Korean Practice Guidelines for Gastric Cancer 2024: An Evidence-based, Multidisciplinary Approach (Update of 2022 Guideline)

Screening

Korea has the highest age-standardized incidence rates of gastric cancer worldwide, but the ratio of gastric cancer-related mortality to cancer incidence is much lower than that of other countries [11]. The Korean National Cancer Screening Program (KNCSP) for gastric cancer appears to have played a crucial role in increasing the number of curable cancers through early detection, and thereby improving overall survival (OS) [12]. The KNCSP for gastric cancer, launched in Korea in 2002, invites all Korean individuals aged 40 or older to undergo endoscopy or upper gastrointestinal series (UGIS) every 2 years. A recent study showed that the screening group had a 41% lower hazard ratio (HR) for gastric cancer mortality compared with the non-screening group [13]. However, the reduction in gastric cancer mortality was significant only in the group that received endoscopic screening, not in the group that received UGIS [14].

Diagnosis and classification of EGC

In the Japanese Classification of Gastric Carcinoma, superficial gastric carcinoma is categorized according to morphologic features; polypoid lesions are classified as type I (protruding), flat lesions as type II (superficial), and ulcerated lesions as type III (excavated) [15]. Type II lesions are further subdivided into 3 groups according to the elevation or depression of the lesion relative to the surrounding mucosa: IIa (superficial elevated), IIb (superficial flat), and IIc (superficial depressed). Tumors elevated by more than 3 mm are classified as type I [15].

Staging by endoscopic ultrasound (EUS)

EUS can be helpful for assessing the depth of local tumor invasion (T stage) and regional lymph node (LN) metastasis [16]. According to the results of a Cochrane review, the summary sensitivity and specificity of EUS in distinguishing T1 and T2 (superficial) vs. T3 and T4 (advanced) gastric carcinomas were 86% (95% confidence interval [CI], 81% to 0.90%) and 90% (95% CI, 87% to 93%), respectively (Table 3) [17]. For the diagnostic capacity of EUS in distinguishing T1 vs. T2 tumors, a meta-analysis of 46 studies (n=2,742) showed that the sensitivity and specificity were 85% (95% CI, 78% to 91%) and 90% (95% CI, 85% to 93%), respectively. For the capacity of EUS to distinguish between T1a (mucosal) vs. T1b (submucosal) cancers, a meta-analysis of 20 studies (n=3,321) showed that sensitivity and specificity were 87% (95% CI, 81% to 92%) and 75% (95% CI, 62% to 84%), respectively. Finally, for assessing metastatic involvement of LNs (N stage), a meta-analysis of 44 studies (n=3,573) showed that sensitivity and specificity were 83% (95% CI, 79% to 87%) and 67% (95% CI, 61% to 72%), respectively. However, the high heterogeneity between studies indicates that the diagnostic accuracy of EUS depends on the operator.

Table 3
Diagnostic accuracy of endoscopic ultrasound (Cochrane review)

• Click for larger image
• Click for full table
• Download as Excel file

UGIS

The UGIS has been used for screening and for evaluating postoperative complications in gastric cancer. Recently, the percentage of participants of the KNCSP who undergo UGIS for gastric cancer screening has decreased [18, 19]. Studies using a large general population cohort participating in hte KNCSP reported that screening by endoscopy reduced gastric cancer mortality, but the effect was not significant in the UGIS group [14, 20]. Therefore, starting in 2018, the program was changed to perform endoscopy as the basic examination, with UGIS being performed optionally in cases where it is difficult to perform endoscopy.

Computed tomography (CT)

CT has been widely used to detect and diagnose gastric cancers, determine the optimal treatment method through accurate staging (cTNM), and assess therapeutic effects after anticancer treatments. Multidetector row CT (MDCT), which has multiple parallel rows of X-ray detectors in the craniocaudal direction (z-direction), enables various high-quality multiplanar reformation (MPR) imaging. Since the introduction of MDCT, the accuracy of gastric cancer staging and the detection of EGCs or small metastatic lesions have improved. Although isolated lung metastasis is uncommon in gastric cancer, chest CT can be helpful in case of esophageal involvement in GEJ cancer [21, 22, 23, 24]. For the CT protocol, an MDCT unit with 16 or more channels is recommended to acquire isotropic imaging with collimation of less than 1.25-mm [25]. The patient should fast for at least 6 hours. Optimal gastric distension is critical for successful CT gastrography and is achieved using either a negative contrast agent (effervescent gas-producing agent) or a neutral contrast agent (water). Anti-peristaltic drugs can help reduce motion artifacts. Patient positioning is determined based on the location of the suspected lesion and the type of oral contrast used (e.g., supine/prone, right decubitus/left posterior oblique). Obtaining images from appropriate positions facilitates evaluation of the entire stomach in a distended state. Portal venous phase images typically provide information on tumor depth, regional LN metastasis, and distant metastasis. Arterial phase images are useful for detecting abnormal gastric wall enhancement and assessing possible anatomic variations in surgically relevant vasculature, such as a replaced left hepatic artery arising from the left gastric artery.

S1. MPR imaging with MDCT in gastric cancer staging accuracy

The staging accuracy of MDCT has been reported to range from 67.1% to 89.1% (median, 78.6%) for T staging and from 49.3% to 79.5% (median, 68.8%) for N staging [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39]. MDCT, which enables faster scanning with thinner slice thicknesses, can generate high-quality reformation images, such as MPR images, CT gastrography, or virtual gastroscopy. In a meta-analysis, the addition of MPR images to axial images improved staging accuracy, particularly for T staging (accuracy difference [95% CI], 0.10 [0.02 to 0.18] for T staging [P=0.01] and 0.04 [−0.04 to 0.13] for N staging [P=0.33]) (Fig. 1) [27, 29, 30]. Three-dimensional reformation images, such as CT gastrography or virtual gastroscopy, can improve the detection rate of EGCs, potentially allowing for more accurate T staging [26, 27, 36]. Regarding the detection of peritoneal metastases, MDCT has been reported to have high specificity, ranging from 57.1% to 100% (median, 96.5%), but low sensitivity, ranging from 25.0% to 90.0% (median, 57.6%) [40, 41, 42, 43, 44, 45].

Fig. 1
Forest plot comparing staging accuracy between MPR plus axial plane vs. axial plane only in multidetector row computed tomography.
MPR = multiplanar reformation; CI = confidence interval.

• Click for larger image

Magnetic resonance imaging (MRI)

Evaluation of liver metastases is one of the most potent applications of MRI in gastric cancer. Many studies on gastrointestinal malignancies, especially colorectal cancer, have demonstrated that liver-specific contrast-enhanced MRI with diffusion-weighted imaging (DWI) is the most sensitive imaging method for the diagnosis of liver metastases [46]. Although no studies have focused exclusively on gastric cancer patients, liver-specific contrast agent-enhanced MRI with DWI is expected to be useful for diagnosing liver metastases in gastric cancer due to its high contrast resolution. A meta-analysis has shown the applicability of MRI in evaluating T stage and peritoneal metastases [47, 48]. However, further investigation is needed to confirm these findings due to the small number of patients included in these analyses.

S2. FDG PET/CT for staging workup

A total of 20 studies were reviewed, 19 studies with 2,195 patients were included in the meta-analysis assessing the diagnostic ability of FDG PET/CT for detecting LN metastasis [55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67] or distant metastasis [60, 62, 64, 68, 69, 70, 71, 72, 73] in gastric cancer patients during staging. The pooled sensitivity and specificity of FDG PET/CT for detecting LN metastasis were 45% (95% CI, 34% to 57%) and 87% (95% CI, 80% to 92%), respectively. For the evaluation of distant metastasis, the pooled sensitivity and specificity were 61% (95% CI, 42% to 78%) and 97% (95% CI, 82% to 99%), respectively.

One possible reason for the low sensitivity of FDG PET/CT in detecting LN or distant metastasis could be the inclusion of diffuse-type (Lauren classification) or signet ring cell type cancers, which generally exhibit lower FDG uptake. However, FDG PET/CT tends to provide more accurate diagnoses for intestinal-type tumors.

FDG PET/CT showed high specificity in detecting LN and distant metastases and can be considered a supplementary diagnostic tool with diagnostic CT for staging the workup in gastric cancer.

S3. FDG PET/CT for the diagnosis of cancer recurrence

A total of 13 studies with 1,567 patients were included in the meta-analysis [74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86]. The pooled sensitivity and specificity of FDG PET/CT for detecting the recurrence of gastric cancer were 81% (95% CI, 71% to 88%) and 88% (95% CI, 80% to 93%), respectively, with an area under the summarized receiver operating characteristic curve (AUC) of 0.91 (95% CI, 0.89 to 0.93).

Of these 13 studies, 5 studies with 438 patients compared the diagnostic ability in detecting recurrence between FDG PET/CT and contrast-enhanced CT [74, 76, 78, 84, 86]. In the meta-analysis of these 5 studies, FDG PET/CT showed a pooled sensitivity of 72% (95% CI, 50% to 87%) and specificity of 89% (95% CI, 69% to 97%) with an AUC of 0.88 (95% CI, 0.85 to 0.90), whereas contrast-enhanced CT revealed a pooled sensitivity of 88% (95% CI, 74% to 95%) and specificity of 83% (95% CI, 65% to 93%) with an AUC of 0.92 (95% CI, 0.90 to 0.94). There was no statistically significant difference in diagnostic accuracy between FDG PET/CT and contrast-enhanced CT (P>0.05). While FDG PET/CT showed higher sensitivity for detecting bone metastasis than contrast-enhanced CT, contrast-enhanced CT showed higher sensitivity for detecting peritoneal metastasis than FDG PET/CT. Due to its high specificity, PET/CT could be helpful for the differential diagnosis of equivocal lesions on contrast-enhanced CT.

Regarding recurrence, 2 studies assessed the diagnostic value of FDG PET/CT for detecting recurrence in 29 patients with elevated levels of serum tumor markers and negative results on conventional radiological imaging [74, 81]. Among these 29 patients, FDG PET/CT detected cancer recurrence in 17 patients (59%).

FDG PET/CT could be useful for detecting recurrence in patients with equivocal findings on contrast-enhanced CT, elevated serum tumor marker levels but negative findings on conventional imaging.

Preparation of the specimens

For resected gastric cancer specimens, the stomach is opened along the greater curvature unless the tumor is located on the greater curvature (in which case, it is opened along the lesser curvature). For endoscopic mucosal resection (EMR)/endoscopic submucosal dissection (ESD) specimens, the specimen is spread out with the mucosal side up and pinned to a flat board. The proximal and distal directions are marked for orientation.

Specimen fixation

After completing the preparation process, the specimens should be immediately immersed in 10% buffered formalin solution (as quickly as possible). The volume of fixative solution should be more than ten times that of the specimen [87]. A proper fixation time (between 24 and 48 hours) at average room temperature is recommended for additional immunohistochemical or genomic evaluation [15, 88].

Macroscopic types

Superficial gastric cancer can be subclassified into 5 categories: protruding (EGC type I), superficial elevated (EGC type IIa), superficial flat (EGC type IIb), superficial depressed (EGC type IIc), and excavated (EGC type III) [15].

Based on Borrmann’s classification, the gross type of advanced gastric cancer (AGC) can be divided into polypoid (type 1), ulcerofungating (type 2), ulceroinfiltrative (type 3), diffuse infiltrative (type 4), and unclassifiable (type 5) [89, 90].

Inspection and sectioning of the specimens

For resected specimens, the location, size (maximum diameter), number, macroscopic types, appearance of the tumor, and length of the closest proximal and distal resection margins should be measured and recorded. The deepest part of the tumor invasion should also be noted. Additionally, any findings other than the tumor lesion, such as congestion, hemorrhage, ulcer, and perforation, should be assessed. For EMR/ESD samples, all specimens should be collected and embedded in blocks. The lateral and basal resection margins should be marked with ink to aid in accurate evaluation of the margins.

For sectioning, EMR/ESD specimens should be sectioned serially at 2-mm intervals parallel to a line that includes the closest lateral margin of the specimen. If the lesion is grossly AGC, at least 4 representative sections should be taken, including the deepest part of the tumor invasion. For grossly EGC lesions, grid mapping should be performed at a width of 4 to 5 mm. If there is suspicion of resection margin involvement with the tumor lesion, additional sections should be taken. In postchemotherapy gastrectomy specimens, representative sections are sufficient if the lesion is grossly obvious. However, the entire tumor bed must be examined microscopically when no residual cancer cells are found in representative section, or when the residual lesion is small or grossly inconspicuous. For multiple tumors or lesions with unusual configurations, appropriate sectioning should be implemented for proper evaluation on a case-by-case basis.

Histologic classification

The World Health Organization (WHO) classification system of digestive tumors, 5th edition, is used for the pathologic classification of gastric carcinoma [91, 92]. In addition, the Lauren classification can be applied in resected specimens, including ESD specimens [93].

Tumor size

Tumor size describes the largest dimension (cm) of the tumor.

Depth of invasion

In the staging of gastric cancer, the pT category is determined by the depth of tumor invasion. Tumors with invasion beyond the proper muscle layer are classified as AGC, while tumors with invasion limited to the mucosal or submucosal layers are classified as EGC. Submucosal invasion depth is further divided into the upper third (sm1), middle third (sm2), and lower third (sm3). When the proper muscle layer is lost at the ulcer site and there is a tumor in that area, it is considered subserosal invasion. Even if there is no tumor cell invasion of the muscle, if the tumor extends below an imaginary line connecting the proper muscle layers, it is classified as invasion of the proper muscle.

For endoscopic resection specimens, submucosal invasion depth is measured from the lowest surface of the muscularis mucosa. When the muscularis mucosa is absent in the area of deepest invasion, the invasion depth is measured from the virtual line that smoothly connects the adjacent normal layers.

