If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Gastroenterology and Hepatology, Amsterdam UMC, University of Amsterdam, Amsterdam, the NetherlandsDepartment of Gastroenterology and Hepatology, St. Antonius Hospital, Nieuwegein, the Netherlands
Reprint requests: J.J. Bergman, MD, PhD, Professor of Gastrointestinal Endoscopy, Director of Endoscopy, Amsterdam UMC, Location Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Affiliations
Department of Gastroenterology and Hepatology, Amsterdam UMC, University of Amsterdam, Amsterdam, the Netherlands
Endoscopic recognition of early Barrett’s neoplasia is challenging. Blue-light imaging (BLI) and linked-color imaging (LCI) may assist endoscopists in appreciation of neoplasia. Our aim was to evaluate BLI and LCI for visualization of Barrett’s neoplasia in comparison with white-light endoscopy (WLE) alone, when assessed by nonexpert endoscopists.
Methods
In this web-based assessment, corresponding WLE, BLI, and LCI images of 30 neoplastic Barrett’s lesions were delineated by 3 expert endoscopists to establish ground truth. These images were then scored and delineated by 76 nonexpert endoscopists from 3 countries and with different levels of expertise, in 4 separate assessment phases with a washout period of 2 weeks. Assessments were as follows: assessment 1, WLE only; assessment 2, WLE + BLI; assessment 3, WLE + LCI; assessment 4, WLE + BLI + LCI. The outcomes were (1) appreciation of macroscopic appearance and ability to delineate lesions (visual analog scale [VAS] scores); (2) preferred technique (ordinal scores); and (3) assessors’ delineation performance in terms of overlap with expert ground truth.
Results
Median VAS scores for phases 2 to 4 were significantly higher than in phase 1 (P < .001). Assessors preferred BLI and LCI over WLE for appreciation of macroscopic appearance (P < .001) and delineation (P < .001). Linear mixed-effect models showed that delineation performance increased significantly in phase 4.
Conclusions
The use of BLI and LCI has significant additional value for the visualization of Barrett’s neoplasia when used by nonexpert endoscopists. Assessors appreciated the addition of BLI and LCI better than the use of WLE alone. Furthermore, this addition led to improved delineation performance, thereby allowing for better acquisition of targeted biopsy samples. (The Netherlands Trial Registry number: NL7541.)
Barrett’s esophagus (BE) is a known precursor of esophageal adenocarcinoma (EAC). Therefore, patients with BE undergo regular endoscopic surveillance to enable timely detection of EAC. When detected at an early stage, EAC can be treated endoscopically, which is associated with excellent outcomes.
Recent guidelines for BE surveillance recommend inspection with high-definition white-light endoscopy (WLE), followed by random biopsies in the absence of any lesions.
However, early Barrett’s lesions are subtle and therefore easily missed, and random biopsies sample only a fraction of the Barrett’s segment and are thus associated with sampling error.
In the last decade, several optical chromoscopy techniques have been developed to improve visualization of Barrett’s neoplasia. These techniques, such as narrow-band imaging (NBI; Olympus, Tokyo, Japan) and blue-light imaging (BLI; FUJIFILM, Tokyo, Japan), are integrated in modern endoscopy systems and use the excitation of blue light to enhance mucosal surface patterns and vascular structures. Until recently, the use of NBI for primary detection of neoplasia was limited by the relative darkness of the image in overview and the loss of resolution of still images due to motion artifacts and interlaced video processing. The latest version of NBI and the recently introduced BLI system have overcome these technical limitations, allowing these techniques to be used for primary detection of neoplasia. One of the features that may lead to the detection of early Barrett’s neoplasia in overview are subtle differences in surface relief (ie, subtle elevations and depressions relative to the normal-appearing flat surrounding mucosa). These differences might be better visualized with optical chromoscopy than with WLE, thereby potentially enhancing the appreciation of lesions by endoscopists, improving their ability to obtain targeted biopsy specimens. Our group recently performed a study that showed that BLI has additional value in the visualization of Barrett’s neoplasia, when assessed by expert endoscopists.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
This study had several limitations. First, these expert endoscopists were all experienced in the use of BLI and might therefore have been biased in their assessment. The appreciation of BLI when assessed by nonexpert endoscopists (ie, those who usually perform surveillance endoscopies) remains unclear. Furthermore, linked-color imaging (LCI; FUJIFILM), the second optical chromoscopy technique integrated in the new FUJIFILM endoscopy system, was not evaluated.
