Surg Endosc DOI 10.1007/s00464-014-3985-9
and Other Interventional Techniques
Efficacy and safety of a novel submucosal lifting gel used for endoscopic submucosal dissection: a study in a porcine model D. W. Scho¨lvinck • L. Alvarez Herrero • O. Goto • S. L. Meijer • H. Neuhaus B. Schumacher • J. J. G. H. M. Bergman • B. L. A. M. Weusten
•
Received: 16 June 2014 / Accepted: 4 November 2014 Ó Springer Science+Business Media New York 2014
Abstract Background and study aims Endoscopic submucosal dissection (ESD) is technically demanding. A viscous gel for submucosal lifting might induce mechanical submucosal dissection facilitating easier and safer ESD. Methods In 12 female pigs (median 64 kg), ESDs of simulated lesions were performed at the posterior wall and greater curvature in the gastric body (one ESD per location) with randomly assigned injection fluids: gel or control fluid (0.9 % saline with hydroxypropyl methylcellulose 3 mg/ml [7:1] and indigo carmine droplets). Additionally, 10 cc gel was injected into the submucosa at the anterior wall without ESD to assess effects of inappropriate injection. Pigs were euthanized at day 0, 3 or 28. In four additional pigs (euthanized day 3 or 28) 10 cc gel was injected into the muscularis propria (MP) after four endoscopic mucosal resections in the gastric body. Results Both fluid groups showed equal ESD-procedure times (28 [gel] vs. 26 min [control]) and complications. Gel-ESDs required less accessory interchanges (3.5 vs. 5.5;
D. W. Scho¨lvinck (&) L. Alvarez Herrero B. L. A. M. Weusten Department of Gastroenterology and Hepatology, St. Antonius Hospital, Koekoekslaan 1, 3435 CM Nieuwegein, The Netherlands e-mail: [email protected] D. W. Scho¨lvinck L. Alvarez Herrero J. J. G. H. M. Bergman B. L. A. M. Weusten Department of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, Netherlands
p = 0.01). Mechanical dissection after circumferential incision was achieved in 25 % of gel-ESDs; none in control-ESDs. The severity of inflammation and fibrosis was equal in both fluid groups. Normal architecture and vital mucosa were found after inappropriate submucosal injection. MP-injections resulted in one transmural hematoma (day 3), and intramuscular encapsulation in 25 % of the sites (day 28). Limitations A pig’s stomach differs from the human stomach. Conclusions The mechanical dissection properties of the gel may reduce the need for submucosal dissection during ESD. The gel is safe when advertently injected in the submucosa and MP. The porcine model appeared suboptimal to evaluate the true mechanical dissection properties of the gel. Keywords Endoscopic submucosal dissection Injection fluid Gastric neoplasia Swine
S. L. Meijer Department of Pathology, Academic Medical Center, Amsterdam, Netherlands H. Neuhaus Department of Gastroenterology, Evangelisches Krankenhaus Du¨sseldorf, Du¨sseldorf, Germany B. Schumacher Department of Internal Medicine and Gastroenterology, Elisabeth-Krankenhaus, Essen, Germany
O. Goto Division of Research and Development for Minimally Invasive Treatment, Cancer Center, School of Medicine, Keio University, Tokyo, Japan
123
Surg Endosc
Abbreviations EMR Endoscopic mucosal resection ESD Endoscopic submucosal dissection MBM Multiband mucosectomy Endoscopic resection allows accurate histopathological assessment and treatment of early neoplastic lesions in the digestive tract. Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are the most commonly used techniques in the esophagus and stomach. For both techniques, the lesion is first delineated with coagulation markings. For EMR, the lesion is then resected with either a lift-suck-and-snare technique (EMR-cap) or a suck-ligateand-snare technique (multiband mucosectomy, MBM). For ESD, the lesion is elevated from the muscularis propria (MP) by submucosal fluid injection. The mucosa is then incised around the delineation, followed by a dissection in the submucosal plane and retrieval of the resection specimen [1]. ESD allows en bloc resection of larger lesions ([2 cm), whereas those lesions require piecemeal resection when using the EMR technique. In addition, local recurrences are less prevalent after ESD than after EMR [2–4]. However, ESD has a higher complication rate compared to EMR [2– 4], is technically more demanding, and carries a long learning curve [5–7]. This technique is well established in Japan and Korea, where early gastric neoplasia is the main indication. ESD of esophageal and colonic early neoplastic lesions, more prevalent in the Europe and Northern America, is extra challenging. For safe and appropriate submucosal dissection, optimal lifting of the lesion by means of submucosal injection is crucial. Normal saline easily diffuses away after injection necessitating repetitive injections. Several agents have been added in order to increase the viscosity of the injection fluid, such as fibrinogen, glycerol, hyaluronic acid, hydroxypropyl methylcellulose, reverse-phase polymers, sodium alginate, autologous blood, and succinylated gelatin [8–16]. So far, no injection fluid is widely accepted as the optimal agent for ESD. Recently, a new gel (Cook Medical, Limerick, Ireland) was developed. The gel is a highly viscous substance offering several potential benefits in ESD. It may limit the need for repetitive injections of lifting fluid, therefore possibly reducing the procedure time. In addition, the high pressure build up by the injection of the gel might result in disruption small vessels and vertical submucosal fibers, disconnecting the submucosa. This mechanical dissection might contribute to the safety of ESD, and might speed up the procedure. Parallel to its beneficial properties, the gel might carry some potential hazards. During ESD, lifting fluids may be inappropriately injected. The submucosal fluid might be injected in areas of which, in the end, the mucosal layer will not be resected. Lifting without subsequent dissection
123
will inevitably occur at the lateral margins of any endoscopic resection: in order to perform a safe resection these margins should be sufficiently lifted. Moreover, accidental injection in the MP may occur. In this study, we therefore evaluated both the efficacy and safety of the novel gel for ESD in the porcine stomach.
Methods Gel injection system The novel viscous gel (Cook Medical, Limerick, Ireland) is a blue transparent lifting substance composed of existing biocompatible components in a patented combination [17]. The syringe containing the gel is placed in the special handle (Cook Medical, Limerick, Ireland). With manual rotation, the gel is pressed from the syringe into the attached needle (Cook Medical, Limerick, Ireland). This 19-gauge needle is made of metal, is thereby pressure resistant, and slides out of the shaft with a pulling mechanism at the attachment side. It contains a pressure monitor to control high pressures due to the viscosity of the gel (Fig. 1). Animal handling and preparation After protocol approval by the Animal Experiment Committee, 16 female pigs (Topigs, Van Beek SPF Swine Breeding BV, Lelystad, The Netherlands) were included in this study. The experiments were performed at the Animal Research Institute AMC, Amsterdam, the Netherlands. Animal care was in accordance with European Union guidelines. For acclimatizing purposes, all survival pigs entered the animal facilities 13 days prior to the start of the
Fig. 1 The Gel Injection System. The Gel Injection System (Cook Medical, Limerick, Ireland) is composed of the 19-gauge gel injection needle with pressure monitor, attached to the syringe (10 mL) with the viscous gel. The syringe is placed in the handle and squeezed out with the rotation device
Surg Endosc
For this experiment, 12 pigs were included: four non-survival, four 3-day survival and four 28-day survival pigs. Two treatment areas (surrogate lesions) with a diameter of
25 mm each were predefined, one at the posterior wall and one at the greater curvature of the body of the stomach using a polypectomy snare (25 mm, Olympus, Tokyo, Japan) as a reference for the size of the lesions. These treatment areas were circumferentially delineated with coagulation markings (Erbotom ICC 200, ERBE Elektromedizin GmbH, Tuebingen, Germany) placed with the tip of a closed Dual Knife (2 mm, Olympus, Tokyo, Japan). The lesions at the greater curvature and posterior wall were then randomly assigned to an ESD procedure with either the gel or control fluid (saline 0.9 % with hydroxypropyl methylcellulose 3 mg/ml [7:1] with a few droplets of indigo carmine). In order to facilitate a proper submucosal injection of the gel, first 1–2 mL of saline was injected with an injection needle (23-gauge Injector Force Max Injection Needle, Olympus, Tokyo, Japan), followed by injection of the gel with the gel needle into the elevated area. Submucosal injection of the control-lifting fluid was performed with the 23-gauge regular injection needle without prior saline injection. When sufficient lifting was achieved, a circumferential mucosal incision 5 mm outside the coagulation
Fig. 2 ESD-procedure flowchart. The surrogate lesions at the posterior wall and greater curvature were randomized to either the viscous gel or a control injection fluid. ESD was further performed according to the
flowchart. At the anterior wall, 10 mL of the gel was injected without further ESD. In four other pigs, 4 9 EMRs were performed followed by injection of 10 mL of the gel into the muscularis propria
experiment; the non-survival pigs entered 6 days prior to the start of the experiment. Pigs were fastened 16 h prior to the experiment with free access to water and lemonade. Sedation was achieved with intramuscular midazolam 1 mg/kg and ketamine 15 mg/kg, followed by endotracheal intubation. After induction with propofol 3 mg/kg, anesthesia was maintained with propofol 6 mg/kg/h and sufentanil 0.0035 mg/kg supported by 2 % isoflurane when necessary. With the pig in left lateral position, the endoscope (GIF-1T140, Olympus, Tokyo, Japan) with a transparent cap (D-201-10704, Olympus, Tokyo, Japan) at its tip was introduced and a thorough gastric lavage was performed. The following procedures are described in Fig. 2. Procedures Experiment 1: endoscopic submucosal dissection
123
Surg Endosc
markers was made with the Dual Knife or Hook Knife (Olympus, Tokyo, Japan) with following diathermic settings (ERBE ICC 200): cutting mode 120 W, coagulation mode 45 W, endocut effect 3 (adjustable at discretion of endoscopist). After completion of the circumferential mucosal incision, three attempts were made to mechanically remove the mucosa: (1) push off the mucosa with the tip of the endoscope; (2) pull off the mucosa with a grasping forceps (2.3-mm rat tooth, Endo Passion, Du¨rbheim, Germany); (3) pull off by cold snaring (AcuSnare Polypectomy Snare, Cook Medical, Limerick, Ireland). For all failed mechanical dissections, a submucosal dissection was performed with the Dual Knife or Hook Knife until the specimen was completely resected and retrieved. During every step of the ESD, additional injection of lifting fluids was allowed. Experiment 2: inappropriate injections of the viscous gel In the 12 pigs of experiment 1, a third surrogate lesion was defined at the anterior wall of the gastric body. One syringe of the gel (10 cc) was injected per surrogate lesion as described in experiment 1. The injection site was inspected, but no subsequent ESD was performed. In the other pigs (n = 4), four EMRs were performed with the MBM technique (DuetteTM MBM Kit, Cook Medical, Limerick, Ireland) at the gastric body in order to expose the MP. In each resection wound, 10 cc of the gel was injected into the MP by sticking the injection needle deep into the resection wound. After retraction of the needle, all lesions were thoroughly checked for any bleeding before ending the procedure. Complications In case of bleeding during any procedure, hemostasis was achieved with either the ESD-device of use or with a Coagrasper (Olympus, Tokyo, Japan; soft coagulation 80 W; effect 3). Bleedings were scored as mild (resolves\5 min), moderate (resolves \15 min), or severe (resolves [15 min). Closing of any perforation was achieved with QuickClips2 (Olympus, Tokyo, Japan). Follow-up Pigs were euthanized with an intravenous overdose of pentobarbital, and the stomach was harvested for histology. Non-survival pigs were immediately euthanized after the experiment. The animals included in the survival experiments were extubated, put on a semi-liquid diet and were placed on a grid for 3 days in order to prevent sawdust perforating the gastric wound after ESD. After 3 days, pigs progressively received a more solid diet.