LN

A sufficient number of regional LN dissections and pathological evaluations are essential for the accurate diagnosis of N staging. The pathologic assessment should include both the total number of nodes and the number of positive nodes. At least 16 local nodes should be assessed to evaluate N3a staging; however, some studies suggest that it is desirable to remove and assess 30 or more nodes [106, 107].

A tumor deposit is defined as a discrete tumor nodule within the lymphatic drainage zone of the primary carcinoma without identifiable LN tissue, blood vessels, or neural structures [106]. Tumor deposits, where metastatic tumor lesions in the subserosal fat are separated from the adjacent primary gastric cancer without evidence of LN tissue, are considered to be local LN metastases.

Resection margin

In gastric cancer, the proximal and distal margin status are described, and, where applicable, the circumferential margin status is additionally described in GEJ cancer. The safety margin describes the distance between the resection margin and the tumor. If there is a discrepancy between the grossly observed safety margin and that observed microscopically, the microscopic findings are reported.

For endoscopic resection specimens, the mucosal resection margins are described by indicating the direction closest to the resection margin and the distance from the resection margin. The deep resection margin is also measured at the closest point from the tumor and described.

Lymphatic invasion, vascular invasion, and perineural invasion

The presence or absence of lymphovascular invasion and perineural invasion should be described. For endoscopic resection specimens, it is recommended to separately report lymphatic invasion and vascular invasion. Immunohistochemical staining (D2-40) could be helpful for identifying lymphatic invasion.

Regression grade

For the grading of primary tumor regression after neoadjuvant therapy, the modified Ryan system is recommended (Table 4) [108].

Table 4
Regression grade

• Click for larger image
• Click for full table
• Download as Excel file

Peritoneal washing

The presence of cancer cells in peritoneal washing cytology is classified as metastatic disease (pM1). There is evidence that positive cancer cells in peritoneal washing cytology among AGC patients are correlated with poor prognosis. Peritoneal washing cytology could be helpful in staging of AGC.

Biomarkers

S4. Endoscopic resection for EGC meeting classical absolute indications

Since the early 2000s, ESD has been used in Korea as a minimally invasive therapy modality for EGC [136, 137]. Data from the Korean National Health Insurance Service System showed 23,828 cases of ESD for EGC between November 2011 and December 2014 [137]. Previous studies have suggested that ESD could be considered as the first-line therapy for mucosal-confined EGC with well or moderately differentiated tubular or papillary adenocarcinoma, with a tumor size ≤2 cm, and no ulcer in the tumor (classical absolute indication) as these characteristics indicate a very low risk of LN metastasis [138]. ESD allows high rates of en bloc curative resection with low adverse event rates [136, 139, 140, 141, 142]. A large Japanese retrospective study including 5,265 patients who underwent gastrectomy with LN dissection for EGC showed that none of the 1,230 well differentiated intramucosal cancers with diameters less than 30 mm were associated with metastases, and none of the 929 lesions without ulceration were associated with LN metastasis regardless of tumor size [138]. For lesions meeting classical absolute indications, the en bloc resection rate was 97.1%–99%, the curative resection rate was 91.5%–96.4%, and the local recurrence rate was 0.2%–1.8% [139].

Few studies have directly compared survival between ESD and gastrectomy for classical absolute indications, as many combine cases with classical absolute and expanded indications. In Korean retrospective cohort studies, patients meeting classical absolute indications showed no significant differences in 5-year OS rates (ESD, 93.6%–96.4% vs. gastrectomy, 94.2%–97.2%) or 10-year OS rates (ESD, 81.9% vs. gastrectomy, 84.9%) between treatment methods [140, 141, 142].

A small Korean study of 35 endoscopic resections and 20 gastrectomies with classical absolute indications, showed no differences in OS (months) (93.4±3.2 [endoscopic resection], 85.8±5.5 [gastrectomy]) or disease-free survival (DFS) (89.7±3.6 [endoscopic resection], 90.4±3.5 [gastrectomy]) [143]. Similar results were reported in a Japanese study, which found no significant OS difference between endoscopic resection and gastrectomy across age groups (<65 years, ≥65 years) for cases meeting classical absolute indications [144].

Five-year metachronous recurrence rates were higher after endoscopic resection (5.8%–10.9%) compared to gastrectomy (0.9%–1.1%) [140, 141, 142]. Close endoscopic surveillance should be performed following ESD for early detection of metachronous cancer.

While endoscopic resection may increase the incidence of metachronous cancer due to preservation of the stomach, it may also offer better quality of life (QOL), shorter hospital stay, lower costs, and fewer treatment-related complication rates compared to gastrectomy [140, 141, 142, 143, 145].

S5. Endoscopic resection for EGC meeting expanded indications

Endoscopic resection for EGC is limited by the inability to perform LN dissection during the procedure. Therefore, to achieve curative resection with survival outcomes comparable to surgery, EGC cases with a very low risk of LN metastasis should be carefully selected. The clinically acceptable threshold of LN metastasis risk may be equivalent to perioperative mortality rates following radical gastrectomy (0.1%–0.3% in high-volume centers in Korea and Japan) [146, 147, 148]. In addition, with endoscopic resection, it is technically feasible to achieve en bloc resection which is important to avoid remnant tumors or local recurrence after the procedure.

When the following criteria were met in the pathologic review of endoscopic resection specimens, the extragastric recurrence (nodal or distant metastasis) rate after endoscopic resection was between 0% and 0.21%, comparable to that of radical gastrectomy: well or moderately differentiated tubular or papillary adenocarcinoma, en bloc resection, negative lateral and vertical resection margins, no lymphovascular invasion, and either 1) tumor size >2 cm, mucosal cancer, and no ulcer in the tumor or 2) tumor size ≤3 cm, mucosal cancer, and ulcer in the tumor [149, 150, 151]. OS was also comparable between patients undergoing endoscopic resection and those treated with radical surgery (93.3%–96.4% vs. 92.0%–97.2%) [140, 145, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162].

Although a number of retrospective cohort studies support ESD, no prospective trials have compared outcomes with those of standard operations based on these criteria, where concerns for node metastases may still exist [150, 163, 164, 165]. Consequently, gastrectomy with LN dissection may also be a valid treatment option, especially in cases of ESD with technical difficulty or when periodic endoscopic follow-up is not be feasible or affordable.

S6. Endoscopic resection for EGC meeting relative indications

EGCs with poorly differentiated tubular and PCC (including SRCC) are associated with a higher risk of LN metastasis than well and moderately differentiated tubular EGCs, making endoscopic resection very cautious consideration.

In previous Japanese Gastric Cancer Guidelines, a literature review of the literature that endoscopic resection could be considered for poorly differentiated tubular adenocarcinoma or PCC (including SRCC) in cases with histologic confirmation from forceps biopsy specimens, endoscopically estimated tumor size ≤2 cm, endoscopically mucosal cancer, and no ulcer in the tumor [105]. When these criteria were met, the risk of LN metastasis was reported to range from 0% to 2.3% [166, 167, 168].

Under the mentioned endoscopic findings, endoscopic resection could be considered for initial treatment. However, when risk factors for LN metastasis (tumor size >2 cm, submucosal invasion, ulcer in the tumor, and lymphovascular invasion) are revealed in pathologic reports, additional gastrectomy may be necessary [169].

In this guideline, we reviewed recent literature published since the previous edition. Currently, no prospective RCTs have compared the long-term OS of endoscopic resection with that of gastrectomy with LN dissection, which is the standard treatment for these indications [170]. Retrospective studies have shown no difference in OS between gastrectomy and endoscopic resection, though endoscopic resection had a higher local recurrence rate in terms of recurrence-free survival (RFS), which is consistent with the findings of previous studies [145, 171, 172]. In a prospective, single-arm, phase III observational study in Japan (JCOG1009/1010), the curative resection rate of the endoscopic resection group in undifferentiated EGC was 71% (195/275). Over a median follow-up period of 69.9 months, the 5-year OS rate was 99.3% (95% CI, 97.1% to 99.8%) and 5-year RFS rate was 98.9% (95% CI, 96.5% to 99.6%) [173]. In Korea, a study on the Comparison of Endoscopic Resection And Surgery for Early Gastric Cancer with undifferentiated histological type: a multicenter RCT (ERASE-GC trial, NCT04890171), is currently ongoing, and its results should be followed-up.

To date, the standard treatment for these criteria has been gastrectomy with LN dissection. Only retrospective cohort studies support these criteria for endoscopic resection, and the data from prospective trials are still lacking. Additionally, a significant portion of cases estimated to meet these criteria in the pre-endoscopic resection workup are found to be out these criteria upon pathologic examination of endoscopic resection specimens. Therefore, standard surgery (gastrectomy with LN dissection) can also be considered for cases meeting these criteria. It is advisable to choose a treatment method after sufficient discussion with the patient about the possibility of LN metastasis, and complications associated with endoscopic procedure and surgery.

S7. Additional surgery after noncurative endoscopic resection for EGC

Endoscopic resection of EGC could be revealed pathologic characteristics that do not meet the criteria for curative resection. Resected tumor characteristics that do not meet the following criteria are considered noncurative: 1) differentiated type (well or moderately differentiated tubular or papillary adenocarcinoma mucosal cancer of any size without ulcer), 2) differentiated type mucosal cancer measuring ≤3 cm with ulcer, 3) differentiated type cancer with minute submucosal invasion (invasion depth ≤500 µm) measuring ≤3 cm, or 4) undifferentiated type (poorly differentiated tubular adenocarcinoma or PCC) mucosal cancer measuring ≤2 cm without ulcer. Lymphovascular invasion and positive vertical margins are also important factors indicating the need for further surgical treatment.

As a result of a literature search for reinforcement of the up-to-date guidelines, a total of 17 studies comparing additional surgery to observation for noncurative endoscopic resection were included in the final table of evidence [150, 164, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186]. Most studies indicated a high risk of bias in terms of participant comparability, as those who did not undergo surgery tended to be older and had higher rates of comorbidities compared to those who underwent additional curative surgery [150, 164, 174, 175, 176, 177, 179, 182, 184, 186]. In addition, there was a significant difference in tumor-related characteristics between groups [150, 164, 174, 176, 177, 178, 180, 181, 183, 184, 185].

The 5-year OS rate across 15 studies was significantly higher in the surgery group than in the observation group [150, 164, 176, 177, 181, 182, 184, 186]. Among 12 studies examining disease-specific survival, all but one study showed a survival benefit for additional surgery [164, 183, 184], although this difference was not statistically significant in several studies [174, 176, 177, 179, 181, 185, 186]. A propensity score matching analysis also showed significantly higher 5-year OS and disease-specific survival rates in the surgery group compared to the observation group (91.0% and 99.0% in the surgery group vs. 75.5% and 96.8% in the observation group) [181].

LN metastasis was found in 2.0%–20.0% of patients who underwent additional surgery following noncurative endoscopic resection [150, 164, 174, 176, 177, 178, 180, 181, 182, 183, 184, 185, 186]. Given the high risk of LN metastasis and the survival benefit associated with curative surgery, additional gastrectomy with LN dissection is recommended when endoscopic resection for EGC does not meet the criteria for curative resection.

However, the survival benefit of additional surgery in older patients (>75 years) remains controversial [165, 179, 180, 184, 187]. For patients with comorbidities or poor general conditions, where curative surgery may not be feasible, observation with regular follow-up could be a valid option, provided the patient gives informed consent after an explanation of the risk of recurrence.

S8. Endoscopic treatment of positive lateral margins after ESD in EGC

It has been reported that there is a minimal risk of LN metastasis with en bloc resection, when only the lateral margin is positive and other criteria for complete resection are met. For differentiated-type EGC with positive lateral margin following ESD, when only close observation was performed, the 5-year local recurrence rate was 11.9% and there was no gastric cancer related mortality [188]. Therefore, close observation, endoscopic treatment (ESD or APC) and gastrectomy are considered as possible treatment options in these cases.

Seven retrospective studies compared recurrence rates among endoscopic treatments (re-ESD and APC), gastrectomy, and close observation [189, 190, 191, 192, 193, 194, 195]. In 6 studies with an average follow-up of 60 months (including both differentiated and undifferentiated cancers), local recurrence rates were as follows: 0% (95% CI, 0% to 0.02%; 0/163) in the gastrectomy group, 1.9% (95% CI, 0.5% to 6.9%; 2/101) in the re-ESD group, 13.4% (95% CI, 7.2% to 23.6%; 9/67) in the APC group, and 23.5% (95% CI, 17.4% to 30.1%; 35/149) in the observation group. In a meta-analysis, endoscopic treatments (re-ESD and APC) significantly reduced recurrence rates compared to close observation (relative risk [RR], 0.22; 95% CI, 0.06 to 0.86; P=0.03) (Fig. 2A), but had higher recurrence rates than gastrectomy (RR, 6.45; 95% CI, 1.17 to 35.52; P=0.03) (Fig. 2B). Local residual cancer was found in 64.7% of gastrectomy cases (95% CI, 56.8% to 71.9%; 97/150), while LN metastasis rate was 0.6% (95% CI, 0.1% to 1.9%; 1/150). However, all local recurrence cases can be successfully managed with further endoscopic treatment or surgery. Gastric cancer-related mortality was reported in 3 studies, and there was no gastric cancer-related death in the endoscopic treatment group or the observation and gastrectomy groups [189, 190, 193]. Therefore, considering QOL and mortality associated with gastrectomy, endoscopic treatment could be considered in patients with a positive lateral margin after ESD in EGC. Considering the recurrence rate of endoscopic treatment (5.8%; 95% CI, 2.29% to 9.21%; 10/174), close follow-up after endoscopic treatment is necessary. Although no gastric cancer-related deaths were reported in any of the 3 groups, the study populations in each study were relatively small, and baseline characteristics differed due to the observational study design. Further research is needed to compare the mortality and survival outcomes of close observation, endoscopic treatment, and gastrectomy in a larger population.

