoscopic submucosal dissection alone (ESD). We compared en bloc resection, pathological complete resection, and complications associated with these methods.

Results

Overall, 206 lesions from 203 patients were included in the study (mean size 25.2 ± 10.1 mm). The number of lesions treated with EMR-P, ESD-S, and ESD was 91 (44.2 %), 57 (27.7 %), and 58 (28.2 %), respectively. There was a significant difference in both the en bloc resection rates (EMR-P, 61.5 %; ESD-S, 64.9 %; ESD, 96.6 %; p = 0.001) and complete resection rates (EMR-P, 51.6 %; ESD–S, 54.4 %; ESD, 75.9 %; p = 0.009). Bleeding and perforation were less frequently observed in the EMR-P group. In the subgroup-analysis of lesions less than 20 mm, however, these differences were not observed.

Conclusions

All endoscopic resection methods, including EMR-P, ESD-S, and ESD, were effective and safe for the treatment of colorectal neoplasms. Technically demanding ESD with high en bloc and complete resection rate should be reserved for the suspicious cancer lesion, which requires the precise histological evaluation. EMR-P with good feasibility can be considered an alternative to ESD for the lesions less than 20 mm.     

Evaluation of Microvessels in Colorectal Tumors by Narrow Band Imaging Magnification: Including Comparison with Magnifying Chromoendoscopy

Background/Aims  

Narrow band imaging (NBI) magnification analysis has entered use in clinical settings to diagnose colorectal tumors. Pit pattern analysis with magnifying endoscopy is already widely used to assess colorectal lesions and invasion depth. Our study compared diagnoses by vascular pattern analysis and pit pattern analysis with NBI magnification.     

Miniprobe Endoscopic Ultrasonography Has Limitations in Determining the T Stage in Early Colorectal Cancer

Background/Aims

Mini-probe endoscopic ultrasonography (mEUS) is a useful diagnostic tool for accurate assessment of tumor invasion. The aim of this study was to estimate the accuracy of mEUS in patients with early colorectal cancer (ECC).

Methods

Ninety lesions of ECC underwent mEUS for pre-treatment staging. We divided the lesions into either the mucosal group or the submucosal group according to the mEUS findings. The histological results of the specimens were compared with the mEUS findings.

Results

The overall accuracy for assessing the depth of tumor invasion (T stage) was 84.4% (76/90). The accuracy of mEUS was significantly lower for submucosal lesions compared to mucosal lesions (p=0.003) and it was lower for large tumors (≥2 cm) (p=0.034). The odds ratios of large tumors and submucosal tumors affecting the accuracy of T staging were 3.46 (95% confidence interval [CI], 1.05 to 11.39) and 6.25 (95% CI, 1.85 to 25.14), respectively. When submucosal tumors were combined with large size, the odds ratio was 14.67 (95% CI, 1.46 to 146.96).

Conclusions

The overall accuracy of T stage determination with mEUS was considerably high in patients with ECC; however, the accuracy decreased when tumor size was >2 cm or the tumor had invaded the submucosal layer.     

Risk factors for delayed bleeding after endoscopic submucosal dissection for colorectal neoplasms

Purpose

Although delayed bleeding is a major complication of endoscopic submucosal dissection (ESD) for colorectal neoplasms, few reports have assessed the risk factors for delayed bleeding after colorectal ESD.

Methods

This study included 363 consecutive patients in whom 377 colorectal neoplasms were resected using ESD between April 2006 and August 2012. We classified patients and lesions into two groups on the basis of presence or absence of delayed bleeding and retrospectively compared the clinicopathological characteristics and clinical outcomes of ESD between the two groups.

Results

Delayed bleeding occurred in 25 (6.6 %) of 377 lesions, and all cases of delayed bleeding were successfully controlled by endoscopic procedures. With respect to patient-related factors, there was no significant difference between the groups in mean age, sex ratio, and current use of antithrombotic agents. With respect to lesion-related factors, there was no significant difference between the groups in mean lesion size, growth pattern, and mean procedure time (p?=?0.6). Lesions located in the rectum (vs colon, p?=?0.0005) and lesions with severe submucosal fibrosis (vs no or mild fibrosis, p?=?0.022) were significantly related to delayed bleeding. Upon multivariate analysis, lesions located in the rectum (vs colon, odds ratio 4.19; p?=?0.0009) were significantly related to delayed bleeding after colorectal ESD.

Conclusions

This study demonstrated that location of lesions in the rectum was a significant independent risk factor for delayed bleeding after ESD for colorectal neoplasms.     

Narrow-Banding Images and Structures of Microvessels of Colonic Lesions

Background  

Magnifying colonoscopy with NBI has been shown to be useful for the differential diagnosis of tumors. However, the relationship between findings on NBI magnification and the microvessel architecture of colorectal lesions remains to be clarified.     