The aim of the current study was to evaluate BLI and LCI as additional tools for visualization of early Barrett’s neoplasia when used by endoscopists with different levels of endoscopic expertise.
Methods
Setting and design
This cohort study was conducted at the Departments of Gastroenterology and Hepatology of the Amsterdam University Medical Centers, Location AMC, the Catharina Hospital Eindhoven, and the University Hospitals Leuven. Endoscopic images were collected prospectively during standard patient care at the Amsterdam UMC, the Catharina Hospital, and the University Hospitals Leuven and anonymized after acquisition.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
Thus, the Medical Research Involving Human Subjects Act did not apply. This study was approved by the Medical Ethics Review Committee of all 3 participating centers (reference numbers W15_068, Amsterdam UMC; 2016-07, Catharina Hospital; and S58340, University Hospitals Leuven) The study was registered in The Netherlands Trial Registry (no. NL7541).
ELUXEO 7000 endoscopy system
The ELUXEO 7000 endoscopy system (FUJIFILM) is a new-generation endoscopy system that enables electronic chromoscopy using a light source consisting of 4 LEDs with different wavelengths. By changing the intensity of these LEDs, WLE, BLI, and LCI mode can be created. BLI has a peak wavelength of 410 ± 10 nm and is therefore more likely to be absorbed by hemoglobin and to be scattered in the surface mucosa. BLI therefore allows better visualization of mucosal and vascular patterns. BLI was introduced several years ago and is now applied for visualization of a wide variety of GI pathologies. LCI has been introduced recently and was initially developed to increase visualization of gastric mucosa, although other applications are currently under investigation.
The LCI mode is created by a combination of light spectrum enhancement and advanced signal processing and was developed to enhance color differences in the mucosa, focusing on increased discrimination of red colors. The CMOS image sensor chip in the tip of the endoscope reduces susceptibility to noise and provides high-resolution images.
Acquisition of endoscopic images
Endoscopic images were collected between 2015 and 2018 from a prospectively established image database that contained images from 68 patients with Barrett’s neoplasia.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
The images of 30 patients were included in the current study. The remaining 38 patients were excluded because of insufficient image quality or lack of similarity among the WLE, BLI, and LCI images. In short, the image acquisition protocol of this prospective study was as follows: patients with BE referred for endoscopic work-up of high-grade dysplasia (HGD) or EAC likely to require endoscopic resection were eligible for this study. All endoscopic procedures were performed by expert endoscopists (J.B., B.W., R.P., W.C., E.S., R.B.). After endoscopic inspection, corresponding still images of the lesion in overview were obtained in WLE, BLI, and LCI mode, with the endoscope in the same position (Fig. 1). All images were digitally recorded and stored in full high-definition format (1280 × 1024 pixels).
Figure 1Exemplary cases of early neoplastic Barrett’s lesions, visualized in overview, by corresponding images in white-light endoscopy, blue-light imaging, and linked-color imaging. All images were delineated independently by 3 expert endoscopists to establish the ground truth. Expert delineations are displayed in blue, pink, and yellow.
For this study, an online proprietary software module (Meducati AB, Göteborg, Sweden) was used. The software allowed endoscopic images to be scored and delineated on a computer screen, and subsequently enabled automatic calculation of surface overlap among delineations for analyses (Fig. 1). Between October 2018 and April 2019, a group of international endoscopists assessed the endoscopic images in 4 separate assessment phases, as outlined below. The assessors originated from the Netherlands, Germany, and Belgium and were grouped according to 3 categories of experience: trainees (fellows in training), junior general gastroenterologists (board certified ≤5 years of practice), and senior general gastroenterologists (≥10 years in practice). None of the assessors were considered to be experts in BE management.
Set-up of the assessment phases
In the first phase, 30 WLE images were displayed to assess the appreciation of WLE images without bias due to comparison of corresponding BLI or LCI images. Assessors were first asked to review the image and subsequently complete the following items per image: (1) estimation of the Paris classification of the lesion; (2) appreciation of the Paris classification of the lesion (visual analog scale (VAS) score 1-10, where 1 reflects the lowest and 10 the highest appreciation); (3) appreciation of the surface relief of the lesion (VAS score 1-10); (4) delineation of the lesion on the screen; and (5) ability to delineate the lesion using WLE (VAS score 1-10).