Fig. 2
Forest plot for comparison of local recurrence. (A) Risk of local recurrence in endoscopic treatment group vs. follow-up without therapy group. (B) Risk of local recurrence in endoscopic treatment group vs. gastrectomy.
Tx = treatment; CI = confidence interval; M-H = Mantel-Haenszel.

• Click for larger image

Three retrospective studies compared gastrectomy and close observation in patients with a positive lateral margin after ESD in differentiated type EGC [189, 190, 193]. The local recurrence rate in the gastrectomy group (0%; 95% CI, 0% to 0.1%; 0/44) was significantly lower than that of the close observation group (19.6%; 95% CI, 12.9% to 28.6%; 19/97). However, cancer-related mortality was zero in both groups. All cases of local recurrence in the observation group can also be managed with additional endoscopic treatment or surgery. Among patients who underwent gastrectomy, local residual cancer was identified in 51.6% of cases (95% CI, 34.8% to 68.0%; 16/31), but the LN metastasis rate was 0% (95% CI, 0% to 0.1%; 0/44). Long-term follow-up studies showed that a cancer-positive lateral margin length longer than 6 mm was significantly associated with local recurrence [188]. Consequently, close observation could be considered a selective treatment option in cases of positive lateral margins in differentiated-type EGC. Recently, a retrospective study was published comparing gastrectomy and non-surgical treatments (endoscopic treatment [6/52] and close observation [46/52]) in undifferentiated-type EGC [195, 196]. The local recurrence rate was 0% in the surgical group and 21.2% (11/52) in the non-surgical group. The 5-year survival rate was 87.8% in the non-surgical group, lower than the 95.0% in the surgical group, though the difference was not statistically significant. Therefore, close observation may be considered in elderly patients or those with high morbidity in undifferentiated-type EGC, but additional studies are needed to confirm these findings.

S9. H. pylori eradication after endoscopic resection for EGC

In 1999, the WHO classified H. pylori as the first-class carcinogen for gastric cancer, and it infects approximately 50% of the global population. Eradication H. pylori has shown significant benefits for asymptomatic infected individuals (pooled incidence rate ratio, 0.62; 95% CI, 0.49 to 0.79) [197], making eradication of H. pylori an important strategy to prevent gastric cancer.

Eradication of H. pylori is also crucial for the prevention of metachronous gastric cancer in patients who have been successfully treated by endoscopic resection of EGC with H. pylori infection. We identified 3 RCTs that observed the incidence of metachronous gastric cancer and precancerous lesions in patients treated with H. pylori eradication vs. those who did not receive eradication treatment following successful endoscopic resection for EGC [198, 199]. The median follow-up periods were 3–5 years. During follow-up, the incidence of metachronous gastric cancer and precancerous lesions was 4.80% (41/856) in the H. pylori eradiation group and 9.75% (87/892) in the non-eradiation group. The risk of developing metachronous gastric cancer and precancerous lesions was significantly lower in the eradication group (HR, 0.45; 95% CI, 0.31 to 0.66) (Fig. 3). Additionally, the study by Fukase et al. [198], which focused solely on metachronous gastric cancer, showed a significant benefit from H. pylori eradication (HR, 0.34; 95% CI, 0.16 to 0.73). Therefore, H. pylori eradication is helpful for the prevention of metachronous gastric cancer in patients who have been successfully treated by endoscopic resection of EGC with H. pylori infection.

Fig. 3
Forest plot for a comparison of the risk of metachronous gastric cancer between Helicobacter pylori eradication vs. no treatment.
SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

S10. Reconstruction methods following distal gastrectomy (DG)

Functional and nutritional outcomes may vary depending on the reconstruction method including BI, BII, and RY [200].

Well-designed studies directly comparing each reconstruction method are rare. In our meta-analysis, which included a limited number of studies, BI showed advantages in operation time (P<0.01), hospital stay (P<0.01), and bile reflux (P<0.03) over BII [201, 202]. There was no significant difference in complications (P=0.10). BI was also more favorable than RY in terms of operation time (P<0.01), complications (P=0.01), and hospital stay (P<0.01) [201, 203, 204, 205, 206]. Other benefits of BI include decreased risk of iron deficiency anemia, preservation of alimentary tract continuity, no risk of Petersen hernia, reduced small bowel adhesions, and easier access to the duodenum and biliary tract in cases of biliary diseases [207, 208, 209, 210] (Fig. 4).

Fig. 4
Forest plots comparing reconstruction methods. (A) Operation time. (B) Complications. (C) Hospital stays. (D) Bile reflux. (E) Esophageal reflux.
SD = standard deviation; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval.

• Click for larger image

RY reconstruction showed advantages in preventing bile and esophageal reflux on endoscopic findings [201, 203, 204, 205, 206, 211]. However, current evidence is insufficient to confirm whether endoscopic bile reflux directly leads to superior QOL or the prevention of metachronous cancer. In a retrospective series, RY with an increased length of limbs after gastrectomy showed favorable metabolic effects for gastric cancer patients with type II diabetes [209, 212].

There was no significant difference in QOL (P=0.290–0.994) or nutritional parameters (e.g., weight loss, albumin) among the 3 reconstruction methods [201, 213]. Additionally, there is insufficient evidence to suggest differences in survival outcomes among reconstruction methods [214].

In summary, the experts in the guideline task force team could not recommend a specific reconstruction method as the best option for all cases, given the unique advantages and disadvantages of each reconstruction method. We recommend that surgeons select the optimal method according to the characteristics of the cancer and the patients.

S11. Resection margin

The impact of microscopically positive margin status varies according to the pathologic status of the cancer [215, 216]. In our review, positive resection margins correlated with inferior survival outcomes compared to negative margins in pathologic T1 cancers (Nagata et al. [217] [68.6% vs. 97.4%, P<0.0001], Sun et al. [216] [66.7% vs. 93.1%, P<0.04]) and T2 cancers (Sun et al. [216] [21.5% vs. 55.2%, P<0.001], Morgagni et al. [218] [8% vs. 64%, P<0.001]).

There is a debate regarding whether achieving an adequate length of margin may influence oncologic outcome in EGC [215, 219, 220, 221]. However, most literature agrees that attaining a negative margin regardless of margin length, is associated with improved survival [220, 222, 223]. Various methods, used alone or in combination, have been introduced to achieve negative margin, including intraoperative frozen section, perioperative gastroscopy, preoperative clipping or dyeing, fluorescence imaging techniques, ultrasonography, and simple X-ray, among others [224, 225, 226, 227, 228, 229, 230, 231, 232].

In cases of EGC, when pathologic results reveal tumor involvement at the resection margin, several studies showed a survival benefit with additional surgery to obtain R0 resection [222, 223, 233]. Therefore, development of this guideline made a consensus to recommend additional surgery when the patient’s condition is favorable and additional surgery is technically feasible.

However, there were reports showing that R1 resection does not always lead to recurrence, potentially due to lack of blood supply on the remnant transection line, discrepancy of true surgical margin from the use of surgical stapler or patients’ immunity [215, 217]. Careful observation with frequent follow-up might be cautiously considered when the extent of the involved margin is minimal or the anticipated risk of reoperation is high. Further investigations are needed to clarify the indications for cases where additional surgery may not be required.

Previous reports have recommended various macroscopic margin lengths (3–8 cm) to ensure pathologic negative resection margins in advanced or infiltrative cancer [231, 234, 235]. Intraoperative frozen section has shown improved accuracy compared to macroscopic margin prediction in achieving R0 resection [228, 231, 236, 237, 238]. The methods mentioned earlier to achieve negative margins can also benefit advanced cancer cases to obtain secure margins and tumor localization.

Unlike in early-stage gastric cancer, many studies have shown that a positive microscopic margin had no prognostic impact when staging was ≥T3 or ≥N2 or ≥IIIa (American Joint Committee on Cancer [AJCC] 7th) [216, 219, 220, 222, 239, 240, 241, 242, 243]. In these cases, achievement of a negative margin has shown limited survival benefits.

Therefore, utilizing various methods including intraoperative frozen section is advisable to achieve R0 resection in advanced cancer. However, in cases of advanced disease (≥pT3, ≥pN2, or ≥Stage IIIa [AJCC 7th]) with an R1 resection, reoperation should be decided cautiously considering the pathologic stage, patient condition, risk of postoperative complications, and the potential delay in systemic therapy.

S12. Proximal gastrectomy (PG) with double tract reconstruction (DTR)

TG has been the standard treatment for upper gastric cancer. Gastric cancer in the upper third of the stomach has limited node metastasis to the lower part of the stomach, making PG an acceptable and oncologically safe option [244, 245]. However, reconstruction after PG has posed challenges due to the high incidence of reflux esophagitis and anastomosis stricture in esophagogastrostomy.

DTR has recently been shown to be feasible in laparoscopic settings. In our systematic review of retrospective studies, significantly fewer patients in the PG-DTR group experienced vitamin B12 deficiency compared to those in the TG group (RR, 0.30; 95% CI, 0.23 to 0.40; P<0.01) [246, 247, 248]. Weight loss after surgery did not differ between the groups (RR, −4.89; 95% CI, −11.75 to 1.97; P=0.16) [248, 249]. Reflux symptoms were also comparable (RR, 1.28; 95% CI, 0.33 to 4.93, P=0.72) [247, 250, 251]. Complications were reported less frequently in the PG-DTR group (RR, 0.61; 95% CI, 0.45 to 0.83; P=0.002) (Fig. 5).

Fig. 5
Forest plots for comparison between PG DTR vs. TG in retrospective studies. (A) Vitamin B12 deficiency. (B) Weight loss. (C) Early complications. (D) Reflux symptom.
PG = proximal gastrectomy; DTR = double tract reconstruction; TG = total gastrectomy; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval; SD = standard deviation.

• Click for larger image

The Korean Laparoendoscopic Gastrointestinal Surgery Study Group conducted a prospective RCT comparing LPG-DTR and laparoscopic TG (KLASS-05) and recently reported early results. The cumulative amount of intramuscular vitamin B12 supplementation required over 2 postoperative years was significantly lower in the PG-DTR group than in the TG group (0.6±2.0 mg vs. 3.4±4.1 mg, P<0.001). The proportion of patients who required vitamin B12 supplementation was also significantly lower in the PG-DTR group (14.7% vs. 58%, P<0.001). However, hemoglobin changes after surgery, one of the primary outcomes, were not significantly different between the groups (5.6%±7.4% vs. 6.9%±8.3, P=0.349). Additionally, the Visick score for reflux symptoms at 2 weeks postoperatively (P=0.793) and postoperative complications (23.5% vs. 17.4%, P=0.373) were not different between the groups [252].

In KLASS-05, the 2-year OS rates and DFS rates of the PG-DTR and TG groups were 98.5% vs. 100% (P=0.330) and 98.5% vs. 97.1%, respectively (P=0.540) [252]. Regarding long-term QOL, the PG-DTR group showed better on the physical functioning (P=0.029) and social functioning (P=0.031) scales of the European Organization for Research and Treatment of Cancer QOL Questionnaire (EORTC QLQ-C30).

Recently, alternative reconstruction techniques such as side overlap esophagogastrostomy, double flap technique (DFT) reconstruction are being explored for better functional outcomes. Some studies show that DFT has better outcomes than TG in terms of morbidity, postoperative hospital stay, reflux esophagitis, and postoperative nutritional status [253]. However, laparoscopic PG with DFT requires more complex intracorporeal suturing technique and longer operation time [254]. Further studies are needed to provide higher level of evidence [255, 256].

S13. Pylorus-preserving gastrectomy (PPG) vs. DG for middle third gastric cancer

PPG preserves the pylorus and distal antrum to prevent rapid food transit into the duodenum and reduce reflux of duodenal contents. Consequently, the reduced postoperative incidence of dumping syndrome and reflux gastritis has been expected to show benefits in nutrition and QOL compared to DG.

Recently, a prospective RCT comparing laparoscopic PPG and laparoscopic DG (KLASS-04) reported its findings [257, 258].

The study found no significant differences in survival outcomes or complications between the PPG and the DG groups [257]. Additionally, the incidence of dumping syndrome one year after surgery was similar between the groups (13.2% vs. 15.8%, P=0.62). However, reflux esophagitis (17.8% vs. 6.3%, P<0.01) and delayed gastric emptying (16.3% vs. 4.0%, P<0.01) were more frequent in the PPG group than in the DG group 3 years postoperatively. Conversely, bile reflux (13.2% vs. 24.4%, P=0.02) and gallstone formation (2.3% vs. 8.7% P=0.03) were less frequent in the PPG group than in the DG group.

Although there was no difference in body weight change after surgery, the total protein levels were better preserved following PPG compared to DG (P<0.01). There were no QOL differences between the groups in terms of the EORTC QLQ-C30 and EORTC-QLQ-Gastric Cancer Module (STO22).

In several studies, there were no differences in survival outcomes or postoperative complications between PPG and DG [258, 259, 260, 261, 262, 263]. The PPG group experienced a lower incidence of postoperative dumping syndrome and reflux [261, 263, 264]. Some studies also reported reduced development of gallstones after PPG, probably due to the preservation of the hepatic branch of the vagus nerve [259, 261]. However, delayed gastric emptying was reported more frequently with PPG than DG [259, 261, 262, 263, 264].

Regarding nutritional status, decreases in serum protein and albumin levels during the first 6 months postoperatively and reductions in abdominal fat area at one year, were lower in the PPG group than in the DG group. Additionally, hemoglobin levels improved more in the PPG group than in the DG group [259, 262, 263, 265].

Two years after surgery, PPG showed trends toward better improvement of QOL and fewer symptoms than DG with BI reconstruction [265]. The PPG group had a better physical functioning score (86.7 vs. 90.0, P=0.032) but reported more pain and reflux than the DG group (median score 8.3 vs. 16.7 in pain, 11.1 vs. 11.1 in reflux, P=0.034 and 0.001, respectively) [263].

