Introduction
Esophageal cancer is the sixth most common cause of cancer-related mortality worldwide
[1]. The overall survival of patients with esophageal cancer remains poor. However,
a favorable prognosis can be expected if this cancer is detected at an early stage
[2]
[3]
[4]. Endoscopic submucosal dissection (ESD) was developed in Japan and has been performed
on many patients with early stage esophageal cancer because it is minimally invasive
and offers excellent results [5]. Subsequently, ESD has been recognized as one of the standard treatments for superficial
esophageal carcinoma.
Strictures are a major complication of esophageal ESD. The frequency of strictures
for a high risk lesion, i. e. a mucosal defect covering more than three-quarters of
the esophageal circumference, is 70 – 90 % [6]
[7]
[8]. It substantially decreases patients’ quality of life and requires multiple endoscopic
balloon dilation sessions. This complication prevents the use of ESD for larger lesions.
Several methods, including local steroid injection, systemic oral steroids, polyglycolic
acid sheet, and tissue-engineered cell sheets, were developed to prevent strictures
[9]
[13]
[14]
[15]
[16]. Although they are partly effective, strictures are still a significant complication
of ESD. If strictures are completely prevented, ESD will be indicated for larger lesions
that were previously treated by esophagectomy.
Post ESD, a stricture develops during the process of scar formation. Scar formation
is thought to be an integral part of wound healing, starting with inflammation and
then proliferation, and remodeling [9]. During the remodeling process, the elasticity and compliance of the esophagus are
reduced by fibrosis, which may lead to stricture formation [10]
[11]
[12]. Previous approaches to prevent strictures were targeted at the process after the
occurrence of inflammation [9]
[13]
[14]
[15]
[16]. For effective prevention of strictures, the process generating inflammation may
also be a good target. To suppress the generation of inflammation, some modification
is required to the ESD process. Electric current is a major factor contributing to
the generation of inflammation during ESD and can be modified by controlling the electrosurgical
unit (ESU). We therefore planned a study to investigate the impact of ESU modes on
the development of strictures.
Methods
Experimental animals and methods
We performed this study in Kobe Medical Device Development Center, after approval
from the animal ethics committee of the Intervention Technical Center (IVTeC) Co.,
Ltd. (Tokyo, Japan) which supports experiments on animals. An in vivo porcine model,
four female adult Landrace-Wide Yorkshire-Duroc, was used to approximate the human
situation. On the day of the procedure, they were fasted but allowed full access to
water. General anesthesia was induced using ketamine hydrochloride (10 mg/kg), atropine
sulfate, xylazine hydrochloride (2 mg/kg), and isoflurane. For each pig, we made virtual
target lesions by marking oval-shaped dots that were about three-quarters of the circumference
of the esophagus and their longitudinal diameters were about 5 mm. Virtual target
lesions were made at sites 32 cm, 37 cm, 42 cm, and 47 cm from the mouth.
Endoscopic submucosal dissection
An endoscope with water-jet function (EVIS GIF-Q260 J, Olympus, Tokyo, Japan) with
a distal attachment cap (D-201-11804; Olympus) was used for all procedures. All procedures
were conducted by one endoscopist (R.I.) who had experienced more than 300 esophageal
ESD procedures. A 0.9 % saline solution was injected into the submucosa using a 23-gauge
endoscopic injection needle (01841, Top Corporation, Tokyo, Japan) to elevate the
lesion. A ball-tip Flush knife (BT) (1.5 mm, DK2618JB; Fujifilm Medical, Tokyo, Japan)
was used for mucosal incision and a Hook knife (KD-620LR; Olympus Medical, Tokyo,
Japan), a monopolar endo-knife, was used for submucosal dissection. A ball-tip Jet
B-knife (1.5 mm, BSJB15B; Zeon Medical, Tokyo, Japan), a bipolar endo-knife, was used
for submucosal dissection.
