Efficacy of scissor-type knife for endoscopic submucosal dissection: a systematic review and meta-analysis

Introduction

Endoscopic submucosal dissection (ESD) enables the operator to achieve an en bloc resection regardless of lesion “size” with complete (R0) resection [1]. This allows for comprehensive pathological tumor staging and treatment of early-stage carcinomas of the gastrointestinal tract [1,2]. Use of ESD in the west, particularly in the United States of America (USA), has remained limited mainly to major academic high-volume centers, where the procedure is performed by therapeutic endoscopists [3,4]. With the establishment of its efficacy, advancements in technology, and accessibility to training, ESD is gaining more acceptance among gastroenterologists in the USA. However, ESD is a highly technical, complex advanced endoscopic resection technique with the potential for severe adverse events, particularly in the early part of the learning curve [5,6]. The risk of adverse events, including perforation and severe bleeding, is highest during the submucosal dissection step, when the endoscopist must perform precise movements to keep the tip moving along the dissection plane at the appropriate pace. Addressing intraprocedural and postprocedural complications when they occur is also an important determinant of successful ESD outcomes. Hence dissection and hemostasis devices play a very crucial role. Several ESD electrosurgical knives are available, with the earlier and conventional ones being needle-type. These include non-insulated tip knives (e.g., HookKnife, Triangle tip knife, DualKnife, FlushKnife) and insulated-tip knives (e.g., IT-2 Knife, IT-2 NanoKnife) [5]. With the evolution of ESDs from a procedure initially limited to gastric lesions, to the more challenging and less forgiving colorectal lesions, a new grasping type or scissor-type ESD (ST-ESD) knife was developed in the late 2000s in Japan to overcome some of the limitations of needle-type knives [7,8]. The ST-ESD knives are unique in their ability to grasp and pull the target tissue away from the deeper muscle layer before electrosurgical cutting or coagulation [6,9]. ST-ESD knives have an external electrically insulated coating, and sharp jagged or smooth, non-insulated inner cutting jaws [7]. This mechanism allows selective delivery of electrosurgical current to the grasped tissue between the upper and lower electrodes while sparing the adjacent regions and deeper muscularis propria. This ST-ESD knife design could decrease adverse effects, primarily when non-experts perform ESDs. The 2 main ST-ESD knives are the “stag beetle (SB) knife” (Sumitomo Bakelite, Tokyo, Japan) and the “ClutchCutter (CC) knife” (Fujifilm Medical, Tokyo, Japan) [7,8]. These are shown in Fig. 1. The first ST knife was the CC knife, introduced in 2007. CC is a thicker knife with a serrated jaw, while SB knives are thinner. This difference makes CC knives more efficient in managing periprocedural bleeding, while SB knives are better suited for sharp mucosal incisions [7]. Multiple studies have evaluated these knives, including studies that compared ST-ESD knives to non-ST (NST)-ESD knives [9-26]. The results have supported the use of ST-ESD knives, even by endoscopists with limited ESD experience. With the growing acceptance and increasing availability of ESD training in the USA, we decided to perform this systematic review and conduct a meta-analysis of all currently available data on the efficacy and safety of ST-ESD knives.

Figure 1 The 2 main scissor-type endoscopic submucosal dissection knives. (A) Stag beetle knife and (B) ClutchCutter knife

Materials and methods

Results

The initial search identified 384 studies, from which 85 relevant articles were reviewed. Data were extracted from 17 studies, comprising 1558 patients, that met the inclusion criteria and were included in the final analysis. The PRISMA diagram showing the details of the review process is given in Fig. 2. All the included studies are available in full-text articles. The characteristics of the included studies are given in Table 1. The quality of studies was good, as evaluated using the Newcastle-Ottawa and Jadad scales. All the pooled estimates given are estimates calculated using the fixed-effects model. The estimates calculated using fixed- and random-effects models were similar. The agreement between reviewers was 1.0, as measured by Cohen’s κ.

Figure 2 Study flow diagram according to the PRISMA guidelines

ESD, endoscopic submucosal dissection

Table 1 Characteristics of studies included in this meta-analysis

The total sample size was 1652 ESD procedures performed in 1558 patients. Mean patient age was 69.62±2.64 years and females constituted 39.53% of the study population. These data included ESDs performed in the esophagus, stomach, colon and rectum. The mean lesion size was 24.24±4.58 mm for esophageal lesions, 35.12±3.78 mm for gastric lesions, and 30.72±3.53 mm for colonic lesions. The ST knives used were the CC knife in 8 studies (25% of ESDs) and the SB Knife/SB Knife Jr in the other 9 studies (75% of ESDs). In 9 of these studies, all the steps using ESD knives were performed using an ST-ESD knife, while argon plasma coagulation and Hook knives were used for marking the circumference in 4. Only 1 study reported using an NST-ESD knife for mucosal incision. All the studies were conducted in various high-volume referral centers across Japan. The ESD operators comprised a combination of experts and non-experts. The definition of experts and non-experts varied among the studies. In studies that evaluated trainees, there was at least one expert for supervision and procedure takeover if needed.