Evaluation of microvessels in colorectal tumors by narrow band imaging magnification

BACKGROUND: Narrow band imaging (NBI) uses optical filters for sequential green and blue illumination and narrows the bandwidth of spectral transmittance. OBJECTIVE: We determined the clinical usefulness of NBI magnification for evaluation of microvascular architecture and qualitative diagnosis of colorectal tumors. DESIGN: This study was a retrospective study. SETTING: Department of Endoscopy, Hiroshima University, Hiroshima, Japan. PATIENTS AND MAIN OUTCOME MEASUREMENTS: A total of 189 colorectal lesions were analyzed. Each lesion was observed by NBI magnifying endoscopy and classified according to microvascular features (ie, thickness and irregularity). Microvessel thickness was classified as invisible, thin, or thick, and microvessel irregularity was classified as invisible, regular, mildly irregular, or severely irregular. NBI endoscopic images were compared with histologic findings. RESULTS: With respect to microvessel thickness, invisible microvessels were found significantly more often in hyperplasia lesions, and thick microvessels were found significantly more often in carcinoma with submucosal massive invasion (sm-m) (P < .01). With respect to microvessel irregularity, invisible microvessels were found significantly often in hyperplasia lesions, and severely irregular microvessels were found significantly often in sm-m lesions (P < .01). Accuracy of diagnosis of sm-m on the basis of thick and severely irregular lesions was 100%. CONCLUSION: Microvascular features determined by NBI magnification are associated with histologic grade and depth of submucosal invasion. These results indicate that NBI magnification is useful for the prediction of histologic diagnosis and selection of therapeutic strategies of colorectal tumors.     

Risk factors for vertical incomplete resection in endoscopic submucosal dissection as total excisional biopsy for submucosal invasive colorectal carcinoma

Purpose

Endoscopic submucosal dissection (ESD) for colorectal tumor is a minimally invasive treatment. Histologic information obtained from the entire ESD specimen is important for therapy selection in submucosal invasive colorectal carcinoma (SMca). This study aimed to identify risk factors for vertical incomplete resection (vertical margin-positive [VM+]) when ESD was performed as total excisional biopsy for SMca.

Methods

From June 2003 through December 2011, 78 SMca cases were resected by ESD at Hiroshima University Hospital. Patient and tumor characteristics, intraoperative variables, and histopathology were compared between the VM+ group and the vertical complete resection (vertical margin-negative) group. The ability of magnifying endoscopy (ME) and endoscopic ultrasonography (EUS) to predict VM+ was assessed.

Results

ESD resulted in VM+ in eight cases (10.3 %), with a greater percentage invading to a depth of ≥2,000 vs. <2,000 μm (P?=?0.047). Severe submucosal fibrosis was found in five of the eight cases (62.5 %, P?=?0.017). Poor differentiation was seen at the deepest invasive portion in six cases (75.0 %), and two of six cases had an invasion depth <2,000 μm. Of 39 EUS cases, 36 not showing deep invasion close to the muscularis propria were completely resected by ESD.

Conclusions

Submucosal fibrosis and poor differentiation at the deepest invasive portion may be risk factors for VM+ in colorectal ESD for tumors with submucosal deep invasion. ME plus EUS is more likely to help determine whether ESD is indicated as complete total excisional biopsy for SMca.     

Factors associated with complete local excision of small rectal carcinoid tumor

Purpose

Although small rectal carcinoid tumors can be treated using local excision, complete resection can be difficult because tumors are located in the submucosal layer. We evaluate the factors associated with pathologically complete local resection of rectal carcinoid tumors.

Methods

Data were analyzed of 161 patients with 166 rectal carcinoid tumors who underwent local excision with curative intent from January 2001 to December 2010. A pathologically complete resection (P-CR) was defined as an en bloc resection with tumor-free lateral and deep margins. The study classified treatments into three categories for analysis: conventional polypectomy (including strip biopsy, snare polypectomy, and hot biopsy), advanced endoscopic techniques (including endoscopic mucosal resection with cap and endoscopic submucosal dissection), and surgical local excision (including transanal excision and transanal endoscopic microsurgery). We evaluated the P-CR rate according to treatment method, tumor size, initial endoscopic impression and the use of endoscopic ultrasound (EUS) or transrectal ultrasound (TRUS).

Results

The mean tumor size was 5.51?±?2.43 mm (range 2–18 mm) and all lesions were confined to the submucosal layer. The P-CR rates were 30.9, 72.0, and 81.8 % for conventional polypectomy, advanced endoscopic techniques, and surgical local excision, respectively. Univariate analysis showed that P-CR was associated with treatment method, use of EUS or TRUS, and initial endoscopic impression. Multivariate analysis showed that only treatment method was associated with P-CR.

