Introduction
Endoscopic mucosal resection (EMR) is a well-established, safe and cost-effective
technique for endoscopic treatment of precancerous gastrointestinal lesions and early-stage
cancer [1 ]
[2 ]. It features a rapid procedure time, a low risk of adverse events (AEs) and a relatively
low technical complexity [3 ]. However, when it comes to larger lesions, EMR is associated with a decreasing rate
of en bloc resections, resulting in increasing rates of recurrence. Particularly for
large sessile or laterally spreading polyps ≥ 2 cm, en bloc EMR is technically almost
impossible [4 ]. At that stage, en bloc resection of colorectal lesions can only be achieved in
about 30 % of cases [3 ]
[5 ]. Alternatively, EMR with a piecemeal technique can be performed, but it is associated
with higher recurrences rates compared to successful en bloc resection [4 ].
Endoscopic submucosal dissection (ESD) is a sophisticated endoscopic technique by
which complex and larger lesions can be successfully addressed [6 ]. ESD was initially developed in Japan for en bloc resection of flat gastric neoplasias
[7 ]
[8 ]. Today, ESD is also increasingly employed in the esophagus and for colorectal lesions
[2 ]
[9 ]. Especially for resection of laterally spreading polyps and flat lesions ≥ 2 cm,
ESD offers a reliable and minimally invasive method.
ESD has several disadvantages, however, as it is associated with a not insignificant
rate of AEs, particularly a higher rate of perforations, although most of them are
microperforations [3 ]. Furthermore, it demands more time, costs and resources, and is technically complex
even for experienced endoscopists. To date, there has been no broad application in
the Western world of ESD and its use is mainly limited to expert centers.
Therefore, novel endoscopic techniques are needed which can be clinically implemented
for en bloc resections of lesions ≥ 2 cm that are feasible and safe to use without
extensive training.
Meeting the mentioned requirements, classical EMR technique can be improved with a
new external additional working channel (AWC, Ovesco Endoscopy, Tuebingen, Germany)
that was recently introduced and termed “EMR+ technique” [10 ]. To this end, the AWC is mounted on a standard endoscope similar to the setup used
with the full-thickness resection device (FTRD) [11 ].
So far, EMR+ has not been systematically evaluated. Therefore, we provide the first
data in terms of usefulness and feasibility of this novel method. The aim of the current
study was to prospectively compare the novel technique EMR+ to the gold standard of
classical EMR using a preclinical ex vivo porcine animal model. We chose en bloc resection
rate, procedure time, and resection area in consideration of the complication rate
as clinical endpoints to assess for which lesions EMR+ would be particularly appropriate.
Materials and methods
The study was designed as a prospective ex vivo trial. It was exempt from IRB review
because no humans or living animals were included. The experiments were conducted
at the research unit of the Department of Gastroenterology and Gastrointestinal Oncology,
University Medical Center Goettingen, Germany.
Pig stomachs were cleaned and frozen. Prior to the procedure, they were defrosted
and placed into the EASIE-R simulator (Endosim, LLC, Hudson, Massachusetts, United
States). The EASIE-R stimulator is an established model for research and interventional
endoscopic training and has already been evaluated for several endoscopic procedures
[12 ]
[13 ].
All interventions (EMR and EMR+) were performed by two well-trained endoscopists with
previous experience in EMR and EMR+ technique in humans and animal models.
Additional working channel
The AWC ([Fig. 1 ]) has a flexible attachment, a shaft with a length of 122 cm (endoscope insertion
length: 103 – 110 cm), an adaptor for fixation at the endoscope handle with Luer-lock,
a valve and a sleeve with adhesive tape. The AWC can be mounted on endoscopes with
a diameter from 8.5 to 13.5 mm. Instruments with an outer diameter of up to 2.8 mm
can be introduced. All AWC procedures were performed with the AWC in the counterpart
position to the working channel.
