Recent Advances in Endoscopic Submucosal Dissection for Gastric Cancer: Focusing on Expanded Indications and Technological Innovations.

INTRODUCTION
Although the incidence of gastric cancer has been declining worldwide [1], it remains the fifth most common cancer and the fifth leading cause of cancer-related mortality [2]. With advances in endoscopic technology and the widespread implementation of medical checkups in Korea and Japan, over half of gastric cancer cases are detected at an early stage [34]. Endoscopic resection (ER) is performed for early gastric cancer (EGC) with a negligible risk of lymph node metastasis (LNM). Traditionally, endoscopic mucosal resection (EMR) was performed, especially for EGCs measuring 2 cm or less in diameter; however, it has largely been replaced by endoscopic submucosal dissection (ESD), especially in East Asian countries [56]. Since its introduction more than 25 years ago [7], ESD technology has continued to evolve, and indications for ER have expanded. In this review, we focus on the expansion of ER indications for EGC and recent technological innovations in gastric ESD.

HISTORY AND CURRENT INDICATION FOR ER AND CURABILITY AFTER ER
There are two key decision points in patients undergoing ER for EGC: determining the indications for ER before the procedure and assessing curability after ER. Numerous clinical practice guidelines for the management of gastric cancer have been published worldwide [891011121314]. We compared the indications for ER and the criteria for curability after ER among the Japanese, Korean, and European guidelines. For the European guidelines, we primarily adopted the recommendations of the European Society of Gastrointestinal Endoscopy (ESGE) [14], rather than those of the European Society for Medical Oncology guidelines [9], because the ESGE guidelines focus specifically on ER.

Indication of ER

Japan
In the Japanese guidelines [13], “a tumor indicated for ER as a standard treatment (absolute indication)” is defined as one with a less than 1% likelihood of LNM. In contrast, “a tumor indicated for ER as an investigational treatment (expanded indication)” refers to lesions with a similarly low risk (<1%) of LNM but lacking sufficient evidence on long-term outcomes after ER. The indications for ER have expanded [15].
According to the pre-2018 guidelines [16], the absolute indication for ER was limited to the following: (i) mucosal cancer, ≤2 cm, differentiated type, no ulceration. Based on two large surgical series [1718], expanded indication for ESD was defined as follows: (ii) mucosal cancer, >2 cm, differentiated, no ulceration; (iii) mucosal cancer, ≤3 cm, differentiated, ulceration; and (iv) mucosal cancer, ≤2 cm, undifferentiated type, no ulceration.
Following two single-arm phase III confirmatory trials in Japan that demonstrated excellent long-term outcomes [1920], the 2018 revision of the Japanese guidelines reclassified categories (ii) and (iii) as absolute indications for ESD [21], and category (iv) was newly added as an absolute indication in the 2021 revision [13]. Category (i) has been defined as an absolute indication for EMR/ESD since 2018 [21]. Since 2018, most lesions outside of categories (i)–(iv) have been designated as relative indications for ER (Table 1).

Korea
In Korea, the terms “absolute indication” and “expanded indication” were previously used [22]. Since 2022, the terminology “classical absolute indication” and “expanded indication” has been adopted [1223]. ER is recommended for category (i). However, ESD and gastrectomy can be indicated for categories (ii) and (iii) (expanded indications), and ER could be cautiously considered for category (iv) (relative indication), suggesting that the Korean guidelines take a more conservative approach than the Japanese guidelines. This caution reflects the lack of prospective studies that directly compare the outcomes between ER and surgical resection. However, a multicenter randomized controlled trial (RCT) (ERASE-GC trial, NCT04890171) focusing on category (iv) is currently ongoing, and its results may influence future indications for ER in Korea.

Europe
In Europe, the ESGE guidelines for ESD in gastric neoplasia were first published in 2015 [24] and updated in 2022 [14]. According to the latest version, EMR should be considered an alternative to ESD for elevated (Paris 0-IIa [25]) lesions of size ≤1 cm with low likelihood of malignancy. ESD is recommended for lesions corresponding to categories (i)–(iii), whereas for category (iv), the recommendation is more cautious than that in the Japanese guidelines; ESD may be considered, but the decision should be individualized.

