Endoscopic Submucosal Dissection of Pseudo-polyps in the Right Colon Using a Flexible Grasper With Articulating Distal Tip and Elbow-Like Bending

Endoscopic submucosal dissection (ESD) is challenging in the right colon. Traction devices can make it technically easier. In this study, we evaluated a exible grasper with articulating tip and elbow-like bending (IgE) through a double-balloon surgical platform (DESP), compared with an earlier generation grasper without elbow-like bending (Ig). The reach of Ig/IgE was investigated at eight locations using a synthetic colon within a 3D model. Using a fresh porcine colorectum, 4 cm pseudo-polyps were created at the posterior wall of the ascending colon. Fifty-four ESD procedures were performed using three techniques: standard ESD (STD), ESD using Ig (DESP+Ig), and ESD using IgE (DESP+IgE). IgE was able to reach the full circumference at all the locations, whereas the medial walls proximal to the descending colon were out of Ig’s reach. Compared with the STD, both DESP+Ig and DESP+IgE showed signicantly shorter procedure time (STD vs. DESP+Ig vs. DESP+IgE = median 48.9 min vs. 38.6 vs. 29.9) and fewer injuries (1.5 vs. 0 vs. 0). Moreover, the DESP+IgE had a shorter procedure time than the DESP+Ig (p=0.0025). The IgE with DESP increased instrument reach compared to Ig, and likely represented a traction tool for excision of large pseudo-polyps in the right colon.


Introduction
Endoscopic submucosal dissection (ESD) has been established as an acceptable technique to treat premalignant colorectal lesions. [1] [2] [3] On the one hand, it avoids the morbidity of a colectomy, and on the other hand, it achieves a higher en bloc resection compared to endoscopic mucosal resection (EMR), facilitating accurate pathological diagnosis, curative resection, and reducing the risk of local recurrence. [4] [5] [6] However, ESD has not been adopted widely due to technical di culties, longer procedure times, higher complication rates, and long learning curves, especially for lesions located in the right colon. [7] [8] [9] As a result, more than 25,000 colectomies for non-malignant polyps are performed annually in the United States. [10] A traction device can increase the e ciency and safety of ESD procedures by maintaining visualization of the submucosal dissection layer. [11] Various traction devices have been proposed to reduce procedure times and complication rates. [12] [13] However, most of devices are limited in their ability to navigate to the right colon or maintain dexterity of the instruments once situated in the right colon. [14] A double-balloon endolumenal surgical platform (DESP, Dilumen TM C 2 , Lumendi LLC, Westport CT USA) is one example of a platform used to advance the endoscope to the lesion, improve stability of the colon, and maintain visualization of the lumen throughout the colon. [15] The endoluminal platform also allows the use of dedicated instruments (e.g. grasper, scissors, and knife) through two external tool channels. The endolumenal interventional grasper with an articulating distal tip (DiLumen Ig™, Lumendi) provides good traction during endolumenal procedures, [16] however, Ig is not necessarily able to reach the lesion or provide traction in all directions due to the limitation of its movement in clinical cases. To resolve these issues, the endolumenal interventional grasper with Elbow (DiLumen IgE™, Lumendi) adds an elbow-like bending section more proximally. In this study, we evaluated the e cacy of a novel exible grasper with an articulating distal tip and elbow-like bending section (IgE) in a preclinical model, compared with Ig.

Materials And Methods
Double-balloon Endolumenal Surgical platform (DESP), Endolumenal Interventional Grasper (Ig), and Endolumenal Interventional Grasper with Elbow (IgE) DESP, Ig, and IgE have been approved by the U.S. Food and Drug Administration. The DESP consists of a exible oversleeve (103cm, 130cm) with two independently in atable balloons and bilateral external tool channels ( Figure. 1a).
[16] The Aft Balloon (AB), which is xed behind the endoscope tip, provides stability for the endoscope. A "therapeutic zone (TZ)" is created by extending the Fore Balloon (FB) beyond the endoscope tip. This adjustable FB improves luminal visualization by attening folds and straightening exures. Both balloons are in ated by a pump on the in ation handle. [15] [16] Ig and IgE instruments consist of a pistol style handle and exible shaft with an articulating jaw at the distal tip ( Figure. 1b).
Both Ig and IgE can be inserted through the external tool channels of the DESP, and the tip can be articulated toward the target using a thumb joy stick. The jaw is closed/opened by squeezing/releasing the trigger and oriented to the desired plane by rotating the jaw rotation knob. The tool channel ports can be secured to the tool mount for additional stability. IgE adds an elbow section, which can bend towards the black line on the shaft by turning the elbow knob ( Figure. 1b, Supplementary Video. S1).

