The use of the submucosal space for NOTES procedures was first reported by Gostout et al. [3]. Full-thickness resection of GI tract lesions through the creation of a submucosal tunnel starts with submucosal injection of normal saline 20–40 mm proximal to the lesion [4]. The mucosa is incised to enter the submucosal space. The next procedural step is the creation of a submucosal tunnel toward the GI tract lesion similar to the technique utilized in per-oral endoscopic myotomy (POEM) [5–10]. The lesion is then resected inside the tunnel and removed, and the mucosal entrance to the tunnel is closed with endoscopic clips or sutures.

This technique can be used for removal of lesions originating from the submucosal or muscularis propria layers of the esophageal and gastric wall, but not the mucosa. Resection of these deeper lesions often results in the entrance into the mediastinum or peritoneal cavity. However, even in the case of resection involving the serosa, the submucosal tunnel separates the site of mucosal entrance from the exit into to the mediastinal or peritoneal space by 20–40 mm, which decreases the chance of post-procedure mediastinitis or peritonitis. After removal of the lesion, the entrance to the submucosal tunnel is usually closed with endoscopic clips or endoscopic sutures, preventing contamination of the mediastinal or peritoneal space by GI tract contents.

To clarify, submucosal tunneling resection can be used to perform deeper resections involving the serosa, and so the convention has emerged that we call these “full-thickness” resections. However, the overlying mucosa must be left intact in this method to prevent perforation. Therefore, the submucosal tunneling approach cannot be used for lesions involving mucosa and is not truly a full-thickness resection. Also, although the submucosal tunneling approach for “full-thickness” resection appears to be safe and effective, it only can be used in certain anatomic locations allowing for the creation of a submucosal tunnel (esophagus, stomach).

“Full-thickness” resection of gastric tumors (i.e., resections involving the serosa) through the submucosal tunnel technique was recently reported by Lee et al. Five lesions (3 gastrointestinal stromal tumors and 2 schwannomas) were accessed via 40-mm submucosal tunnels and resected en bloc [11]. All procedures were performed under conscious sedation, and the mean tumor size was 20.8 ± 3.27 mm. The mean hospital stay post-endoscopic submucosal tunnel full-thickness resection of gastric lesions was only 3.8 ± 1.48 days, and there were no procedure-related complications (bleeding, peritonitis, etc.) reported.

A “closed” approach to full-thickness resection of GI tract lesions implies plication of the GI tract wall, with subsequent resection of the lesion without entrance into the peritoneal cavity. The first clinical application of the “closed” technique was reported by Kaehler et al. [12, 13] for full-thickness resection of gastric lesions. He used a flexible endoscopic stapler to create a plication of the gastric wall, and then, the lesion was cut out of the inside of the stomach and removed without entrance into the peritoneal cavity. Unfortunately, the flexible endoscopic stapler is no longer available.

Another application of the “closed” technique was recently reported for a full-thickness resection of colonic lesions [14]. The authors used a specially designed, large OTSC^® over-the-scope clip (Ovesco Endoscopy, Tübingen, Germany) mounted on the distal end of the endoscope. Colonic lesions were pulled into the full-thickness resection device; then, the clip was closed, isolating the lesion from the peritoneal cavity. The lesion could then be resected. The initial pilot study (25 patients) reported technical success in 83.3% (20/24) with complete (R0) resection in 75% (18/24) and full-thickness resection in 87.5% (21/24). Minor adverse events (bleeding, postpolypectomy syndrome) were observed in 3 patients (12%). Follow-up endoscopy revealed residual adenoma in 4 (16%) and local recurrence in 1 patient (4%).

Following this pilot study, preliminary results of a large, prospective, multicenter trial were reported during Digestive Disease Week in May 2016. Nine academic referral centers in Germany recruited 180 patients from February 2015 through April 2016 [15]. Mean lesion size was 17.6 mm. Technical success was achieved in 88.3% of patients (159/180), with complete resection (R0) in 78.1% (139/178), and confirmed full-thickness resection in 78.4% (120/153). Adverse events were observed in 9.4% of patients (17/180), including 12 major adverse events (bleeding in 5, perforation in 5, and appendicitis in 2). Surgical correction was needed in 7.2% of patients (13/180). However, follow-up in 3 months demonstrated residual or recurrent lesions in 10% (9/90) patients.

Despite the theoretical attractiveness of the “closed” full-thickness resection technique, available studies demonstrated its obvious limitations: Only relatively small lesions (mean 17.6 mm) located in favorable positions can be removed, but with a high rate of adverse events (9.4%) and high rate of residual or recurrent lesions (10.0%) [12, 13, 15].

The first endoscopic full-thickness resection of GI tract lesions was described in a live porcine model by Ikeda et al. [16]. The authors performed full-thickness resection of the gastric wall in 4 acute and 8 survival animal experiments. Large, full-thickness defects in the gastric wall were closed endoscopically with T-tags inserted into the peritoneal cavity through blind puncture of the gastric wall with a hollow needle. Unfortunately, this closure technique could potentially damage adjacent intraperitoneal organs and is now practically abandoned [17].