Curability after ER

Japan
Until 2018, curability after ER was categorized into curative resection, curative resection for expanded indications, or noncurative resection [16]. However, these terms were updated in the 2018 revision of the Japanese guidelines [21] to endoscopic curability (eCura), which is now classified from eCuraA to eCuraC-2. Specifically, curative resection and curative resection for expanded indications were redefined as eCuraA and eCuraB, respectively, whereas noncurative resection was subdivided into eCuraC-1 and eCuraC-2.
In the latest version of the Japanese guidelines [26], which are now available only in Japanese, eCuraA is defined as follows: (I) mucosal cancer, ≤2 cm, differentiated type-dominant, no ulceration, no lymphovascular invasion (LVI), and negative resection margin; (II) mucosal cancer, >2 cm, differentiated type-dominant, no ulceration, no LVI, and negative resection margin; (III) mucosal cancer, ≤3 cm, differentiated type-dominant, ulceration, no LVI, and negative resection margin; (IV) mucosal cancer, ≤2 cm, undifferentiated type-dominant, no ulceration, no LVI, and negative resection margin. Furthermore, (V) cancer with submucosal invasion of < 500 µm, ≤3 cm, differentiated type-dominant, no LVI, and negative resection margin was reclassified from eCuraB to eCuraA in the latest version, based on the favorable long-term prognosis and the extremely low rate of metastatic recurrence (0.5%) after ESD demonstrated in a large-scale prospective cohort study in Japan [27]. However, when an undifferentiated component is present in the submucosal invasive portion, the lesion is regarded as eCuraC-2 because of the increased risk of LNM [2829]. eCuraC-1 corresponds to differentiated-type cancers with unsatisfactory curative factors for piecemeal resection or en bloc resection with a positive horizontal margin, whereas eCuraC-2 corresponds to others.
Follow-up endoscopy at 12-month intervals is recommended for patients with eCuraA, whereas follow-up using both endoscopy and abdominal imaging (computed tomography or ultrasonography) at 6- to 12-month intervals is recommended for those with eCuraB. For eCuraC-1, the risk of LNM is low. Thus, one of the following alternatives could be selected in accordance with institutional policy after obtaining patient consent: repeat ESD, surgical resection, close observation to assess the burn effect of the initial ESD, and endoscopic coagulation using a laser or argon plasma coagulator. Additional surgery is recommended as the standard of care for eCuraC-2 (Table 2).

Korea
The Korean guidelines also use the terms “curative resection” and “noncurative resection” [122223]. According to the latest Korean guidelines [12], a lesion is curatively resected if it meets any of the criteria (I)–(V), with the threshold for submucosal invasion defined as ≤500 μm (rather than <500 μm in Japan). Observation is recommended in such cases. Lesions that do not meet these criteria are categorized as noncurative resections, for which additional surgery is recommended. Endoscopic treatments such as ESD and argon plasma coagulation and surgical resection are considered possible treatment options for EGCs that have only positive horizontal margins and meet all other criteria for curative resection because of the minimal risk of LNM (0.6%) [12].
The Korean guidelines for ER [30] recommend endoscopic surveillance every 6–12 months after curative resection, and abdominopelvic computed tomography is recommended at 6- to 12-month intervals.

Europe
The criteria for curability after ER in the European guidelines [1424] are largely consistent with those in the Japanese and Korean guidelines. However, the terms for the ER type in the latest European guidelines include very low-risk (curative) resection, low-risk (curative) resection, local-risk resection, and high-risk (noncurative) resection [14]. Very low-risk (curative) resection, which corresponds to categories (I)–(III), is defined as a lesion with <1% risk of LNM. Low-risk (curative) resection, corresponding to categories (IV) and (V) (using ≤500 μm rather than <500 μm for submucosal invasion), is defined as a lesion with <3% risk of LNM. Local-risk resection is defined as an LNM risk of <3% but a local recurrence risk of 10%–30%, corresponding to eCuraC-1 in the Japanese system. Lesions that do not meet the criteria for very low-, low-, or local-risk resection are classified as high-risk (noncurative) resections.