123
Histology For experiment 1, the retrieved ESD resection specimens were pinned down on paraffin, and length and width were measured for calculating the specimen circumference (ellipse circumference = 2 9 p 9 H([length2 ? width2]/2) and surface (ellipse surface = [ 9 length] 9 [ 9 width] 9 p). After termination of all pigs in experiment 1 & 2, the wound beds were resected from the harvested stomachs and inspected for any transmural damage. Ulcer sizes were measured with a ruler and the wound beds were pinned down on paraffin. Photographs of all lesions and any abnormalities were taken. All pinned down specimens were placed in 10 % formalin. After fixation for at least 24 h, 4-mm slices were cut at , the center, and of each specimen and wound bed. Each slice was then embedded in paraffin and cut in 4-lm slides for standard haematoxylin & eosin staining. Histological evaluation was performed by an expert gastro-intestinal pathologist (SM) blinded for the injection fluid. All ESD resection specimens (exp. 1) were evaluated for the thickness of the remaining submucosa with the software program LabSens 1.1 (Olympus Soft Imaging Solutions, GmbH, Hamburg, Germany). In the wound beds of the pigs in experiment 1, the severity of inflammation and necrosis (3-day and 28-day survival) and the degree of fibrosis (28-day survival) was scored. The specimens of pigs with injection without subsequent ESD (exp. 2) were evaluated for the presence of vertical submucosal fibers (0 days), severity of inflammation (3 & 28 days), vitality of the mucosa (3 & 28 days), any remaining architectural changes (28 days), and hematomas (all survival groups). In the pigs with gel injection into the MP (exp. 2), the severity of inflammation and necrosis in the MP, the presence of any serosal damage and the presence of signs of perforation, was assessed. Aforementioned scoring of the severity of inflammation and fibrosis in all histopathological slides was based on a 4-point scale [none—little—medium— severe].
Endpoints Primary outcome parameter was the difference in the duration of the ESD-procedure with the gel compared to the control fluid. Secondary outcome parameters were as follows: the ability to mechanically dissect the specimen after submucosal injection; the number of interchanges between accessories needed; the number of bleedings and perforations per procedure; the (variation in) thickness of the resection specimen; any histopathological differences between gel- and control-ESDs; histological changes after submucosal injection of the gel without subsequent ESD; histological changes after injection of the gel into the MP.
Surg Endosc
Sample size estimation & statistical analysis Assuming that a standard ESD in this setting would take 45 min (SD ± 7.5 min) and expecting the gel to reduce this procedure time by 25 %, the power calculation (twosided test) resulted in 11 lesions per injection fluid group when using a significance level of 0.05 and a power of 90 %. With the chance of a drop-out due to perforation, 12 pigs were included (4 non-survival, four 3-day survival, and four 28-day survival) for experiment 1. Statistical analysis was performed with the Statistical Software Package version 19.0.0.1. for Windows (SPSS, Chicago, Illinois, USA). For descriptive statistics, the median with interquartile range (IQR) was used for variables with a skewed distribution. As both injection fluids were compared per stomach, and stomach-related conditions such as thickness and elasticity of the tissue may have been of influence, paired analyses were performed. Categorical data were compared with the McNemar test. Continuous data that was not normally distributed were analyzed with the Wilcoxon sign rank test.
Results All 16 pigs survived the predetermined period without any complications or signs of illness. Experiment 1: endoscopic submucosal dissection Twelve pigs (median weight of 64 kg [IQR 59–68 kg]) underwent one ESD at the posterior wall and one ESD at greater curvature of the body of the stomach with either the gel or control fluid (12 vs. 12 lesions). The procedural characteristics are listed in Table 1. En bloc resection of ESD-lesions was achieved in 6/8 (83 %) with the gel and in 8/8 (100 %) with the control fluid. One lesion at the greater curvature broke into three pieces during the attempt of mechanical pulling off using a cold snare. The other lesion located at the posterior wall was resected in two pieces: a piece of the specimen was ruptured by mechanical dissection with the grasping forceps, the other part by regular submucosal dissection. The total procedure time was comparable in both groups (median 28 min [gel] vs. 26 min [control]). Yet, the gel required more time for submucosal injection (3.46 vs. 1.59 min; p = 0.01) and circumferential incision (17.07 vs. 13.27 min; p = 0.48). Complete mechanical dissection was achieved in 3 gel lesions using a cold snare (2 en bloc, 1 piecemeal) before submucosal dissection by an ESDknife. In the gel lesions in which the mechanical dissection did not succeed, there was no difference in submucosal dissection time compared to the control lesions (6.54 vs. 8.44 min, respectively; p = 0.26).
Table 1 Endoscopic parameters during endoscopic submucosal dissection (ESD) Viscous gel (n = 12)
Control fluid (n = 12)
p value
En bloc resection, n (%)
10 (83)
12 (100)
0.50
Size specimen, median cm2 (IQR)
8.5 (7.3–10.5)
11.2 (8.2–15.1)
0.04
Procedure
Mechanical submucosal dissections By scraping with endoscope
0 (0 %)
0 (0 %)
1
By using grasping forceps
0 (0 %)
0 (0 %)
1
By using cold snare
3 (25 %)
0 (0 %)
0.25
Total (min) (IQR)
28.25 (20.34–32.21)
26.03 (16.26–46.37)
0.94
Submucosal injection (min) (IQR)
03.46 (02.06–08.12)
01.59 (01.44–02.34)
0.01
Circumferential incision (min) (IQR)
17.07 (11.41–19.48)
13.27 (06.56–28.11)
0.48
Submucosal dissection (min) (IQR)
06.54 (05.10–12.05)
08.44 (06.47–19.39)
0.26
Procedure time
Excluding mechanical dissection* Procedure time (corrected for specimen size) Circumferential incision (sec/ cm) (IQR)
97 (69–115)
64 (39–135)
0.18
Submucosal dissection (sec/ cm2) (IQR)*
55 (32–76)
53 (40–77)
0.86
16.5 (10.5–29.25)
43.0 (34.5–69.0)
0.002
6 (4–7)
13 (9–24)
0.002
3.5 9 (2.25–4.75)
5.5 9 (4–7.75)
0.01
Mild-resolves \5 min
0
4
Moderateresolves \15 min
1
0
Severe-resolves [15 min
0
0
Injection fluid Total amount (ml) Total number of injections Interchanges between accessories Complications Bleedings
123
Surg Endosc Table 1 continued
Perforation
Viscous gel (n = 12)
Control fluid (n = 12)
1
1
p value
* n = 9 (successful mechanical dissections using cold snare excluded)
As the median size of the specimen differed slightly between the two injection fluids (median 8.5 cm2 (IQR 7.3–10.5) [gel] vs. 11.2 cm2 (IQR 8.2–15.1) [control]), times were corrected, further decreasing the differences between the gel and control fluid for circumferential incision time and the submucosal dissection (Table 1). For the total ESD-procedure, the number of injections (p = 0.002) and the total injected amount (p = 0.002) of the gel was significantly lower compared to the control fluid. After initial submucosal injection for sufficient lifting, additional injections were needed during circumferential incision (the gel in 1/12 vs. control fluid in 4/12), before submucosal dissection (the gel in 1/12 vs. control fluid in 11/12) and during submucosal dissection (the gel in 0/12 vs. control fluid in 5/12). This difference in number of additional injections resulted in significantly less interchanges between accessories in the gel lesions compared to the control fluid lesions (p = 0.01). Bleeding was seen during four ESD procedures. In one procedure, a moderate bleeding occurred during the circumferential incision which was treated with the Coagrasper. In three procedures with the control fluid, 4 mild bleedings occurred (3 during circumferential incision, 1 during submucosal dissection) which were treated with the ESD-device (1x) or the Coagrasper (3x). Two small perforations occurred during submucosal dissection, one in a gel lesion (acute experiment) and one in a control fluid (3day survival) lesion. Both lesions were clipped and antibiotics were administered. Histopathological evaluation showed the maximum thickness of the submucosa in the resection specimen to be significantly larger in the lesions resected with the gel (median 649 lm [IQR 520–827 lm]) compared to the control-lifting fluid (480 lm [IQR 374–622 lm]), p = 0.005. No distinct histopathological differences appeared between the resection wounds of the gel and the control fluid lesions after ESD. Ulcer sizes were identical in both groups after 3 days (median 12.0 in gel vs. 12.4 cm2 in control) and after 28 days (median 0.30 in gel vs. 0.18 cm2 in control). At 3 days, severe inflammation was seen in the submucosa and MP in both groups. At 28 days, little inflammation was seen in the submucosa and MP, and little-to-medium fibrosis was present in both groups.