A VIO electrosurgical generator (VIO 300D; ERBE, Tübingen, Germany), one of the most
popular ESUs, was used for all ESD procedures [17]. ENDO CUT mode was used for marginal incision in all lesions. In the current study,
we selected the ESU modes commonly used for submucosal dissection. The settings used
for submucosal dissection with the monopolar endo-knife were: ENDO CUT; SWIFT COAG;
FORCED COAG E2; FORCED COAG E3; FORCED COAG E4; and SPRAY COAG ([Table 1]). The setting used for submucosal dissection with the bipolar endo-knife was FORCED
COAG. The detailed settings and peak voltages of these modes are shown in [Table1]. The esophageal location where each ESU mode was applied is shown in [Table2]. When hemorrhage was observed, the bleeding point was coagulated with Coagrasper
(FD-411QR; Olympus Medical, Tokyo, Japan), using the SOFT COAG mode (Effect 6, 80 W).
Table 1
Details and peak voltage of each electrosurgical unit (ESU) mode.
|
ESU mode
|
Details of settings
|
Peak voltage, V
|
|
Monopolar
|
|
|
ENDO CUT I, Effect2 Duration2 Interval3
|
550
|
|
|
SWIFT COAG, Effect3, 40 W
|
990
|
|
|
FORCED COAG, Effect2, 40 W
|
1100
|
|
|
FORCED COAG, Effect3, 40 W
|
1430
|
|
|
FORCED COAG, Effect4, 40 W
|
1800
|
|
|
SPRAY COAG, Effect1, 40 W
|
3520
|
|
Bipolar
|
|
|
Bipolar, FORCED COAG, Effect2, 50 W
|
1100
|
Table 2
Esophageal location where each electrosurgical unit (ESU) mode was applied.
|
Distance from the mouth
|
Pig 1
|
Pig 2
|
Pig 3
|
Pig 4
|
|
32 cm
|
FORCED COAG E4
|
SPRAY COAG
|
FORCED COAG E2
|
Bipolar, FORCED COAG
|
|
37 cm
|
FORCED COAG E3
|
FORCED COAG E2
|
SWIFT COAG
|
ENDO CUT
|
|
42 cm
|
FORCED COAG E2
|
FORCED COAG E3
|
ENDO CUT
|
SWIFT COAG
|
|
47 cm
|
SPRAY COAG
|
FORCED COAG E4
|
Bipolar, FORCED COAG
|
FORCED COAG E2
|
Assessment of stricture and fibrosis
One month after ESD, all pigs were examined by endoscopy to review the resection site
and then humanely killed. After esophagectomy by a senior laboratory animal technician,
the esophagus was opened longitudinally. The length of the normal part of the esophagus
and the length of the ESD scars were both measured in the minor axis direction of
the esophagus ([Fig. 1]). The quotient of stricture (%) was calculated using the formula: 100 × (Maximum
length of the normal part of each esophagus – Length of ESD scar)/Maximum length of
the normal part of each esophagus. The quotient of stricture was compared among electrocautery
modes.
Fig. 1 Macroscopic view of the whole esophagus from Pig 3. The red line is the maximum length
of the normal part of the esophagus and the yellow lines are the lengths of ESD scars,
all measured in the minor axis direction of the esophagus.
After the assessment of stricture, the esophageal tissue was stretched, pinned onto
a corkboard, and preserved in formalin. Then esophageal scars were cut up in the minor
axis direction of the esophagus and stained using hematoxylin and eosin (HE) stain,
desmin stain, and azan stain. Fibrosis of the muscle layer was assessed and compared
among electrocautery modes. The quotient of fibrosis (%) was calculated using the
formula: 100 × maximal thickness of fibrosis in the muscle layer/thickness of the
muscle layer.
Data are expressed as mean (standard deviation). The association of stricture with
various factors such as location, resected specimen size, frequency of hemostasis,
and procedure time was tested using the Spearman correlation coefficient.