Considerable factors after noncurative resection

Risk-scoring systems for LNM or gastric cancer-specific mortality
As described earlier, additional surgery with lymph node dissection after noncurative resection is recommended as the standard treatment in all Korean, Japanese, and European guidelines [121314]. A recent meta-analysis reported an LNM rate of 8.0% (95% confidence interval, 7.1%–8.9%) following noncurative resection [31].
Several risk-scoring systems have been developed to predict LNM or gastric cancer-specific mortality after noncurative resection. The eCura system, based on five pathological factors—lymphatic invasion (3 points), tumor size >3 cm (1 point), submucosal invasion ≥500 µm (1 point), vascular invasion (1 point), and positive vertical margin (1 point)—categorizes patients into three risk groups [32]. The corresponding LNM rates were 2.5%, 6.7%, and 22.7% in the low-, intermediate-, and high-risk groups, respectively. The 5-year disease-specific survival rates without additional treatment were 99.6%, 96.0%, and 90.1%, respectively. This scoring system has been externally validated in Korea [33], China [34], Japan [35], and Europe [36], demonstrating good discrimination between Korea and Japan and excellent discrimination between China and Europe. Moreover, disease-specific survival was similar between patients with and without additional surgery in the low-risk eCura group (99.6% vs. 99.7% at 5 years) [37].
The W-eCura score is a modified version of the eCura system, changing the cutoff for deep submucosal invasion from 500 µm to 1,000 µm [36]. External validation showed similar discriminative performance between the W-eCura and eCura systems (c-statistic: 0.750 vs. 0.745) [35]. Another variant, the eCuraU system, was developed for undifferentiated-type EGC [38]. It assigns 3 points for LVI, 2 points for submucosal invasion ≥500 µm, and 1 point each for tumor size >2 cm, ulceration, positive vertical margin, and submucosal invasion <500 µm. Although the eCuraU system showed a higher discriminative ability (c-statistic: 0.723 vs. 0.687), the difference was not statistically significant. The characteristics of the three scoring systems are summarized in Table 3.

Older patients
With the rapidly aging population, the number of older patients with gastric cancer has increased in East Asian countries [394041]. Treatment decisions after noncurative resection in these patients are often more complex because of reduced physiological function, multiple comorbidities, and other age-related factors. In clinical practice, additional surgery is performed after noncurative resection in 42.6%–54.5% of patients across all age groups [334243]; however, the rate varies substantially by age group. In one Japanese study, additional surgery was performed in 70.0% of patients aged <70 years but in only 20.1% of those aged ≥80 years (Fig. 1) [44]. Another study involving patients aged ≥85 years demonstrated that only 11.1% underwent additional surgery, with selection rates varying by risk category in the eCura system (4.8%, 5.3%, and 25.0% in the low-, intermediate-, and high-risk categories, respectively) (Fig. 1) [45].
It is crucial to balance the risks of gastric cancer-specific mortality, non-gastric cancer-related mortality, and quality of life when selecting the most appropriate treatment [4647]. Most older patients die of non-gastric cancer-related causes regardless of whether additional surgery is performed [444548]. Although the indications for ER and curability criteria may be expanded to older patients, current evidence remains insufficient to establish specific recommendations. Several retrospective studies have investigated the factors associated with all-cause mortality in older patients [495051]; however, a comprehensive geriatric assessment may provide a more accurate evaluation of the risk of all-cause and non-gastric cancer-related mortality. Comprehensive geriatric assessment is a multidimensional, multidisciplinary process; however, it is time-consuming and not always feasible in routine clinical practice [5253]. Therefore, a simple and practical prognostic assessment tool is required. Prospective studies on the management of older patients after noncurative resection are ongoing, and their findings are expected to provide valuable evidence for optimizing treatment strategies for this growing patient population.