Fig. 1 Principles of the EMR+ procedure. a Target lesion. b Submucosal injection. c Positioning of snare and grasper. d Elevation of the lesion and snare closure. e Pushback of the grasper while snare stays closed followed by resection. (Source:
with permission from Ovesco Endoscopy AG, Tuebingen, Germany)
EMR and EMR+ procedure
We selected standardized lesions, measuring 1 cm, 2 cm, 3 cm or 4 cm. Prior to intervention,
lesions in the corpus of the porcine stomach were manually marked with coagulation
dots inside a template. Then, the stomach was transferred into the EASIE-R model.
The porcine esophagus and stomach were fixed to the modelʼs plastic shell [13 ].
After submucosal injection (hydroxyethyl starch with methylene blue), in all resections,
a 33-mm snare (Boston Scientific, Malborough, Massachusetts, United States) was positioned
around the lesion. A grasper (Ovesco Endoscopy, Tuebingen, Germany) was used for EMR+,
which was introduced via the AWC. This can also be set the other way with the snare
introduced via the original working channel and the grasper in the AWC as demonstrated
in [Fig. 1 ].
If adequate positioning of a lesion required more than 15 minutes, the procedure was
stopped and results were classified as failed en bloc resection. [Fig. 2 ] illustrates the use of the EMR+ technique and a resected specimen in the ex-vivo
model.
Fig. 2 Application of EMR+ in the ex-vivo model. a Lesion to be targeted after submucosal injection. b Positioning of snare and grasper. c Elevation of the lesion and snare closure. d Post-interventional site of resection. d Resected specimen after EMR+.
Data collection
The following parameters were recorded by an independent observer: Prepared lesion
size (1 cm, 2 cm, 3 cm or 4 cm), rate of en bloc resection (R0, defined by all marking
dots retrieved on the resected specimen), time of EMR and EMR+ procedure, resection
area (in cm2 ), AEs (perforations).
After each EMR and EMR+, the specimens were spread out, pinned on cork plates and
photographed. EMR and EMR+ procedure time was defined from submucosal injection to
complete resection of the lesion. Following visual evaluation of each resection site,
an insufflation test was performed to determine the presence of a perforation.
Statistical analysis
Data analysis was performed using SPSS (IBM). Chi-square-test was used for statistical
analysis of en bloc resection rates. Mann-Whitney-U-Test was used for the analysis
of time of procedure and resection area. P < 0.05 was considered statistically significant and is marked by *.
Results
In both the EMR and EMR+ groups, lesions with four different sizes were set with a
diameter of 1 cm (n = 12 per group), 2 cm (n = 22 per group), 3 cm (n = 22 per group)
and 4 cm (n = 20 per group). In total, 152 endoscopic procedures (76 EMR, 76 EMR+)
were performed in the ex vivo porcine model by two experienced endoscopists. Overall,
we used 15 stomachs, each with eight to 12 lesions, dependent on stomach and lesion
sizes. An overview of the study design is presented in [Fig. 3 ].
Fig. 3 Study design.
Rate of en bloc resection
In 1-cm lesions, EMR achieved an en bloc resection rate (R0) of 100 % (12 of 12).
In 2-cm lesions, it was 54.55 % (12 of 22), decreasing to 18.18 % (4 of 22) in 3-cm
lesions and ending with 0 % (0 of 20) in 4-cm lesions. In 1-cm lesions, EMR+ resulted
in an en bloc resection rate of 100 % (12 of 12). Compared to EMR, EMR+ en bloc resection
rate was significantly higher in 2-cm lesions (95.44 % (21 of 22) vs. 54.55 %, P = 0.02*), in 3-cm lesions (86.36 % (19 of 22) vs. 18.18 %, P < 0.01*) and also in 4-cm lesions (60.00 % (12 of 20) vs. 0 %, P < 0.01*). These data are presented in [Fig. 3 ] and [Fig. 4 ].
Fig. 4 Rate of en bloc resection (R0).
Procedure time
In 1-cm lesions, median procedure time in the EMR-group was 2.50 minutes (SD 2.45).
