Introduction
The growth of diagnostic endoscopy has been facilitated by numerous advances in imaging
technology. These include evolution from fiberoptic imaging to generation of images
using a charge-coupled device and high definition digital imaging. Enhanced endoluminal
imaging techniques have included chromoendoscopy and modalities that aspire to optical
biopsy.
Among techniques for enhanced optical diagnosis, narrow band imaging (NBI) is a proprietary
imaging modality in which the endoscope processor filters standard white light to
specific wavelengths in the blue – green spectrum (415 nm and 540 nm). NBI thereby
capitalizes on the peak absorption of hemoglobin and has the ability to accentuate
visualization of the mucosal vasculature [1]. Proposed clinical applications of NBI include endoscopic evaluation of Barrett’s
esophagus and endoscopic diagnosis of colorectal polyps. Studies of NBI in endoscopic
inspection of Barrett’s esophagus have demonstrated high sensitivity of NBI in detection
of Barrett-associated high grade dysplasia [2], and the ability of NBI to detect dysplasia in a higher proportion of patients with
fewer biopsy samples compared with standard white light endoscopy [3]. With respect to colorectal polyps, an NBI-based classification scheme has been
developed which may accurately distinguish adenomatous from hyperplastic polyps [4], although it is not certain that use of NBI improves polyp detection rates [5]
[6]
[7].
Despite data regarding potential gastrointestinal endoscopic applications of NBI,
the degree to which NBI use has been adopted into clinical practice is unknown. As
with any new medical technology, legitimate questions exist regarding the degree to
which efficacy of NBI as demonstrated in clinical studies will translate into effective
use of NBI in actual practice [8]. The aim of this study, therefore, was to prospectively define the rate of NBI use
among patients referred to a large group endoscopy practice for diagnostic endoscopy
(esophagogastroduodenoscopy and colonoscopy), and to identify procedural factors associated
with NBI use.
Methods
Approval to conduct this study as a quality assurance protocol was granted by the
Institutional Review Board at the study institution.
This study was conducted at the endoscopy center of a tertiary care academic center,
where both inpatient and outpatient procedures are performed in a hospital-based suite.
Each procedure room is equipped with NBI capability (180 series gastroscope or colonoscope,
CV-180 Evis Exera II video processor, and CLV-180 light source; Olympus Medical Systems,
Tokyo, Japan).
Elective diagnostic procedures were prospectively observed over a 2-week period. Each
consecutive esophagogastroduodenoscopy (EGD) and colonoscopy was directly observed
by in-room endoscopy technicians, who assist the endoscopist with equipment setup
and when endoscopic accessory use (i. e. biopsy forceps, polypectomy snare, endoscopic
hemostatic device, etc.) is required. A technician is present in each procedure room
for the entire duration of each procedure. For each eligible procedure, the technician
documented whether or not NBI was used. Faculty endoscopists observed during this
time period were unaware of the study. Procedures performed by the study author were
not included; otherwise all faculty endoscopists were eligible for observation.
The following endoscopic procedures were excluded from the analysis: EGD or colonoscopy
with therapeutic intent, specifically endoscopic dilation, endoluminal stent maneuvers,
or delivery of endoscopic hemostatic therapy; EGD for nasoenteral or percutaneous
feeding tube placement; balloon- or spiral-assisted small-bowel enteroscopy; endoscopic
retrograde cholangiopancreatography; EGD with endoscopic ultrasound; ileoscopy and
pouchoscopy.
Following completion of the 2-week observational study period, operative reports including
reports generated by Olympus Endoworks software and dictated operative notes were
manually reviewed by the study author, who was not blinded to study design or intent.
The following data were extracted: procedure type, procedure indication, identification
of attending endoscopist, trainee involvement, patient type (inpatient versus outpatient),
procedure start time (a.m. vs. p.m.), sedation-type (endoscopist-directed conscious
or deep sedation vs. monitored anesthesia care with anesthesia staff support); performance
of endoscopic polypectomy, either by snare or forceps biopsy; performance of endoscopic
tissue biopsy for intent other than polypectomy.
Extracted data were stored in a Microsoft Excel spreadsheet, and statistical analyses
were performed using JMP 10.0.0 software (SAS Institute, Cary, NC, USA). The chi-squared
or Fisher’s exact test was used for comparison of categorical variables. Two-sided
P values of < 0.05 were considered statistically significant.
