Author's personal copy Int J Colorectal Dis DOI 10.1007/s00384-012-1531-6
ORIGINAL ARTICLE
Possibility of ex vivo animal training model for colorectal endoscopic submucosal dissection Naohisa Yoshida & Nobuaki Yagi & Yutaka Inada & Munehiro Kugai & Kazuhiro Kamada & Kazuhiro Katada & Kazuhiko Uchiyama & Takeshi Ishikawa & Tomohisa Takagi & Osamu Handa & Hideyuki Konishi & Satoshi Kokura & Ken Inoue & Naoki Wakabayashi & Yasuhisa Abe & Akio Yanagisawa & Yuji Naito
Accepted: 27 June 2012 # Springer-Verlag 2012
K. Inoue Department of Gastroenterology, Kyoto Yosanoumi Hospital, Kyoto, Japan
Methods Harvested porcine cecum, rectum, and stomach and bovine cecum and rectum were analyzed regarding ease of mucosal injection, degree of submucosal elevation, and status of the proper muscle layer. Ex vivo animal model with blood flow was made using the bovine cecum. The vessel around the cecum was detached, and red ink was injected. Endoscopic hemostasis for perioperative hemorrhage and endoscopic closure for perforation were performed in this model. Results Mucosal injection was easily performed in the bovine cecum and rectum. Submucosal elevation was low in the bovine cecum, while the proper muscle layer was not tight in the porcine rectum and bovine cecum. Endoscopic hemostasis were accomplished in six (60 %) out of ten procedures of the ex vivo blood flow model. In two non-experts, the completion rates of endoscopic closure were 40 and 60 % in the first five procedures. These rates became 100 % in the last five procedures. Conclusions We have evaluated the characteristics of various ex vivo animal models and shown the possibility of training for endoscopic hemostasis and endoscopic closure in the ex vivo animal model.
N. Wakabayashi Department of Gastroenterology, Otsu City Hospital, Shiga, Japan
Keywords Endoscopic submucosal dissection . Animal model . Hemorrhage . Perforation . Training
Y. Abe Johnson & Johnson, Tokyo, Japan
Introduction
A. Yanagisawa Department of Surgical Pathology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
Early colorectal cancers must be diagnosed carefully by endoscopic examination, including chromoendoscopy and imageenhanced endoscopy techniques such as narrow band imaging
Abstract Purpose Colorectal endoscopic submucosal dissection (ESD) has not been standardized due to technical difficulties and requires extensive training for reliability. Ex vivo animal model is convenient, but has no blood flow. The objective of this study is to evaluate the characteristics of various ex vivo animal models including a blood flow model for colorectal ESD training and the usefulness of practicing endoscopic hemostasis and closure using an animal model. N. Yoshida (*) : N. Yagi : Y. Inada : M. Kugai : K. Kamada : K. Katada : K. Uchiyama : T. Ishikawa : T. Takagi : O. Handa : H. Konishi : S. Kokura : Y. Naito Department of Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, 465 Kajii-cho, Kawaramachi-Hirokoji, Kamigyo-ku, Kyoto 602-8566, Japan e-mail:
[email protected]
Author's personal copy Int J Colorectal Dis
and flexible spectral imaging color enhancement [1–3]. Improvements in endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) have facilitated the removal of large, early colorectal cancers [4–7]. In particular, the rate of en bloc resection of colorectal tumors more than 20 mm in diameter has been reported to be 84.0–98.9 % when ESD is used [8–13]. The ESD procedure has not been standardized due to technical difficulties. The perforation rate for ESD is reported to be higher than that for EMR [4, 5, 14]. A safe strategy, a suitable knife, and the adoption of other equipment are necessary to prevent the associated complications, including perforation of ESD. In general, endoscopists should acquire extensive experience with gastric ESD before performing colorectal ESD. However, different training for colorectal ESD is required when the number of patients with early gastric cancer is low, such as in the Western world. In this environment, observing and visiting ESD experts in their native institutions is an important component for the training of less experienced endoscopists. Another expected component of ESD training is extensive practice using animal models [15–17]. Both in vivo animal models and ex vivo animal models using harvested organs have been used. Porcine and canine in vivo models have been reported to be useful for ESD training [15, 17]. However, in vivo animal models are expensive and inconvenient and require that the animal is euthanized afterward. In contrast, ex vivo animal models are inexpensive and convenient. Hon et al. demonstrated the usefulness of a porcine colon ex vivo animal model for training in colorectal ESD [17]. However, one of the weaknesses of ex vivo animal models is the lack of blood flow. Prevention of perioperative hemorrhaging and rapid hemostasis are important techniques in clinical ESD. Recently, an ex vivo animal model with blood flow was developed by Johnson & Johnson. We have helped contribute to this along with many other Japanese endoscopists. This model has been used subsequently in academic seminars for the training of less experienced endoscopists in colorectal ESD. However, the characteristics and efficacy of various ex vivo animal models (including the blood flow model) have not yet been evaluated in detail. In the current study, we evaluated the characteristics of various ex vivo animal models including the blood flow model for colorectal ESD training and the possibility of practicing endoscopic hemostasis and endoscopic closure using an ex vivo animal model.