123
At necropsy, the stomach of one 3-day survival pig showed an abscess at the serosal side of the greater curvature (gel lesion) (Fig. 3). We speculate that transmural injection of the gel must have been the cause. This pig survived the predetermined period without any clinical signs of illness. At day 28, the amount of inflammation and fibrosis was identical in both groups. Experiment 2: inappropriate injections In the pigs with submucosal gel injection but no subsequent dissection (n = 12), the center of all four specimens from the non-survival pigs showed full disruption of the mucosa from the MP. No intact vertical fibers connecting the mucosa and MP were visible, and no signs of major bleeding were present (Fig. 4A). In the 3-day survival group in all specimen, the lifting was still present (median length and width were 5.5 and 5.5 cm, respectively). One of four lesions contained a mixture of blood and liquid content; the other three had a transparent liquid content. Histology showed some inflammation, starting encapsulation with granulation tissue, and a vital mucosa in all four specimens (Fig. 4B). At 28 days, one injection site was still lifted (4.5 9 3.0 cm). In the submucosa, an elevated area of submucosal granulation tissue was seen (Fig. 4D). The other specimens were flat with little submucosal granulation tissue and fibrosis. No major architectural changes remained (Fig. 4C). The mucosa was vital in all specimens. In the pigs with gel injection into the gastric MP (n = 4), eight lesions were created per survival group
Fig. 3 Transmural injection for ESD with the viscous gel in 3-day survival pig. In a 3-day survival pig, an abscess was found at the serosal side of the gastric wall containing a mix of the gel and granulocytes. At histopathological evaluation, the content was identified as the gel, probably due to accidental transmural injection
Surg Endosc
Fig. 4 Histopathological effects of the viscous gel injection without ESD. (A) Injection of the viscous gel without ESD in acute experiment showing disruption of vertical submucosal fibers. (B) Injection of the viscous gel without ESD resulted in gel-filled
submucosal cavities in 3 of 4 lesions after 3 days survival. After 28 days survival, injection of the viscous gel without ESD resulted in little remaining damage in 3 of 4 lesions (C), whereas in one an elevated area with granulation tissue remained (D)
(3-day and 28-day). In the 3-day survival group, submucosal cavities with bloody content next to the ulcers were macroscopically seen in 3/8 lesions (Fig. 5A, B). Six of eight lesions showed severe submucosal inflammation, whereas in 2/8 lesions, severe inflammation and necrosis of the MP were seen. In one lesion a hematoma at the serosal side was found. At histology, this lesion showed very severe necrosis of the MP and serosal inflammation. In the 28-day survival group, the EMR-ulcers had healed in all lesions with vital overlying mucosa. Two of the eight lesions still showed an encapsulated cavity with mediumgrade inflammation (Fig. 5D). Severity of submucosal fibrosis was classified as medium in 2, and as little in 6 lesions. One lesion showed little fibrosis at the serosal side, but no other signs of perforation were observed.
gel, the procedure was split up into three steps: the submucosal injection, the circumferential mucosal incision, and the submucosal dissection. Injection with the gel required more time to achieve proper submucosal lifting compared to the control fluid. However, during 25 % of the procedures with the gel no submucosal dissection was required as grabbing and pulling the specimen with a snare resulted in a complete mechanical dissection. This balance equalized the total procedure time. Several properties of the gel play a role in these observations. The main difference between the gel and the control fluid is the viscosity of the injection medium. This feature of the gel has several potential benefits. First, it more easily disrupts the submucosal fibers and small vessels, resulting in a (partial) mechanical separation of the mucosal layer and the MP. As shown in our data, this allowed a mechanical dissection with cold snare in 3 of 12 lesions with the gel. Second, the viscosity of the gel ensures that the injection fluid will not diffuse into the surrounding tissue. During the entire procedure, fewer injections are required to maintain the lifting of the lesion. Consequently, lower volumes of injection fluid and fewer interchanges between accessories are necessary. These characteristics may facilitate an easier ESD-procedure. On the other hand, the viscosity does have some disadvantages as well. First, given the high pressures needed for pressing the gel through a small needle, a specially
Discussion This study assessed the efficacy and safety of a new viscous submucosal lifting fluid for ESD in the stomach of a living porcine model. Our data suggest that the gel is effective and safe for this purpose. The main outcome parameter of this study was the time for the total procedure. We found no difference in the total procedure time when the gel was compared to our standard viscous solution. To evaluate the potential benefits of the
123
Surg Endosc
Fig. 5 Histopathological effects of intentional injection of viscous gel into the muscularis propria after EMR with multiband mucosectomy. After 3 days, deep injections resulted in cavities with bloody liquid content (A, B) in 3 of 8 lesions. In the other lesions, only the
ulcer of the EMR (C) was seen. After 28 days survival, 2/8 injection sites consisted of an encapsulated cavity without extensive inflammatory response (D). The other 6 injection sites healed without any architectural changes (E)
designed metal needle is required. Injection of the gel through the needle takes more time than conventional injection with a standard needle and a less viscous substance. In addition, although not tested in this experiment, the stiffness of this needle might hamper submucosal injection with the endoscope in the retroflex position. Second, due to the high viscosity of the gel, a little momentum is necessary for direct lifting. This is achieved by injecting a small volume of saline before injection of the gel. This preceding saline injection requires the use of an additional regular injection needle, as the gel needle needs to be primed and ready for use. Both these disadvantages may explain the prolonged time required for the submucosal injection with the gel compared to the control fluid in this study. Future modifications of the needle will likely speed up the submucosal injection procedure. Moreover, the use of the gel may harbor a learning curve and more extensive experience with its properties, as currently with the control fluid, may possibly lead to a faster procedure.