Results
During the two-week observational study period in July – August 2013, data regarding
NBI use were recorded for 318 elective endoscopic procedures, consisting of 106 EGDs
and 212 colonoscopies. Additional procedural data are summarized in [Table 1].
Table 1
Procedure data for 318 elective endoscopic procedures in a large group endoscopy practice
|
n (%)
|
|
Procedure type
|
|
|
EGD
|
106 (33)
|
|
Colonoscopy
|
212 (67)
|
|
Sedation type
|
|
|
Conscious sedation
|
197 (62)
|
|
MAC
|
121 (38)
|
|
Patient type
|
|
|
Outpatient
|
305 (96)
|
|
Inpatient
|
13 (4)
|
|
Procedure start time
|
|
|
a.m.
|
195 (61)
|
|
p.m.
|
123 (39)
|
|
Trainee involvement
|
|
|
Yes
|
14 (4)
|
|
No
|
304 (96)
|
Data presented as N (%).
EGD, esophagogastroduodenoscopy; MAC, monitored anesthesia care.
The most common indications for EGD were evaluation of abdominal pain/dyspepsia (22 %)
and evaluation of gastroesophageal reflux disease (19 %). Additional indications included
evaluation of weight loss, anemia or suspected gastrointestinal bleeding, and diarrhea.
The most common indication for colonoscopy was screening/surveillance for colorectal
cancer (56 %). Additional indications included evaluation of suspected or established
inflammatory bowel disease, anemia or suspected gastrointestinal bleeding, and diarrhea.
NBI use was observed in 6.6 % (21/318) of procedures. This included use of NBI in
4.7 % (5/106) of EGDs and 7.5 % (16/212) of colonoscopies (P = 0.47 for comparison). No difference in rate of NBI use was found when comparing
EGD with or without biopsy (5.5 % [3/55] vs. 3.9 % [2/51]; P = 1), or when comparing colonoscopy with or without biopsy (10.5 % [8/76] vs. 5.9 %
[8/136]; P = 0.28). NBI use was significantly higher in colonoscopy with polypectomy when compared
with colonoscopy without polypectomy (13 % [10/77] vs. 4.4 % [6/135]; P = 0.03) ([Fig. 1]).
Fig. 1 Rates of narrow band imaging (NBI) use for all study procedures, esophagoduodenoscopy
(EGD) with and without biopsy, colonoscopy with and without polypectomy. NS, not significant.
On univariate analysis, there was no association between patient type (outpatient
vs. inpatient), procedure start time, sedation type, or trainee involvement and use/non-use
of NBI. For both EGD and colonoscopy, there was no association between procedure indication
and use/non-use of NBI. NBI use was observed in zero of seven EGDs with a documented
primary procedural indication of screening for or surveillance of Barrett’s esophagus.
NBI use was observed in 2 of 9 (22 %) colonoscopies with a documented primary indication
of surveillance in the setting of chronic ulcerative colitis. No use of an alternative
imagine modality, such as methylene blue chromoendoscopy or high magnification endoscopy
was described in any operative report.
Observed endoscopic procedures were performed by 23 faculty endoscopists, with a procedure
volume ranging from 1 to 58 per endoscopist during the 2-week study period. The rate
of NBI use ranged from 0 to 100 % of procedures per endoscopist. After excludsion
of endoscopists who performed fewer than five endoscopies during the study period,the
rate of NBI use for the remaining 15 endoscopists ranged from 0 to 23 %. There was
a significant difference in rate of NBI use (P < 0.01 for overall comparison) amongst these 15 endoscopists. NBI use among the 5
highest volume endoscopists ranged from 0 to 16 %.
Among cases with observed NBI use, photodocumentation of NBI use was present in 24 %
(5/21) of operative reports. There was no text or written documentation of NBI use
in any operative report.
Discussion
This observational study detected use of NBI in 6.6 % of elective endoscopic procedures.
The highest rate of NBI use was observed in colonoscopies with polypectomy (13 %),
and this rate was significantly higher than that observed in colonoscopies without
polypectomy. This study also detected a significant difference in the rate of NBI
use in global comparison of individual endoscopists. Specific factors that influence
the adoption of NBI use into routine practice by individual endoscopists were not
examined in this study.