RECENT ADVANCES IN THE TECHNOLOGIES FOR GASTRIC ESD

Prevention of delayed bleeding
Delayed bleeding is one of the major adverse events in gastric ESD, with an incidence ranging from 4.4% to 5.1% [425455]. To stratify the risk of delayed bleeding, a predictive model, known as the BEST-J score, was recently developed [56]. This model comprises ten factors (4 points each for warfarin and direct oral anticoagulant; 3 points for chronic kidney disease with hemodialysis; 2 points each for P2Y12 receptor antagonist and aspirin; 1 point each for cilostazol, tumor size >3 cm, lower third tumor location, and presence of multiple tumors; and 1 point for interruption of each kind of antithrombotic agent). The delayed bleeding rates in the low- (0–1 point), intermediate- (2 points), high- (3–4 points), and very high-risk (≥5 points) groups were 2.8%, 6.1%, 11.4%, and 29.7%, respectively. Preventive measures are crucial for patients classified as high- or very high-risk.
Acid-suppressive agents are known to reduce the risk of delayed bleeding after gastric ESD, and their efficacy appears to depend on the potency of acid suppression. A recent large-scale database study showed that proton pump inhibitors (PPIs) reduce the risk of delayed bleeding by more than 60% [57]. An RCT demonstrated that PPIs are superior to histamine-2 receptor antagonists in reducing delayed bleeding [58]. Vonoprazan, a potassium-competitive acid blocker with stronger acid suppression than PPIs [59], had a reduced effect on delayed bleeding in gastric ESD compared with PPIs in large database studies [6061]. Although acid-suppressive therapy can significantly decrease this risk, it does not completely prevent delayed bleeding, highlighting the need for additional preventive strategies.
Several interventions have been proposed to prevent delayed bleeding; however, none have been definitively proven effective in RCTs (Table 4). Second-look endoscopy, once routinely performed after gastric ESD, did not reduce the incidence of delayed bleeding in RCTs [6263]. Thus, second-look endoscopy after gastric ESD is not routinely recommended in European or Japanese guidelines [7273], and its ineffectiveness is acknowledged in Korean guidelines [30]. A large-scale study showed no significant difference in delayed bleeding rates between patients with and without second-look endoscopy, regardless of the risk categories of the BEST-J score [56]. Fibrin glue and polyglycolic acid sheets have been considered promising for preventing delayed bleeding [7475]. However, two RCTs failed to demonstrate a preventive effect of fibrin glue in high-risk patients [64] or all patients [65], and another RCT found no significant benefit of polyglycolic acid sheets combined with fibrin glue in high-risk patients [66]. A hemostatic gel (TDM-621) has been developed [76], but its efficacy in preventing delayed bleeding after gastric ESD remains uncertain. However, an RCT evaluating this gel after EMR for large duodenal or colorectal lesions found no reduction in delayed bleeding [77]. Similarly, the use of hemostatic powder did not show significant efficacy in preventing delayed bleeding in gastric ESD in an RCT [67].
Several closure techniques have recently been developed for gastric ESD (Table 4). Closures using endoloops and clips did not reduce delayed bleeding in a retrospective study [68]. In contrast, the endoscopic O-ring closure method and reopenable-clip over-the-line method significantly reduced delayed bleeding in retrospective studies using propensity score matching [6970]. Moreover, a phase II study demonstrated promising results for endoscopic hand suturing with a reduced rate of delayed bleeding after gastric ESD [71]. Although these findings are encouraging, none have been validated in RCTs.
Overall, many preventive techniques for delayed bleeding have shown apparent efficacy in retrospective and non-randomized prospective studies; however, their benefits have not been confirmed in RCTs. Therefore, large-scale randomized studies with adequate sample sizes are warranted to establish these methods as a part of routine clinical practice.