Histopathological evaluation of the resection specimen showed a significant difference in submucosal thickness between both fluid groups. This was not due to a deeper resection, but more a result of the effects of the gel. In the gel specimen, the tissue and structures were more spread out with gel in between, whereas in the control group the structures were more compact. This confirmed the hypothesis that the gel causes mechanical disruption. Evaluation of the ESD wounds showed no difference in amount of inflammation, necrosis, and fibrosis in both groups in the acute setting, or at short- and long-term survival. However, in one ESD lesion (3-day survival) an abscess was found in a pocket at the serosal side, although clinically the pig did not have any signs of infection. As neither endoscopically nor at histopathological assessment a perforation was observed, it is likely that this resulted from inadvertent deep injection during submucosal lifting. Ulceration has been reported to occur just by submucosal injection of fluids even without subsequent dissection, especially in hypertonic solutions such as hypertonic
123
Surg Endosc
dextrose (C20 %), hypertonic sodium chloride (3.75 %), or hydroxypropyl methylcellulose [18, 19]. In case of inappropriate submucosal injections of the gel, one might postulate the possibility of ulceration or even necrosis of the overlying mucosa because of mechanical disruption of vertical strands including the feeding vessels. In clinical practice, this might lead to an extended ulcer after ESD since the lifted area is often much larger than the actual resected specimen. Our results show that after 3 days the gel had not dissipated, but no severe inflammation or necrosis was seen. Considering the fact that the inflammatory response is most severe at 3–5 days after the procedure [20], safety and wound healing do not seem to be impaired by submucosal gel injection. After 28 days, the lifted tissue consisted of granulation tissue and no major architectural changes seemed to remain. These long-term results are in line with the persistent submucosal cushions without signs of mucosal ischemia or ulceration found by Chandrasekhara et al. [17]. Although the gel remains present for much longer compared to other injection agents, it seems safe when it comes to long-term histopathological outcomes if injected at locations that in the end will not be resected. Deep gel injections into the MP after initial EMR in the second part of this study did not cause any striking complications. Histopathological evaluation showed healing and vitality of all wall layers with no significant remaining architectural abnormalities. Some inflammation and fibrosis at the serosal side of the MP was observed, but no transmural necrosis was seen. Two of the total number of 16 lesions in the four pigs with deep injection did show inflammation and fibrosis at the serosal side of the gastric wall. It is conceivable that these serosal changes are caused by transmural instead of intramural injections. However, no gel was found in these specimens at necropsy and the pigs showed no clinical signs of illness. Therefore, the gel seems safe even when injected transmurally, as may be expected from its biocompatible characteristics. The major limitation in this experiment is the pig’s stomach as a study model. In the body of a pig’s stomach a relatively thick and stiff mucosa-submucosa layer is present. In fact, the stiffness of the intact gastric wall is determined by the stiffness of the mucosa-submucosa layer [21]. This characteristic of our model may have influenced our experiment, as proper lifting was not achieved as easily as anticipated. Higher pressures and often multiple injections were needed to create proper submucosal lifting. Consequently, the success rate of mechanical dissection, which is essentially the main quality of the gel, was lower than expected based on earlier results in a porcine model by Khashab et al., where only a single injection with the gel led to full mechanical dissection [22]. In the study of Khashab
et al., however, a different injection technique was applied. In contrast to the 2 cc injected in our study used to create a small cushion for safer gel injection, Khashab aimed to achieve complete lifting of the surrogate lesion by injection of 10 cc of saline prior to gel injection. Possibly, this might have created a pre-defined submucosal space for the gel to be more easily injected, thereby obviating the need for repetitive gel injections as appeared to be necessary in our study. Still, in our study in 25 % of the lesions dissection of the mucosa by cold snaring was successful. Furthermore, for all the other gel lesions the time needed for submucosal dissection was shorter, although not significantly, compared to the lesions lifted with the control fluid. This is suggestive for at least a partial mechanical dissection by the gel despite the model and possibly inferior injection technique used. Given all this, we anticipate that the use of the gel for ESD in the human stomach using the injection technique described by Khashab et al. might compare favorably with our results [17, 23, 24]. Second, due to failed mechanical dissection two gelESDs were not resected en bloc, which may hinder optimal histopathological evaluation and is associated with a higher chance for local recurrence [2–4]. However, these accidental piecemeal resections are more likely to be attributed to the difference in wall layers within the porcine model than to the gel itself, and may be different in humans. The final limitation of this study is the relatively low number of experiments which may underestimate the risk for uncommon complications. In conclusion, the viscous gel seems effective and safe for ESD in the stomach of a porcine model. Complete mechanical submucosal dissection was achieved in 25 % of cases. Gel ESD did not result in shorter procedure times, although this might—at least in part—be attributed to the limitations of the model used. Further assessment of the use of gel in humans using a more optimal injection technique may elucidate the potential beneficial features of the gel in a better way. Acknowledgments The authors thank Elly van Zwol, Sanne Hackmann, Alwin Blankestijn, and Hester de Bruin for their biotechnical assistance. This study was supported by Cook Medical, Limerick, Ireland. Prof. J.J.G.H.M. Bergman is a consultant for Cook Medical and received speaker fees from Cook Medical. Prof. H. Neuhaus is a consultant for Cook Medical. Disclosures D.W. Scholvinck, L. Alvarez Herrero, O. Goto, S.L. Meijer, H. Neuhaus, B. Schumacher, J.J.G.H.M. Bergman, and B.L.A.M. Weusten have no conflicts of interest or financial ties to disclose.
References 1. Kodashima S, Fujishiro M, Yahagi N, Kakushima N, Omata M (2006) Endoscopic submucosal dissection using flexknife. J Clin Gastroenterol 40(5):378–384
123
Surg Endosc 2. Park YM, Cho E, Kang HY, Kim JM (2011) The effectiveness and safety of endoscopic submucosal dissection compared with endoscopic mucosal resection for early gastric cancer: a systematic review and metaanalysis. Surg Endosc 25(8):2666–2677 3. Cao Y, Liao C, Tan A, Gao Y, Mo Z, Gao F (2009) Meta-analysis of endoscopic submucosal dissection versus endoscopic mucosal resection for tumors of the gastrointestinal tract. Endoscopy 41(9):751–757 4. Lian J, Chen S, Zhang Y, Qiu F (2012) A meta-analysis of endoscopic submucosal dissection and EMR for early gastric cancer. Gastrointest Endosc 76(4):763–770 5. Kakushima N, Fujishiro M, Kodashima S, Muraki Y, Tateishi A, Omata M (2006) A learning curve for endoscopic submucosal dissection of gastric epithelial neoplasms. Endoscopy 38(10): 991–995 6. Hotta K, Oyama T, Shinohara T, Miyata Y, Takahashi A, Kitamura Y, Tomori A (2010) Learning curve for endoscopic submucosal dissection of large colorectal tumors. Dig Endosc 22(4):302–306 7. Probst A, Golger D, Anthuber M, Markl B, Messmann H (2012) Endoscopic submucosal dissection in large sessile lesions of the rectosigmoid: learning curve in a European center. Endoscopy 44(7):660–667 8. Giday SA, Magno P, Buscaglia JM, Canto MI, Ko CW, Shin EJ, Xia L, Wroblewski LM, Clarke JO, Kalloo AN, Jagannath SB, Kantsevoy SV (2006) Is blood the ideal submucosal cushioning agent? A comparative study in a porcine model. Endoscopy 38(12):1230–1234 9. Lee SH, Park JH, Park do H, Chung IK, Kim HS, Park SH, Kim SJ, Cho HD (2006) Clinical efficacy of EMR with submucosal injection of a fibrinogen mixture: a prospective randomized trial. Gastrointest Endosc 64(5):691–696 10. Uraoka T, Fujii T, Saito Y, Sumiyoshi T, Emura F, Bhandari P, Matsuda T, Fu KI, Saito D (2005) Effectiveness of glycerol as a submucosal injection for EMR. Gastrointest Endosc 61(6): 736–740 11. Yamamoto H, Yahagi N, Oyama T, Gotoda T, Doi T, Hirasaki S, Shimoda T, Sugano K, Tajiri H, Takekoshi T, Saito D (2008) Usefulness and safety of 0.4 % sodium hyaluronate solution as a submucosal fluid ‘‘cushion’’ in endoscopic resection for gastric neoplasms: a prospective multicenter trial. Gastrointest Endosc 67(6):830–839 12. Lenz L, Di Sena V, Nakao FS, Andrade GP, Rohr MR, Ferrari AP Jr (2010) Comparative results of gastric submucosal injection with hydroxypropyl methylcellulose, carboxymethylcellulose and normal saline solution in a porcine model. Arq Gastroenterol 47(2):184–187 13. Feitoza AB, Gostout CJ, Burgart LJ, Burkert A, Herman LJ, Rajan E (2003) Hydroxypropyl methylcellulose: a better
123
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
submucosal fluid cushion for endoscopic mucosal resection. Gastrointest Endosc 57(1):41–47 Akagi T, Yasuda K, Tajima M, Suzuki K, Inomata M, Shiraishi N, Sato Y, Kitano S (2011) Sodium alginate as an ideal submucosal injection material for endoscopic submucosal resection: preliminary experimental and clinical study. Gastrointest Endosc 74(5):1026–1032 Moss A, Bourke MJ, Metz AJ (2010) A randomized, double-blind trial of succinylated gelatin submucosal injection for endoscopic resection of large sessile polyps of the colon. Am J Gastroenterol 105(11):2375–2382 Fernandez-Esparrach G, Shaikh SN, Cohen A, Ryan MB, Thompson CC (2009) Efficacy of a reverse-phase polymer as a submucosal injection solution for EMR: a comparative study (with video). Gastrointest Endosc 69(6):1135–1139 Chandrasekhara V, Sigmon JC Jr, Surti VC, Kochman ML (2013) A novel gel provides durable submucosal cushion for endoscopic mucosal resection and endoscopic submucosal dissection. Surg Endosc 27(8):3039–3042 Sato T (2006) A novel method of endoscopic mucosal resection assisted by submucosal injection of autologous blood (blood patch EMR). Dis Colon Rectum 49(10):1636–1641 Fujishiro M, Yahagi N, Kashimura K, Mizushima Y, Oka M, Enomoto S, Kakushima N, Kobayashi K, Hashimoto T, Iguchi M, Shimizu Y, Ichinose M, Omata M (2004) Comparison of various submucosal injection solutions for maintaining mucosal elevation during endoscopic mucosal resection. Endoscopy 36(7):579–583 Mitchell RN, Ramzi SC (2003) Tissue repair: cell regeneration and fibrosis. In: Kumar V, Ramzi SC, Robbins SL, Saunders WB (eds) Robbins basic pathology, 7th edn. Saunders, Philadelphia, pp 61–64 Zhao J, Liao D, Chen P, Kunwald P, Gregersen H (2008) Stomach stress and strain depend on location, direction and the layered structure. J Biomech 41(16):3441–3447 Khashab MA, Saxena P, Sharaiha RZ, Chavez YH, Zhang F, Kord Valeshabad A, Aguila G, Canto MI, Pasricha PJ, Kalloo AN (2013) A novel submucosal gel permits simple and efficient gastric endoscopic submucosal dissection. Gastroenterology 144(3):505–507 Saxena P, Simons BW, Gabrielson K, Haito-Chavez Y, Valeshabad AK, Kumbhari V, Singh VK, Lennon AM, Canto MI, Pasricha PJ, Kalloo A, Khashab MA (2014) A novel viscous dissecting gel for endoscopic submucosal dissection: a prospective survival study in a porcine model. Endoscopy 46(7):605–609 Samarasena JB, Shinoura S, Nakai Y, Choi KD, Lee JG, Chang KJ (2013) New submucosal lifting gel for endoscopic submucosal dissection of gastrointestinal lesions: ESD made easy with ‘‘needle-free auto-dissection. Gastrointest Endosc 77(5):AB189– AB190
and Other Interventional Techniques
Efficacy and safety of a novel submucosal lifting gel used for endoscopic submucosal dissection: a study in a porcine model D. W. Scho¨lvinck • L. Alvarez Herrero • O. Goto • S. L. Meijer • H. Neuhaus B. Schumacher • J. J. G. H. M. Bergman • B. L. A. M. Weusten