Since that time, several studies demonstrated the feasibility of endoscopic full-thickness resection of GI tract lesions using the “open” technique. Huang et al. [18] reported endoscopic full-thickness resection of 13 gastric stromal tumors over 20 mm in size originating from the muscularis propria. All defects in the gastric wall were closed with endoscopic clips without any complications. Shi et al. [19] also reported full-thickness resection of gastric submucosal tumors (mean lesion size 14.7 ± 7.2 mm, range 4–30 mm) originating from the muscularis propria layer in 20 patients. All lesions were resected en bloc (100%), with subsequent closure of full-thickness gastric wall defects by a combination of endoloops (Olympus, Tokyo, Japan) and endoscopic metallic clips. There were no serious complications (peritonitis, abdominal abscess, etc.) after full-thickness resection, although 5 patients developed fever and slight abdominal pain on the first day after the procedure. The largest series (30 patients) of “open” full-thickness resection of gastric submucosal tumors was published by Fan et al. [20]. The mean size of the resected lesions was 19 mm, and all full-thickness defects in the gastric wall were closed with the combination of an endoloop and endoscopic clips. There were no serious complications during or after the procedure, and follow-up endoscopy 1 month after resection confirmed complete healing of the gastric wall defect.

Zhang et al. compared endoscopic “open” full-thickness resection (n = 22) of gastric stromal tumors over 20 mm in size to laparoscopic resection (n = 20) [21]. Endoscopic full-thickness resection was similar to laparoscopic resection in mean operative times (90 ± 17 min vs. 95 ± 21 min, p > 0.05), complete resection rates (100% vs. 95%, p > 0.05), and mean length of hospital stay (6.0 ± 1.8 days vs. 7.3 ± 1.7 days, p > 0.05). None of the patients in the endoscopic full-thickness resection group had any complications, whereas one patient in the laparoscopic resection group required conversion to laparotomy, and one experienced postoperative gastroparesis. No recurrences were observed in either group. The authors optimistically concluded that endoscopic full-thickness resection may replace surgical or laparoscopic procedures for removal of gastric stromal tumors.

Endoscopic “open” full-thickness resection of colonic lesions in a live animal model produced very heterogeneous results. Brigic et al. [22] analyzed animal studies devoted to endoscopic full-thickness resection of colonic lesions published between 1990 and 2012. A total of 113 procedures was performed on 99 porcine models, with an overall success rate of 89% and 4% mortality. The intra-operative complication rate was prohibitively high (mean = 22%; range = 0–67%). Significant heterogeneity was observed in procedure duration (median or mean 3–233 min) and size of the excised specimen (median or mean 1.7–3.6 cm). Defect closure techniques included endoscopic stapling devices, T-tags, compression closure, or closure with laparoscopic assistance. There was a high rate of failure to close the defect (5–55%) and a high incidence of abnormal findings at postmortem examination (84%). This systematic analysis clearly demonstrated that reliable closure of full-thickness colonic defects after resection could not be achieved with the previously used endoscopic closure techniques (T-tags, stapling, compression devices, etc.).

An endoscopic suturing device (OverStitch™, Apollo EndoSurgery, Austin, TX, USA) has become available for clinical use in 2011 [23–25]. It is a very versatile and user-friendly device which allows creation of separate stitches or a running suture line of variable length [26–30]. The device has already been successfully used for closure of full-thickness, colonic wall perforations with superior results compared to traditional endoscopic, through-the-scope clips [31, 32]. We have previously reported our preliminary results demonstrating safety and feasibility of “open” full-thickness endoscopic resection of GI tract lesions on a live animal model (90-kg domestic pigs) [2, 33]. Purely endoscopic full-thickness resection of the gastric and colonic wall was successfully achieved in all 9 animals [2, 34]. All defects after resection were easily and reliably closed using the OverStitch. Postmortem examination revealed good full-thickness healing of the GI tract wall at the sites of resection. After successful completion of initial animal experiments, we performed purely endoscopic full-thickness resection of an actively bleeding 2 × 4 cm colon cancer located at the hepatic flexure. In another patient, we accomplished purely endoscopic, full-thickness resection of a 2 × 5 cm recurrent gastric stromal tumor [2, 33, 34]. In both patients, colonic or gastric wall defects were completely closed with a continuous suture line. Both patients tolerated EFTR well, had no pain post-procedure, and were discharged home in 3 days. Follow-up endoscopy at 3, 6, and 12 months revealed good healing of resection sites without any residual lesions or strictures.

Our technique of NOTES “open” full-thickness resection of GI tract lesions consists of several consecutive steps (Figs. 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8, 5.9, 5.10, 5.11, 5.12, 5.13, 5.13, 5.14, 5.15 and 5.16) [35, 36]. We start en bloc removal of the lesion (Fig. 5.2) with endoscopic submucosal dissection (ESD) utilizing dedicated ESD accessories—DualKnife™ and HookKnife™ (Olympus America, Center Valley, PA) or Hybrid Knife® (ERBE USA Inc, Marietta, GA). If the submucosal space is obliterated by extensive fibrosis (after previous unsuccessful attempts of piecemeal resection), or if cancer involves the muscularis layer, we convert ESD into a full-thickness resection (Figs. 5.3 and 5.4). After en bloc full-thickness resection of the lesion (Fig. 5.5), a double-channel upper endoscope (GIF 2T180, Olympus America, Center Valley, PA, USA) preloaded with the OverStitch is advanced toward the full-thickness resection site (Fig. 5.6). The suturing arm of the device is opened and then closed delivering a needle through the proximal edge of the full-thickness colonic wall defect (Fig. 5.7). The needle is transferred back onto the suturing arm (Fig. 5.8), the suturing arm is reopened (Fig. 5.9), and the device is moved toward the distal edge of the full-thickness defect. Then, the suturing arm is closed again, advancing the needle through the distal edge of the full-thickness colonic wall defect, and the needle reloaded back onto the suturing arm (Fig. 5.10). The needle is driven sequentially through the proximal (Fig. 5.11) and distal (Fig. 5.12) edges of the full-thickness defect until it is fully closed (Figs. 5.13 and 5.14). The special cinching mechanism is then deployed (Fig. 5.15), tightening and finishing the creation of the continuous suture line, completely closing the full-thickness colonic wall defect (Fig. 5.16).