Methods Preparation and evaluation of various ex vivo animal models Harvested porcine cecum, rectum, and stomach and bovine cecum and rectum were obtained from fresh meat markets for use in this study. In preparation for training in colorectal
ESD, an endoscopic overtube (TOP, Tokyo, Japan) was attached to one end of the organ for insertion of the endoscope (Fig. 1a). For organs other than the cecum, the other end was ligated. To mimic the abdomen, the stomach was fixed to a plastic box and the cecum and rectum were fixed to a square board (Fig. 1a). The patient plate electrode was attached below the organs. Bovine cecum was used for the ex vivo animal model with blood flow. To mimic blood flow, red ink was used. The artery around the bovine cecum was detached (Fig. 1b). A plastic needle was inserted into the artery and a syringe containing red ink was connected. The red ink was injected until the mucosal vasculature could be visualized by endoscopy (Fig. 2a, b). If the injection pressure was excessive, the other vessel was detached and used. Thereafter, red ink was maintained by intermittent injection during ESD. After the preparation, all ex vivo models were used five times to perform ESD to normal mucosa with 20–30 mm in diameter, including mucosal injection, mucosal incision, and submucosal dissection. We evaluated ease of mucosal injection, submucosal elevation during dissection, and status of the proper muscle layer in all ex vivo animal models. The ease of mucosal injection was evaluated by the number of injections required to achieve adequate submucosal elevation. Five procedures were performed and the mean number of injections was calculated. Submucosal elevation was classified as “high” or “low” compared to the human colorectum in clinical ESD (Fig. 3a, b). The status of the proper muscle layer was classified as “tight” or “not tight” by the evaluation of the muscle layer during or after ESD (Fig. 3c, d). Evaluations of submucosal elevation and the status of the proper muscle layer were performed by three endoscopists (N.Y., M.K., and Y.I.), and the majority of evaluation of these three endoscopists was adopted. The ex vivo blood flow model using bovine cecum was evaluated ten times using ten blood flow models using bovine cecum. The visibility of submucosal vessels during submucosal dissection and the number of perioperative hemorrhages in mucosal injection, mucosal incision, and submucosal dissection were analyzed. Moreover, the ability to perform training in endoscopic hemostasis for perioperative hemorrhages during ESD was analyzed. The endoscopic hemostasis was achieved by coagulation by an ESD knife or by special hemostat forceps (Coagrasper, Olympus Medical Co., Tokyo, Japan) and the number of hemostats was counted [14]. We also analyzed the efficacy of in vivo animal models using living minipig colon (Specific Pathogen Free Swine: Sanesu Grand Parent Stock (Secondary SPF)), for comparison to the ex vivo blood flow animal model. A minipig was fasted overnight and was placed in the left lateral decubitus position after tracheal intubation with general anesthesia. After the preparation, this model was used five times to perform ESD to normal mucosa with 20–30 mm in diameter. The visibility of submucosal vessels during submucosal dissection and the
Author's personal copy Int J Colorectal Dis
Fig. 1 Preparation of the ex vivo blood flow animal model. a The ex vivo blood flow animal model was made using bovine cecum. An endoscopic overtube was attached to one end of the organ for insertion
of the endoscope. A small syringe of red ink was connected to the model. b The artery around the bovine cecum was detached. A plastic needle was inserted into the artery
number of perioperative hemorrhage in mucosal injection, mucosal incision, and submucosal dissection were evaluated.