•
Received: 16 June 2014 / Accepted: 4 November 2014 Ó Springer Science+Business Media New York 2014
Abstract Background and study aims Endoscopic submucosal dissection (ESD) is technically demanding. A viscous gel for submucosal lifting might induce mechanical submucosal dissection facilitating easier and safer ESD. Methods In 12 female pigs (median 64 kg), ESDs of simulated lesions were performed at the posterior wall and greater curvature in the gastric body (one ESD per location) with randomly assigned injection fluids: gel or control fluid (0.9 % saline with hydroxypropyl methylcellulose 3 mg/ml [7:1] and indigo carmine droplets). Additionally, 10 cc gel was injected into the submucosa at the anterior wall without ESD to assess effects of inappropriate injection. Pigs were euthanized at day 0, 3 or 28. In four additional pigs (euthanized day 3 or 28) 10 cc gel was injected into the muscularis propria (MP) after four endoscopic mucosal resections in the gastric body. Results Both fluid groups showed equal ESD-procedure times (28 [gel] vs. 26 min [control]) and complications. Gel-ESDs required less accessory interchanges (3.5 vs. 5.5;
D. W. Scho¨lvinck (&) L. Alvarez Herrero B. L. A. M. Weusten Department of Gastroenterology and Hepatology, St. Antonius Hospital, Koekoekslaan 1, 3435 CM Nieuwegein, The Netherlands e-mail: [email protected] D. W. Scho¨lvinck L. Alvarez Herrero J. J. G. H. M. Bergman B. L. A. M. Weusten Department of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, Netherlands
p = 0.01). Mechanical dissection after circumferential incision was achieved in 25 % of gel-ESDs; none in control-ESDs. The severity of inflammation and fibrosis was equal in both fluid groups. Normal architecture and vital mucosa were found after inappropriate submucosal injection. MP-injections resulted in one transmural hematoma (day 3), and intramuscular encapsulation in 25 % of the sites (day 28). Limitations A pig’s stomach differs from the human stomach. Conclusions The mechanical dissection properties of the gel may reduce the need for submucosal dissection during ESD. The gel is safe when advertently injected in the submucosa and MP. The porcine model appeared suboptimal to evaluate the true mechanical dissection properties of the gel. Keywords Endoscopic submucosal dissection Injection fluid Gastric neoplasia Swine
S. L. Meijer Department of Pathology, Academic Medical Center, Amsterdam, Netherlands H. Neuhaus Department of Gastroenterology, Evangelisches Krankenhaus Du¨sseldorf, Du¨sseldorf, Germany B. Schumacher Department of Internal Medicine and Gastroenterology, Elisabeth-Krankenhaus, Essen, Germany
O. Goto Division of Research and Development for Minimally Invasive Treatment, Cancer Center, School of Medicine, Keio University, Tokyo, Japan
123
Surg Endosc
Abbreviations EMR Endoscopic mucosal resection ESD Endoscopic submucosal dissection MBM Multiband mucosectomy Endoscopic resection allows accurate histopathological assessment and treatment of early neoplastic lesions in the digestive tract. Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are the most commonly used techniques in the esophagus and stomach. For both techniques, the lesion is first delineated with coagulation markings. For EMR, the lesion is then resected with either a lift-suck-and-snare technique (EMR-cap) or a suck-ligateand-snare technique (multiband mucosectomy, MBM). For ESD, the lesion is elevated from the muscularis propria (MP) by submucosal fluid injection. The mucosa is then incised around the delineation, followed by a dissection in the submucosal plane and retrieval of the resection specimen [1]. ESD allows en bloc resection of larger lesions ([2 cm), whereas those lesions require piecemeal resection when using the EMR technique. In addition, local recurrences are less prevalent after ESD than after EMR [2–4]. However, ESD has a higher complication rate compared to EMR [2– 4], is technically more demanding, and carries a long learning curve [5–7]. This technique is well established in Japan and Korea, where early gastric neoplasia is the main indication. ESD of esophageal and colonic early neoplastic lesions, more prevalent in the Europe and Northern America, is extra challenging. For safe and appropriate submucosal dissection, optimal lifting of the lesion by means of submucosal injection is crucial. Normal saline easily diffuses away after injection necessitating repetitive injections. Several agents have been added in order to increase the viscosity of the injection fluid, such as fibrinogen, glycerol, hyaluronic acid, hydroxypropyl methylcellulose, reverse-phase polymers, sodium alginate, autologous blood, and succinylated gelatin [8–16]. So far, no injection fluid is widely accepted as the optimal agent for ESD. Recently, a new gel (Cook Medical, Limerick, Ireland) was developed. The gel is a highly viscous substance offering several potential benefits in ESD. It may limit the need for repetitive injections of lifting fluid, therefore possibly reducing the procedure time. In addition, the high pressure build up by the injection of the gel might result in disruption small vessels and vertical submucosal fibers, disconnecting the submucosa. This mechanical dissection might contribute to the safety of ESD, and might speed up the procedure. Parallel to its beneficial properties, the gel might carry some potential hazards. During ESD, lifting fluids may be inappropriately injected. The submucosal fluid might be injected in areas of which, in the end, the mucosal layer will not be resected. Lifting without subsequent dissection
123
will inevitably occur at the lateral margins of any endoscopic resection: in order to perform a safe resection these margins should be sufficiently lifted. Moreover, accidental injection in the MP may occur. In this study, we therefore evaluated both the efficacy and safety of the novel gel for ESD in the porcine stomach.
Methods Gel injection system The novel viscous gel (Cook Medical, Limerick, Ireland) is a blue transparent lifting substance composed of existing biocompatible components in a patented combination [17]. The syringe containing the gel is placed in the special handle (Cook Medical, Limerick, Ireland). With manual rotation, the gel is pressed from the syringe into the attached needle (Cook Medical, Limerick, Ireland). This 19-gauge needle is made of metal, is thereby pressure resistant, and slides out of the shaft with a pulling mechanism at the attachment side. It contains a pressure monitor to control high pressures due to the viscosity of the gel (Fig. 1). Animal handling and preparation After protocol approval by the Animal Experiment Committee, 16 female pigs (Topigs, Van Beek SPF Swine Breeding BV, Lelystad, The Netherlands) were included in this study. The experiments were performed at the Animal Research Institute AMC, Amsterdam, the Netherlands. Animal care was in accordance with European Union guidelines. For acclimatizing purposes, all survival pigs entered the animal facilities 13 days prior to the start of the
Fig. 1 The Gel Injection System. The Gel Injection System (Cook Medical, Limerick, Ireland) is composed of the 19-gauge gel injection needle with pressure monitor, attached to the syringe (10 mL) with the viscous gel. The syringe is placed in the handle and squeezed out with the rotation device
Surg Endosc
For this experiment, 12 pigs were included: four non-survival, four 3-day survival and four 28-day survival pigs. Two treatment areas (surrogate lesions) with a diameter of
25 mm each were predefined, one at the posterior wall and one at the greater curvature of the body of the stomach using a polypectomy snare (25 mm, Olympus, Tokyo, Japan) as a reference for the size of the lesions. These treatment areas were circumferentially delineated with coagulation markings (Erbotom ICC 200, ERBE Elektromedizin GmbH, Tuebingen, Germany) placed with the tip of a closed Dual Knife (2 mm, Olympus, Tokyo, Japan). The lesions at the greater curvature and posterior wall were then randomly assigned to an ESD procedure with either the gel or control fluid (saline 0.9 % with hydroxypropyl methylcellulose 3 mg/ml [7:1] with a few droplets of indigo carmine). In order to facilitate a proper submucosal injection of the gel, first 1–2 mL of saline was injected with an injection needle (23-gauge Injector Force Max Injection Needle, Olympus, Tokyo, Japan), followed by injection of the gel with the gel needle into the elevated area. Submucosal injection of the control-lifting fluid was performed with the 23-gauge regular injection needle without prior saline injection. When sufficient lifting was achieved, a circumferential mucosal incision 5 mm outside the coagulation
Fig. 2 ESD-procedure flowchart. The surrogate lesions at the posterior wall and greater curvature were randomized to either the viscous gel or a control injection fluid. ESD was further performed according to the
flowchart. At the anterior wall, 10 mL of the gel was injected without further ESD. In four other pigs, 4 9 EMRs were performed followed by injection of 10 mL of the gel into the muscularis propria
experiment; the non-survival pigs entered 6 days prior to the start of the experiment. Pigs were fastened 16 h prior to the experiment with free access to water and lemonade. Sedation was achieved with intramuscular midazolam 1 mg/kg and ketamine 15 mg/kg, followed by endotracheal intubation. After induction with propofol 3 mg/kg, anesthesia was maintained with propofol 6 mg/kg/h and sufentanil 0.0035 mg/kg supported by 2 % isoflurane when necessary. With the pig in left lateral position, the endoscope (GIF-1T140, Olympus, Tokyo, Japan) with a transparent cap (D-201-10704, Olympus, Tokyo, Japan) at its tip was introduced and a thorough gastric lavage was performed. The following procedures are described in Fig. 2. Procedures Experiment 1: endoscopic submucosal dissection
123
Surg Endosc
markers was made with the Dual Knife or Hook Knife (Olympus, Tokyo, Japan) with following diathermic settings (ERBE ICC 200): cutting mode 120 W, coagulation mode 45 W, endocut effect 3 (adjustable at discretion of endoscopist). After completion of the circumferential mucosal incision, three attempts were made to mechanically remove the mucosa: (1) push off the mucosa with the tip of the endoscope; (2) pull off the mucosa with a grasping forceps (2.3-mm rat tooth, Endo Passion, Du¨rbheim, Germany); (3) pull off by cold snaring (AcuSnare Polypectomy Snare, Cook Medical, Limerick, Ireland). For all failed mechanical dissections, a submucosal dissection was performed with the Dual Knife or Hook Knife until the specimen was completely resected and retrieved. During every step of the ESD, additional injection of lifting fluids was allowed. Experiment 2: inappropriate injections of the viscous gel In the 12 pigs of experiment 1, a third surrogate lesion was defined at the anterior wall of the gastric body. One syringe of the gel (10 cc) was injected per surrogate lesion as described in experiment 1. The injection site was inspected, but no subsequent ESD was performed. In the other pigs (n = 4), four EMRs were performed with the MBM technique (DuetteTM MBM Kit, Cook Medical, Limerick, Ireland) at the gastric body in order to expose the MP. In each resection wound, 10 cc of the gel was injected into the MP by sticking the injection needle deep into the resection wound. After retraction of the needle, all lesions were thoroughly checked for any bleeding before ending the procedure. Complications In case of bleeding during any procedure, hemostasis was achieved with either the ESD-device of use or with a Coagrasper (Olympus, Tokyo, Japan; soft coagulation 80 W; effect 3). Bleedings were scored as mild (resolves\5 min), moderate (resolves \15 min), or severe (resolves [15 min). Closing of any perforation was achieved with QuickClips2 (Olympus, Tokyo, Japan). Follow-up Pigs were euthanized with an intravenous overdose of pentobarbital, and the stomach was harvested for histology. Non-survival pigs were immediately euthanized after the experiment. The animals included in the survival experiments were extubated, put on a semi-liquid diet and were placed on a grid for 3 days in order to prevent sawdust perforating the gastric wound after ESD. After 3 days, pigs progressively received a more solid diet.