Endoscopists were not asked to self-report NBI use and were not informed of the study.
The study was designed in this fashion to avoid the potential of a Hawthorne effect,
wherein performance may be influenced in subjects who are aware that they are being
observed. The endoscopists were blinded to the study, making the study design and
findings unique. Published data regarding rate of NBI use in actual practice are virtually
nonexistent. A survey of European university hospitals reported use of NBI or alternative
commercial enhanced-imaging technology in 67 % of institutions in the evaluation of
Barrett’s neoplasia [9], but did not report a per-case use rate.
In the current study, there was no a priori hypothesis regarding the rate of NBI use
expected in this observational study, as there are limited standard recommendations
or guidelines for routine NBI use in diagnostic endoscopy, and existing guidelines
are open to flexible interpretation. For instance, the American Gastroenterological
Association medical position statement on Barrett’s esophagus suggests that chromoendoscopy
or electronic chromoendoscopy is not necessary in the routine endoscopic surveillance
of Barrett’s esophagus, but may be helpful in guiding biopsies for patients with dysplasia
or visible mucosal abnormalities [10]. The rate of NBI use in this study can therefore not be subjectively designated
as high or low, but instead serves as a baseline metric in this study setting against
which future NBI use can be measured.
Adoption of NBI use in routine diagnostic endoscopy may have implications for application
of future enhanced-imaging technologies. In the hands of the endoscopist, NBI is a
fast and efficient enhanced-imaging technology. No additional equipment is required
other than an existing video monitor and NBI-equipped scope and processor. An endoscopist
can toggle between NBI and standard white light literally in seconds, without use
of additional endoscopic devices, accessories, or medications. And while there may
be a learning curve for interpretation of NBI images [11]
[12], endoscopist interpretation of endoluminal NBI images would seem less a departure
from white light endoluminal images than interpretation of images generated by other
optical techniques, such as endomicroscopy, that focus on cellular structures. Based
on these criteria, one would speculate that adoption and use of imaging technologies
requiring additional equipment or capital investment, additional time with respect
to procedure duration, and/or increasingly complex image analysis would be lower than
that for NBI.
Finally, it is worthwhile to note that documentation of NBI use was uncommon. Photodocumentation
of NBI use was provided in 24 % (5/21) of operative reports from cases in which NBI
use was observed. There was no text or written documentation of NBI use in any operative
report. Future consideration may be warranted as to whether documentation guidelines
should exist for procedural use of NBI or other adjunct imaging modalities. An American
Society of Gastrointestinal Endoscopy Preservation and Incorporation of Valuable endoscopic
Innovations (PIVI) statement on real-time endoscopic assessment of histology of colon
polyps suggests that lesion photocumentation is necessary if a resect-and-discard
strategy is to be implemented [13].
While extending the study duration to increase sample size would increase statistical
power and eliminate the possibility of type II error in examining factors associated
with NBI use, the study as completed is likely to have strong internal validity, in
that the observed range of endoscopic procedures during the 2-week time period is
likely to offer an accurate representation of the range of endoscopic practice in
this setting. It may not be possible, however, to generalize the study findings to
other institutions or endoscopy settings which may have differing practice patterns.
Limitations of the current study include the potential for misclassification of NBI
use. Instances of NBI use may not have been documented if not observed or not recognized
by the endoscopy technician. An alternative would have been to video record the entire
procedure for subsequent review; however maintenance of adequate blinding with this
approach would be challenging. Documented cases of NBI use cannot distinguish between
intentional and unintentional NBI use – for instance, if the endoscopist had inadvertently
pressed the button on the scope handle for application of NBI when he/she had intended
instead to press the button for image capture. In addition, there was limited heterogeneity
in some of the procedural variables (e. g. trainee involvement) to adequately assess
for potential association with NBI use/non-use. A post hoc survey of participating
endoscopists might offer insight as to whether their opinions on the role of NBI match
their actual practice, but has no bearing on the aim of this study, which was to objectively
document rate of NBI use.
In summary, NBI use was observed in 6.6 % of elective endoscopic procedures. Use of
NBI was highest in colonoscopies with polypectomy. Rate of NBI use varied significantly
among endoscopists. Additional large-scale prospective data are needed to assess the
magnitude of impact of NBI on routine endoscopic practice.