clips (EZclip: HX-610-135 and HX-110QR, Olympus Medical System Co., Ltd., Tokyo, Japan) (Fig. 4b). The over-the-scope clip probably enabled us to close this size of the hole; however, it was not approved by Japanese health insurance. Thus, we used the EZclip system [20]. The endoscopic closure was performed ten times by each of the three endoscopists, one expert (N.Y.) and two non-experts (M.K., Y.I.). The expert had performed more than 300 colorectal ESD procedures and the two non-experts had each performed fewer than ten colorectal ESD procedures. The procedure time for each endoscopic closure was calculated to determine the learning curve. The completion of endoscopic closure was confirmed by cessation of the air leaking out of the cecum.
ESD For ESD, we used a lower gastrointestinal endoscope with a single channel (EC-590MP; Fujifilm Medical Co., Tokyo, Japan). The injection solution used in this study was 0.4 % hyaluronic acid (800 kDa preparation: Mucoup; Johnson & Johnson, Tokyo, Japan; Seikagaku Corporation, Tokyo, Japan) with a small amount of indigo carmine [18]. A 25gauge needle 3 mm in length (01885: TOP Co., Tokyo, Japan) and a small syringe were used. The flush knife BT (Fujifilm Medical Co., Tokyo, Japan) and clutch cutter (Fujifilm Medical Co., Tokyo, Japan) were used as the ESD knife [13, 19]. ESD was performed as previously described [14]. Practicing endoscopic closure The animal model with perforation was developed using the bovine cecum. After EMR or ESD, the endoscopic knife was used to make a 2–3-mm hole in the proper muscle layer of the ulceration (Fig. 4a). The perforation was confirmed as a small hole on the outside of the model. The endoscopic closure of the hole was performed with three endoscopic Fig. 2 The mucosa of the ex vivo blood flow animal model. a The mucosa of the ex vivo bovine cecum before the injection of red ink. No blood flow was visible in the mucosa. b The mucosa showing “blood” flow after red ink was injected
Results For mucosal injection of various ex vivo animal models, the mean number of injections was less than 2 for both the bovine cecum (1.5) and rectum (1.3) (Table 1). In contrast, the number of injections was greater than 2 for the porcine cecum (2.6), rectum (2.5), and stomach (2.4). Submucosal elevation was high in the porcine cecum, rectum, and stomach and the bovine rectum but low in the bovine cecum. The status of the proper muscle layer was not tight in the porcine rectum
Author's personal copy Int J Colorectal Dis Fig. 3 Submucosal elevation and state of proper muscle layer of various ex vivo animal models. a Submucosal elevation in the bovine cecum. This was evaluated as “low.” b Submucosal elevation in the porcine stomach. This was evaluated as “high.” c The status of the proper muscle layer of the bovine rectum. This was classified as “tight” during and after ESD (black arrow). There was no slit in the proper muscle layer. d The status of the proper muscle layer of the bovine cecum. This was classified as “not tight.” The blue injection liquid was seen in the slit in the proper muscle layer
and the bovine cecum and tight in the porcine cecum and stomach and the bovine rectum. In the ex vivo blood flow model, the perioperative hemorrhage during mucosal injection and mucosal incision was observed in eight (80 %) out of ten ex vivo animal model procedures. On the other hand, that was observed in two (40 %) of five in vivo animal model procedures. Submucosal vessels were detectable by endoscopy in nine (90 %) out of ten ex vivo animal model procedures (Table 2). On the other hand, submucosal vessels were detectable in two (40 %) out of five in vivo animal model procedures. The perioperative hemorrhage during submucosal dissection was detected in seven (70 %) of the ten ex vivo animal model procedures (Fig. 5a, b). Moreover, endoscopic hemostasis in mucosal incision and submucosal dissection was needed and achieved in seven (60 %) of the ten ex vivo animal model procedures. On the Fig. 4 Ex vivo animal model with perforation for training of endoscopic closure. a After ESD, the endoscopic knife was used to make a 2–3-mm hole in the proper muscle layer of the ulceration. b The endoscopic closure of the hole was performed with three endoscopic clips
other hand, in the in vivo animal model procedure, perioperative hemorrhage due to submucosal dissection was detected in two (40 %) out of five procedures and endoscopic hemostasis was not needed in any of the procedures. In the animal model with perforation, endoscopic closure using three endoscopic clips was achieved by all three endoscopists in similar situations. The expert endoscopist completed all ten procedures with a mean procedure time of 2.9 min (range 2.5–3.5 min) (Table 3). One non-expert endoscopist (M.K.) completed seven (70 %) out of ten procedures with a mean procedure time of 4.7 min (range 3.8–5.8 min). The other non-expert endoscopist (Y.I.) completed eight (80 %) out of ten procedures with a mean procedure time of 3.9 min (range 3.0–4.5 min). One nonexpert endoscopist (M.K.) failed to complete three of the first five procedures while the other non-expert endoscopist