In 2-cm lesions, it was 6.00 minutes (SD 2.62), in 3-cm lesions, it was 12.50 minutes
(SD 3.89) and in 4-cm lesions, median procedure time was 15.00 minutes (SD 0.58).
The difference in procedure times with EMR+ in 1-cm and 2-m lesions did not reach
significance in the statistical analysis (median 2.50 minutes (SD 0.90) vs. 2.50 minutes
(SD 2.45, P = 0.74; median 4.00 minutes (SD 2.34) vs. 6.00 minutes (SD 2.62), P = 0.16). In contrast, procedure time in 3-cm and 4-cm lesions was significantly lower
in the EMR+ group compared to EMR (5.00 minutes (SD 3.38) vs. 12.50 minutes, P < 0.01*; 5.50 minutes (SD 2.69) vs. 15.00 minutes, P < 0.01*). These findings are presented in [Fig. 5 ].
Fig. 5 Duration of EMR and EMR+ procedure.
Resection area
Median EMR resection area for 1-cm lesions was 3.14 cm2 (SD 1.19). In 2-cm lesions, it was 3.30 cm2 (SD 1.55). At 3 cm it was 1.50 cm2 (SD 2.81) and in 4-cm lesions, median resection area was 4.02 cm2 (SD 1.61) for EMR. In lesions of all sizes, the resection area was significantly
larger in the EMR+ groups. In 1-cm lesions, the median resection area was 4.44 cm2 for EMR+ (SD 1.56, P = 0.012*). It was 5.94 cm2 (SD 3.91, P < 0.01*) at 2 cm and 9.62 cm2 (SD 3.98, P < 0.01*) at 3 cm. In 4-cm lesions, EMR+ median resection area was 13.37 cm2 (SD 7.67, P = 0.03*) ([Fig. 6 ]).
Fig. 6 Resection area (in cm2 ).
Perforations
In all EMR groups, no perforations occurred, whereas in the EMR+ group only in 4-cm
lesions, three perforations were observed (15 %, 3 of 20).
Discussion
With its broad availability, EMR is a well-established minimally invasive technique
for treatment of dysplastic and early malignant gastrointestinal lesions. However,
when it comes to treatment of large, particularly flat lesions, EMR has certain limitations.
For potentially challenging en bloc resection of flat lesions ≥ 2 cm, ESD offers a
reliable and oncologically convincing method [6 ]. But ESD comes with a not insignificant rate of AEs, especially perforations [3 ]. ESD is expensive, time-consuming and technically complex as it involves a long
learning curve even for experienced endoscopists [3 ]. For these reasons, in the Western world, ESD is mostly limited to expert centers
[3 ]. EFTR has especially advanced management of complex colorectal precancerous and
malignant lesions without lifting sign, e. g. due to scarring [14 ]
[15 ]. However, its applicability is practically limited to the lower gastrointestinal
tract and its diameter is not appropriate for flat and spreading lesions, which are
object of this study [16 ]
[17 ]
[18 ].
Bearing this in mind, EMR+ was recently developed and is based on an additional working
channel. An effective grasp and snare technique using a dual-channel endoscope has
already been described [19 ]
[20 ]
[21 ]. However, its practicality is limited due to the close and fixed distance between
the two working channels, which results in a lack of sufficient triangulation, flexibility,
and overview. Furthermore, a dual-channel endoscope is an expensive investment for
endoscopy units and is consequently not always available.
In analogy to the FTRD [11 ], the AWC is mounted at the tip of a standard gastroscope or pediatric colonoscope,
thus making a dual-channel endoscope dispensable. By turning its cap, more variable
and wider positions of both working channels (AWC plus standard channel) can be achieved
in contrast to a dual-channel endoscope [10 ]. If required, this leads to better visibility and more flexible triangulation of
the instruments.
First, the results of our prospective ex vivo study represent the frequently described
advantages and limitations of classical EMR, e. g. implemented in the European guidelines
[1 ]. In analogy to the clinical situation, also in our ex-vivo porcine animal model,
EMR is very reliable for lesions of 1 cm with an en bloc resection rate of 100 %.