Traction device
Several traction devices are currently available (Fig. 2), and some RCTs have recently been published. In these studies, the use of dental floss clip traction or EndoTrac traction in gastric ESD did not significantly shorten the overall procedure time [7879]. However, it appeared to reduce the procedure time for lesions located on the greater or lesser curvature of the upper or middle stomach. Moreover, this may reduce the risk of perforation [78]. Another RCT demonstrated that the spring-and-loop with clip traction technique effectively shortened the procedure time for gastric ESD [80]. However, these devices require the withdrawal and reinsertion of the endoscope, which can interrupt the procedure. The multiloop traction device, a through-the-scope traction system, eliminates the need for endoscope reinsertion [81]. Although no RCTs have been conducted on gastric ESD, an RCT on colorectal ESD showed that a multiloop traction device did not significantly increase dissection speed [82]. However, a meta-analysis reported that traction-assisted ESD significantly shortened the mean procedure time compared with conventional ESD, particularly for lesions located on the greater curvature and in the upper or middle stomach [83]. Based on these findings, ESGE recommends the use of traction methods for ESD of gastric lesions located in the greater curvature of the upper or middle stomach (particularly if >2 cm) and if difficult access to the submucosa is encountered during ESD [72].

Novel image-enhanced endoscopies for gastric ESD
Red dichromatic imaging (RDI) is a novel image-enhanced endoscopic technique that uses two long-wavelength lights (600 nm and 630 nm) [84]. It enhances the visualization of thick vessels in deeper mucosal or submucosal layers [84] and improves the visibility of blood vessels and bleeding points (Fig. 3A and B) [85]. In an RCT, hemostatic procedures performed using RDI did not shorten hemostasis time but reduced psychological stress during hemostasis [86]. However, further studies are required to confirm its clinical utility.
More recently, another image-enhanced endoscopy technology, amber-red color imaging (ACI), has been developed [87]. ACI enhances the visualization of blood flow during bleeding by using amber and orange light; unlike RDI, it preserves other color tones, particularly blue hues, similar to those in white-light imaging (Fig. 3C and D). A recent study demonstrated that ACI displays the submucosa in blue only when an adequate solution is injected, which may help operators recognize the dissection line during gastric ESD [88]. Although these early findings are promising, further studies, including RCTs, are required to validate the utility of ACI for gastric ESD.

Emerging technologies for gastric ESD
Several RCTs have recently been conducted to improve ESD techniques and shorten procedure time. The multibending endoscope (Fig. 4), which has two bending sections at its tip, facilitates access to areas that are difficult to reach using a conventional single-bending endoscope. An RCT demonstrated that its use significantly reduced the procedure time and minimized muscle layer injury [89]. However, because the multibending endoscope is heavier than standard endoscopes, its routine use in all gastric ESDs is not recommended; rather, it is useful in selected cases.
The spray coagulation mode reduced the need for hemostatic forceps compared with the forced coagulation mode in an RCT [90]. Hybrid ESD has been introduced to shorten the procedure time for small gastric lesions. Hybrid ESD is a modified endoscopic technique that combines a mucosal incision with partial submucosal dissection (ESD portion), followed by planned snaring (EMR portion), and can be performed using a single device, SOUTEN (ST1850-20; Kaneka Medix, Tokyo, Japan) [91]. An RCT demonstrated a significantly shorter procedure time for hybrid ESD than for conventional ESD for lesions with absolute indications for EMR/ESD [92].

CONCLUSIONS
The indications for ER and curability criteria for EGC have expanded as long-term evidence has accumulated, although these criteria differ among countries. Furthermore, the risk of LNM after noncurative resection was stratified. Continuous advances in ESD techniques and technologies, including traction methods, multibending endoscopes, optimized coagulation modes, and emerging image-enhanced endoscopy, have contributed to safer and more reliable procedures. However, several significant challenges remain unresolved. Management after noncurative resection, particularly in older patients with substantial competing mortality risks, requires more individualized decision-making supported by better prognostic tools. Additionally, although multiple closure and hemostatic approaches have shown potential for reducing delayed bleeding, definitive benefits have not yet been established in RCTs. Continued research that addresses these challenges is essential for establishing optimal management approaches for patients with EGC.