In summary, PPG showed benefits in reducing gallstone formation, bile reflux and preserving serum total protein, while survival, postoperative complications and QOL were similar to DG as shown in KLASS-04. Observational studies further reported possible nutritional and functional benefits of PPG. For EGC located at least 5 cm proximal to the pylorus, PPG could be a viable option; however, delayed gastric emptying should be carefully considered in decision-making.

S14. Prophylactic splenectomy for splenic hilar LN dissection in proximal-third AGC

The standard surgical procedure for proximal-third gastric carcinoma is TG with appropriate LN dissection. Therapeutic splenectomy may be necessary if the tumor directly invades the spleen or if LN metastasis around the splenic hilum is suspected. However, there is debate regarding whether splenic resection with LN dissection of the splenic hilum should be performed in the absence of direct invasion of the spleen, splenic hilum, or greater curvature of the stomach.

Three prospective RCTs have evaluated the survival benefit of prophylactic splenectomy in proximal-third gastric carcinoma [266, 267, 268]. Our meta-analysis showed no survival difference between groups (HR, 0.91; 95% CI, 0.68 to 1.21; P=0.51), but significantly fewer postoperative complications in the group without splenectomy (HR, 0.42; 95% CI, 0.31 to 0.59; P<0.01) (Fig. 6).

Fig. 6
Forest plot for a comparison between no splenectomy vs. splenectomy. (A) Survival. (B) Complications.
M-H = Mantel-Haenszel; CI = confidence interval.

• Click for larger image

The studies included in our meta-analysis excluded cases of advanced cancer with gross involvement of the greater curvature or the gastrosplenic ligament, where the metastasis rate of LN#10 is relatively high, and splenic hilar dissection with splenectomy is necessary for standard treatment [269, 270, 271].

In all situations, splenectomy is associated with increased postoperative complications and mortalities. To address this, operative techniques for LN#10 dissection around the splenic hilum without splenectomy has been reported; however, its oncologic outcomes are still under investigation [272, 273, 274].

S15. PG for GEJ adenocarcinoma

TG is the standard treatment for AGC in the upper part of the stomach [109, 235, 275]. However, some studies have questioned whether removing the entire stomach along with the perigastric tissues and LNs may be unnecessary in selected cases [244, 276, 277].

Our meta-analysis included 5 retrospective studies that investigated LN mapping of proximal gastric cancer following TG [276, 277, 278, 279, 280]. In the pooled data, the metastasis rates to distal LN station #4d, #5, and #6 (distal LN) were analyzed. The metastasis rates in pT2 cancer were very low: #4d (0/359), #5 (1/425), and #6 (0/359) (distal LN). The risk ratio of metastasis in the distal LN for pT3 was 2.48 (CI, 0.86 to 7.15) compared to pT2 (P=0.09), and the risk ratio for pT4 was 10.88 (CI, 4.95 to 23.91) compared to pT2 (P<0.01) (Fig. 7).

Fig. 7
Forest plot, LN metastasis rates of distal stomach according to the depth of tumor. (A) T2 vs. T3. (B) T2 vs. T4.
M-H = Mantel-Haenszel; LN = lymph node; CI = confidence interval.

• Click for larger image

In a retrospective study in Korea involving 873 patients [244], multivariable analysis identified a distance greater than 30 mm from the GEJ to the tumor epicenter, tumor size >70 mm, macroscopic Bormann type IV tumor, and serosa invasion as risk factors for LN metastasis to the distal stomach. In patients without these risk factors, the LN metastasis rates at stations #4d, #5, and #6 were 0.0%, 0.4%, and 0.4%, respectively. The therapeutic value index (TVI) of LN #4d, #5, and #6 were 0, 0.4, and 0.4, respectively (TVI is calculated by incidence of LN metastasis and 5-year survival after removing corresponding LN and can be used for evaluating necessity of dissection of specific LN stations) [270].

These results suggest that PG without dissection of LN stations #4d, #5, and #6 could be considered in selected cases of AGC with adenocarcinoma histology located in the GEJ (Siewert II/III) without serosal invasion. However, further data are needed to determine the detailed indications for PG and to evaluate the clinical outcomes of PG.

S16. Lower mediastinal LN dissection and surgical approach for GEJ adenocarcinoma

The definition and extent of surgery around the GEJ have not been solidly established. The most frequently used classification is the Siewert classification, which defines GEJ carcinoma as a tumor with its epicenter within 5 cm proximal and distal of the anatomical cardia and categorizing it into 3 types: type 1 (lower esophageal cancer), type 2 (true GEJ cancer), and type 3 (subcardial cancer) [281]. In Japan, GEJ carcinoma is defined as a tumor according to the Japanese classification system, regardless of histological type, when its epicenter is located within 2 cm proximal or distal to the GEJ [15].

It is generally acknowledged that Siewert type I and type III carcinomas are usually treated as esophageal and gastric tumors, respectively [282, 283]. Siewert type II adenocarcinoma, located 1 cm above to 2 cm below the GEJ, represents adenocarcinoma that arises from the epithelium of the cardia or short segments of intestinal metaplasia, primarily seen in Western countries. There has been considerable controversy about whether Siewert type II carcinoma is esophageal or gastric cancer and about the extent of LN dissection.

In a Korean retrospective analysis including 672 patients who underwent radical TG with lymphadenectomy without lower mediastinal LN dissection for GEJ carcinoma type II, type III, or upper third of the stomach [284], suggested that lower mediastinal LN dissection may not be necessary for early-stage cancer based on excellent survival regardless of the location (93.2% vs. 96.7% vs. 98.7% for Siewert type II, III, and upper-third gastric cancer, P=0.158). However, for advanced cancer, the survival was worse in Siewert type II than that in Siewert type III cancer (47.9% vs. 75.4% vs. 71.8% in Siewert type II, III, and upper-third gastric cancer, P<0.001), which result implies the necessity of mediastinal LN dissection.

Conversely, another retrospective analysis in Korea that reviewed 125 type II and 338 type III GEJ cancer patients demonstrated that there was no increase of recurrence in the mediastinal LNs without complete mediastinal LN dissection, regardless of the type. This led to the suggestion that a TH approach without complete mediastinal LN dissection can be acceptable for common types of GEJ cancer in Korea [285].

A prospective nationwide multicenter study in Japan reviewed the frequency of LN metastasis of GEJ tumors with cT2-T4 stages and recommended lower mediastinal LN dissection (especially station 110) if the length of esophageal invasion was more than 2 cm [23].

In summary, lower mediastinal LN dissection seems to be not essential for early GEJ cancers. For advanced GEJ cancers, lower mediastinal LN dissection may be needed to sufficiently remove possible metastatic LNs in case esophageal involvement is more than 2 cm; however, further data are needed to determine its impact on local recurrence and survival benefits.

Two RCTs on the optimal surgical approach for GEJ adenocarcinoma (Siewert type II, III) compared the surgical and oncological outcomes between the TH and TT approaches (one left thoracoabdominal approach and one right thoracotomy) [286, 287]. Our meta-analysis of these RCTs (Fig. 8) found comparable in-hospital mortality (P=0.10) and anastomosis leakage (P=0.57) between the TT and TH approaches. Although the Japanese RCT (JCOG9502) comparing the left thoracoabdominal and the TH approaches was stopped after interim analysis, the 5-year OS of the TH approach was not inferior to the TT approach (HR, 0.74; CI, 0.42 to 1.32; P=0.31) in both RCTs [286, 287].

Fig. 8
Forest plots for comparisons between the TH abdominal approach (Experimental) vs. TT approach (Control). TT approaches in the observational studies included. In-hospital mortality: (A) RCTs, (B) Observational studies. Anastomosis leakage: (C) RCTs, (D) Observational studies. Pulmonary complications: (E) RCTs, (F) Observational studies. Five-year survival: (G) RCTs, (H) Observational studies.
TH = transhiatal; TT = transthoracic; M-H = Mantel-Haenszel; CI = confidence interval; RCT = randomized controlled trial.

• Click for larger image

In our meta-analysis with observational studies including right and left thoracotomy and right thoracoscopic approaches, the TT approach was associated with a higher incidence of pulmonary complications (P=0.0002), a higher in-hospital mortality rate (P=0.02), and similar anastomosis leakage (P=0.57) compared to the TH approach. The TH approach was also not inferior to the TT approach in terms of 5-year OS (HR, 0.80; CI, 0.59 to 1.11; P=0.18) [23, 288, 289, 290, 291, 292, 293, 294].

Given the higher rate of surgical complications and no difference in 5-year OS, the TH approach is recommended rather than the TT approach for resectable adenocarcinoma invading the GEJ.

S17. D1+ dissection for EGC

D2 dissection has been considered the standard LN dissection based on long-term survival data of the Dutch trial [295]. However, the necessity of standard D2 dissection for EGC has been questioned, especially in laparoscopic surgery. There have been no studies comparing D2 dissection and less extensive dissection in EGC patients, but D1+ dissection has been widely accepted by surgeons in Korea and Japan considering the spatial information about LN metastasis, TVI calculated by the frequency of LN metastasis and 5-year survival rate after removing each LN station, and the Maruyama index, which evaluates the adequacy of LN dissection [270, 296, 297].

An Italian study suggested D2 dissection due to the presence of significant LN metastasis [298]. Other studies reported 10-year survival rates of 95% and 87.5% following standard D2 and D1 dissection, respectively, in EGC patients, with no statistically significant difference (P=0.80) [299]. In a Japanese report, the 5- and 10-year survival rates were 97% and 91% for standard D2 dissection and 98% and 91% for modified D2 (D1+) dissection. There were no cases with metastasis to second-tier LNs in patients with cT1N0 or cT1N1 disease [300].

Although evidence for comparing D1+ and D2 dissection is limited, we referenced the excellent survival outcomes from RCTs comparing laparoscopic and open gastrectomy in EGC, in which less than D2 dissection was performed, and strongly suggested that D1+ dissection can be performed for EGC [301, 302, 303].

S18. Sentinel node navigation surgery (SNNS)

To date, numerous feasibility studies for the sentinel node concept have shown a high detection rate and acceptable sensitivity of sentinel node mapping [304, 305]. Previous meta-analyses reported pooled detection rates and sensitivity rates of 93.7%–97% and 80.8%–89%, respectively, with higher sensitivity associated with clinical T1 tumors, dual tracers, submucosal injection, and IHC examinations [304, 305].

The oncologic safety and clinical benefits of SNNS have been evaluated only in a limited number of studies [306, 307, 308, 309, 310, 311]. In these studies, SNNS was performed on small sized early lesions less than 3 cm in diameter. Two case-control studies and one RCT compared 5-year OS rates [307, 310, 311]. Our meta-analysis of 2 case-control studies showed that there is no significant difference of OS between the SNNS and conventional LG groups (Fig. 9). In the RCT (SENORITA trial), the primary endpoint of 3-year DFS did not show non-inferiority of SNNS compared to LG due to a higher incidence of metachronous cancer (91.8 vs. 95.5%, SNNS vs. LG) [311].

Fig. 9
Forest plots for comparison between the SNNS vs. conventional surgery (Conventional) in observational studies. (A) Overall survival. (B) Body weight: percentages compared to preoperative body weight.
SE = standard error; SNNS = sentinel node navigation surgery; SD = standard deviation; IV = interval variable; CI = confidence interval.

• Click for larger image

However, secondary endpoints including 5-year DFS, OS, and disease-specific survival, were not different after rescue surgery in cases of recurrence or metachronous gastric cancer; 88.9% vs. 80.7% for DFS, 97.3% vs. 88.3% for OS, and 99.2% vs. 99.3% for disease-specific survival in SNNS vs. LG, respectively (P=0.056, 0.74, and 0.959) [312].

In terms of nutritional outcomes, SNNS was associated with less body weight loss and higher hemoglobin levels than LG [306, 308, 311]. QOL, assessed using various tools, showed better QOL in some subscales for SNNS [306, 308, 309, 311].

There are still controversies over technical issues for sentinel node such as type of tracer, detection method, and practical pathological examination method. It should be noted that the SENORITA trial had a rigorous protocol including a dual tracer composed of radioactive isotope (Technetium-99 m) and indocyanine green, sentinel basin dissection rather than pick-up biopsy, intraoperative frozen examination for sentinel LNs with 2 mm-interval cutting, 4-direction resection margins, and cytokeratin IHC for permanent histological evaluation. Therefore, SNNS should be performed under a strict protocol including indication criteria, detection methods, and follow-up plans, and treatment decisions should be made after sufficient discussion with the patient regarding the possibility of metachronous cancer and rescue surgery. Under these conditions, SNNS could be a treatment option for EGC, offering potential benefits in nutritional outcomes and QOL.

S19. Laparoscopic vs. open approach in DG

Since the first clinical trial in the early 2000s, several pivotal trials comparing LDG and ODG for early or locally AGC have been published [313].

Three prospective RCTs (KLASS-01, COACT 0301, and JCOG0912) were conducted to evaluate the noninferiority of LDG for EGC [148, 314, 315]. Our meta-analysis demonstrated longer operation times of laparoscopic surgery (P<0.01) but better surgical outcomes, such as less operative blood loss (P<0.001), fewer postoperative complications (P<0.001), or shorter hospital stay (P<0.001), compared with ODG (Fig. 10). The long-term survival in LDG was not inferior to that of the ODG group in all 3 trials (HR, 0.90; CI, 0.69 to 1.16; P=0.42) [301, 302, 315].

Fig. 10
Forest plots for comparisons between laparoscopic (Laparoscopy) and open (Conventional) distal gastrectomy in cStage I gastric cancer. (A) Overall survival. (B) Complications. (C) Intraoperative blood loss.
SE = standard error; SD = standard deviation; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval.