123
Histology For experiment 1, the retrieved ESD resection specimens were pinned down on paraffin, and length and width were measured for calculating the specimen circumference (ellipse circumference = 2 9 p 9 H([length2 ? width2]/2) and surface (ellipse surface = [ 9 length] 9 [ 9 width] 9 p). After termination of all pigs in experiment 1 & 2, the wound beds were resected from the harvested stomachs and inspected for any transmural damage. Ulcer sizes were measured with a ruler and the wound beds were pinned down on paraffin. Photographs of all lesions and any abnormalities were taken. All pinned down specimens were placed in 10 % formalin. After fixation for at least 24 h, 4-mm slices were cut at , the center, and of each specimen and wound bed. Each slice was then embedded in paraffin and cut in 4-lm slides for standard haematoxylin & eosin staining. Histological evaluation was performed by an expert gastro-intestinal pathologist (SM) blinded for the injection fluid. All ESD resection specimens (exp. 1) were evaluated for the thickness of the remaining submucosa with the software program LabSens 1.1 (Olympus Soft Imaging Solutions, GmbH, Hamburg, Germany). In the wound beds of the pigs in experiment 1, the severity of inflammation and necrosis (3-day and 28-day survival) and the degree of fibrosis (28-day survival) was scored. The specimens of pigs with injection without subsequent ESD (exp. 2) were evaluated for the presence of vertical submucosal fibers (0 days), severity of inflammation (3 & 28 days), vitality of the mucosa (3 & 28 days), any remaining architectural changes (28 days), and hematomas (all survival groups). In the pigs with gel injection into the MP (exp. 2), the severity of inflammation and necrosis in the MP, the presence of any serosal damage and the presence of signs of perforation, was assessed. Aforementioned scoring of the severity of inflammation and fibrosis in all histopathological slides was based on a 4-point scale [none—little—medium— severe].
Endpoints Primary outcome parameter was the difference in the duration of the ESD-procedure with the gel compared to the control fluid. Secondary outcome parameters were as follows: the ability to mechanically dissect the specimen after submucosal injection; the number of interchanges between accessories needed; the number of bleedings and perforations per procedure; the (variation in) thickness of the resection specimen; any histopathological differences between gel- and control-ESDs; histological changes after submucosal injection of the gel without subsequent ESD; histological changes after injection of the gel into the MP.
Surg Endosc
Sample size estimation & statistical analysis Assuming that a standard ESD in this setting would take 45 min (SD ± 7.5 min) and expecting the gel to reduce this procedure time by 25 %, the power calculation (twosided test) resulted in 11 lesions per injection fluid group when using a significance level of 0.05 and a power of 90 %. With the chance of a drop-out due to perforation, 12 pigs were included (4 non-survival, four 3-day survival, and four 28-day survival) for experiment 1. Statistical analysis was performed with the Statistical Software Package version 19.0.0.1. for Windows (SPSS, Chicago, Illinois, USA). For descriptive statistics, the median with interquartile range (IQR) was used for variables with a skewed distribution. As both injection fluids were compared per stomach, and stomach-related conditions such as thickness and elasticity of the tissue may have been of influence, paired analyses were performed. Categorical data were compared with the McNemar test. Continuous data that was not normally distributed were analyzed with the Wilcoxon sign rank test.
Results All 16 pigs survived the predetermined period without any complications or signs of illness. Experiment 1: endoscopic submucosal dissection Twelve pigs (median weight of 64 kg [IQR 59–68 kg]) underwent one ESD at the posterior wall and one ESD at greater curvature of the body of the stomach with either the gel or control fluid (12 vs. 12 lesions). The procedural characteristics are listed in Table 1. En bloc resection of ESD-lesions was achieved in 6/8 (83 %) with the gel and in 8/8 (100 %) with the control fluid. One lesion at the greater curvature broke into three pieces during the attempt of mechanical pulling off using a cold snare. The other lesion located at the posterior wall was resected in two pieces: a piece of the specimen was ruptured by mechanical dissection with the grasping forceps, the other part by regular submucosal dissection. The total procedure time was comparable in both groups (median 28 min [gel] vs. 26 min [control]). Yet, the gel required more time for submucosal injection (3.46 vs. 1.59 min; p = 0.01) and circumferential incision (17.07 vs. 13.27 min; p = 0.48). Complete mechanical dissection was achieved in 3 gel lesions using a cold snare (2 en bloc, 1 piecemeal) before submucosal dissection by an ESDknife. In the gel lesions in which the mechanical dissection did not succeed, there was no difference in submucosal dissection time compared to the control lesions (6.54 vs. 8.44 min, respectively; p = 0.26).
Table 1 Endoscopic parameters during endoscopic submucosal dissection (ESD) Viscous gel (n = 12)
Control fluid (n = 12)
p value
En bloc resection, n (%)
10 (83)
12 (100)
0.50
Size specimen, median cm2 (IQR)
8.5 (7.3–10.5)
11.2 (8.2–15.1)
0.04
Procedure
Mechanical submucosal dissections By scraping with endoscope
0 (0 %)
0 (0 %)
1
By using grasping forceps
0 (0 %)
0 (0 %)
1
By using cold snare
3 (25 %)
0 (0 %)
0.25
Total (min) (IQR)
28.25 (20.34–32.21)
26.03 (16.26–46.37)
0.94
Submucosal injection (min) (IQR)
03.46 (02.06–08.12)
01.59 (01.44–02.34)
0.01
Circumferential incision (min) (IQR)
17.07 (11.41–19.48)
13.27 (06.56–28.11)
0.48
Submucosal dissection (min) (IQR)
06.54 (05.10–12.05)
08.44 (06.47–19.39)
0.26
Procedure time
Excluding mechanical dissection* Procedure time (corrected for specimen size) Circumferential incision (sec/ cm) (IQR)
97 (69–115)
64 (39–135)
0.18
Submucosal dissection (sec/ cm2) (IQR)*
55 (32–76)
53 (40–77)
0.86
16.5 (10.5–29.25)
43.0 (34.5–69.0)
0.002
6 (4–7)
13 (9–24)
0.002
3.5 9 (2.25–4.75)
5.5 9 (4–7.75)
0.01
Mild-resolves \5 min
0
4
Moderateresolves \15 min
1
0
Severe-resolves [15 min
0
0
Injection fluid Total amount (ml) Total number of injections Interchanges between accessories Complications Bleedings
123
Surg Endosc Table 1 continued
Perforation
Viscous gel (n = 12)
Control fluid (n = 12)
1
1
p value
* n = 9 (successful mechanical dissections using cold snare excluded)
As the median size of the specimen differed slightly between the two injection fluids (median 8.5 cm2 (IQR 7.3–10.5) [gel] vs. 11.2 cm2 (IQR 8.2–15.1) [control]), times were corrected, further decreasing the differences between the gel and control fluid for circumferential incision time and the submucosal dissection (Table 1). For the total ESD-procedure, the number of injections (p = 0.002) and the total injected amount (p = 0.002) of the gel was significantly lower compared to the control fluid. After initial submucosal injection for sufficient lifting, additional injections were needed during circumferential incision (the gel in 1/12 vs. control fluid in 4/12), before submucosal dissection (the gel in 1/12 vs. control fluid in 11/12) and during submucosal dissection (the gel in 0/12 vs. control fluid in 5/12). This difference in number of additional injections resulted in significantly less interchanges between accessories in the gel lesions compared to the control fluid lesions (p = 0.01). Bleeding was seen during four ESD procedures. In one procedure, a moderate bleeding occurred during the circumferential incision which was treated with the Coagrasper. In three procedures with the control fluid, 4 mild bleedings occurred (3 during circumferential incision, 1 during submucosal dissection) which were treated with the ESD-device (1x) or the Coagrasper (3x). Two small perforations occurred during submucosal dissection, one in a gel lesion (acute experiment) and one in a control fluid (3day survival) lesion. Both lesions were clipped and antibiotics were administered. Histopathological evaluation showed the maximum thickness of the submucosa in the resection specimen to be significantly larger in the lesions resected with the gel (median 649 lm [IQR 520–827 lm]) compared to the control-lifting fluid (480 lm [IQR 374–622 lm]), p = 0.005. No distinct histopathological differences appeared between the resection wounds of the gel and the control fluid lesions after ESD. Ulcer sizes were identical in both groups after 3 days (median 12.0 in gel vs. 12.4 cm2 in control) and after 28 days (median 0.30 in gel vs. 0.18 cm2 in control). At 3 days, severe inflammation was seen in the submucosa and MP in both groups. At 28 days, little inflammation was seen in the submucosa and MP, and little-to-medium fibrosis was present in both groups.