Author's personal copy Int J Colorectal Dis Table 1 The characteristics of various ex vivo animal models Number of mucosal injection
Submucosal elevation
Status of proper muscle layer
Porcine cecum
2.6
High
Tight
Porcine rectum Porcine stomach
2.5 2.4
High High
Not tight Tight
Bovine cecum
1.5
Low
Not tight
Bovine rectum
1.3
High
Tight
(Y.I.) failed to complete two of the first five procedures. However, both non-expert endoscopists completed all of their last five procedures. The median procedure times of the two non-expert endoscopists were 4.7 min (M.K.) and 3.9 min (Y.I.), which were longer than that of the expert endoscopist. Comparison between procedure times of the first five procedures and the last five procedures in the two non-expert endoscopists was as follows: 5.1 and 4.2 min for M.K. and 4.2 and 3.6 min for Y.I.
Discussion In the current study, we have evaluated the characteristics of various ex vivo animal models and shown the possibility of training for endoscopic hemostasis and endoscopic closure in ex vivo animal model. Mucosal injection was difficult in the porcine cecum, rectum, and stomach, possibly because Table 2 The status of perioperative hemorrhage of ex vivo blood flow animal model using bovine cecum compared to that of the in vivo model
the porcine rectum and cecum are thinner than the human colorectum [17]. Conversely, mucosal injection was difficult in the porcine stomach due to the thickness of the wall. Mucosal injection was easy in the bovine cecum and rectum, which might therefore be appropriate for training beginners, although mucosal injection is sometimes difficult in human colorectal ESD. In submucosal elevation, high submucosal elevation, which facilitated submucosal dissection, was detected in the porcine cecum, rectum, and stomach and the bovine rectum. In contrast, low submucosal elevation was observed in the bovine cecum, complicating submucosal dissection. The status of the proper muscle layer was not tight in the porcine rectum and the bovine cecum. The proper muscle layer is sometimes not tight in human colorectal ESD. However, the bovine cecum was not tight greatly, making ESD difficult. The widths of the organs were also different. The porcine cecum, porcine stomach, and bovine cecum were wider than the human colon and rectum, while the porcine rectum and bovine rectum were similar to the human rectum. There were no folds in the porcine rectum, which made ESD too easy. The bovine cecum and rectum are expected for ESD training for the lesion on the folds. The bovine cecum was ultimately chosen for the ex vivo blood flow animal model, because the vessels were more easily detached than in the other animal models. The vessels in the porcine cecum were thin and there was a great deal of fat around the porcine and bovine rectums. Thus, these models were not appropriate for the blood flow model. Various ex vivo animal models had characteristic features, making it possible to choose a suitable animal model
The number of hemorrhages due to mucosal injection and mucosal incision Ex Ex Ex Ex Ex
Visibility of submucosal vessels
The number of hemorrhages due to submucosal dissection
The number of endoscopic hemostasis
vivo 1st procedure vivo 2nd procedure vivo 3rd procedure vivo 4th procedure vivo 5th procedure
1 1 1 2 1
Yes Yes Yes Yes Yes
1 2 0 0 1
1 2 0 0 0
Ex vivo 6th procedure Ex vivo 7th procedure Ex vivo 8th procedure Ex vivo 9th procedure Ex vivo 10th procedure In vivo 1st procedure In vivo 2nd procedure In vivo 3rd procedure In vivo 4th procedure In vivo 5th procedure
0 1 0 1 2 1 0 0 1 0
Yes Yes No Yes Yes No Yes Yes No No
1 2 0 2 1 0 1 1 0 0
1 1 0 2 2 0 0 0 0 0
Author's personal copy Int J Colorectal Dis Fig. 5 Ex vivo blood flow animal model for training of endoscopic hemostasis. a In the ex vivo blood flow model using bovine cecum, the submucosal vessels were visible (white arrow). b The occurrence of perioperative hemorrhage was accomplished after knife electric cutting of the vessel
according to the skill level of the endoscopist. Based on its various characteristics, we recommend the bovine rectum for training beginners in colorectal ESD. In Japan, colorectal ESD training is a step-by-step system starting with observing and assisting in ESD procedures performed by experts. Next, animal model training is performed to the extent possible. Finally, clinical practice is performed under the supervision of instructors. Generally, clinical practice training proceeds according to the difficulty of the procedure, beginning with gastric ESD, then rectal ESD, and finally colonic ESD. Regarding animal training in Table 3 The procedure time and completion in endoscopic closure No. of procedure
1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th
Procedure time and completion Expert (N.Y.)