• Click for larger image

Three prospective RCTs (KLASS-02, CLASS-01, and JLSSG0901) were conducted to evaluate the oncologic safety of LDG compared to ODG in the treatment of locally AGC [316, 317, 318]. These trials enrolled cases preoperatively diagnosed with AGC (cT2-T4a): KLASS-02 included cases without LN metastasis or with limited metastasis to the left gastric artery or perigastric LNs, CLASS-01 recruited all cN0-3 cases, and JLSSG0901 included cases with N0-2 without bulky node metastasis [319, 320, 321]. Surgical outcomes were reported in operation times, intraoperative blood loss, complications, hospital stays [320, 322, 323], and oncological outcomes were reported in 5-year OS rates [316, 317, 318].

In our meta-analysis, LDG showed less intraoperative blood loss and longer operating time compared to ODG (both P<0.01, respectively). There were no significant differences in hospital stay (P=0.10) or complication rates (P=0.35). Five-year OS rate was also not different between the LDG and ODG groups (HR, 1.05; CI, 0.88–1.26, P=0.61) (Fig. 11).

Fig. 11
Forest plots for comparisons between laparoscopic (Laparoscopy) and open (Conventional) distal gastrectomy in cT2-4a gastric cancer. (A) Overall survival. (B) Complications. (C) Hospital stays (day). (D) Operation time (minutes). (E) Intraoperative blood loss (mL).
SE = standard error; SD = standard deviation; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval.

• Click for larger image

LDG can be an optimal treatment option for AGC as well as ODG. However, further studies are needed on application of laparoscopic approach to far-AGC such as cT4b cancers, which requires multivisceral resection.

S20. Robot gastrectomy (RG)

RG has some technical advantages compared to LG, including surgeon-controlled vision, tremor filter, and ergonomic wrist motion instruments [324, 325, 326]. Our meta-analysis included 2 RCTs and 8 retrospective to compare RG and LG [325, 327, 328, 329, 330, 331, 332, 333, 334, 335].

In our meta-analysis of the 2 RCTs [327, 328], RG was associated with fewer postoperative complications than LG (RR, 0.49; 95% CI, 0.31 to 0.78; P=0.003). The incidence of pancreatic fistula was not different in both RCT (P=0.20) and non-RCT (P=0.58) analyses. There was no significant difference in reoperation rates (P=0.25) and hospital stay duration (P=0.11) between RG and LG. However, RG was associated with a longer operation time (mean difference, 47.04 minutes; 95% CI, 30.67 minutes to 63.41 minutes; P<0.01) compared to LG (Fig. 12).

Fig. 12
Forest plots for comparisons between daVinci™ robot gastrectomy (Robotic) vs. laparoscopic gastrectomy (Laparoscopic). (A) Complications (RCTs). (B) Pancreatic fistula (RCTs and observational studies). (C) Overall survival (observational studies). (D) Operation time (observational studies).
SE = standard error; SD = standard deviation; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval; RCT = randomized controlled trial.

• Click for larger image

In the retrospective studies, there were no differences in either the 5-year OS (HR, 0.84; 95% CI, 0.57 to 1.24; P=0.38) or the 5-year RFS rate (HR, 0.98; 95% CI, 0.71 to 1.34; P=0.88) between the groups.

In selected reports, the additional total hospital cost for RG ranged from $3,000 to $5,000 compared to LG [325, 326, 327].

Based on the current evidence, the guideline task force team decided the recommendation as “conditional for” because of noninferior perioperative and survival outcomes and fewer postoperative complications of RG compared to LG. However, further investigation is needed to identify the potential benefit of RG to justify longer operation time and higher cost, which should be discussed with the patient as part of the shared decision-making process.

S21. Partial omentectomy (PO) for AGC

TO is regarded as a mandatory treatment for AGC, though high-level evidence to support this is lacking [105]. TO can be challenging and time-consuming during LG [336]. PO has been widely adopted as a safe procedure for EGC based on excellent survival outcomes observed in RCTs where PO was predominantly performed. Therefore, our meta-analysis focused on the applicability of PO for AGC [148, 303].

Our meta-analysis included 5 retrospective studies comparing PO and TO [337, 338, 339, 340, 341]. Among these, 5 studies with propensity-matched survival data were selected to minimize selection bias. PO was found to be noninferior to TO in progression-free survival (PFS) (HR, 0.89; CI, 0.74 to 1.07; P=0.20) and OS (HR, 0.82; CI, 0.67 to 1.00; P=0.06). Seven studies were combined for complication meta-analysis, which showed no differences in overall complications (P=0.10) or serious complications (P=0.92) between PO and TO (Fig. 13) [337, 338, 339, 340, 341, 342, 343].

Fig. 13
Forest plots for comparisons between PO vs. TO in advanced gastric cancer in observational studies. (A) Overall survival (propensity score matched). (B) Relapse-free survival (propensity score matched). (C) Complications.
PO = partial omentectomy; TO = total omentectomy; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

Seven previous meta-analyses evaluated the oncologic feasibility of PO [344, 345, 346, 347, 348, 349, 350]. All showed that PO was not inferior to TO regarding OS and relapse-free survival. Additionally, PO required less operation time and resulted in lower blood loss.

Although these studies included a significant proportion of patients with serosal invasion (T4a), the development working group of guidelines expressed concern about the possibly insufficient extent of resection in locally far advanced cancer, including cT4a cases. They recommended careful consideration of PO in such cases while waiting for more confirmative results.

S22. Ursodeoxycholic acid (UDCA) treatment after gastrectomy

Gallstone formation is known as one of the long-term complications following gastrectomy [208]. Denervation of the vagus nerve, obesity, rapid weight loss, and TG increase both the incidence and severity of gallstone formation [207, 351, 352]. Given the positive effects of prophylactic UDCA on reducing gallstones after bariatric surgery, UDCA can be as effective for patients undergoing gastrectomy for gastric cancer.

One RCT studied the use of prophylactic UDCA after gastrectomy in patients with gastric cancer [353]. Patients were randomized into 3 groups: placebo, 300 mg UDCA, or 600 mg UDCA, administered for one year. Compared to placebo, both 300 mg (odds ratio [OR], 0.27; 95% CI, 0.12 to 0.62; P<0.002) and 600 mg (OR, 0.20; 95% CI, 0.08 to 0.50; P<0.001) of UDCA showed decreased gallstone formation. A daily dose of 300 mg seems to be sufficient as the protective effect did not differ between 300 mg and 600 mg doses.

Considering that there are some risk factors for gallstone formation and that the PPG or DG with preservation of the hepatic branch of vagus nerve rarely experiences gallstone formation, further studies are needed to determine which patients may particularly benefit from UDCA prophylaxis and to assess the long-term effect over one year postoperatively.

S23. Adjuvant chemotherapy following curative gastrectomy

Surgical resection with D2 LN dissection is the standard treatment for gastric cancer. However, the prognosis for AGC remains poor [354, 355]. Two phase III RCTs conducted in Asia showed significant survival benefits for adjuvant chemotherapy over observation after curative gastrectomy with D2 LN dissection in patients with gastric cancer [356, 357]. In the adjuvant chemotherapy trial of S-1 for gastric cancer (ACTS-GC) in Japan, 1,059 patients with stage II (excluding T1) or III gastric cancer (by Japanese Classification, 2nd English edition) who underwent D2 gastrectomy received adjuvant S-1 for 1 year or were observed after surgery [357, 358]. The 3-year DFS rates were 72.2% in the S-1 group and 59.6% in the surgery-only group (HR, 0.62; 95% CI, 0.50 to 0.77; P<0.001), and the 3-year OS rates were 80.1% and 70.1%, respectively (HR, 0.68; 95% CI, 0.52 to 0.87; P=0.003). In the trial of capecitabine and oxaliplatin study in stomach cancer (CLASSIC) conducted in South Korea, China, and Taiwan, 1,035 patients with stage II–IIIB gastric cancer (by AJCC 6th edition [359]) who underwent D2 gastrectomy received either CAPOX for 6 months or were observed [356]. The 3-year DFS rates were 74% in the chemotherapy and surgery groups and 59% in the surgery-only group (HR, 0.56; 95% CI, 0.44 to 0.72; P<0.001). The 5-year follow-up data in these studies confirmed these findings [360, 361]. Our meta-analysis further showed that adjuvant chemotherapy significantly improved OS and DFS compared to surgery alone (OS: HR, 0.66; 95% CI, 0.56 to 0.78; DFS: HR, 0.62; 95% CI, 0.54 to 0.72; P<0.001) (Fig. 14). Based on these results, both chemotherapy regimens (S-1 or CAPOX) are currently accepted as standard treatments in patients with pathological stage II or III gastric cancer after curative gastrectomy. It should be noted, however, that there is currently no evidence to support adjuvant chemotherapy for patients who fall into the category of stage IB by the AJCC 6th edition but stage IIA by the AJCC 7th and 8th editions (pT1N2M0 and pT3N0M0) [362].

Fig. 14
Forest plots for comparisons between AC vs. surgery only and doublet vs. S1 monotherapy. (A) Overall survival for AC vs. surgery only (Surgery only). (B) DFS for AC vs. surgery only (Surgery only). (C) DFS for oral fluoropyrimidine-based doublet (Doublet) vs. S1 monotherapy (TS-1).
SE = standard error; IV = interval variable; CI = confidence interval; AC = adjuvant chemotherapy; DFS = disease-free survival.

• Click for larger image

Although the survival benefit of adjuvant S-1 for 1 year in gastric cancer patients was demonstrated, the optimal duration of adjuvant S-1 for gastric cancer was unclear. In the randomized phase III noninferiority trial (OPAS-1) in Japan, 590 patients with stage II (excluding T1N2–3 and T3N0) gastric cancer (by Japanese Classification, 3rd English edition) who underwent gastrectomy with D2 LN dissection (with D1 plus dissection being allowed for clinical stage IA) received 8-courses (12 months) or 4-courses (6 months) of S-1 [15, 363]. At the first planned interim analysis, this study was terminated early because the HR for DFS of the 4-course group compared with the 8-course group exceeded 1.37 and met the prespecified criteria for early termination. The 3-year DFS rate was 93.1% for the 8-course group and 89.8% for the 4-course group (HR, 1.84; 95% CI, 0.93 to 3.6; noninferiority margin for HR, 1.37), and the 3-year OS rates were 96.1% and 92.6%, respectively (HR, 3.37; 95% CI, 1.23 to 9.19). Therefore, S-1 for one year remains the standard adjuvant treatment for pathological stage II gastric cancer.

Despite the benefit of adjuvant S-1 for gastric cancer in the ACTS-GC trial, there was a question about a lack of efficacy in more advanced stages [357, 361]. In the JACCRO GC-07 trial, adjuvant chemotherapy of the S-1 plus docetaxel showed a survival benefits compared to the S-1 monotherapy in patients with stage III gastric cancer (by Japanese Classification, 3rd English edition) who underwent D2 gastrectomy [15, 364]. The 3-year DFS rates were 67.7% in the S-1 plus docetaxel group and 57.4% in the S-1 group (HR, 0.72; 95% CI, 0.59 to 0.87; P<0.001), and the 3-year OS rates were 77.7% and 71.2%, respectively (HR, 0.74; 95% CI, 0.60 to 0.92; P=0.008). Similarly, the ARTIST-2 trial for stage II or III gastric cancer with positive LNs (by AJCC 7th edition [365]) also showed the superiority of S-1 plus oxaliplatin (SOX; 74.3%) to S-1 monotherapy (64.8%) for the 3-year DFS rate (HR, 0.69; 95% CI, 0.41 to 0.99; P=0.042) [366]. Our meta-analysis, including data from JACCRO GC-07 and ARTIST-2 (Fig. 14), showed that adjuvant oral pyrimidine-based doublet regimens improved DFS compared to S-1 monotherapy (HR, 0.71; 95% CI, 0.59 to 0.85; P=0.0001). Furthermore, the subgroup analysis of the CLASSIC trial [356] revealed that the efficacy of CAPOX was maintained even in a more advanced stage, which was not observed in the ACTS-GC trial. These findings suggest that oral pyrimidine-based doublet regimens, such as CAPOX, can be a more favorable treatment option than S-1 monotherapy for pathological stage II with positive LN or stage III gastric cancer.

Recent advances have demonstrated the clinical benefits of combining immunotherapy with chemotherapy as a standard treatment option for unresectable or metastatic gastric cancer. Building on this progress, the ATTRACTION-5 study explored the addition of nivolumab, an anti-PD-1 monoclonal antibody, to adjuvant chemotherapy for resected stage III (by AICC 7th edition [365]) gastric or GEJ cancer [367]. This study showed that the addition of nivolumab to adjuvant chemotherapy did not significantly improve the primary end point of RFS compared to chemotherapy alone, with 3-year RFS rates of 68.4% in the nivolumab plus chemotherapy group and 65.3% in the chemotherapy group (HR, 0.90; 95% CI, 0.69 to 1.18; P=0.44). Similarly, there was no significant improvement in OS, with an HR of 0.88 (95% CI, 0.66 to 1.17). These results suggest that the addition of immunotherapy has not been established as an adjuvant treatment strategy in gastric cancer.

MSI-H was associated with a better prognosis in resected gastric cancer, but has not shown a survival benefit from adjuvant chemotherapy in the post hoc analysis of the CLASSIC trial [368] and a multinational meta-analysis [369] of individual patient data from the MAGIC [370], CLASSIC [356], ARTIST [371], and ITACA-S [372] trials. Meanwhile, previous retrospective study reported the conflicting result of adjuvant chemotherapy in resectable MSI-H/dMMR gastric cancer [373]. Currently, the benefit of adjuvant chemotherapy in resectable MSI-H/dMMR gastric cancer remains unclear due to conflicting data and the small proportion of this population included in these studies. Therefore, further investigation is warranted to clarify the role of adjuvant chemotherapy in resected MSI-H gastric cancer.

