Guideline development

This CPG was developed in accordance with the World Health Organization’s recommended process and with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) for assessing evidence quality, and using the Evidence to Decision framework to formulate clinical recommendations. This process minimized bias and increased the transparency of the process. Quality of evidence indicates the degree of certainty of the findings. GRADE categorizes the quality of evidence into high, moderate, low, and very low by assessing various aspects of the body of evidence, namely risk of bias, inconsistency, indirectness, imprecision, and publication bias. The strengths of the recommendations in this CPG are categorized as strong, weak, and conditional. The factors that promoted strong recommendation were high certainty of evidence, similarity in stakeholders’ values and preferences, cost effectiveness, and sharp contrast between benefit and harm.^[5]

The GDG identified important clinical questions through discussion, which were later converted into research questions using the problem/patient/population, intervention/exposure, comparisons, outcomes (PICO) format to pave the way for systematic reviews. The GDG held several meetings between 2022 and 2023 to review the evidence under each PICO question, and to reach a consensus on the corresponding recommendations. Consensus was reached in each case through open discussion and voting, where 70% defined the threshold to pass a recommendation.

The full CPG report was sent for review to external guideline methodologists and clinicians without direct involvement in the current CPG development. The feedback was collected and incorporated, as appropriate. We referenced Appraisal of Guidelines for Research and Evaluation II (AGREE II) before and during the creation of this CPG to ensure quality, and we followed the Reporting Items for practice Guidelines in HealThcare (RIGHT) statement for reporting.^[6,7]

Part 2. Screening

Clinical question 2.1: Does endoscopic screening, which is based on gastric cancer risk stratification, achieve health economic benefits?

Recommendation 2.1: We suggest endoscopic screening in accordance with Li et al’s prediction model to obtain health economic benefits (weak recommendation; very low certainty of evidence).

Endoscopy combined with histopathological biopsy is the most important and reliable method for gastric cancer screening. A population-based multicenter cohort study from China showed that endoscopic screening significantly reduced gastric cancer mortality in high-risk areas.^[30] However, large-scale screening in China is difficult to perform because of the excessive work required of doctors. Currently, a number of gastric cancer risk stratification methods are used as the initial prescreening tool before endoscopy, and one of the most effective methods was proposed by Li et al.^[31,32] However, few studies have investigated the health economic benefits of endoscopic screening in China. A Japanese study showed that the “ABC method” combined with endoscopic screening was cost-effective.^[33]

Clinical question 2.2: What is the best risk stratification method for gastric cancer screening?

Recommendation 2.2: Patients over 40 years of age with H. pylori infection and other risk factors for gastric cancer (male, smoking, pernicious anemia, immediate relatives with gastric cancer, severe atrophic gastritis, intestinal metaplasia, and adenoma) are recommended for gastric cancer screening. H. pylori detection with pepsinogen (PG) I measurement and the PG I/II ratio are recommended as gastric cancer screening indicators (strong recommendation, low certainty of evidence).

H. pylori infection is one of the most common chronic infections in humans. The numbers of new cases and deaths from gastric cancer in China annually are approximately half of all patients with gastric cancer worldwide. Approximately 90% of noncardiac cancers are attributed to H. pylori infection.^[34]H. pylori was classified as a Class I (definite) carcinogen in human gastric cancer by the International Agency for Research on Cancer of the World Health Organization in 1994.^[35] PG is a precursor of pepsin, a gastric mucosa-specific functional enzyme, with PG I and PG II subgroups. The serum PG concentration can reflect the morphology and function of the gastric mucosa in different parts of the stomach, and combined PG I measurement with the PG I/II ratio is useful as a “serological biopsy” of the gastric mucosa.^[36] The combined detection of PG and H. pylori is of great importance in the early detection and treatment of gastric cancer and is worthy of clinical application. The Consensus Opinions on Early Gastric Cancer Screening and Endoscopic Diagnosis and Treatment in China (April 2014, Changsha)^[37] suggested that the risk of gastric cancer could be effectively stratified based on the results of serum PG and H. pylori antibody measurement, and further examination strategies could be determined.

Clinical question 2.3: Gastroscopy is the most effective method for gastric cancer screening. Are other noninvasive techniques, including magnetically controlled capsule gastroscopy, also effective methods for gastric cancer screening?

Recommendation 2.3: Magnetically controlled capsule gastroscopy, as a beneficial supplement and the best alternative to conventional gastroscopy, can be used for gastric cancer screening. For certain populations (i.e., people who decline or cannot tolerate conventional gastroscopy, or those who have a high risk of complications during gastroscopy), magnetically controlled capsule gastroscopy is recommended for gastric cancer screening after exclusion of any contraindications (strong recommendation, moderate quality of evidence).