123
At necropsy, the stomach of one 3-day survival pig showed an abscess at the serosal side of the greater curvature (gel lesion) (Fig. 3). We speculate that transmural injection of the gel must have been the cause. This pig survived the predetermined period without any clinical signs of illness. At day 28, the amount of inflammation and fibrosis was identical in both groups. Experiment 2: inappropriate injections In the pigs with submucosal gel injection but no subsequent dissection (n = 12), the center of all four specimens from the non-survival pigs showed full disruption of the mucosa from the MP. No intact vertical fibers connecting the mucosa and MP were visible, and no signs of major bleeding were present (Fig. 4A). In the 3-day survival group in all specimen, the lifting was still present (median length and width were 5.5 and 5.5 cm, respectively). One of four lesions contained a mixture of blood and liquid content; the other three had a transparent liquid content. Histology showed some inflammation, starting encapsulation with granulation tissue, and a vital mucosa in all four specimens (Fig. 4B). At 28 days, one injection site was still lifted (4.5 9 3.0 cm). In the submucosa, an elevated area of submucosal granulation tissue was seen (Fig. 4D). The other specimens were flat with little submucosal granulation tissue and fibrosis. No major architectural changes remained (Fig. 4C). The mucosa was vital in all specimens. In the pigs with gel injection into the gastric MP (n = 4), eight lesions were created per survival group
Fig. 3 Transmural injection for ESD with the viscous gel in 3-day survival pig. In a 3-day survival pig, an abscess was found at the serosal side of the gastric wall containing a mix of the gel and granulocytes. At histopathological evaluation, the content was identified as the gel, probably due to accidental transmural injection
Surg Endosc
Fig. 4 Histopathological effects of the viscous gel injection without ESD. (A) Injection of the viscous gel without ESD in acute experiment showing disruption of vertical submucosal fibers. (B) Injection of the viscous gel without ESD resulted in gel-filled
submucosal cavities in 3 of 4 lesions after 3 days survival. After 28 days survival, injection of the viscous gel without ESD resulted in little remaining damage in 3 of 4 lesions (C), whereas in one an elevated area with granulation tissue remained (D)
(3-day and 28-day). In the 3-day survival group, submucosal cavities with bloody content next to the ulcers were macroscopically seen in 3/8 lesions (Fig. 5A, B). Six of eight lesions showed severe submucosal inflammation, whereas in 2/8 lesions, severe inflammation and necrosis of the MP were seen. In one lesion a hematoma at the serosal side was found. At histology, this lesion showed very severe necrosis of the MP and serosal inflammation. In the 28-day survival group, the EMR-ulcers had healed in all lesions with vital overlying mucosa. Two of the eight lesions still showed an encapsulated cavity with mediumgrade inflammation (Fig. 5D). Severity of submucosal fibrosis was classified as medium in 2, and as little in 6 lesions. One lesion showed little fibrosis at the serosal side, but no other signs of perforation were observed.
gel, the procedure was split up into three steps: the submucosal injection, the circumferential mucosal incision, and the submucosal dissection. Injection with the gel required more time to achieve proper submucosal lifting compared to the control fluid. However, during 25 % of the procedures with the gel no submucosal dissection was required as grabbing and pulling the specimen with a snare resulted in a complete mechanical dissection. This balance equalized the total procedure time. Several properties of the gel play a role in these observations. The main difference between the gel and the control fluid is the viscosity of the injection medium. This feature of the gel has several potential benefits. First, it more easily disrupts the submucosal fibers and small vessels, resulting in a (partial) mechanical separation of the mucosal layer and the MP. As shown in our data, this allowed a mechanical dissection with cold snare in 3 of 12 lesions with the gel. Second, the viscosity of the gel ensures that the injection fluid will not diffuse into the surrounding tissue. During the entire procedure, fewer injections are required to maintain the lifting of the lesion. Consequently, lower volumes of injection fluid and fewer interchanges between accessories are necessary. These characteristics may facilitate an easier ESD-procedure. On the other hand, the viscosity does have some disadvantages as well. First, given the high pressures needed for pressing the gel through a small needle, a specially
Discussion This study assessed the efficacy and safety of a new viscous submucosal lifting fluid for ESD in the stomach of a living porcine model. Our data suggest that the gel is effective and safe for this purpose. The main outcome parameter of this study was the time for the total procedure. We found no difference in the total procedure time when the gel was compared to our standard viscous solution. To evaluate the potential benefits of the
123
Surg Endosc
Fig. 5 Histopathological effects of intentional injection of viscous gel into the muscularis propria after EMR with multiband mucosectomy. After 3 days, deep injections resulted in cavities with bloody liquid content (A, B) in 3 of 8 lesions. In the other lesions, only the
ulcer of the EMR (C) was seen. After 28 days survival, 2/8 injection sites consisted of an encapsulated cavity without extensive inflammatory response (D). The other 6 injection sites healed without any architectural changes (E)
designed metal needle is required. Injection of the gel through the needle takes more time than conventional injection with a standard needle and a less viscous substance. In addition, although not tested in this experiment, the stiffness of this needle might hamper submucosal injection with the endoscope in the retroflex position. Second, due to the high viscosity of the gel, a little momentum is necessary for direct lifting. This is achieved by injecting a small volume of saline before injection of the gel. This preceding saline injection requires the use of an additional regular injection needle, as the gel needle needs to be primed and ready for use. Both these disadvantages may explain the prolonged time required for the submucosal injection with the gel compared to the control fluid in this study. Future modifications of the needle will likely speed up the submucosal injection procedure. Moreover, the use of the gel may harbor a learning curve and more extensive experience with its properties, as currently with the control fluid, may possibly lead to a faster procedure.