In the second phase, 15 cases were displayed in WLE and BLI in a side-to-side display, and 15 other cases in WLE and LCI. Assessors were asked to review both images and complete the following items: (1) Paris classification; (2) appreciation of the Paris classification of the lesion using both techniques (VAS scores); (3) appreciation of the surface relief of the lesion using both techniques (VAS scores); (4) choice of the best technique to appreciate the macroscopic appearance using an ordinal scale, ranging from −2 (BLI/LCI is much worse than WLE), −1 (BLI/LCI is a little worse than WLE), 0 (BLI/LCI is the same as WLE), +1 (BLI/LCI is a little better than WLE) to +2 (BLI/LCI is much better than WLE); (5) delineation of the lesion on the preferred image (WLE or BLI/LCI); (6) ability to delineate the lesion using both techniques (VAS scores); and (7) choice of the best technique to delineate the lesion (BLI/LCI or WLE) using the ordinal scale mentioned above.
The set-up of the third assessment phase was the same as for the second phase; however, the cases assessed in WLE and BLI in phase 2 were now displayed in WLE and LCI (n = 15), and the WLE and LCI cases were now displayed in WLE and BLI (n = 15).
The assessment phases were structured so that phases 2 and 3 both contained 15 cases of WLE + BLI and 15 cases of WLE + LCI. However, in our analyses, we restructured these 2 assessment phases so that assessment phase 2 contained all 30 cases of WLE + BLI, and assessment phase 3 contained all 30 cases of WLE + LCI. This enabled specific evaluation of the additive value of BLI and LCI, respectively. Therefore, in our analyses, phase 2 is considered as WLE + BLI and phase 3 as WLE + LCI.
In the fourth assessment phase, all 30 cases were displayed in WLE, BLI, and LCI in a side-to-side display. Assessors were asked to review all 3 images and complete the following items: (1) Paris classification; (2) appreciation of the Paris classification of the lesion using all 3 techniques (VAS scores); (3) appreciation of the surface relief of the lesion using all 3 techniques (VAS scores); (4) choice of best technique for assessment of macroscopic appearance (WLE, BLI, or LCI); (5) delineation of the lesion on the preferred image (WLE, BLI, or LCI); (6) ability to delineate the lesion using all 3 techniques (VAS scores); and (7) choice of the best technique to delineate the lesion (WLE, BLI, or LCI).
The 4 assessment phases were separated by a washout period of at least 2 weeks. The order of the images was randomized among assessment rounds and among assessors. Images were locked directly after assessment. Assessors were therefore not able to go back to previous images, and each assessment phase had to be completed in a single session.
Ground truth development: assessment by experts
In order to establish a ground truth for delineation, all 4 assessment phases were also completed by 3 experts (R.P., W.C., and R.B.; Fig. 2). In contrast to the assessments by nonexperts, experts delineated the lesions on all imaging modalities that were displayed in assessment phases 2 to 4, instead of a single delineation using the preferred modality. This approach was vital to enable ground truth comparison with each nonexpert delineation, regardless of preferred modality.
Figure 2Neoplastic lesion with ground truth delineations by 3 experts (blue, yellow, pink), visualization of the sweet spot (red line), assessor delineation (green), and the corresponding AND and OR areas (black lines) that were used to measure delineation performance.
To establish a combined ground truth for each image, the area with 3 overlapping expert delineations was labeled as “the sweet spot” and indicated the area with the highest likelihood of containing neoplasia (Fig. 2).
Quantification of nonexpert Assessor delineations
In order to develop a quantitative score for nonexpert delineations, the expert sweet spot was used as the ground truth. First, we calculated the percentage of the expert sweet spot that was included by the assessors’ delineation. Then, we calculated the “AND/OR” ratio for the sweet spot, as created in previous studies by our group.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
In this ratio, the AND area is defined as the area where the nonexpert delineation overlaps with the expert sweet spot. The OR area is defined as the remaining area of the nonexpert delineation without overlap with the sweet spot, plus the remaining sweet spot without overlap with the assessor delineation. Thereby, the AND/OR ratio is a quantitative measure for the extent of overlap with ground truth. Perfect similarity would thus lead to an AND/OR ratio of 1 and lack of any similarity to 0. Figure 2 shows a graphical display of these 2 areas.