With conventional gastroscopy and histopathological examination as the gold standards, a number of studies have shown that magnetically controlled capsule gastroscopy has high sensitivity, specificity, and overall accuracy in the diagnosis of gastric diseases, such as gastric cancer, and is better tolerated by people. Therefore, magnetically controlled capsule gastroscopy can be used as a beneficial supplement and best alternative to conventional gastroscopy for gastric cancer screening.^{[38,39,40,41]} Magnetically controlled capsule endoscopy has been recommended as a screening modality for gastric cancer in asymptomatic populations by the China Experts Consensus on the Protocol of Early Gastric Cancer Screening (2017, Shanghai) and the Chinese Guideline on Magnetically Controlled Capsule Gastroscopy (2021, Shanghai).^[42,43] Other noninvasive detection methods for gastric cancer screening, namely serum tumor marker testing, X-rays, barium swallow radiography, and spiral CT, have limited value in early gastric cancer screening and are not recommended.^[44]

Part 3. Diagnosis

Clinical question 3.1: Should mucolytic agents and defoaming agents be used before gastroscopy to improve the visibility of the gastric mucosa and increase the detection rate of early gastric cancer?

Recommendation 3.1: It is recommended to use mucolytic agents (such as pronase or N-acetylcysteine) to dissolve and remove gastric mucus before gastroscopy, and to use defoaming agents (such as simethicone or dimethicone) to reduce gastric foam. These methods can improve the visibility of the mucosa and may increase the detection rate of early gastric cancer (strong recommendation, moderate certainty of evidence).

A clear visual field is the prerequisite for high-quality endoscopic examination.^[45] The presence of foam and mucus can prolong examination time and may affect the detection of gastric mucosal lesions. Administering defoaming and mucolytic agents before endoscopy can remove bubbles and mucus in the stomach, improve gastric visibility scores, and decrease the need for water lavage, with no increase in endoscopic examination time.^[46] A clear image is conducive to the detection of minimal gastric mucosal lesions, including early gastric cancer.^[47,48]

There are conflicting results regarding whether premedication with defoaming and mucolytic agents before gastroscopy increases the detection rate of early gastric cancer. One study showed that defoaming and mucolytic agents increase the detection rate of precancerous lesions and early gastric cancer (36.4% in the test group vs. 26.8% in the control group; P = 0.000).^[49] Two other multicenter studies showed that premedication with pronase alone or combined with simethicone may not increase the detection rate of early gastric cancer.^[50,51]

Clinical question 3.2: In patients with severe gastric peristalsis, which interferes with clear observation, can antispasmolytic drugs that inhibit gastric peristalsis (such as butylscopolamine, glucagon, menthol oil, and L-menthol) be considered? Can gastric peristalsis inhibitors improve the detection rate of early gastric cancer?

Recommendation 3.2: For patients with severe gastric peristalsis, which interferes with clear observation, L-menthol spray on the gastric mucosa is suggested to inhibit gastric peristalsis (weak recommendation, moderate certainty of evidence).

L-menthol is a nontoxic substance that relaxes smooth muscle in the gastrointestinal tract by blocking calcium ions from entering the smooth muscle cells. Multiple meta-analyses and clinical trials have demonstrated that L-menthol sprayed directly on the gastrointestinal mucosa significantly inhibits gastrointestinal peristalsis with minimal adverse effects.^[52,53,54] Intravenous injections of butylscopolamine or glucagon have also been used to inhibit gastrointestinal peristalsis. However, this treatment lacks meta-analysis support, and there are safety concerns.^[55,56] Studies have shown that butylscopolamine may have adverse effects on the ophthalmic, urinary, and cardiovascular systems, and may lead to severe anaphylactic shock.^[57,58,59] Additionally, glucagon may cause nausea, vomiting, delayed hypoglycemia, and allergic reactions.^[60,61] A previous clinical trial showed that gastric peristalsis inhibitors did not improve the detection rate of early gastric cancer.^[62]

Clinical question 3.3: Can painless endoscopy improve the detection rate of early gastric cancer?

Recommendation 3.3: Sedatives and analgesics are suggested for patients who are extremely anxious and/or unable to cooperate during gastroscopy. Painless endoscopy may improve the detection rate of early gastric cancer (weak recommendation, low certainty of evidence).