S24. NCT for resectable gastric cancer

Adjuvant chemotherapy following D2 gastrectomy has been the standard treatment for pathological stage II or III gastric cancer in Asia. However, survival outcomes with adjuvant chemotherapy remain suboptimal, particularly for patients with stage III disease. Additionally, adjuvant chemotherapy is often delayed following surgical resection due to surgical morbidities, and administrating chemotherapy after gastrectomy is associated with more frequent adverse events. In this context, NCT can be considered to intensify chemotherapy and initiate chemotherapy earlier when patients are more medically fit.

Three prospective RCTs in Asia have evaluated the clinical benefit of NCT as part of perioperative chemotherapy for locally AGC. The PRODIGY study in Korea investigated whether NCT with docetaxel, oxaliplatin, and S-1 (DOS) followed by surgery and adjuvant S-1 could improve outcomes compared to upfront surgery followed by adjuvant S-1 in patients with locally AGC with clinical T2/3N+ or cT4Nany disease. Neoadjuvant DOS achieved a higher rate of complete (R0) resection compared to upfront surgery (95% vs. 84%, P<0.001). It also significantly improved PFS compared to upfront surgery followed by adjuvant S-1 (adjusted HR, 0.70; 95% CI, 0.52 to 0.95; stratified log-rank P=0.023) [374]. In the long-term follow-up, neoadjuvant DOS significantly prolonged OS (adjusted HR, 0.72; 95% CI, 0.54 to 0.96; stratified log-rank P=0.027) [375]. The RESOLVE study compared perioperative SOX vs. upfront surgery followed by adjuvant CAPOX in patients with cT4aN+ or cT4bNany disease. Perioperative SOX significantly improved DFS compared with adjuvant CAPOX (HR, 0.77; 95% CI 0.61 to 0.97, P=0.027) [376]. Long-term follow-up confirmed a significant OS benefit with perioperative SOX (HR, 0.79; 95% CI, 0.62 to 1.00; P=0.049) [377]. The JCOG0501 study evaluated the efficacy of neoadjuvant S-1 plus cisplatin followed by gastrectomy and adjuvant chemotherapy vs. upfront surgery and adjuvant S-1 in patients with Borrmann type 4 or large (≥8 cm) type 3 gastric cancer [378]. However, NCT with S-1 plus cisplatin did not demonstrate a survival benefit. Our meta-analysis of these Asian trials showed that NCT provided a clinical benefit compared to upfront surgery with improved DFS (HR, 0.81; 95% CI, 0.68 to 0.97; P=0.02) (Fig. 15).

Fig. 15
Forest plot for a comparison between NAC as part of perioperative chemotherapy vs. AC). (A) Overall survival. (B) Progression free survival.
SE = standard error; IV = interval variable; CI = confidence interval; NAC = neoadjuvant chemotherapy; AC = adjuvant chemotherapy.

• Click for larger image

On the other hand, it should be noted that the standard perioperative treatment for locally AGC varices by region [379]. Indeed, the efficacy and safety of perioperative chemotherapy regimens commonly used in Western countries such as fluorouracil (5-FU) plus leucovorin, oxaliplatin, and docetaxel (FLOT), have not been investigated in Asian populations. Therefore, FLOT should not be adopted in Asia without further evidence.

Given its OS benefit, NCT as part of perioperative chemotherapy can be considered a viable therapeutic option for patients with resectable, locally AGC in Korea. Clinical decision to proceed with NCT should be made based on a careful discussion considering various factors including clinical stage as well as its potential advantages and limitations over upfront surgery (either followed by adjuvant chemotherapy or not according to the pathological stage). A multidisciplinary team approach is recommended to guide these treatment decisions.

Recently, several studies have investigated the addition of ICIs to NCT. In the phase III KEYNOTE-585 study, pembrolizumab, an anti-PD-1 monoclonal antibody, combined with chemotherapy didn’t showed event-free survival benefit, although this study showed an improvement in pathologic complete response (pCR) rate [380]. In a pre-planned analysis of the phase III MATTERHORN study, the combination of durvalumab, an anti-PD-L1 monoclonal antibody, with FLOT significantly increased pCR rate from 7% to 19% [381]. While the survival results of the MATTERHORN study are still pending, there is currently no established clinical benefit to adding ICIs to perioperative treatment regimens.

S25. Palliative first-line systemic therapy for HER2-negative disease

Nivolumab is the first successful ICI used in combination with chemotherapy as palliative first-line systemic therapy for HER2-negative locally advanced, unresectable or metastatic gastric cancer. The global phase III CheckMate-649 trial demonstrated a significant improvement in OS with the addition of nivolumab to capecitabine or 5-FU/oxaliplatin compared to chemotherapy alone across all randomized patients (median OS, 13.8 vs. 11.6 months; HR, 0.80; 99.3% CI, 0.68 to 0.94; P=0.0002), with even greater benefit observed in patients with PD-L1 CPS ≥5 (median OS, 14.4 vs. 11.1 months; HR, 0.71; 98.4% CI, 0.59 to 0.86; P<0.01) [123, 382]. PD-L1 IHC in this study was performed using the Dako PD-L1 IHC 28-8 pharmDx assay (Dako, Santa Clara, CA, USA). The ATTRACTION-4 phase III trial, conducted in Japan, Korea, and Taiwan, also evaluated palliative first-line nivolumab combined with chemotherapy in HER2-negative gastric cancer. Nivolumab plus capecitabine or S-1/oxaliplatin significantly improved PFS (median PFS, 10.5 vs. 8.3 months; HR, 0.68; 95% CI, 0.51 to 0.90; P<0.01), though no OS benefit was observed [383].

Pembrolizumab has also been investigated as a first-line palliative systemic therapy in 2 phase III trials. In the KEYNOTE-062 phase III trial, pembrolizumab plus chemotherapy (5-FU or capecitabine and cisplatin) did not demonstrate superiority over chemotherapy for OS in patients with HER2-negative and PD-L1 CPS ≥1 gastric cancer (median OS, 12.5 vs. 11.1 months; HR, 0.85; 95% CI, 0.70 to 1.03; P=0.05) [384]. PD-L1 IHC for KEYNOTE-062 trial was performed using the PD-L1 IHC 22C3 pharmDx assay (Agilent Technologies, Carpinteria, CA, USA). However, the phase III KEYNOTE-859 trial demonstrated a significant improvement in OS with pembrolizumab combined with 5-FU/cisplatin (FP) or CAPOX as palliative first-line systemic therapy. In the intention-to-treat population, which included patients regardless of their PD-L1 status, the median OS was 12.9 months for pembrolizumab plus chemotherapy compared to 11.5 months for chemotherapy alone (HR, 0.78; 95% CI, 0.70 to 0.87; P<0.01). The benefit of adding pembrolizumab was more pronounced in patients with PD-L1 CPS ≥1 (median OS, 13.0 vs. 11.4 months; HR, 0.74; 95% CI, 0.65 to 0.84; P<0.01), and CPS ≥10 (median OS, 15.7 vs. 11.8 months; HR, 0.65; 95% CI, 0.53 to 0.79; P<0.01) [385]. PD-L1 IHC was also performed using the 22C3 pharmDx assay in KEYNOTE-859.

In our meta-analysis, the combination of anti-PD-1 antibody with chemotherapy showed superior outcomes in OS (HR, 0.81; 95% CI, 0.76 to 0.86), PFS (HR, 0.78; 95% CI, 0.73 to 0.83; P<0.001), objective response rate (ORR), and disease control rate (DCR) compared to chemotherapy alone (Fig. 16). Based on the CheckMate-649 and KEYNOTE-859 trials, nivolumab combined with chemotherapy (CAPOX or 5-FU, leucovorin, and 5-FU [FOLFOX]) and pembrolizumab combined with chemotherapy (FP or CAPOX) are currently approved as palliative first-line systemic therapy in Korea, regardless of PD-L1 status. Our meta-analysis also revealed that the benefit of anti-PD-1 antibody plus chemotherapy was more pronounced in the PD-L1-positive group (Fig. 17). Therefore, nivolumab with chemotherapy is recommended for patients with a PD-L1 CPS ≥5, and pembrolizumab with chemotherapy is recommended for those with a PD-L1 CPS ≥1.

Fig. 16
Forest plots for comparisons between palliative first-line ICI+Chemo vs. Chemo in human epidermal growth factor receptor 2 negative patients. (A) Overall survival. (B) Progression-free survival. (C) Objective response rates (D) Disease control rates.
Chemo = chemotherapy; SE = standard error; IV = interval variable; M-H = Mantel-Haenszel; CI = confidence interval; ICI = immune checkpoint inhibitor.

• Click for larger image

Fig. 17
Forest plots for comparisons between palliative first-line ICI+Chemo vs. Chemo in human epidermal growth factor receptor 2 negative patients according to PD-L1 expression level. (A) Overall survival in patients with PD-L1 CPS ≥1 (B) Progression-free survival in patients with PD-L1 CPS ≥1 (C) Overall survival in patients with PD-L1 CPS ≥5. (D) Progression-free survival in patients with PD-L1 CPS ≥5. (E) Overall survival in patients with PD-L1 CPS ≥10. (F) Progression-free survival in patients with PD-L1 CPS ≥10.
Chemo = chemotherapy; ICI = immune checkpoint inhibitor; SE = standard error; IV = interval variable; CI = confidence interval; PD-L1 = programmed cell death ligand-1; CPS = combined positive score.

• Click for larger image

In metastatic gastric cancer, MSI-H/dMMR tumors are rare, occurring in less than 5% of cases, making it challenging to conduct RCTs specifically for this subgroup. Nevertheless, anti-PD1 antibodies have demonstrated remarkable clinical benefits, especially in patients with MSI-H/dMMR tumors. Subgroup analyses from the CheckMate-649, KEYNOTE-062, and KEYNOTE-859 trials have consistently showed significant advantages when anti-PD-1 antibodies were combined with chemotherapy as first-line palliative systemic therapy for MSI-H/dMMR gastric cancer [382, 384, 385, 386]. This consistent and significant survival benefit underscores the robust efficacy of anti-PD-1 antibodies for MSI-H/dMMR tumors in gastric cancer, positioning them as a highly effective treatment approach for this unique subset.

Tislelizumab, an anti-PD-1 monoclonal antibody, has also shown clinical benefit when combined with chemotherapy as a first line palliative systemic therapy. The global phase III RATIONALE-305 trial compared tislelizumab plus chemotherapy vs. chemotherapy alone. This study met its primary endpoints, showing significant improvements in OS across all randomized patients (median OS, 15.0 vs. 12.9 months; HR, 0.80; 95% CI, 0.70 to 0.92; P=0.001), and in patients with PD-L1 tumor area positivity score ≥5 (median OS, 17.2 vs. 12.6 months; HR, 0.74; 95% CI, 0.59 to 0.94) [387]. PD-L1 IHC was performed using the Ventana PDL1 (SP263) assay. Currently, tislelizumab has not yet been approved for the treatment of gastric cancer in Korea.

Zolbetuximab is a first-in-class immunoglobulin G1 monoclonal antibody targeting CLDN18.2, a tight junction protein exclusively expressed in normal gastric mucosal cells and retained in most gastric adenocarcinomas. The phase III SPOTLIGHT and GLOW trials demonstrated the efficacy of zolbetuximab plus modified FOLFOX6 (mFOLFOX6) or CAPOX as first-line treatment in patients with HER2-negative and CLDN18.2-positive AGC [125, 126]. CLDN18.2-positive was defined as ≥75% of tumor cells showing moderate-to-strong membranous CLDN18 staining, as determined by central IHC using the investigational VENTANA CLDN18 [43-14A] RxDx Assay (Roche Diagnostic Solutions, Tucson, AZ, USA) [125, 126].

The SPOTLIGHT phase III trial showed that the addition of zolbetuximab to mFOLFOX6 significantly improved PFS (median, 11.0 vs. 8.9 months; HR, 0.73; 95% CI, 0.59 to 0.91; P<0.01) and OS (median, 18.2 vs. 15.6 months; HR, 0.78; 95% CI, 0.64 to 0.95; P<0.01), Recent longer-term follow-up data confirm the persistence of these clinical benefits [125, 388]. The GLOW phase III trial showed that the addition of zolbetuximab to CAPOX significantly improved PFS (median, 8.3 vs. 6.8 months; HR, 0.68; 95% CI, 0.55 to 0.85; P<0.01) and OS (median, 14.3 vs. 12.2 months; HR, 0.77; 95% CI, 0.62 to 0.95; P<0.01), with persistent results upon longer follow-up [126, 389]. The most common treatment-emergent adverse events (TEAEs) with zolbetuximab plus mFOLFOX6 or CAPOX were nausea, vomiting, and decreased appetite, and incidence of serious TEAEs was similar between treatment arms in both trials [388, 389]. In our meta-analysis, zolbetuximab combined with chemotherapy showed the clinical benefit of OS and PFS (HR, 0.85; 95% CI, 0.75 to 0.97; P=0.01) in patients with CLDN18.2-positive AGC (Fig. 18).

Fig. 18
Forest plots for comparisons between palliative first-line ZOL+Chemo vs. Chemo in human epidermal growth factor receptor 2 negative patients. (A) Overall survival. (B) Progression-free survival. (C) Objective response rates. (D) Disease control rates.
ZOL = zolbetuximab; Chemo = chemotherapy; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

Therefore, zolbetuximab combined with mFOLFOX6 or CAPOX is recommended as first-line treatment in patients with HER2-negative and CLDN18.2-positive locally advanced unresectable or metastatic gastric cancer.