The pooled en bloc resection rate and R0 resection rate were 97.94% (95% confidence interval [CI] 97.20-98.57) and 94.32% (95%CI 93.11-95.43), respectively. The forest plot showing individual study estimates and the pooled estimate for en bloc resection is shown in Fig. 3. There was no publication bias according to the Egger bias indicator, -0.24 (95%CI -1.30-0.80, P=0.62), and the Harbord bias indicator, 0.71 (95%CI -0.62-2.04, P=0.32). The forest plot for R0 resection is shown in Fig. 4. The Harbord bias indicator calculated for R0 resection was -0.41 (95%CI -1.83-0.99), showing no publication bias. The pooled en bloc resection rate was 98.20% (95%CI 97.48-98.81) and the pooled R0 resection rate was 94.45% (95%CI 93.24-95.55) on a leave-out sensitivity analysis that excluded high-RoB studies. A ”per-knife” sensitivity analysis was also performed on the primary outcomes by calculating the rates of en bloc and R0 resection based on whether the SB knife or CC knife was used for performing ESD. The pooled rate of en bloc resection using the SB knife alone was 97.66% (95%CI 96.76-98.41) while the pooled rate of en bloc resection using CC knife alone was 98.77% (95%CI 97.42-99.63). Pooled rates of R0 resection were 93.86% (95%CI 92.45-95.12) and 96.13% (95%CI 93.60-98.04) using the SB knife and CC knife, respectively. Outcomes based on the location of lesions showed that the en bloc resection rates for esophageal, gastric and colorectal lesions were 100% (95%CI 97-100), 98.60% (95%CI 96.88-99.64), and 97.49 (95%CI 96.55-98.29), respectively. The R0 resection rates were 92.68% (95%CI 87.99-96.28) for esophageal lesions and 93.65% (95%CI 92.20-94.96) for colorectal lesions. The main adverse events were perforation and delayed postprocedural bleeding, with pooled rates of 1.07% (95%CI 0.63-1.62) and 1.86% (95%CI 1.26-2.56), respectively. The forest plot showing the individual study estimates and the pooled estimate for perforation rate is shown in Fig. 5. There was no heterogeneity, with an I² score of 0% (95%CI 0-44.50%). The pooled mean procedure time for the entire cohort was 67.45 min (95%CI 58.01-76.89). Procedure time was also calculated based on the location of the lesion resected and was 93.95±30.85 min for esophageal ESDs, 73.51 ± 17.27 min for gastric ESDs, and 62.89±14.35 min for colorectal ESDs.

Figure 3 Forest plot showing individual and pooled estimates of en bloc resection rate with the scissor-type endoscopic submucosal dissection knife

Figure 4 Forest plot showing individual and pooled estimates of R0 resection rate with scissor-type endoscopic submucosal dissection knife

Figure 5 Forest plot showing individual and pooled estimates of perforation with scissor-type endoscopic submucosal dissection knife

Discussion

The results of this meta-analysis show that ST-ESD knives can consistently achieve good en bloc and R0 resection rates comparable to those of conventional needle-type ESD knives. Achieving an en bloc and potentially R0 resection, regardless of the lesion size, is the principal advantage of ESD over other endoscopic resection techniques, such as endoscopic mucosal resection [1,4]. In one of the largest prospective studies of colonic ESDs by Saito et al, which involved 1111 colonic lesions, the reported en bloc and R0 resection rates were 88% and 89%, respectively [27]. A meta-analysis by De Ceglie et al, involving 66 studies, showed an en bloc resection rate of 91% with colonic ESD [28]. The en bloc resection rate of 98% and an R0 resection rate of 94% found in our study are comparable to, if not better than those reported in prior studies that primarily used NST-ESD knives. Additionally, 9 studies included in this meta-analysis reported using an ST-ESD knife for all the steps involving an electrosurgical knife [10,12,14,16,18,22,24-26], with only 1 study reporting the use of an NST-ESD knife to perform the mucosal incision step [19].