Outcome measures
Primary outcome measures
Assessors’ appreciation of lesions and imaging modalities
•
Appreciation of macroscopic appearance in terms of surface relief and Paris classification (VAS scores; ranging from 1 to 10);
•
Choice of best technique to appreciate the macroscopic appearance (ordinal scores; ranging from −2 to +2);
•
Appreciation of ability to delineate the lesion (VAS scores; ranging from 1 to 10);
•
Choice of best technique for delineation (ordinal scores; ranging from −2 to +2).
Secondary outcome measures
Assessors’ delineation performance
•
Outcome 1: percentage of the expert sweet spot that was delineated by the assessor;
•
Outcome 2: AND/OR scores with the expert sweet spot.
Statistical analyses
Statistical analysis was performed using SPSS Statistical software package for Windows (version 25, SPSS Inc, Chicago, Ill, USA) and R version 3.4.0. For descriptive statistics, normally distributed data are shown as means (± standard deviation) and variables with skewed distribution are shown as medians (interquartile range). Because this was the first study to evaluate the role of BLI and LCI for the visualization of Barrett’s neoplasia by nonexpert endoscopists, no formal sample size calculation was conducted. To test differences in outcome parameters, paired t tests, Wilcoxon signed-rank tests, and McNemar tests were performed. Linear mixed-effect models were performed to evaluate the effect of the assessment phase on the delineation performance of the assessors. Image case numbers were included in the model as a random effect to take into account the correlation among measurements of the same image. To control for potential confounding, models were adjusted for the effect of country of origin and endoscopic experience.
Results
Thirty cases of early neoplasia in 30 patients were included in the online assessment. These were selected from a total of 68 collected cases. Thirty-eight cases were excluded because of insufficient image quality or lack of similarity among the WLE, BLI, and LCI images. All lesions were visualized in WLE, BLI, and LCI and were subsequently removed with endoscopic resection. Histologic evaluation of the endoscopic resection specimens showed HGD or EAC in all cases.
In total, 95 assessors from 3 countries started the assessment (Netherlands, 42; Germany, 39; Belgium, 14). These included trainees (27), junior general gastroenterologists (38), and senior general gastroenterologists (30). Nineteen assessors dropped out during one of the assessment phases and were therefore excluded from further participation. A total of 76 assessors (21 fellows, 32 juniors, 23 seniors) completed all 4 assessment phases and were included in the analyses.
Assessors’ appreciation of imaging modalities
Median VAS scores in assessment phase 2 (WLE + BLI), phase 3 (WLE + LCI), and phase 4 (WLE + BLI + LCI) were significantly higher compared with assessment phase 1 (WLE only) for appreciation of the Paris classification, the surface relief, and the ability to delineate the lesion (P < .001 for all outcome parameters) (Table 1).
Table 1Median visual analog scale scores of assessment phase 1 compared with assessment phases 2, 3, and 4 (Wilcoxon signed-rank tests)
Assessment phase 1: WLE
Assessment phase 2: WLE + BLI
Assessment phase 3: WLE + LCI
Assessment phase 4: WLE + BLI + LCI
VAS score (IQR)
VAS score (IQR)
P value
VAS score (IQR)
P value
VAS score (IQR)
P value
Appreciation Paris Classification
6 (4-7)
7 (5-8)
<.001
7 (5-8)
<.001
7 (6-8)
<.001
Appreciation surface relief
6 (4-7)
6 (5-8)
<.001
6 (5-8)
<.001
7 (6-8)
<.001
Ability to delineate the lesion
5 (4-7)
6 (5-7)
<.001
6 (5-8)
<.001
7 (5-8)
<.001
WLE, White-light endoscopy; BLI, blue-light imaging; LCI, linked-color imaging; VAS, visual analog scale; IQR, interquartile range.
In phases 2 and 3, assessors had to indicate their preferred technique for appreciation of macroscopic appearance and delineation when these techniques were visualized in a side-to-side display (WLE vs BLI, and WLE vs LCI). Assessors preferred BLI over WLE for appreciation of the macroscopic appearance and delineation of lesions (62% vs 12%, P < .001; and 66% vs 13%, P < .001, respectively). Assessors also preferred LCI over WLE for both outcomes (56% vs 13%, P < .001; and 62% vs 13%, P < .001, respectively) (Table 2).