Conclusion

Pathologically complete resection of small rectal carcinoid tumors was more likely to be achieved when using advanced endoscopic techniques or surgical local excision rather than conventional polypectomy.     

Does Magnifying Narrow-Band Imaging or Magnifying Chromoendoscopy Help Experienced Endoscopists Assess Invasion Depth of Large Sessile and Flat Polyps?

Background

Distinguishing deep submucosa (SM) from superficial SM cancer in large sessile and flat colorectal polyps (>2 cm) is crucial in making the most appropriate therapeutic decision. We evaluated the additional role of magnifying narrow-band imaging (NBI) and magnifying chromoendoscopy (MCE) in assessing the depth of invasion in large sessile and flat polyps in comparison to morphological evaluation performed by experienced endoscopists.

Methods

From May 2011 to December 2011, a total of 85 large sessile and flat polyps were analyzed. Endoscopic features of the polyps were independently evaluated by experienced endoscopists. Subsequently, the polyps were observed using magnifying NBI and MCE.

Results

A total of 58 intramucosal lesions and 27 SM cancers (five superficial and 22 deep) were identified. The diagnostic accuracy of the experienced endoscopists, NBI, and MCE were 92.9, 90.6, and 89.4 %, respectively, for deep SM cancer. In combination with NBI or MCE, the diagnostic accuracy of the experienced endoscopists did not change significantly for deep SM cancer, with an accuracy of 95.3 % for both NBI and MCE.

Conclusions

Conventional colonoscopy can differentiate superficial from deep SM cancers with an accuracy of as high as 92.9 % in large sessile and flat polyps. Further diagnostic strategies are required in order to precisely assess the depth of invasion, especially in large colorectal polyps.     

Factors associated with technical difficulties and adverse events of colorectal endoscopic submucosal dissection: retrospective exploratory factor analysis of a multicenter prospective cohort

Background

Colorectal endoscopic submucosal dissection (C-ESD) is a promising but challenging procedure. We aimed to evaluate the factors associated with technical difficulties (failure of en bloc resection and procedure time, ≥2 h) and adverse events (perforation and bleeding) of C-ESD.

Methods

We conducted a retrospective exploratory factor analysis of a prospectively collected cohort in 15 institutions. Eight-hundred sixteen colorectal neoplasms larger than 20 mm from patients who underwent C-ESD were included. We assessed the outcomes of C-ESD and risk factors for technical difficulties and adverse events.

Results

Of the 816 lesions, 767 (94 %) were resected en bloc, with a median procedure time of 78 min. Perforation occurred in 2.1 % and bleeding in 2.2 %. Independent factors associated with failure of en bloc resection were low-volume center (<30 neoplasms), snare use, and poor lifting after submucosal injection. Factors significantly associated with long procedure time (≥2 h) were large tumor size (≥4 cm), low-volume center, less-experienced endoscopist, CO₂ insufflation, and use of two or more endoknives. Poor lifting was the only factor significantly associated with perforation, whereas rectal lesion and lack of a thin-type endoscope were factors significantly associated with bleeding. Poor lifting after submucosal injection occurred more frequently for nongranular-type laterally spreading tumors (LST) and for protruding and recurrent lesions than for granular-type LST (LST-G).

Conclusions

Poor lifting after submucosal injection was the risk factor most frequently associated with technical difficulties and adverse events on C-ESD. Less experienced endoscopists should start by performing C-ESDs on LST-G lesions.     

Differences of image enhancement in image-enhanced endoscopy: narrow band imaging versus flexible spectral imaging color enhancement

Background  

Narrow band imaging (NBI) can emphasize images of the surface microvasculature of lesions, because the central wavelengths of the NBI filter are 415 and 540 nm and these wavelengths are well absorbed by hemoglobin. Flexible spectral imaging color enhancement (FICE) increases the contrast in depictions of mucosal lesions. However, quantitative evaluation of the image enhancement shown by NBI and FICE has not been reported. The aim of this study was to measure and compare the degrees of image enhancement in NBI and FICE.     

Usefulness and safety of 0.4% sodium hyaluronate solution as a submucosal fluid "cushion" for endoscopic resection of colorectal mucosal neoplasms: A prospective multi-center open-label trial

Background

Sodium hyaluronate (SH) solution has been used for submucosal injection in endoscopic resection to create a long-lasting submucosal fluid "cushion". Recently, we proved the usefulness and safety of 0.4% SH solution in endoscopic resection for gastric mucosal tumors. To evaluate the usefulness of 0.4% SH as a submucosal injection solution for colorectal endoscopic resection, we conducted an open-label clinical trial on six referral hospitals in Japan.