Non-expert (M.K.)
Non-expert (Y.I.)
3.5 min Complete 2.9 min
5.8 min Incomplete 5.2 min
4.0 min Incomplete 4.5 min
Complete 3.0 min Complete 2.9 min Complete 3.2 min Complete
Complete
Complete
4.6 min Incomplete 5.0 min Incomplete 5.0 min Complete
4.5 min Complete 4.0 min Incomplete 4.0 min Complete
3.0 min Complete 2.5 min Complete 2.9 min Complete 2.5 min Complete
4.4 min Complete 4.2 min Complete 4.4 min Complete 4.2 min Complete
4.2 min Complete 3.9 min Complete 3.4 min Complete 3.0 min Complete
2.9 min Complete
3.8 min Complete
3.4 min Complete
ESD, there are many reports on ex vivo animal models for gastric ESD [16, 21, 22] but only one on an ex vivo animal model for colorectal ESD [17]. Repeated animal model training procedures have recently been proven to decrease procedure time [16, 17]. Our study also showed that repeated endoscopic closure improved the non-experts’ completion rates and decreased their procedure times. The importance of experience was shown by studies that reported that a beginner must perform about 20–40 gastric ESD procedures to gain early proficiency [23, 24]. For colorectal ESD, Hotta et al. showed that approximately 40 procedures were sufficient to acquire skill in avoiding perforations, and the perforation rate in the first 40 cases was about 12.5 % [25]. We believe that experience obtained by training on an animal model would also improve performance of clinical colorectal ESD. Perioperative hemorrhage is one of the major complications of ESD. Yamamoto et al. indicated that improvements in submucosal dissection technique, especially in hemostasis, might have contributed to the observed increase in completion rate and shortening of operation time [26]. Training in endoscopic hemostasis is difficult in conventional ex vivo animal models. However, the ex vivo blood flow animal model in the current study would allow endoscopist to gain experience with perioperative hemorrhage. Moreover, this model was both easier to set up and less expensive than an in vivo animal model. We also analyzed the perioperative hemorrhage about in vivo animal model using live minipig colon. The rate of perioperative hemorrhage due to mucosal injection and mucosal incision was higher in the ex vivo animal model (80 %, eight out of ten procedures) than that in the in vivo animal model (40 %, two out of five procedures). The rate of perioperative hemorrhage due to submucosal dissection was also higher in the ex vivo animal model (70 %, seven out of ten procedures) than that in the in vivo animal model (40 %, two out of five procedures). Moreover, endoscopic hemostasis was experienced in six (60 %) out of ten ex vivo procedures. In contrast, endoscopic hemostasis was not experienced in all five in vivo animal model procedures because all hemorrhage in the in vivo animal
Author's personal copy Int J Colorectal Dis
models stopped spontaneously. One of the possible reasons for this was that the live minipig used for this study was small. In detail, the weight of the minipig (about 30 kg) was smaller than that of the ex vivo bovine model (more than 150 kg). The vessels in the minipig’s colon were possibly smaller than those of bovine models. However, the preparation of a bigger live minipig is not convenient. In view of these points, we consider that the ex vivo blood flow animal model is more suitable for training of endoscopic hemostasis than the in vivo animal model. Moreover, the perioperative hemorrhage was able to become massive by increasing the pressure of injection of red ink in this ex vivo animal model. Pulsed hemorrhage like arterial hemorrhage was also possible by pulsed injection of red ink, using this model. One of the most serious complications of colorectal ESD is perforation. If perforation occurs, endoscopic closure using clips is effective [26]. In clinical colorectal ESD, the rate of