S26. Palliative first-line systemic therapy for HER2-positive disease

The Trastuzumab for Gastric Cancer (ToGA) phase III trial demonstrated the efficacy of trastuzumab, a monoclonal antibody targeting HER2, plus fluoropyrimidine/cisplatin as a first-line treatment in patients with HER2-positive gastric cancer [110]. The addition of trastuzumab to capecitabine or FP improved OS (median, 13.8 vs. 11.1 months; HR, 0.74; 95% CI, 0.60 to 0.91; P=0.0046). The survival benefit was more pronounced in patients with IHC 3+ or IHC 2+ and fluorescence ISH + (median OS, 16.0 vs. 11.8 months; HR, 0.65; 95% CI, 0.51 to 0.83; P<0.01). PFS was also significantly extended in the trastuzumab plus chemotherapy group (median, 6.7 vs. 5.5 months; HR, 0.71; 95% CI, 0.59 to 0.85; P=0.036). The ORR was higher with trastuzumab plus chemotherapy compared to chemotherapy alone (ORR, 47% vs. 35%; OR, 1.70; 95% CI, 1.22 to 2.38; P=0.002). Similarly, the DCR was superior in the trastuzumab plus chemotherapy group (DCR, 75% vs. 70%; OR, 1.66; 95% CI, 1.14 to 2.41; P=0.008) (Fig. 19).

Fig. 19
Forest plots for comparisons between palliative first-line T+Chemo vs. Chemo in ToGA trial. (A) Overall survival in all patients. (B) Overall survival in in patients with HER2 2+/FISH+ or IHC 3+ and IHC 0 or 1+/FISH+. (C) Progression-free survival in all patients. (D) Objective response rates in all patients (E) Disease control rates in all patients.
T = trastuzumab; Chemo = chemotherapy; SE = standard error; IV = interval variable; CI = confidence interval; HER2 = human epidermal growth factor receptor 2; IHC = immunohistochemistry; FISH = fluorescence in situ hybridization.

• Click for larger image

For dual blockade of HER2, the JACOB phase III trial evaluated the efficacy of pertuzumab, a monoclonal antibody that interferes with HER2 heterodimerization with other EGFR family members, combined with trastuzumab and chemotherapy as a first-line therapy, compared to trastuzumab and chemotherapy [390]. Although PFS was improved with the addition of pertuzumab to trastuzumab plus fluoropyrimidine/cisplatin (median, 8.5 vs. 7.2 months; HR, 0.73; 95% CI, 0.62 to 0.85; P<0.01), there was no statistically significant improvement in OS for the primary endpoint in patients with HER2-positive gastric cancer (median, 17.5 vs. 14.2 months; HR, 0.84; 95% CI, 0.71 to 1.00, P=0.057). Lapatinib, a small-molecule tyrosine kinase inhibitor of EGFR and HER2, was evaluated in the LOGiC phase III trial, where lapatinib and CAPOX as first-line therapy did not significantly improve OS (median, 12.2 vs. 10.5 months; HR, 0.91; 95% CI, 0.73 to 1.12; P=0.349) compared to CAPOX in HER2-amplified gastric cancer. Therefore, trastuzumab plus capecitabine or FP is recommended in patients with HER2-positive locally advanced unresectable or metastatic gastric cancer [391].

Integrating ICI into the standard chemotherapy regimen for first-line treatment of HER2-negative gastric cancer has improved survival outcomes, and these results have been extended to HER2-positive gastric cancer as well. The randomized, phase III KEYNOTE-811 trial showed the efficacy of combined pembrolizumab and trastuzumab with standard chemotherapy (fluoropyrimidine/platinum-based therapy) vs. trastuzumab with standard chemotherapy in first-line treatment in patients with HER2-postivie gastric cancer [392]. At the third interim analysis, PFS was improved in the pembrolizumab group compared to the placebo group in the intention-to-treat population (median 10.0 vs. 8.1 months; HR, 0.73; 95% CI, 0.61 to 0.87, P=0.0005), and OS was also significantly improved in pembrolizumab group (median, 20.0 vs. 16.8 months; HR, 0.84; 95% CI, 0.70 to 1.01; P=0.06). In patients with PD-L1 CPS ≥1, the pembrolizumab group showed a more pronounced improvement in PFS (median, 10.9 vs. 7.3 months; HR, 0.72; 95% CI, 0.60 to 0.87; P<0.01), and OS (median, 20.1 vs. 15.7 months; HR, 0.79; 95% CI, 0.66 to 0.95; P=0.006). The ORR was also higher in the pembrolizumab group (ORR, 72.6% vs. 60.1%; OR, 1.76; 95% CI, 1.28 to 2.42; P<0.001), with a better DCR (91.7% vs. 87.4%; OR, 1.60; 95% CI, 0.98 to 2.63; P=0.06) (Fig. 20). Based on these results, the addition of pembrolizumab to trastuzumab and chemotherapy is recommended for patients with HER2-positive and PD-L1 CPS ≥1.

Fig. 20
Forest plots for comparisons between palliative first-line P+T+Chemo vs. T+Chemo in human epidermal growth factor receptor 2 positive patients. (A) Overall survival in all patients. (B) Overall survival in patients with PD-L1 CPS ≥1 and <1. (C) Progression-free survival in all patients. (D) Progression-free survival in patients with PD-L1 CPS ≥1 and <1. (E) Objective response rates in all patients. (F) Disease control rates in all patients.
P = pembrolizumab; T = trastuzumab; Chemo = chemotherapy; SE = standard error; IV = interval variable; CI = confidence interval; PD-L1 = programmed cell death ligand-1; CPS = combined positive score.

• Click for larger image

S27. Palliative second-line systemic therapy

RCTs and previous meta-analyses have demonstrated the survival benefit of second-line palliative chemotherapy (with irinotecan or taxanes) compared to best supportive care alone for patients with locally advanced unresectable or metastatic gastric cancer [393, 394, 395, 396]. In the present meta-analysis, second-line systemic therapy significantly improved OS compared to best supportive care (HR, 0.69; 95% CI, 0.59 to 0.82; P<0.001) (Fig. 21). Weekly paclitaxel was associated with similar survival outcomes to biweekly irinotecan in previous phase III trials [397, 398]. Meanwhile, ramucirumab monotherapy, a monoclonal antibody targeting vascular endothelial growth factor receptor-2, significantly improved OS and PFS compared to placebo in the REGARD trial [399]. Furthermore, the addition of ramucirumab to weekly paclitaxel significantly prolonged OS (median, 9.6 vs. 7.4 months; HR, 0.807; 95% CI, 0.68 to 0.97; P=0.017) and PFS (median, 4.4 vs. 2.9 months; HR, 0.635; 95% CI, 0.53 to 0.76; P<0.0001) compared to paclitaxel plus placebo in the RAINBOW trial (Fig. 22) [400].

Fig. 21
Forest plots for comparisons between palliative second-line systemic Tx vs. BSC or placebo. (A) Overall survival. (B) Progression-free survival. (C) Objective response rates. (D) Disease control rates.
Tx = therapy; BSC = best supportive care; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

Fig. 22
Forest plots for comparisons between palliative second-line Ram+PTX vs. PTX. (A) Overall survival. (B) Progression-free survival. (C) Objective response rates. (D) Disease control rates.
Ram = ramucirumab; PTX = paclitaxel; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

Based on previous trials, ramucirumab in combination with paclitaxel is recommended as the preferred second-line treatment. Other agents, including irinotecan, docetaxel, paclitaxel can also be considered second-line options if not previously administered in the first-line treatment. In addition, various investigational agents are currently being tested in combination with paclitaxel or paclitaxel plus ramucirumab as second-line treatment [401, 402, 403].

Pembrolizumab failed to provide a significant survival benefit compared to paclitaxel [404, 405, 406], however, it was effective in patients with solid tumors characterized as MSI-H, dMMR, or TMB-high (≥10 mutations/megabase) [407, 408]. In Korea, pembrolizumab was approved in patients with several inoperable or metastatic solid tumors, including gastric cancer with MSI-H or dMMR, who have progressed following prior treatment and who have no satisfactory alternative treatment options.

Trastuzumab deruxtecan, a HER2-directed antibody and topoisomerase inhibitor conjugate, was approved by the Food and Drug Administration for the treatment of patients with HER2-positive gastric cancer who have received a prior trastuzumab-based regimen. The phase II trial of trastuzumab deruxtecan as second-line treatment provided clinical evidence in the Western population only [409]. Currently, a global phase III trial (DESTINY-Gastric04), which includes Asian patients, is ongoing to compare the survival outcomes of trastuzumab deruxtecan to ramucirumab plus paclitaxel, the current standard second-line treatment, in patients with HER2-positive gastric cancer.

S28. Palliative third-line systemic therapy

For patients with preserved performance status who have disease progression after second-line systemic therapy, palliative third-line systemic therapy is recommended. In our present meta-analysis, third-line systemic therapy significantly improved OS compared to best supportive care (HR, 0.68; 95% CI, 0.61 to 0.76, P<0.001) (Fig. 23).

Fig. 23
Forest plots for comparisons between palliative third-line systemic Tx vs. BSC or placebo. (A) Overall survival. (B) Progression-free survival. (C) Objective response rate. (D) Disease control rate.
Tx = therapy; BSC = best supportive care; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

Cytotoxic agents, such as docetaxel or irinotecan, can be recommended as palliative third-line therapy. A randomized phase III trial showed a survival benefit with docetaxel or irinotecan (median OS, 5.3 vs. 3.8 months; HR, 0.66; 95% CI, 0.49 to 0.89; P=0.007) [393]. Several phase II and retrospective studies investigating taxane- or irinotecan-based chemotherapy as third-line treatments also have shown consistent results [410, 411, 412]. A randomized phase III trial of trifluridine/tipiracil significantly improved OS compared to placebo (median OS, 5.7 vs. 3.6 months; HR, 0.69; 95% CI, 0.56 to 0.85; P=0.00058) in gastric cancer patients who had received at least 2 previous systemic treatments [413].

Nivolumab also showed a survival benefit over placebo in heavily pre-treated patients who had received 2 or more previous systemic therapies in the randomized phase III ATTRACTION-2 trial (median OS, 5.3 vs. 4.1 months; HR, 0.63; 95% CI, 0.51 to 0.78; P<0.0001) [414]. Two-year updated data from ATTRACTION-2 confirmed the long-term survival benefit of nivolumab regardless of PD-L1 expression status [415]. Based on these results, nivolumab is recommended as a palliative third-line therapy for gastric cancer patients who are naïve to ICI therapy.

In the phase II DESTINY-Gastric01 trial conducted in Japan and South Korea, trastuzumab deruxtecan was compared to the physician’s choice of irinotecan or paclitaxel in patients with HER2-positive gastric cancer who had received at least 2 prior palliative systemic treatments including trastuzumab [416]. In this trial, trastuzumab deruxtecan was associated with significant improvements in the ORR (51% vs. 14%, P<0.001) and OS (median OS, 12.5 vs. 8.4 months; HR, 0.59; 95% CI, 0.39 to 0.88; P=0.01) compared to the physician’s choice of irinotecan or paclitaxel. The positive results from the DESTINY-Gastric01 trial led to the approval of trastuzumab deruxtecan in patients with HER2-positive gastric cancer who have received a prior trastuzumab-based regimen as a second-line or later treatment in the US and a third-line or later treatment in Korea.

In the double-blinded, placebo-controlled phase III trial (INTEGRATE IIa), regorafenib plus best supportive care was compared to a placebo plus best supportive care in a 2:1 randomization among patients who had failed or were intolerant to at least 2 or more prior lines of therapy including a platinum agent plus fluoropyrimidine [417]. In this trial, regorafenib was associated with significant improvements in the OS (median, 4.5 vs. 4.0 months; HR, 0.70; 95% CI, 0.53 to 0.92; P=0.011) and PFS (median, 1.8 vs. 1.6 months; HR, 0.52; 95% CI, 0.40 to 0.69; P<0.0001) compared to the placebo. The ongoing international randomized phase III INTEGRATE IIb trial is evaluating regorafenib combined with nivolumab compared to standard chemotherapy in pre-treated patients with AGC (NCT0487936).

S29. Adjuvant chemoradiation

A total of 6 RCTs were included in the present meta-analysis, including 2 recent RCTs published after the Korean Practice Guideline for Gastric Cancer 2018: An Evidence-based, Multidisciplinary Approach [366, 371, 418, 419, 420, 421]. The target volume of radiation therapy was generally similar in these trials, including the tumor bed, anastomotic site and/or stump, and regional LN stations.

In the meta-analysis, the addition of adjuvant chemoradiation therapy (CRT) reduced locoregional recurrence compared to chemotherapy alone (HR, 0.62; 95% CI, 0.48 to 0.81; P=0.0004) with no significant difference in grade 3 or higher toxicities (HR, 0.85; 95% CI, 0.63 to 1.13; P=0.26). Adjuvant CRT showed superior outcomes compared to adjuvant chemotherapy alone in terms of DFS (HR, 0.85; 95% CI, 0.713 to 0.98; P=0.03). However, when compared to platinum-based combination chemotherapy, there was no benefit in terms of DFS (HR, 0.91; 95% CI, 0.78 to 1.07; P=0.25) and OS (HR, 1.03; 95% CI, 0.87 to 1.23; P=0.70) (Fig. 24).

Fig. 24
Forest plots for comparison between adjuvant concurrent CRT vs. adjuvant platinum-based combination CA. (A) Overall survival. (B) Disease-free survival. (C) Locoregional recurrence.
CRT = chemoradiation therapy; CA = chemotherapy alone; SE = standard error; IV = interval variable; CI = confidence interval; M-H = Mantel-Haenszel.

• Click for larger image

Based on these studies, the addition of adjuvant CRT is not recommended in gastric cancer patients after complete resection with D2 lymphadenectomy. Further prospective trials should focus on identifying potential candidates who might benefit from adjuvant CRT.