Results
Clinical course during and after ESD
The mean weight of the four pigs was 35.2 kg (32.9 – 36.2 kg). A total of 16 ESD,
four ESD in each pig, was conducted without any complications ([Fig. 2]). Vital status was stable during ESD in all pigs. All pigs resumed water on the
day of the procedure and were placed on fluid food from postoperative day 1 and on
solid food from postoperative day 3. All pigs were kept alive with an uneventful clinical
course in terms of food intake, liquid intake, and physical activity. Endoscopy at
1 month, just before euthanasia, showed scar formation in all pigs ( [Fig. 3]). An endoscope (EVIS GIF-Q260 J), 9.9 mm in diameter, could not pass the stricture
on the oral side in Pigs 1, 2, and 4, and the stricture on the anal side in Pig 3.
Fig. 2 Endoscopic view just after ESD shows a lesion 32 cm from the mouth of Pig 3.
Fig. 3 Endoscopic view 4 weeks after ESD shows a scar 32 cm from the mouth of Pig 3.
Assessment of strictures and fibrosis
Resected esophagus showed complete mucosal regrowth and scar formation in all pigs.
Mild to severe stricture was observed at the site of endoscopic resection ( [Fig.1]). The quotient of stricture is summarized in [Table 3]. ENDO CUT mode showed the lowest mean quotient of stricture among all modes (16 ± 1 %).
SWIFT COAG mode showed a slightly lower quotient of stricture (28 ± 28 %) than the
other coagulation modes. The severity of fibrosis in the muscle layer is summarized
in [Table 3]. ENDO CUT mode showed very mild fibrosis (7 ± 5 %) ( [Fig. 4a]). Mild to severe fibrosis was observed at all resection sites ([Fig. 4b]). SWIFT COAG mode showed slightly lower fibrosis (28 ± 4 %) than the other coagulation
modes. Bipolar FORCED COAG (100 ± 0 %) mode and SPRAY COAG (63 ± 53 %) mode showed
severe fibrosis. The quotients of stricture for lesions located 32, 37, 42, and 47 cm
from the mouth were 49.3 %, 24.4 %, 31.6 %, and 38.6 %, respectively, and the quotients
of fibrosis were 67.5 %, 19.5 %, 23.8 %, and 51.3 %, respectively. Lesion location
showed no significant correlation with stricture (r = −0.14, P = 0.61) or fibrosis (r = −0.09, P = 0.75), but both strictures and fibrosis were slightly less severe when the lesions
were located 37 and 42 cm from the mouth. We therefore conducted subgroup analysis
to reduce the impact of location. For lesions located 37 and 42 cm from the mouth,
the quotients of stricture were 16 %, 28 %, 32 %, 23 %, and 51 % with ENDO CUT, SWIFT
COAG, FORCED COAG E2, FORCED COAG E3, and FORCED COAG E4 modes, respectively, while
the quotients of fibrosis were 7 %, 28 %, 23 %, 20 %, and 40 %, respectively. For
lesions located 32 and 47 cm from the mouth, the quotients of stricture were 43 %,
44 %, 50 %, 39 %, and 47 % with FORCED COAG E2, FORCED COAG E3, FORCED COAG E4, SPRAY
COAG, and Bipolar FORCED COAG mode, respectively, and the quotients of fibrosis were
40 %, 40 %, 30 %, 63 %, and 100 %, respectively. ENDO CUT mode was thus associated
with a lower quotient of stricture and fibrosis after adjusting for location. The
mean (SD) size of the resected specimen was 20.4 (3.1) mm, and the size of the resected
specimen showed no significant correlation with stricture (r = −0.42, P = 0.10) or fibrosis (r = −0.36, P = 0.17) formation. Hemostasis with forceps was only conducted twice (once during
ENDO CUT and once during Bipolar FORCED COAG mode), and it was therefore not possible
to investigate the relationship between hemostasis and stricture or fibrosis formation.