Reach of Ig/IgE
The colonoscope (PCF-H180AL, Olympus), equipped with DESP, was navigated through a 3-D synthetic colon model (Colon-rectum tube, Kyoto Kagaku America Inc, Torrance CA USA) ( Figure. 2). Eight target points were selected to evaluate the reach of the Ig and IgE instruments; the ascending colon (site 1), beyond the hepatic exure (2), the transverse colon (3), beyond the splenic exure (4), the descending colon (5), beyond the sigmoid-descending colon (6), the sigmoid colon (7), and the rectum (8). At each point, the FB was deployed to straighten the colon and create the TZ. The reach of Ig/IgE was investigated when the synthetic colon was distended to its maximum using the endoscope insu ation.

Ex-vivo setting
In this study, we purchased and used resected porcine colons and no live animals were used. Using electrocautery, 4 cm pseudo-polyp was created in a fresh porcine colorectum and the intestine was positioned within an established 3D colon model ( Figure.  Six endoscopists who have used this surgical platform several times performed the ESD procedures by these three techniques in a sequential alternating fashion. Clinically, two endoscopists perform colorectal ESD and one endoscopist performs EMR (>100 procedures); these endoscopists were de ned as the experienced group. The other three endoscopists only had ESD experience in ex-vivo models (>10) and were de ned as the novice group.
Navigation to the cecum A straight transparent cap (D-201-12704, Olympus) was attached to a pediatric colonoscope. The colonoscope, equipped with/without DESP, was advanced to the cecum. Abdominal pressure and use of the balloon technique were applied to reduce loops and shorten the colon at the endoscopist's discretion.

ESD procedures using Ig/IgE with DESP
After navigation to the cecum, the FB and AB were deployed proximal and distal to the lesion. The FB was docked during the procedure at the endoscopist's discretion. Normal saline with 0.04% methylene blue was used for submucosal injection by injection needle (23G Interject TM , Boston Scienti c). The same settings (Endo Cut Q mode) of the ERBE electrosurgical unit (Erbe VIO 300D, Erbe USA) was used for all parts of the ESD procedures. Using a DualKnife (KD-650U, Olympus), a circumferential mucosal incision was made around the lesion with a 5 mm margin ( Figure. 3a). Once the circumferential incision was completed, the grasping instrument, Ig or IgE, was inserted through the external channel. The Ig or IgE grasped the distal edge of the lesion, retracting it to provide tension to visualize the appropriate plane of dissection ( Figure.

Outcome measurements
The primary outcomes were the reach of the Ig/IgE and the total procedure time for ESD procedures. Total procedure time was de ned as the time from submucosal injection to the removal of the specimen. Secondary outcomes included the cecal intubation time, resected specimen size, grasping time (for DESP + Ig and DESP + IgE groups), submucosal dissection time/speed, the number of muscularis propria injuries, and occurrence of perforations. Submucosal dissection speed was calculated as follows: (major radius) x (minor radius) π/ (dissection time). Grasping time was de ned as the time from Ig/IgE insertion through the tool channel to grasping of the lesion.

Statistical analysis
All variables are expressed as a median. The associations between continuous parameters were evaluated by the Mann-Whitney U test to compare two groups and by the Kruskal-Wallis test to compare more than two groups. Variables with p value < 0.05 in the Kruskal-Wallis test were subjected to multiple comparisons using the Steel-Dwass method. All analyses were performed using JMP Pro 14 statistical DiscoveryTM (SAS Institute Inc., Cary NC, USA). p values < 0.05 were considered to be signi cant.