For 2-cm lesions, that drops to 55 %, whereas classical EMR does not provide satisfying
resection rates for 3-cm or 4-cm lesions (18 % and 0 %, respectively). Thus, our model
seems to represent the clinical reality with a good discriminatory power in terms
of the gold standard EMR, showing insufficient results in lesions ≥ 2 to 4 cm [1 ].
In defined 1-cm lesions, there was no difference between EMR and EMR+ in terms of
en bloc resection rate. However, en bloc resection rates with conventional EMR decreased
rapidly for defined 2-cm lesions. Lesions measuring 3 cm showed a poor en bloc resection
rate for EMR compared to EMR+, the latter revealing satisfying results at 86 %. In
4-cm lesions, the en bloc resection rate for EMR+ dropped to 60 %, whereas it was
0 % in the EMR group. Obviously, 4-cm lesions basically cannot be addressed with conventional
EMR. Hence in our animal model, the critical size of lesions to be successfully removed
by conventional EMR corresponds to the data in humans [1 ]
[3 ]
[4 ]
[5 ].
Furthermore, procedure time was not longer if EMR+ was used. Conversely, starting
at a size of 2 cm, procedure time for lesions was significantly shorter in the EMR+
subgroups compared to conventional EMR. Based on the technical aspects of EMR+, larger
resection areas were achieved in all groups, starting with 1-cm lesions.
In 4-cm lesions, few perforations occurred in the EMR+ group. Because conventional
EMR becomes completely insufficient in lesions of this size, they cannot be compared
in terms of use of EMR versus EMR+. However, our data suggest that 4-cm lesions also
seem to be too big for EMR+.
Our prospective study was conducted in an established and well-evaluated ex vivo animal
model. However, there are certain limitations that may limit the transferability from
the porcine ex vivo model to humans. First, to ensure a standardized experimental
setup, we consistently used a 33-mm snare for both EMR and EMR+ for all lesions. Snare
size potentially affects en bloc resection rate. Second, the porcine stomach has higher
mucosal rigidity than human gastric mucosa, which affects the technical opportunities
for EMR and EMR+. Obviously, bleeding, tissue movement, histopathological evaluation
and other physiological factors cannot be reproduced in our ex vivo model. Principally,
damage may occur to the specimen as a result of grabbing the lesion with the forceps.
Our data clearly demonstrate that EMR+ is a novel resection technique with potential
to improve en bloc resection rates for snare-based polypectomy in flat lesions larger
than 1 cm. In terms of the high number of neoplastic flat lesions seen in the right
colon, future reports on how AWC will work in difficult colonoscopy cases (tortuous
sigmoid colon; severe diverticulosis; long, mobile, looping colons) are needed.
Conclusion
The newly developed EMR+ technique with AWC facilitates en bloc resection of larger
lesions compared to conventional EMR, the well-described gold standard for lesions
< 2 cm. Consistently, we did not observe an additional benefit of EMR+ in these lesions
whereas beginning at 2 cm, EMR+ had advantages over conventional EMR, reaching its
best discriminatory power over conventional EMR at 3 cm. Because EMR+ reaches its
inherent limits at 4 cm, accompanied by a rising risk of perforations, ESD or surgery
should be considered in lesions larger than that size.
In our ex-vivo porcine model, we could show that EMR+ works and can significantly
and relevantly increase the size of lesions to be successfully resected en bloc. Consequently,
using EMR+, lesions can be resected en bloc that otherwise could only be addressed
via ESD or surgery. Therefore, EMR+ could help to close a therapeutic gap with manageable
technical complexity, time, and costs.
The AWC device allows easy transformation of a standard single-channel endoscope to
double-channel functionality, leading to good opportunities for bimanual work by triangulation
with better intraluminal resections accompanied by more efficient tissue traction.
The AWC could also be used for other indications, e. g. for ESD procedures (“ESD+”).