Table 3
Mean quotient of stricture for each electrosurgical unit (ESU) mode and histopathological
mean quotient of fibrosis in the muscle layer.
|
Modes
|
Quotient of stricture, %
|
SD
|
Quotient of fibrosis in the proper muscle layer, %
|
SD
|
|
ENDO CUT
|
16
|
1
|
7
|
5
|
|
SWIFT COAG
|
28
|
28
|
28
|
4
|
|
FORCED COAG, E2
|
38
|
20
|
31
|
10
|
|
FORCED COAG, E3
|
33
|
15
|
30
|
14
|
|
FORCED COAG, E4
|
51
|
1
|
35
|
7
|
|
SPRAY COAG
|
39
|
39
|
63
|
53
|
|
Bipolar, FORCED COAG
|
47
|
1
|
100
|
0
|
SD, standard deviation.
Fig. 4 Histopathologic findings in resected specimens (Azan stain). Fibrous tissue is dyed
blue. a In ENDO CUT mode, fibrosis was observed in the submucosa (yellow arrows). No fibrosis
was observed in the muscle layer. b In FORCED COAG E2 mode, about 40 % of the muscle layer was damaged and replaced by
fibrosis (area within blue ellipse). This is a histologic view of the ESD scar 32 cm
from the mouth of Pig 3.
The mean (SD) resection time was 18.8 (4.2) minutes. Resection time was significantly
correlated with stricture (r = 0.52, P = 0.04) and marginally correlated with fibrosis (r = 0.48, P = 0.06) formation. We therefore conducted subgroup analysis to account for the impact
of procedure time. For lesions treated within the mean procedure time, the quotients
of stricture were 16 %, 28 %, 33 %, 33 %, and 46 % using ENDO CUT, SWIFT COAG, FORCED
COAG E2, FORCED COAG E3, and Bipolar FORCED COAG mode, respectively, while the quotients
of fibrosis were 7 %, 28 %, 28 %, 30 %, and 100 %, respectively. The quotients of
stricture for lesions treated with longer procedure times using FORCED COAG E2, FORCED
COAG E4, SPRAY COAG, and Bipolar FORCED COAG were 52 %, 51 %, 39 %, and 48 %, respectively,
while the quotients of fibrosis were 40 %, 35 %, 63 %, and 100 %, respectively. ENDO
CUT mode was thus associated with lower quotients of stricture and fibrosis after
adjusting for procedure time.
Discussion
This study compared the effect of electrocautery mode on fibrosis of the muscle layer
and stricture after ESD. ENDO CUT mode showed the lowest quotient of stricture and
the lowest fibrosis among all ESU modes.
ESU has two basic electrocautery patterns, one is cut current and the other is coagulation
current. Cut current consists of a continuous low-voltage sinusoidal wave with no
inactive period. The continuous sinusoidal wave causes very rapid heating with formation
of sparks and gives sharp cutting and little coagulation. Coagulation current consists
of a very short high-voltage sinusoidal wave (6 – 10 % of cycle) with long inactive
periods (90 – 94 % of cycle). The long inactive period facilitates coagulation. The
depth or severity of tissue injury and the coagulation effect mainly depend on electrical
discharge, which is regulated by the height of the peak voltage [18]. Coagulation current, which has a higher peak voltage than cut current, theoretically
has a higher coagulation effect and causes deeper and more severe damage to the surrounding
tissue than cut current. Deep and severe damage probably results in extensive fibrosis
and stricture. Various coagulation modes exist in the VIO 300 D ESU such as FORCED
COAG, SWIFT COAG, and SPRAY COAG. FORCED COAG is frequently used for submucosal dissection;
however, other modes such as SWIFT COAG, SPRAY COAG, and ENDO CUT offer alternative
modes to FORCED COAG [19]. The coagulation effect increases with increasing peak voltage, in the order SWIFT
COAG, FORCED COAG, and SPRAY COAG. However, the degree of fibrosis and stricture did
not differ among coagulation modes. This probably means that the peak voltage of all
coagulation modes is sufficiently high to cause significant damage to the muscle layer
and form strictures.