In the United States, more than 98% of gastroscopies are performed with sedation; however, in China, many routine gastroscopies are performed without sedation. Painless gastroscopy can relieve patients’ anxiety and discomfort, and enable their cooperation to improve the quality of the procedure.^[63,64,65] Sedation may improve the detection rate of early gastric cancer, likely owing to enhancing the use of accessary endoscopic techniques, prolonging observation time, and permitting more biopsies in different locations. A few systematic reviews have analyzed the effect of sedation on the performance of endoscopy and found that moderate sedation can provide a high level of physician and patient satisfaction without increasing the incidence of serious or life-threatening adverse events.^[63,66,67] A multicenter retrospective study investigated the influence of sedation on the endoscopic detection rate of upper gastrointestinal early cancer and precancerous lesions and found that sedation statistically increased the detection rate of early gastric cancer (0.12% in the nonsedation group vs. 0.16% in the sedation group; P = 0.02).^[68]

Clinical question 3.4: The duration of observation in the stomach is related to the detection rate of early gastric cancer; therefore, it is recommended that the stomach be observed for a sufficient length of time. What is the exact length of time?

Recommendation 3.4: Adequate time for gastroscopy is helpful in the detection of early gastric cancer. We recommend no less than 7 minutes for a complete esophagogastroduodenal examination (strong recommendation, low certainty of evidence).

To improve the detection rate of early gastric cancer, systematic observation should be performed in the stomach. To date, four papers have been published about the relationship between the duration of upper gastrointestinal endoscopy and the detection rate of early gastric cancers.^{[69,70,71,72]} The detection of early gastric cancer is related to the observation time in the stomach. Sufficient examination time is helpful for lesion detection, and the stomach should be systematically observed to ensure that there are no blind spots; however, extremely long examination time is unnecessary.

Clinical question 3.5: Can imaging, namely computed tomography (CT), magnetic resonance imaging (MRI), endoscopic ultrasonography (EUS), and positron emission tomography (PET), be used to diagnose early gastric cancer?

Recommendation 3.5.1: EUS is recommended to identify the clinical tumor stage (T-stage) of early gastric cancer (strong recommendation, moderate certainty of evidence). The diagnostic accuracy of both multidetector CT (MDCT) and MRI to determine the T-stage varies compared with EUS; moreover, less experience has been gained with MDCT or MRI (weak recommendation, moderate certainty of evidence). PET/CT is not suggested as a routine imaging modality to determine gastric cancer T-stage (weak recommendation, moderate certainty of evidence).

The clinical T-stage of gastric cancer can be identified based on the intramural depth of tumor invasion detected by EUS, CT, or MRI. One meta-analysis of 46 studies involving 2742 patients demonstrated that EUS can distinguish T1 and T2 gastric cancer, with a pooled sensitivity and specificity of 85% (95% CI: 78%–91%) and 90% (95% CI: 85%–93%), respectively.^[73] Moreover, another meta-analysis indicated that stage T1a or T1b gastric cancer can be differentiated by EUS, with a pooled sensitivity and specificity of 87% (95% CI: 81%–92%) and 75% (95% CI: 62%–84%), respectively.^[73] Regarding the identification of T1 stage gastric cancer, one meta-analysis showed that the sensitivity of EUS (82%) was significantly higher than that of MDCT (41%), and the specificity of EUS and MDCT was 89% and 97%, respectively, with no significant difference.^[74] Although one meta-analysis showed that the accuracy of MRI in the identification of T1 stage gastric cancer was 86.3%, the study involved only 109 patients.^[75] The ability of PET to determine early gastric cancer T-stage cannot be confirmed.^[75] These studies indicate that the most experience has been gained with EUS for the accurate diagnosis of early gastric cancer T-stage. Few MDCT studies and even fewer MRI studies are available for meta-analysis.

Recommendation 3.5.2: MDCT is recommended to identify the clinical lymph node stage (N-stage) in patients with early gastric cancer (strong recommendation, moderate certainty of evidence). EUS could be used to detect regional metastatic lymph nodes (weak recommendation, moderate certainty of evidence). MRI and PET can provide supplemental information to determine the N-stage of early gastric cancer when MDCT cannot confirm a final decision (strong recommendation, moderate certainty of evidence).

The identification of gastric cancer N-stage using EUS, MDCT, and MRI is based mainly on the short-axis diameter of the lymph nodes. MDCT is recommended as the first choice to determine the N-stage of early gastric cancer. Compared with other imaging methods, EUS is more accurate in the diagnosis of regional metastatic lymph nodes because it can detect the detailed structure of the lymph nodes surrounding the stomach. The accuracy of MRI to determine the N-stage can be improved by adding diffusion-weighted imaging (DWI) because metastatic lymph nodes show higher signal intensity.^[76] The higher signal intensity is induced by denser cellularity compared with that of benign lymph nodes.^[76] Additionally, regarding determination of the early gastric cancer N-stage, PET has higher specificity compared with MDCT and MRI owing to the higher glucose metabolism of metastatic lymph nodes compared with benign lymph nodes.