Reach of Ig/IgE
Ig was able to reach the full circumference of the synthetic colon at the rectum and the sigmoid colon via both the right and left channels ( Figure. 2). However, the posterior wall beyond the sigmoid-descending colon and the medial wall at the other ve locations, especially in the locations immediately beyond the splenic exure and the hepatic exure, were out of Ig's reach. On the contrary, IgE was able to reach the full circumference at all eight locations by using the elbow bending section, regardless of which external channel was used. ESD procedures of 4cm lesions at the posterior wall of the ascending A total of 54 ex-vivo ESD procedures were performed with three different techniques (STD (n=18), DESP + Ig (n=18), and DESP + IgE (n=18)). All lesions were successfully resected en bloc without perforations. There was no difference in the resected specimen size. Total procedure time was signi cantly reduced in the DESP + Ig and DESP + IgE groups, compared to the STD group (STD vs. DESP + Ig vs. DESP + IgE = 48.9 vs. 38.6 vs. 29.9 min); this reduction was driven by signi cantly reduced submucosal dissection time (STD vs. DESP + Ig vs. DESP + IgE = 31.9 vs. 22.9 vs. 16.1 min, Table. 2). As expected, the shorter procedures in the DESP + Ig and DESP + IgE groups used a smaller volume of submucosal injection uid. Moreover, DESP + Ig and DESP + IgE groups created fewer injuries to the muscularis propria (STD vs. Regarding the difference between DESP + Ig and DESP + IgE groups, the DESP + IgE group showed signi cantly shorter total procedure times (p=0.025), shorter submucosal dissection times (p=0.022), and smaller volumes of submucosal injection uid (p=0.0069) ( Table. 2). For all procedures, the distal edge of lesions was grasped, and there was no difference in grasping time between the two groups (DESP + Ig vs.