ENDO CUT is a unique mode which is characterized by alternating cutting and coagulation
cycles. Coagulation current employed in this mode has the lowest peak voltage (SOFT
COAG) among various coagulation currents, which may attenuate the heat damage to surrounding
tissue. This mode is designed to realize submucosal dissection with lowest tissue
damage by combining cutting current and lowest peak voltage coagulation current. Bahin
et al. examined inflammation after endoscopic mucosal resection and showed that ENDO
CUT mode resulted in less inflammation of the submucosal layer and muscle layer than
COAG mode [20]. Similar to their result, in our study, lower fibrosis of the muscle layer was observed
in ENDO CUT mode than in all coagulation modes. Because of its electrical characteristics,
ENDO CUT mode probably causes less inflammation and fibrosis which may result in a
lower degree of stricture after ESD. Considering the characteristics of ENDO CUT mode
and the results of this study, ENDO CUT mode is a promising ESU mode to prevent strictures
after ESD.
ESU setting is also divided into two types based on the electrodes. Monopolar electrocautery
has one small active electrode and one large dispersive plate electrode. The monopolar
endo-knife used in this study had a small active electrode at the tip of the knife
and a large dispersive electrode placed at the back of the pig. By using a large dispersive
electrode and small active electrode, electric discharge is radiated from the tip
of the monopolar knife. Radiated electrical discharge of COAG modes caused severe
muscle layer fibrosis and strictures in this study. Bipolar electrocautery has two
small active electrodes very close to each other (active and return electrode). The
bipolar endo-knife used in this study had an active electrode at the tip and return
electrode at the base of the 1.5 mm knife. Localized and concentrated electrocautery
flows between active and return electrode in the bipolar knife and the localized electrocautery
may theoretically attenuate the damage to surrounding tissue. However, in this study,
bipolar coagulation mode had similar muscle layer fibrosis and stricture formation
to monopolar coagulation modes. There are two possible explanations for this result.
One explanation is that the dissection process of ESD is conducted just above the
muscle layer. Even localized electric current may damage the muscle layer and thus
cause strictures when the knife is used adjacent to the muscle layer. Another explanation
is that concentrated, probably strong, electric current around the bipolar knife caused
more severe damage to the surrounding tissue than expected. Considering the concept
of the bipolar knife, it has the potential to reduce strictures after ESD. We have
to optimize the settings and usage of this knife to reduce the damage to muscle layer
and formation of strictures.
A limitation of the study is the use of a porcine model which may differ from human
in certain aspects. For example, the submucosa and muscle layer are thinner in pigs
than in the human esophagus [20]. Damage to the muscle layer, caused by electrocautery, may differ in the human esophagus.
However, when we consider that the effects of ESU modes were compared among various
modes and that a large difference was observed between ENDO CUT mode and coagulation
modes, our results may be transferable to human esophageal disease. Lesion location,
resected specimen size, frequency of hemostasis, and procedure time may influence
stricture and fibrosis formation. In this study, procedure time was significantly
associated with stricture formation. Although it was not possible to conduct multivariate
analysis because of the limited sample size, subgroup analyses confirmed that ENDO
CUT mode was associated with lower quotients of stricture and fibrosis than other
ESU modes, after adjusting for procedure time. Another limitation of the study is
the small number of pigs. More pigs would be required to detect a statistically significant
difference among the various modes. There are many modes in ESU such as FORCED COAG
mode, SWIFT COAG mode, SPRAY COAG mode, and ENDO CUT mode. Comparing all modes in
terms of muscle layer damage and stricture may be too costly and difficult. We therefore
decided to perform this study as an initial-phase test to compare various modes and
identified promising modes to suppress stricture formation.
In this study, ENDO CUT mode showed promising results to attenuate fibrosis and strictures
after esophageal ESD. Further study, which focuses on the effect of ENDO CUT mode,
would provide more information for the prevention of strictures.