S30. Neoadjuvant chemoradiation (NCRT)

NCRT is mainly studied for cancer of the esophagus, GEJ, and/or gastric cardia, where achieving a complete R0 resection is challenging and the locoregional relapse is high. The MAGIC trial showed that perioperative chemotherapy significantly improved OS over surgery alone for distal esophageal and gastric cardia adenocarcinoma [370]. Studies have focused on assessing whether adding radiation therapy to NCT would provide benefits in gastric cancer. Two RCTs have been conducted, and one RCT is ongoing to compare the outcomes of NCRT vs. NCT alone in resectable cancer of the GEJ or stomach [422, 423, 424, 425, 426].

Final treatment outcomes were reported in the POET and NeoRes trials [421, 424]. In our meta-analysis (Fig. 25), the pCR rate (23.6% in NCRT vs. 6.3% in NCT) and pathologic N0 rate (69.9% in NCRT vs. 45.7% in NCT) were significantly higher in the NCRT group. Local PFS was reported only in the POET trial, showing a significant improvement with NCRT in long-term follow-up (HR, 0.37; 95% CI, 0.16 to 0.85). However, these improved pathologic responses did not lead to a significant OS benefit (HR, 0.85; 95% CI, 0.63 to 1.15). In addition, there were no significant differences in the R0 resection rate (74% in NCRT vs. 66% in NCT). PFS was reported only in the NeoRes trial, with no significant difference in the 3-year PFS rate (44% in NCRT vs. 44% in NCT). An interim analysis of the TOPGEAR study showed no significant difference in adverse events or surgical complications [423]. In the pooled analysis of the TOPGEAR and NeoRes studies, there was no significant difference in severe gastrointestinal toxicity (15.3% in NCRT vs. 13.2% in NCT).

Fig. 25
Forest plots for comparisons between NCRT compared to NCT. (A) Overall survival. (B) Pathologic complete response. (C) Pathologic complete nodal regression. (D) R0 resection.
NCRT = neoadjuvant chemoradiation; NCT = neoadjuvant chemotherapy; SE = standard error; IV = interval variable; CI = confidence interval; M-H = Mantel-Haenszel.

• Click for larger image

It is noteworthy that these studies were conducted mainly in patients with esophageal and/or GEJ cancer, which is more common in Western countries. Most studies evaluating the efficacy of NCRT for gastric cancer (mainly GEJ cancer) have also been conducted in Western populations. In addition, in the NeoRes trial, some patients with esophageal squamous cell carcinoma were included. Evidence remains limited, and further prospective studies including Asian populations and nonjunction cancers are needed to establish more robust evidence.

S31. Endoscopic stenting (ES) vs. surgical gastrojejunostomy (GJ) for malignant gastric outlet obstruction (GOO)

In patients with advanced or metastatic gastric cancer, GOO often manifests with symptoms such as nausea, vomiting, dehydration, and malnutrition, significantly deteriorating patients’ QOL. Given that radical surgery is not indicated in patients with incurable gastric cancer, palliative treatments are required to relieve symptoms of GOO and restore the ability to consume an oral diet.

Surgical GJ and ES are palliative treatments for GOO caused by unresectable gastric cancer. GJ has been the standard for effective symptom relief in GOO; however, it carries a risk of substantial early major complications and procedure-related mortality [427, 428]. ES, with its shorter procedure time, quicker resumption of oral intake, and shorter duration of hospital stay than GJ, offers a less invasive alternative [427, 428]. However, ES is associated with higher rate of complications, reinterventions, and recurrent obstructions compared to GJ [427, 428].

In this clinical guideline, we compared the outcomes of ES with GJ through a meta-analysis of 1,637 articles [187, 429, 430, 431, 432, 433, 434, 435, 436, 437, 438, 439, 440, 441], ultimately including 15 studies (12 observational and 3 RCTs), with 5 studies conducted in Korea [187, 434, 437, 438, 439]. Of the 1,286 patients, 818 patients received ES, and 468 patients underwent GJ. The overall certainty for outcomes from RCTs was rated as low, and very low for observational studies due to a limited number of events and bias from confounding and selection of participants. Regarding procedure outcomes, there was no significant difference between the ES and GJ groups in terms of technical success (OR, 1.33; 95% CI, 0.37 to 4.73; P=0.66) and clinical success (OR, 0.68; 95% CI, 0.41 to 1.13; P=0.14) (Fig. 26A and B). There was no significant difference in procedure-related mortality (OR, 0.64; 95% CI, 0.26 to 1.63; P=0.35) (Fig. 26C). For postoperative outcomes, patients in the ES group had a faster resumption of oral intake (mean duration, −3.94 days; 95% CI, −4.01 to −3.88 days; P<0.001) and a shorter duration of hospital stay (mean duration, −6.56 days; 95% CI, −7.20 to −5.92 days; P<0.001) (Fig. 27A and B). The rate of minor complications was not significantly different between the 2 groups (OR, 0.52; 95% CI, 0.25 to 1.10; P=0.09) (Fig. 27C). However, major complications (OR, 1.81; 95% CI, 1.10 to 2.96; P=0.02) and reintervention rates (OR, 3.83; 95% CI, 2.40 to 6.12; P<0.001) were significantly higher in the ES group than in the GJ group (Fig. 27D and E). Additionally, ES was significantly associated with a shorter patency duration (mean duration, −4.97 months; 95% CI, −6.42 to −3.51 months; P<0.001) (Fig. 27F). However, OS was not significantly different between the ES and GJ groups (mean duration, 0.12 months; 95% CI, −0.48 to 0.72 months, P=0.69) (Fig. 27G).

Fig. 26
Forest plots of procedure outcomes between stent insertion (Stenting) and surgery (Surgery). (A) Technical success. (B) Clinical success. (C) Procedure related mortality.
CI = confidence interval; M-H = Mantel-Haenszel.

• Click for larger image

Fig. 27
Forest plots of post procedure or operative outcomes between stent insertion (Stenting) and surgery (Surgery). (A) Resumption of oral intake. (B) Duration of hospital stay. (C) Minor complications. (D) Major complications. (E) Re-intervention. (F) Patency duration. (G) Overall survival.
SD = standard deviation; IV = interval variable; CI = confidence interval; M-H = Mantel-Haenszel.

• Click for larger image

Both GJ and ES are effective palliative treatments for GOO caused by unresectable gastric cancer. Our results suggest that ES may be associated with more favorable for patients who are poor surgical candidates with relatively short life expectancy and who prioritize resuming an oral diet and early discharge, while GJ is preferable in patients with a longer prognosis and good performance status. Stent insertion by radiologic intervention showed similar efficacy to endoscopic stent insertion in a single-center study [439], and radiologic stent insertion can also be considered for GOO when available.

S32. Surgical resection for metastatic gastric cancer

The REGATTA trial, the only phase III RCT comparing gastrectomy with D1 dissection followed by chemotherapy vs. chemotherapy alone, focused on pure reduction surgery without metastasectomy [439]. Gastric cancer patients with a single non-curable factor were enrolled. Reduction surgery showed no survival benefit compared to chemotherapy alone, and the trial was terminated after the first interim analysis due to the lack of benefit in the surgery group (HR, 1.08; 95% CI, 0.74 to 1.58; P=0.66). Based on these findings, it was concluded that reductive gastrectomy is not justified in patients with metastatic gastric cancer.

Our meta-analysis included 3 retrospective studies and one prospective study. The data on OS showed better survival outcomes in stage IV gastric cancer patients who underwent systemic chemotherapy followed by radical surgery compared to those who received chemotherapy alone (HR, 0.58; 95% CI, 0.42 to 0.80; P<0.001) (Fig. 28).

Fig. 28
Forest plot for overall survival between CS vs. CA in metastatic gastric cancer.
CS = conversion surgery after chemotherapy; CA = chemotherapy alone; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

A large retrospective review showed that stage IV cancer patients who were responsive to chemotherapy and underwent R0 resection following chemotherapy had better survival than those in the R1 and R2 resection groups [442]. In a prospective nonrandomized trial, surgery after chemotherapy, particularly in R0 resection, was associated with a survival benefit for gastric cancer patients with limited distant metastasis [443]. However, as most studies included are retrospective and potentially subject to selection bias, the evidence is limited for making stronger recommendations.

While the meta-analysis suggests possible benefits, the role of conversion surgery and detailed indications are inconclusive due to inherent selection biases in observational studies comparing surgery and systemic therapy for stage IV gastric cancer. Additionally, it remains uncertain whether advancements in systemic therapies will enhance or reduce the relevance of conversion surgery.

The recently reported results of the RENAISSANCE prospective RCT (AIO-FLOT5) [444] showed no significant difference in OS between the surgical resection after systemic therapy group and the systemic therapy-only group (median OS, 18.5 vs. 23.6 months; HR, 1.037; 95% CI, 0.691 to 1.556; P=0.861), thus not meeting the primary endpoint. The limitation of the study was considered due to non-favorable surgical outcomes; morbidity up to 60%, 20% reoperation rate within 90 days, and postoperative mortality of approximately 8% in the surgery group. This high morbidity rate not only impacted survival but also prevented many patients from completing postoperative systemic therapy, unlike those who remained on continuous FLOT. Subgroup analysis indicated a need for strategies to address early mortality in the surgery group, with a significant survival benefit observed in the retroperitoneal LN group but significantly worse survival in the peritoneal disease subgroup. The implementation of safe local treatment was important in the continuation of systemic therapy.

At the KINGCA 2024 conference, an expert consensus meeting on conversion surgery was held. A panel of 17 experts, consisting of 10 surgeons and 7 medical oncologists from both domestic and international institutions, participated in discussions focused on 9 key topics. The results of this consensus meeting are described in Table 5 and the details will be reported in a separate report [445].

Table 5
Consensus rate from the expert consensus meeting on conversion therapy at KINGCA 2024

• Click for larger image
• Click for full table
• Download as Excel file

S33. Surgical resection plus systemic therapy for gastric cancer patients with single-organ oligometastasis

Traditionally, oligometastasis has been defined as an intermediate state between localized and widespread systemic disease with the presence of fewer than 5 metastases [446]. However, the definition remains unclear, and oligometastasis generally defined as fewer than 3 or 5 metastatic lesions involving one or 2 organs [447, 448].

For hepatic oligometastasis in gastric cancer, 2 retrospective studies were included in our meta-analysis (Fig. 29) [449, 450]. The meta-analysis showed that radical gastrectomy with hepatectomy plus chemotherapy provided a survival benefit compared to chemotherapy alone (HR, 0.27; 95% CI, 0.12 to 0.62; P=0.002). Local treatments, such as transarterial chemoembolization (TACE), radiofrequency ablation, and hepatic arterial infusion, have also been reported to provide potential survival benefits [451, 452, 453, 454]. However, most of the related studies were single-arm or compared outcomes with hepatectomy, so they could not be included in the meta-analysis. Liu et al. [454] retrospectively compared radical gastrectomy with TACE plus chemotherapy with chemotherapy alone, including all types of liver metastasis (H1, H2, H3) in the study. The median OS was 14 months in the surgery group and 8 months in the chemotherapy groups (P<0.001) [453].

Fig. 29
Forest plot for comparison of overall survival between surgery, mastectomy and gastrectomy with chemotherapy (Surgery) vs. CA in gastric cancer with oligometastasis confined to liver from observational studies.
CA = chemotherapy alone; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

There is some promising evidence that resection of liver oligometastasis may provide a survival benefit. However, due to the small sample sizes and retrospective nature of the studies, the evidence is weak, so candidates for liver resection should be selected cautiously. Further evidence is needed for generalization.

For ovarian metastasis, 3 retrospective studies were included in the meta-analysis, showing better survival in the metastasectomy group (HR, 0.45; 95% CI, 0.34 to 0.59, P<0.001) [455, 456, 457] (Fig. 30). Cheong et al. [458] reported that Krukenberg tumors were often accompanied by peritoneal dissemination, which was associated with a significantly worse prognosis (HR, 1.74; 95% CI, 1.28 to 2.36; P<0.001), and only when curative resection was achieved was the median OS time longer in the resection group than in the non-resection group (17 vs. 3 months, P<0.001).

Fig. 30
Forest plot for comparison of overall survival between surgery, oophorectomy with chemotherapy (Surgery) vs. CA in gastric cancer with oligometastasis confined to liver from observational studies.
CA = chemotherapy alone; SE = standard error; IV = interval variable; CI = confidence interval.

• Click for larger image

For para-aortic LNs, only 3 prospective nonrandomized studies evaluated the response rate of preoperative chemotherapy and the efficacy of subsequent D2 LN dissection plus para-aortic LN dissection; they did not show favorable survival outcomes [459, 460, 461].

S34. Intraperitoneal (IP) chemotherapy for patients with peritoneal carcinomatosis

Peritoneal metastasis is known to be less responsive to systemic therapy and is associated with a worse prognosis than hematogenous or lymphatic metastasis [462, 463]. One reason is the limited delivery of anticancer drugs to the peritoneum due to the peritoneum-plasma barrier. To directly target cancer cells on the peritoneum, IP infusion using anticancer drugs with lower systemic absorption and toxicity, such as paclitaxel and docetaxel, has been studied in various cancers. It is considered effective and safe in treating peritoneal cancer dissemination from ovarian cancer [464, 465, 466, 467, 468, 469, 470].

In gastric cancer patients with peritoneal metastasis, phase I and II studies showed improved survival for patients in the IP (paclitaxel + docetaxel) plus systemic chemotherapy group compared the systemic chemotherapy alone group, with median OS of 24.6 vs. 15.1 months and 1-year survival rates of 78% vs. 70.4% [463].

However, in a phase III trial conducted by Ishigami et al. [471], IP plus systemic chemotherapy did not show a significant improvement in survival outcome compared to systemic chemotherapy alone (HR, 0.72; 95% CI, 0.49 to 1.04; stratified log-rank P=0.080). The authors suggested that patient withdrawal and protocol violations may have led to an underestimation of the true effect of IP therapy.

Currently, phase I, II, and III studies are ongoing in Korea. Further investigation is required for recommendation, and until trial outcome data become available, IP chemotherapy should be applied only for investigational purposes.