Discussion
This study demonstrated that the use of DESP and the exible grasper (Ig/IgE) allow for the removal of large polyps in the right colon with a signi cantly shorter dissection time, total procedure time, and fewer injuries than standard ESD technique in an ex vivo porcine model. Moreover, IgE provided a broader reach of the grasper and decreased ESD procedural times and submucosal injection volumes. Finally, this surgical platform and these tools may help bridge the gap of skills in ESD procedures between the expert and the novice.
In order to access the more proximal colon, endoscopes and devices pass along the outer side of the intestinal curves and loops, leading to limited access to inner lesions, especially lesions located just beyond the splenic exure and the hepatic exure. [17] In this study, the Ig instrument could not reach behind each exure. Moreover, the same di culty also occurred in the ascending colon, the transverse colon, and the descending colon since the overtube device itself made the wide intestinal arch. The elbow bending function can resolve this issue, allowing the IgE to reach everywhere in the large intestine.
In this study, each procedure was able to successfully reach the cecum. Navigation of the colonoscope equipped with DESP may be more di cult as it required assistance with abdominal pressure, unlike standard colonoscopy. However, the simpli ed colon model does not allow demonstration of the advantages of the balloon technique that assists navigation under more challenging situations. [18] Similarly, the advantages of the double-balloon platform to improve the endoscope operability in the right colon were not demonstrated in this simpli ed model. [15] [19] Traction devices have been introduced to provide tension and counter-tension to facilitate resection of larger lesions, more complex lesions, and lesions located in di cult locations. However, few traction devices can be used anywhere in the large intestine, provide real-time dynamic traction in all directions, and move independently from the endoscope, allowing for continuous direct visualization of the target plane of submucosa. The S-O clip (Zeon Medical, Tokyo, Japan), traction wire (ProdiGI TM Traction Wire, Medtronic, Ireland) and elastic band (Elastic Traction Device; Micro-Tech Endoscopy USA Inc, Ann Arbor, Mich, USA) can be used in the right colon and provide traction in all directions, but it cannot be repositioned during the procedure. [20] [21] The clip-nylon method and the suture/clip method utilizing the FB of the double-balloon interventional platform (DEIP) can provide real-time traction, but only in a single dimension, proximal and distal/back and forth. [19] [22] [23] Robotic platforms have not been applied to the right colon yet, though new technology might solve this limitation. [24] The dexterity of IgE led to signi cantly shorter dissection time and shorter total procedure time compared to using Ig. First, IgE was able to reach the targeted site, even the inner side at the ascending colon which was out of the Ig's reach ( Figure. 3c). Second, despite the slightly longer grasping time of IgE (Ig vs. IgE = 1.5 vs. 1.9 min), IgE could provide more traction per grasp using both elbow bending and tip articulation, leading to better visualization and reducing the number of regrasping. This novel traction method may also be particularly helpful for lesions suspected to have submucosal invasions, mucosal lesions with brosis, residual/recurrence lesions after incomplete endoscopic resection, and non-polypoid colorectal dysplasia with in ammatory bowel disease; these lesions are di cult to lift due to their submucosal brosis. [14] [25] This exible grasper was not used during mucosal incision; in fact, there was no signi cant difference in mucosal incision time between the three techniques. This grasper has the potential to make mucosal incision easier by providing stability and tension, as observed with endoscopic models of robotic surgery. [24] Using the grasper during mucosal incision may further emphasize the bene ts of this instrument. Moreover, the use of an additional knife/scissors in combination with IgE might also be helpful. These multi-instrument techniques should be investigated further. After tumor resection, closure of large mucosal defects (>20 mm) in the right colon is recommended. [3] We believe the ease of mucosal closure would also greatly facilitated using the grasper to approximate the two edges of the defect. [26] In the United States, ESD has been less widespread, compared to East Asian countries, since the relative lack of gastric tumor cases forces endoscopists to start with the more challenging colorectal tumors, leading to the increased risk and a longer time to acquire skills in ESD. In our study, there was no difference in ESD procedural outcomes using DESP + IgE between the experienced and the novice groups.
The novice could perform ESD e ciently and safely as well as the expert. This surgical platform has the potential to bridge the gap between the expert-novice differences, and shorten the learning curve of ESD.
There are several limitations in this study.
(1) A synthetic colon, which keeps the shape compared to human colons, was used to investigate the reach of Ig/IgE. The function of Ig/IgE might be underestimated since we could collapse the colon clinically in grasping the tissue. Nevertheless, we experienced Ig wasn't able to reach the lesion in certain locations.
(2) The ex-vivo model in this study has no peristalsis or bleeding, so performing ESD may be more di cult clinically. On the other hand, the simplicity of the model may underestimate the bene ts of DESP and IgE grasper. A similar double-balloon platform, DEIP, is clinically used in the right colon and has demonstrated the bene ts of the double-balloon platform in navigation. Traction devices have also been used for ESD and demonstrated clinical advantages.
(3) All lesions were 4 cm in diameter and placed at the posterior wall of the ascending colon to limit variability. The use of this traction device will need to be evaluated for resection of other challenging lesions and lesions at other challenging locations, such as circumferential lesions or lesions at the hepatic exure or under a fold.
(4) No other traction comparator was done. In addition, normal saline was used as a submucosal injection uid, not a long-lasting gel. The need for repeated injections with longer procedures is exaggerated with the use of short-acting injection uid. Additionally, the need for repeated injections further increases the total procedure time. [27] (5) All participants are advanced endoscopists and participants familiar with the ex-vivo research model and using this endolumenal platform. Therefore, this study didn't investigate the learning curve of this platform. [9] Our previous study showed the experience in ve ESD procedures might be enough in an exvivo model. [16] (6) This was a preclinical study. Larger human feasibility studies are warranted to validate these results.
In conclusion, we demonstrated a novel exible grasper with DESP likely represents an important traction tool for removal of large pseudo-polyps in the right colon. This new technology has the potential to expand the capacity for endolumenal resection, leading to lower colectomy rates for complex polyps.    (5), beyond the sigmoid-descending colon (6), the sigmoid colon (7), and the rectum (8). The range in blue indicated the reach when Ig/ IgE was inserted through the right channel, the range in red through the left channel, and the range in black means out of reach.