perforation has been reported to be 1.4–10.4 %, and perforation is associated with large tumor size and the presence of fibrosis [14]. The coagulation by knife is the most frequent cause of perforation [27]. The scissor-shaped knife such as a clutch cutter is useful for preventing perforations [19]. The scissor-shaped knife is especially safe because it can grasp and cut a piece of tissue without unintentional incision. Saito et al. showed that perforation was related to the number of ESD procedures, with higher risk when the endoscopist had performed fewer than 100 procedures [28]. However, the perforation rate did not decrease to zero even if the skill level improved greatly. Therefore, we believe that the endoscopist must also obtain expertise in endoscopic closure. Small perforations can be closed by endoscopic clipping [27, 29]. However, endoscopic clipping requires a high level of endoscopic skill and experience, and perforation is relatively rare in clinical medicine, making it difficult to gain experience in the endoscopic clipping technique in clinical practice. On the other hand, frequent practices of endoscopic clipping for perforation cannot be experienced in an in vivo animal model, because massive perforation and incomplete endoscopic clipping might lead to the death of the in vivo animal model. That is why our ex vivo animal model with perforation is more useful for training in endoscopic closure than the in vivo animal model. The limitation of this study is that the benefit of training is assessed only in the animal model. There are several differences between the harvested animal model and the human colorectum that could produce different results. Other limitation is that the success of making this model is due to freshness of organs and selection of the artery. If the selection of artery and freshness of organs are accomplished, 20–40 % of the model fails due to unknown reasons. We are trying to investigate the unknown reasons.
Acknowledgments We thank Dr. Noriya Uedo, Dr. Toshio Uraoka, Dr. Ken Ohata, Dr. Shinji Tanaka, Dr. Kiyoaki Homma, Dr. Hirohisa Machida, Dr. Yoshinori Morita, and Dr. Naohisa Yahagi for providing helpful advice for developing the ex vivo animal model with blood flow. Conflict of interest None.
References 1. Kudo S, Hirota S, Nakajima T et al (1994) Colorectal tumours and pit pattern. J Clin Pathol 47:880–885 2. Tobaru T, Mitsuyama K, Tsuruta O et al (2008) Sub-classification of type VI pit patterns in colorectal tumors: relation to the depth of tumor invasion. Int J Oncol 33:503–508 3. Yoshida N, Naito Y, Kugai M et al (2011) Efficacy of magnifying endoscopy with flexible spectral imaging color enhancement in the diagnosis of colorectal tumors. J Gastroenterol 46:65–72 4. Tanaka S, Haruma K, Oka S et al (2001) Clinicopathological features and endoscopic treatment of superficially spreading colorectal neoplasms larger than 20 mm. Gastrointest Endosc 54:62–66 5. Saito Y, Fukuzawa M, Matsuda T et al (2010) Clinical outcome of endoscopic submucosal dissection versus endoscopic mucosal resection of large colorectal tumors as determined by curative resection. Surg Endosc 24:343–352 6. Yoshida N, Naito Y, Yagi Y et al (2012) Importance of histological evaluation in endoscopic submucosal dissection and endoscopic mucosal resection for early colorectal cancer. World J Gastrointest Pathophysiol 3:44–59 7. Yoshida N, Naito Y, Kugai M et al (2011) Efficacy of hyaluronic acid in endoscopic mucosal resection for colorectal tumors. J Gastroenterol Hepatol 26:286–291 8. Tanaka S, Oka S, Kaneko I et al (2007) Endoscopic submucosal dissection for colorectal neoplasia: possibility of standardization. Gastrointest Endosc 66:100–107 9. Saito Y, Uraoka T, Matsuda T et al (2007) Endoscopic treatment of large superficial colorectal tumors: a case series of 200 endoscopic submucosal dissections (with video). Gastrointest Endosc 66:966– 973 10. Fujishiro M, Yahagi N, Kakushima N et al (2007) Outcomes of endoscopic submucosal dissection for colorectal epithelial neoplasms in 200 consecutive cases. Clin Gastroenterol Hepatol 5:678–683 11. Yoshida N, Naito Y, Sakai K et al (2010) Outcome of endoscopic submucosal dissection for colorectal tumors in elderly people. Int J Colorectal Dis 25:455–461 12. Takeuchi Y, Uedo N, Ishihara R et al (2010) Efficacy of an endo-knife with a water-jet function (flushknife) for endoscopic submucosal dissection of superficial colorectal neoplasms. Am J Gastroenterol 105:314–322 13. Toyonaga T, Man-I M, Morita Y et al (2009) The new resources of treatment for early stage colorectal tumors: EMR with small incision and simplified endoscopic submucosal dissection. Dig Endosc 21 (Suppl 1):S31–S37 14. Yoshida N, Yagi N, Naito Y et al (2010) Safe procedure in endoscopic submucosal dissection for colorectal tumors focused on preventing complications. World J Gastroenterol 16:1688–1695 15. Tanimoto MA, Torres-Villalobos G, Fujita R et al (2010) Endoscopic submucosal dissection in dogs in a World Gastroenterology Organisation training center. World J Gastroenterol 16:1759–1764 16. Parra-Blanco A, Arnau MR, Nicolás-Pérez D et al (2010) Endoscopic submucosal dissection training with pig models in a Western country. World J Gastroenterol 16:2895–2900
Author's personal copy Int J Colorectal Dis 17. Hon SS, Ng SS, Lee JF et al (2010) In vitro porcine training model for colonic endoscopic submucosal dissection: an inexpensive and safe way to acquire a complex endoscopic technique. Surg Endosc 24:2439–2443 18. Hirasaki S, Kozu T, Yamamoto H et al (2009) Usefulness and safety of 0.4 % sodium hyaluronate solution as a submucosal fluid “cushion” for endoscopic resection of colorectal mucosal neoplasms: a prospective multi-center open-label trial. BMC Gastroenterol 9:1 19. Akahoshi K, Motomura Y, Kubokawa M et al (2009) Endoscopic submucosal dissection of a rectal carcinoid tumor using grasping type scissors forceps. World J Gastroenterol 15:2162–2165 20. Schurr MO, Hartmann C, Ho CN et al (2008) An over-the-scope clip (OTSC) system for closure of iatrogenic colon perforations: results of an experimental survival study in pigs. Endoscopy 40:584–588 21. Vazquez-Sequeiros E, de Miquel DB, Olcina JR et al (2009) Training model for teaching endoscopic submucosal dissection of gastric tumors. Rev Esp Enferm Dig 101:546–552 22. Yamamoto H (2007) Technology insight: endoscopic submucosal dissection of gastrointestinal neoplasms. Nat Clin Pract Gastroenterol Hepatol 4:511–520
23. Gotoda T, Friedland S, Hamanaka H et al (2005) A learning curve for advanced endoscopic resection. Gastrointest Endosc 62:866–867 24. Choi IJ, Kim CG, Chang HJ et al (2005) The learning curve for EMR with circumferential mucosal incision in treating intramucosal gastric neoplasm. Gastrointest Endosc 62:860–865 25. Hotta K, Oyama T, Shinohara T et al (2010) Learning curve for endoscopic submucosal dissection of large colorectal tumors. Dig Endosc 22:302–306 26. Yamamoto S, Uedo N, Ishihara R et al (2009) Endoscopic submucosal dissection for early gastric cancer performed by supervised residents: assessment of feasibility and learning curve. Endoscopy 41:923–928 27. Yoshida N, Wakabayashi N, Kanemasa K et al (2009) Endoscopic submucosal dissection for colorectal tumors: technical difficulties and rate of perforation. Endoscopy 41:758–761 28. Saito Y, Uraoka T, Yamaguchi Y et al (2010) A prospective, multicenter study of 1111 colorectal endoscopic submucosal dissections (with video). Gastrointest Endosc 72:1217–1225 29. Fujishiro M, Yahagi N, Kakushima N et al (2006) Successful nonsurgical management of perforation complicating endoscopic submucosal dissection of gastrointestinal epithelial neoplasms. Endoscopy 38:1001–1006