Histopathological evaluation of the resection specimen showed a significant difference in submucosal thickness between both fluid groups. This was not due to a deeper resection, but more a result of the effects of the gel. In the gel specimen, the tissue and structures were more spread out with gel in between, whereas in the control group the structures were more compact. This confirmed the hypothesis that the gel causes mechanical disruption. Evaluation of the ESD wounds showed no difference in amount of inflammation, necrosis, and fibrosis in both groups in the acute setting, or at short- and long-term survival. However, in one ESD lesion (3-day survival) an abscess was found in a pocket at the serosal side, although clinically the pig did not have any signs of infection. As neither endoscopically nor at histopathological assessment a perforation was observed, it is likely that this resulted from inadvertent deep injection during submucosal lifting. Ulceration has been reported to occur just by submucosal injection of fluids even without subsequent dissection, especially in hypertonic solutions such as hypertonic
123
Surg Endosc
dextrose (C20 %), hypertonic sodium chloride (3.75 %), or hydroxypropyl methylcellulose [18, 19]. In case of inappropriate submucosal injections of the gel, one might postulate the possibility of ulceration or even necrosis of the overlying mucosa because of mechanical disruption of vertical strands including the feeding vessels. In clinical practice, this might lead to an extended ulcer after ESD since the lifted area is often much larger than the actual resected specimen. Our results show that after 3 days the gel had not dissipated, but no severe inflammation or necrosis was seen. Considering the fact that the inflammatory response is most severe at 3–5 days after the procedure [20], safety and wound healing do not seem to be impaired by submucosal gel injection. After 28 days, the lifted tissue consisted of granulation tissue and no major architectural changes seemed to remain. These long-term results are in line with the persistent submucosal cushions without signs of mucosal ischemia or ulceration found by Chandrasekhara et al. [17]. Although the gel remains present for much longer compared to other injection agents, it seems safe when it comes to long-term histopathological outcomes if injected at locations that in the end will not be resected. Deep gel injections into the MP after initial EMR in the second part of this study did not cause any striking complications. Histopathological evaluation showed healing and vitality of all wall layers with no significant remaining architectural abnormalities. Some inflammation and fibrosis at the serosal side of the MP was observed, but no transmural necrosis was seen. Two of the total number of 16 lesions in the four pigs with deep injection did show inflammation and fibrosis at the serosal side of the gastric wall. It is conceivable that these serosal changes are caused by transmural instead of intramural injections. However, no gel was found in these specimens at necropsy and the pigs showed no clinical signs of illness. Therefore, the gel seems safe even when injected transmurally, as may be expected from its biocompatible characteristics. The major limitation in this experiment is the pig’s stomach as a study model. In the body of a pig’s stomach a relatively thick and stiff mucosa-submucosa layer is present. In fact, the stiffness of the intact gastric wall is determined by the stiffness of the mucosa-submucosa layer [21]. This characteristic of our model may have influenced our experiment, as proper lifting was not achieved as easily as anticipated. Higher pressures and often multiple injections were needed to create proper submucosal lifting. Consequently, the success rate of mechanical dissection, which is essentially the main quality of the gel, was lower than expected based on earlier results in a porcine model by Khashab et al., where only a single injection with the gel led to full mechanical dissection [22]. In the study of Khashab
et al., however, a different injection technique was applied. In contrast to the 2 cc injected in our study used to create a small cushion for safer gel injection, Khashab aimed to achieve complete lifting of the surrogate lesion by injection of 10 cc of saline prior to gel injection. Possibly, this might have created a pre-defined submucosal space for the gel to be more easily injected, thereby obviating the need for repetitive gel injections as appeared to be necessary in our study. Still, in our study in 25 % of the lesions dissection of the mucosa by cold snaring was successful. Furthermore, for all the other gel lesions the time needed for submucosal dissection was shorter, although not significantly, compared to the lesions lifted with the control fluid. This is suggestive for at least a partial mechanical dissection by the gel despite the model and possibly inferior injection technique used. Given all this, we anticipate that the use of the gel for ESD in the human stomach using the injection technique described by Khashab et al. might compare favorably with our results [17, 23, 24]. Second, due to failed mechanical dissection two gelESDs were not resected en bloc, which may hinder optimal histopathological evaluation and is associated with a higher chance for local recurrence [2–4]. However, these accidental piecemeal resections are more likely to be attributed to the difference in wall layers within the porcine model than to the gel itself, and may be different in humans. The final limitation of this study is the relatively low number of experiments which may underestimate the risk for uncommon complications. In conclusion, the viscous gel seems effective and safe for ESD in the stomach of a porcine model. Complete mechanical submucosal dissection was achieved in 25 % of cases. Gel ESD did not result in shorter procedure times, although this might—at least in part—be attributed to the limitations of the model used. Further assessment of the use of gel in humans using a more optimal injection technique may elucidate the potential beneficial features of the gel in a better way. Acknowledgments The authors thank Elly van Zwol, Sanne Hackmann, Alwin Blankestijn, and Hester de Bruin for their biotechnical assistance. This study was supported by Cook Medical, Limerick, Ireland. Prof. J.J.G.H.M. Bergman is a consultant for Cook Medical and received speaker fees from Cook Medical. Prof. H. Neuhaus is a consultant for Cook Medical. Disclosures D.W. Scholvinck, L. Alvarez Herrero, O. Goto, S.L. Meijer, H. Neuhaus, B. Schumacher, J.J.G.H.M. Bergman, and B.L.A.M. Weusten have no conflicts of interest or financial ties to disclose.
References 1. Kodashima S, Fujishiro M, Yahagi N, Kakushima N, Omata M (2006) Endoscopic submucosal dissection using flexknife. J Clin Gastroenterol 40(5):378–384
123
Surg Endosc 2. Park YM, Cho E, Kang HY, Kim JM (2011) The effectiveness and safety of endoscopic submucosal dissection compared with endoscopic mucosal resection for early gastric cancer: a systematic review and metaanalysis. Surg Endosc 25(8):2666–2677 3. Cao Y, Liao C, Tan A, Gao Y, Mo Z, Gao F (2009) Meta-analysis of endoscopic submucosal dissection versus endoscopic mucosal resection for tumors of the gastrointestinal tract. Endoscopy 41(9):751–757 4. Lian J, Chen S, Zhang Y, Qiu F (2012) A meta-analysis of endoscopic submucosal dissection and EMR for early gastric cancer. Gastrointest Endosc 76(4):763–770 5. Kakushima N, Fujishiro M, Kodashima S, Muraki Y, Tateishi A, Omata M (2006) A learning curve for endoscopic submucosal dissection of gastric epithelial neoplasms. Endoscopy 38(10): 991–995 6. Hotta K, Oyama T, Shinohara T, Miyata Y, Takahashi A, Kitamura Y, Tomori A (2010) Learning curve for endoscopic submucosal dissection of large colorectal tumors. Dig Endosc 22(4):302–306 7. Probst A, Golger D, Anthuber M, Markl B, Messmann H (2012) Endoscopic submucosal dissection in large sessile lesions of the rectosigmoid: learning curve in a European center. Endoscopy 44(7):660–667 8. Giday SA, Magno P, Buscaglia JM, Canto MI, Ko CW, Shin EJ, Xia L, Wroblewski LM, Clarke JO, Kalloo AN, Jagannath SB, Kantsevoy SV (2006) Is blood the ideal submucosal cushioning agent? A comparative study in a porcine model. Endoscopy 38(12):1230–1234 9. Lee SH, Park JH, Park do H, Chung IK, Kim HS, Park SH, Kim SJ, Cho HD (2006) Clinical efficacy of EMR with submucosal injection of a fibrinogen mixture: a prospective randomized trial. Gastrointest Endosc 64(5):691–696 10. Uraoka T, Fujii T, Saito Y, Sumiyoshi T, Emura F, Bhandari P, Matsuda T, Fu KI, Saito D (2005) Effectiveness of glycerol as a submucosal injection for EMR. Gastrointest Endosc 61(6): 736–740 11. Yamamoto H, Yahagi N, Oyama T, Gotoda T, Doi T, Hirasaki S, Shimoda T, Sugano K, Tajiri H, Takekoshi T, Saito D (2008) Usefulness and safety of 0.4 % sodium hyaluronate solution as a submucosal fluid ‘‘cushion’’ in endoscopic resection for gastric neoplasms: a prospective multicenter trial. Gastrointest Endosc 67(6):830–839 12. Lenz L, Di Sena V, Nakao FS, Andrade GP, Rohr MR, Ferrari AP Jr (2010) Comparative results of gastric submucosal injection with hydroxypropyl methylcellulose, carboxymethylcellulose and normal saline solution in a porcine model. Arq Gastroenterol 47(2):184–187 13. Feitoza AB, Gostout CJ, Burgart LJ, Burkert A, Herman LJ, Rajan E (2003) Hydroxypropyl methylcellulose: a better
123
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
submucosal fluid cushion for endoscopic mucosal resection. Gastrointest Endosc 57(1):41–47 Akagi T, Yasuda K, Tajima M, Suzuki K, Inomata M, Shiraishi N, Sato Y, Kitano S (2011) Sodium alginate as an ideal submucosal injection material for endoscopic submucosal resection: preliminary experimental and clinical study. Gastrointest Endosc 74(5):1026–1032 Moss A, Bourke MJ, Metz AJ (2010) A randomized, double-blind trial of succinylated gelatin submucosal injection for endoscopic resection of large sessile polyps of the colon. Am J Gastroenterol 105(11):2375–2382 Fernandez-Esparrach G, Shaikh SN, Cohen A, Ryan MB, Thompson CC (2009) Efficacy of a reverse-phase polymer as a submucosal injection solution for EMR: a comparative study (with video). Gastrointest Endosc 69(6):1135–1139 Chandrasekhara V, Sigmon JC Jr, Surti VC, Kochman ML (2013) A novel gel provides durable submucosal cushion for endoscopic mucosal resection and endoscopic submucosal dissection. Surg Endosc 27(8):3039–3042 Sato T (2006) A novel method of endoscopic mucosal resection assisted by submucosal injection of autologous blood (blood patch EMR). Dis Colon Rectum 49(10):1636–1641 Fujishiro M, Yahagi N, Kashimura K, Mizushima Y, Oka M, Enomoto S, Kakushima N, Kobayashi K, Hashimoto T, Iguchi M, Shimizu Y, Ichinose M, Omata M (2004) Comparison of various submucosal injection solutions for maintaining mucosal elevation during endoscopic mucosal resection. Endoscopy 36(7):579–583 Mitchell RN, Ramzi SC (2003) Tissue repair: cell regeneration and fibrosis. In: Kumar V, Ramzi SC, Robbins SL, Saunders WB (eds) Robbins basic pathology, 7th edn. Saunders, Philadelphia, pp 61–64 Zhao J, Liao D, Chen P, Kunwald P, Gregersen H (2008) Stomach stress and strain depend on location, direction and the layered structure. J Biomech 41(16):3441–3447 Khashab MA, Saxena P, Sharaiha RZ, Chavez YH, Zhang F, Kord Valeshabad A, Aguila G, Canto MI, Pasricha PJ, Kalloo AN (2013) A novel submucosal gel permits simple and efficient gastric endoscopic submucosal dissection. Gastroenterology 144(3):505–507 Saxena P, Simons BW, Gabrielson K, Haito-Chavez Y, Valeshabad AK, Kumbhari V, Singh VK, Lennon AM, Canto MI, Pasricha PJ, Kalloo A, Khashab MA (2014) A novel viscous dissecting gel for endoscopic submucosal dissection: a prospective survival study in a porcine model. Endoscopy 46(7):605–609 Samarasena JB, Shinoura S, Nakai Y, Choi KD, Lee JG, Chang KJ (2013) New submucosal lifting gel for endoscopic submucosal dissection of gastrointestinal lesions: ESD made easy with ‘‘needle-free auto-dissection. Gastrointest Endosc 77(5):AB189– AB190