Methods

A prospective multi-center open-label study was designed. A total of 41 patients with 5–20 mm neoplastic lesions localized in the colorectal mucosa at six referral hospitals in Japan in a single year period from December 2002 to November 2003 were enrolled and underwent endoscopic resection with SH. The usefulness of 0.4% SH was assessed by the en bloc complete resection and the formation and maintenance of mucosal lesion-lifting during endoscopic resection. Safety was evaluated by analyzing adverse events during the study period.

Results

The usefulness rate was high (82.5%; 33/40). The following secondary outcome measures were noted: 1) steepness of mucosal lesion-lifting, 75.0% (30/40); 2) intraoperative complications, 10.0% (4/40); 3) time required for mucosal resection, 6.7 min; 4) volume of submucosal injection, 6.8 mL and 5) ease of mucosal resection, 87.5% (35/40). Two adverse events of bleeding potentially related to 0.4% SH were reported.

Conclusion

Using 0.4% SH solution enabled sufficient lifting of a colorectal intramucosal lesion during endoscopic resection, reducing the need for additional injections and the risk of perforation. Therefore, 0.4% SH may contribute to the reduction of complications and serve as a promising submucosal injection solution due to its potentially superior safety in comparison to normal saline solution.     

Endoscopic Tri-Modal Imaging Improves Detection of Gastric Intestinal Metaplasia Among a High-Risk Patient Population in Singapore

Background

Detection of pre-neoplastic gastric mucosal changes and early gastric cancer (EGC) by white-light endoscopy (WLE) is often difficult. In this study we investigated whether combined autofluorescence imaging (AFI) and narrow band imaging (NBI) can improve detection of pre-neoplastic lesions and early gastric cancer in high-risk patients.

Patients and Methods

Chinese patients who were 50-years-old or above with dyspepsia were examined by both high-resolution WLE and combined AFI followed by NBI (AFI–NBI), consecutively in a prospective randomized cross-over setting, by two experienced endoscopists. The primary outcome was diagnostic ability of the two methods for patients with pre-neoplastic lesions such as intestinal metaplasia (IM) and mucosal atrophy.

Results

Sixty-five patients were recruited. One patient with large advanced gastric cancer was found and excluded from the analysis. Among the remaining 64 patients, 38 (59 %) had IM; of these, 26 (68 %) were correctly identified by AFI–NBI (sensitivity 68 %, specificity 23 %) and only 13 (34 %) by WLE (sensitivity 34 %, specificity 65 %). AFI–NBI detected more patients with IM than did WLE (p = 0.011). Thirty-one patients (48 %) had mucosal atrophy. Ten patients (32 %) were identified by AFI–NBI (sensitivity 32 %, specificity 79 %) and four patients (13 %) by WLE (sensitivity 13 %, specificity 88 %) (p = 0.100). No dysplasia or EGC was found.

Conclusion

AFI–NBI identified significantly more patients with IM than did WLE. Our result warrants further studies to define the role of combined AFI–NBI endoscopy for detection of precancerous conditions.     

Differential diagnosis of colorectal polyps with respect to similar capabilities of white light colonoscopy and narrow-band imaging

Purpose

The responses of polyps to light essentially determine the diagnostic capability of an endoscopy system in differentiating adenomas from hyperplastic polyps. Compared with white light colonoscopy (WLC), narrow-band imaging (NBI) is expected to improve the diagnostic capability. The diagnostic capabilities of WLC and NBI are evaluated and compared based on the polyp responses.

Methods

The following WLC and NBI images were retrospectively reviewed and categorized: 195 images and polyps (89 WLC, 106 NBI) with the best visual quality were categorized in the best image group (BG), and 484 images of 242 polyps (both WLC and NBI) were categorized in the paired image group (PG). For each reflection of light used for WLC or NBI, the polyp responses were objectively expressed as reflection features. The reflection features were then used to establish a classification model for identifying adenomas. The diagnostic capability of reflection feature or classification model was measured by the area under the receiver operating characteristic curve (AUC).

Results

In both image groups, the diverse and heterogeneous features of the polyp responses enabled accurate identification of adenomas, regardless of the light source used for WLC and NBI. For differential diagnosis of adenomas and hyperplastic polyps, the WLC and NBI did not significantly differ in BG (AUC, 0.905 and 0.922, respectively; P?=?0.690) or in PG (AUC, 0.782 and 0. 769, respectively; P?=?0.755).

Conclusions

Using WLC and NBI as classification models is effective in differential diagnosis of colorectal polyps and exhibited similar capabilities.