Our pooled analysis included ESDs performed by experts, gastroenterologists with varying levels of ESD experience, and trainees with no previous ESD experience. Dohi et al, in their study comparing ST-ESD knives to NST-ESD knives in gastric ESDs, found no significant difference in R0 resection between the 2 groups when experts performed the procedure [21]. However, ST-ESD knives showed significant superiority in both self-completion rates and mean procedure time when non-experts performed the procedures [21]. A study by Yamashina et al also found a better self-completion rate when trainees used ST-ESD knives [17]. Two other studies also compared procedure completion time between experts and non-experts when ST-ESD knives were used. Esaki et al found no significant difference in median procedure completion time, while Inoue et al observed that, among less-experienced endoscopists, the mean procedure time was significantly shorter in the ST-ESD knife group than in the NST-ESD knife group [22,24]. Other studies have not consistently shown this difference, but data are limited, and head-to-head comparison of these 2 types of ESD knives in trainees is scarce, as most available studies are drawn from expert experience. Our data showing a good aggregate en bloc and R0 resection from a pool that included different tiers of experience warrant further studies directly comparing these 2 types of knives, mainly when used by trainees.

Another significant finding from this meta-analysis is the low rates of delayed bleeding and perforation with ST-ESD knives. Bleeding is the most common complication of ESD, and can be immediate or delayed [5]. Delayed bleeding is defined as a drop in hemoglobin by 2 g or more after being stable for the initial 24 h [5]. Delayed bleeding rates can differ significantly based on the location of the lesion, with studies reporting rates of 4-9% for gastric ESDs [29-31] and 0.5-2.70% for colonic ESDs [32-34]. The pooled delayed bleeding rate of 1.80% from this meta-analysis falls within the lower range of those reported in previous studies. The reason may be ST-ESD knives’ ability to grasp vasculature and function as a coagulation grasper, utilizing soft coagulation settings without the need for switching devices. This may encourage the more robust use of coagulation settings to treat vessels in the resection bed effectively. Perforation is more common with ESD than other endoscopic resection techniques [5]. It can be intraprocedural or delayed. Tumor size, location, fibrosis, procedure duration and operator inexperience are known risk factors for perforation [35]. Previous studies have reported varying rates of perforation, from 4.30-8.20%, with gastric ESDs [30,35], around 4% with esophageal ESDs [36], and about 5% with colonic ESDs [27,33,34,37]. Duodenal ESDs carry the highest risk for perforation, with wide variability in the reported rates [38]. The pooled perforation rate of 1% noted in this study is lower than those typically reported for ESDs and suggests that ST-ESD knives could potentially lower the risk of perforation. This difference is probably related to the ability of an ST-ESD knife to grasp and pull tissue away from the underlying muscularis propria before delivering the electrosurgical current. In addition, the electrically insulated outer coating also further protects the underlying and adjacent tissue from unintended thermal injury. Kuwai et al, in the largest study comparing ST-ESD knives to NST-ESD knives, involving 1221 propensity score-matched colorectal ESDs, found that the rates of perforation and delayed bleeding were significantly lower in the ST group [25]. Other studies involving trainees, however, have not consistently noted this difference. However, it is also important to note that none of the studies reported a greater incidence of adverse events with ST-ESD knives, including when used by less experienced endoscopists.

The findings from this meta-analysis thus suggest a need for further prospective studies to clarify whether there are any safety advantages with the use of ST-ESD knives, especially in new adopters. To the best of our knowledge, this is the first systematic review and meta-analysis of all currently available data on the efficacy of ST knives for ESD. The findings from this study show that ST-ESD knives can be a feasible alternative or adjunct to conventional needle-type or blade-type ESD knives. Further studies are needed to define the role of this relatively new style of ESD knife and evaluate whether there is a significant difference in the learning curve if ST-ESD knives are used in the early part of ESD training.

There are a few limitations to this study. All the studies evaluated were performed in Japan, where the technique of ESD was pioneered and has been in practice for over 2 decades. There is also a risk of selection bias in some of the included studies, as only 3 were randomized controlled trials, whereas the others were single or multicenter prospective or retrospective studies. This could have affected patient selection for the performance of ESD using ST knives. Furthermore, the results of this meta-analysis include a combination of data from studies of the CC knife and SB knife (which includes both SB knife and SB Knife Jr). Although they are all ST knives, they have differences, which could affect some of the outcomes evaluated and may have introduced moderate heterogeneity to our findings. However, the similar findings for en bloc and R0 resection rates seen on a sensitivity analysis, excluding high RoB studies, and a per-knife sensitivity analysis further indicate that the findings from this study are robust.

Our analysis shows that the ST knife delivers consistent performance across a wide range of ESD procedures in various locations of the gastrointestinal tract. ST-ESD knives have a good safety profile, comparable to conventional needle-type ESD knives. Additionally, there seems to be a potential for reduced rates of adverse events and perhaps greater ease of use in the hands of less experienced operators during ESD procedures. These findings suggest that ST-ESD knives could be an alternative or adjunct to traditional NST knives for ESDs. More head-to-head studies are warranted, particularly including less experienced operators of ESD.