Table 2Assessors’ ordinal scores in assessment phases 2 and 3: WLE versus BLI (phase 2) and WLE versus LCI (phase 3) (McNemar tests)
Assessment phase 2, ordinal score (%)
Assessment phase 3, ordinal score (%)
WLE
BLI
No preference
P value
WLE
LCI
No preference
P value
Best technique for appreciation of macroscopic appearance
In phase 4, assessors again had to indicate their preferred technique when all 3 modalities were displayed side to side (WLE vs BLI vs LCI). In 51% of all assessments, assessors preferred BLI for appreciation of macroscopic appearance, rather than WLE (19%) or LCI (31%). For delineation, assessors preferred to use BLI rather than WLE (13%) or LCI (38%) in 49% of assessments.
Assessors’ delineation performance
Ground truth delineations
In 1 case, no expert ground truth was established (ie, no overlap among all 3 experts delineations, and therefore no sweet spot). This case was therefore excluded from the analyses of the delineation performance of the assessors.
Outcome 1: Percentage of sweet spot delineated by the assessor
A linear mixed-effect model indicated that the combined addition of BLI and LCI to WLE in phase 4 was significantly associated with an increase in how much of the sweet spot was delineated by assessors (β = 0.024, P <.001). The addition of only BLI or LCI next to WLE was not significantly associated with an increase in scores. Adjusting for the effect of endoscopic expertise and country of origin did not change the results (Table 3).
Paired comparison between phase 1 and phase 4 showed a significant increase in scores, yet this increase was minimal (absolute difference 2%, P < .001) (Table 4).
Table 4Mean scores for both delineation outcome measures
Again, a linear mixed-effect model indicated that the combined addition of BLI and LCI to WLE in phase 4 was significantly associated with an increase in AND/OR scores (β = 0.078, P < .001). However, the addition of BLI and LCI individually was also significantly associated with this increase. Again, adjusting for the effect of endoscopic experience or country of origin did not change the results (Table 3).
Paired comparison between phase 1 and phase 4 showed a significant increase in scores (absolute difference 8%, P < .001) (Table 4).
Discussion
In this article, we describe the evaluation of the additional value of BLI and LCI for the visualization of Barrett’s neoplasia when assessed by nonexpert endoscopists. We previously described the evaluation of BLI by expert endoscopists, concluding that “BLI has additional value when used next to WLE for the visualization and delineation of BE neoplasia.”
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
To date, the value of BLI for visualization of Barrett’s neoplasia and the value of the recently introduced LCI technique, when assessed by nonexpert endoscopists, have not been interrogated.
Therefore, in the current study, 76 international nonexpert endoscopist assessors evaluated 30 cases of Barrett’s neoplasia visualized in WLE, BLI, and LCI in 4 separate assessment rounds. These assessors were shown to appreciate BLI and LCI, when visualized next to WLE, significantly better than WLE alone. VAS scores in assessment phases 2 to 4 were significantly higher than in phase 1 for all outcome parameters, indicating the preference of endoscopists to have the addition of optical chromoscopy techniques next to WLE. These results were supported by the strong preference for both BLI and LCI when the assessors were asked to choose among these techniques and for WLE for appreciation of macroscopic appearance and delineation.
Assessors were asked to delineate all lesions on the endoscopic images, using an online software tool enabling quantification of delineations. This allowed us to compare assessors’ delineations with the ground truth. Ideally, the ground truth on the endoscopic images is histopathologically correlated to the true demarcation line of a lesion. Although pathology reports showed all lesions to contain HGD or EAC, pixel-precise correlation with endoscopic images was not feasible. We reasoned that multiexpert input would provide the most reliable approximation of the true demarcation line. We therefore established ground truth by the creation of the expert sweet spot, a well-established method previously described by our group.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
The sweet spot was considered to have the highest likelihood of containing neoplasia, because all 3 experts indicated this area as neoplastic. We reasoned that, for an imaging technique to have a clinically relevant contribution, at least a large part of the sweet spot should be recognized by the assessor. We therefore developed outcome 1, the percentage of sweet spot that was delineated by the assessor. However, in theory, when an assessor would delineate the entire endoscopic image, he/she would reach maximum performance without actually being able to identify the sweet spot. Therefore, outcome 2, the sweet spot AND/OR score displays the ratio between the assessor correctly identifying the sweet spot versus delineating the surrounding mucosa that is less likely to contain neoplasia (ie, the area outside the sweet spot). A high AND/OR score indicates that a large portion of an assessor’s delineation overlaps with the sweet spot.
The results show that the addition of BLI and LCI next to WLE in phase 4 is associated with a significant increase in both delineation outcome measures compared with WLE alone in phase 1. Although statistically significant, this increase in scores was only modest in outcome 1. The absolute increase in AND/OR scores (outcome 2) was, however, substantial. These results indicate that the addition of BLI and LCI next to WLE leads to a higher AND/OR score compared with the use of WLE alone. The information provided by the addition of BLI and LCI allows for improved differentiation of the truly neoplastic tissue from its surrounding mucosa. When only BLI or LCI was used next to WLE, the ability to discriminate abnormal tissue from its surroundings also improved.
What are the clinical consequences of these delineation results? This is the first study to describe a quantifiable diagnostic additive value of optical chromoscopy for visualization of Barrett’s neoplasia, next to the perceived qualitative benefit. Although the minor increase in the percentage of delineated sweet spots is of marginal clinical relevance, the substantial increase in AND/OR scores is clinically relevant. Because these were nonexpert assessors, often not involved in treatment of neoplasia, the precise assessment of the true demarcation line is clinically of less importance. This is often done with the use of magnified endoscopy in expert centers. However, identification of the most abnormal area is of major clinical importance, because this allows the acquisition of well-placed biopsy samples, targeting the area most suspicious for neoplasia and thereby reducing the risk of sampling error. The significant increase in AND/OR scores shows that the addition of BLI, LCI, or BLI and LCI enables better identification of the most abnormal area of a lesion, thereby increasing diagnostic performance.
In this study, we have evaluated the use of BLI and LCI for the appreciation of macroscopic appearance of lesions and the ability to localize the most relevant parts of neoplastic lesions with optical chromoscopy. We therefore only included images with Barrett’s neoplasia. We did not aim to assess any increase in the primary detection rate by the use of optical chromoscopy because primary detection is mostly conducted in high-definition WLE.
We did not evaluate BLI or LCI alone (ie, without WLE). Our rationale was that in clinical practice, these techniques are always used in addition to WLE. We did, however, evaluate BLI and LCI next to WLE individually in assessment phases 2 and 3 to establish the effect of these techniques separately and attribute any significant effect in assessment phase 4 more specifically to either BLI or LCI. A clear effect of either one of these techniques, explaining the significant increase in scores in phase 4, was not found. We therefore reasoned that the combined information from BLI and LCI, next to the information in the WLE image, led to improved delineation performance. This improved performance might also be partly attributed to the availability of additional visual information, regardless of the specific optical chromoscopy technique.
This study has several limitations. First, selection bias might have occurred because all images were pre-selected and suitability for inclusion depended on quality and similarity among the 3 imaging techniques. Second, assessors were not completely naive to optical chromoscopy, because all modern-day endoscopes are equipped with some form of optical chromoscopy. Therefore, they might have had a pre-existing preference for one of these techniques. Third, this study relied on endoscopic images evaluated using an online software module. The additive value of optical chromoscopy techniques in a real-life endoscopic setting has not been interrogated in this study.
In conclusion, this study shows that the use of optical chromoscopy techniques has significant additional value for the visualization of Barrett’s neoplasia when used by nonexpert endoscopists. Assessors appreciated the addition of both BLI and LCI better than the use of WLE alone for assessment of macroscopic appearance and delineation. Furthermore, the addition of BLI and LCI next to WLE led to quantifiable improved delineation performance by increasing the ability to discriminate neoplastic tissue from its surroundings. This might allow for better targeted acquisition of biopsy samples in daily practice.
Acknowledgments
This study was supported by an unrestricted research grant from FUJIFILM Europe (FUJIFILM Europe GmbH, Düsseldorf, Germany), which had no involvement in the design, recruitment, data collection, analysis, interpretation, or writing of the manuscript.
References
Pech O.
May A.
Manner H.
et al.
Long-term efficacy and safety of endoscopic resection for patients with mucosal adenocarcinoma of the esophagus.
Blue-light imaging has an additional value to white-light endoscopy in visualization of early Barrett's neoplasia. an international multicenter cohort study.
DISCLOSURE: Dr Bergman has received researchsupportand speakerfeesfrom FUJIFILM. Dr Bisschops has received researchsupport, consultingfees, and speakerfeesfrom FUJIFILM. All other authors disclosed no financial relationships.
32579-9&id=gr1.jpg){kind=link}
32579-9&id=gr2.jpg){kind=link}