One meta-analysis of 10 studies and 708 patients demonstrated that MDCT could identify gastric cancer N-stage, with a pooled sensitivity and specificity of 80.0% (95% CI: 62.5%–91.9%) and 77.8% (50.0%–87.9%), respectively.^[77] Regarding EUS in the identification of gastric cancer N-stage, one meta-analysis of 44 studies and 3573 patients indicated that the sensitivity and specificity were 83% (95% CI: 79%–87%) and 67% (95% CI: 61%–72%), respectively.^[73] Moreover, better results appeared not to be achievable with MRI or PET compared with MDCT.^[75] Another meta-analysis of 12 retrospective studies and 3 prospective studies that involved 1301 patients aimed to compare the accuracy of DWI with PET/CT in the identification of gastric cancer N-stage. The pooled accuracy, sensitivity, and specificity of DWI and PET/CT in the meta-analysis were 79% (95% CI: 73%–85%) and 69% (95% CI: 61%–77%); 81% (95% CI: 77%–84%) and 52% (95% CI: 39%–64%); and 88% (95% CI: 61%–97%) and 66% (95% CI: 62%–70%), respectively.^[78]

Recommendation 3.5.3: MDCT is recommended to identify the clinical metastatic stage (M-stage) in patients with early gastric cancer (strong recommendation, high certainty of evidence). When MDCT cannot be used to confirm liver metastases, MRI is recommended to confirm the diagnosis (strong recommendation, high certainty of evidence). If distant metastases are suspected clinically, but neither MDCT nor MRI findings are sufficient to support a diagnosis, PET is recommended to provide additional information in the identification of M-stage (strong recommendation, moderate certainty of evidence).

The identification of M-stage in patients with early gastric cancer is a key point in determining the treatment strategy. The guidelines for gastric cancer recommend MDCT as the first choice in the determination of the M-stage, including liver and peritoneal metastases. If MDCT cannot diagnose liver metastases, MRI combined with DWI and the administration of gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (EOB-Gd-DTPA) is strongly recommended to make the final decision. MDCT tends to misdiagnose micro- or miniperitoneal metastases, especially those close to the organs. DW-MRI or PET/CT is best to identify peritoneal metastases.

One meta-analysis evaluated liver metastases assessed by MDCT, MRI, and PET/CT and included 36 studies.^[79] The pooled sensitivity of MDCT (11 studies involving 2151 lesions), MRI (12 studies involving 2301 lesions), and PET/CT (13 studies involving 1846 lesions) in the analysis were 82.1% (95% CI: 74.0%–88.1%), 93.1% (95% CI: 88.4%–96.0%), and 74.1% (95% CI: 62.1%–83.3%), respectively. The pooled specificity of MDCT, MRI, and PET/CT were 73.5% (95% CI: 53.7%–86.9%), 87.3% (95% CI: 77.5%–93.2%), and 93.9% (95% CI: 83.9%–97.8%), respectively. MRI with Gd-EOB-DTPA showed higher sensitivity in the diagnosis of liver metastases compared with MDCT and PET/CT, with a similar specificity to that of PET/CT. Additionally, DWI to diagnose liver metastases had a pooled sensitivity and specificity of 87% (95% CI: 0.84%–0.89%) and 90% (95% CI: 0.87%–0.93%),^[80] respectively. Another meta-analysis of peritoneal metastases included 20 studies of MDCT, 10 studies of PET/CT, and 7 studies of DWI. The sensitivity, specificity, and diagnostic odds ratio (OR) of CT, PET/CT, and DWI were 68% (95% CI: 46%–84%), 88% (95% CI: 81%–93%), and 15.9 (95% CI: 4.4–58.0); 80% (95% CI: 57%–92%), 90% (95% CI: 80%–96%), and 36.5 (95% CI: 6.7–199.5); and 92% (95% CI: 84%–96%), 85% (95% CI: 78%–91%) and 63.3 (95% CI: 31.5–127.3), respectively.^[81]

Part 4. Treatment

Clinical question 4.1: What are the indications for endoscopic treatment for early gastric cancer?

Recommendation 4.1: Generally, patients are suggested to undergo endoscopic treatment when there is an extremely low probability of lymph node metastasis, and when the size and site of the lesion indicate that en bloc resection is feasible (weak recommendation, low certainty of evidence).

Once a patient has been diagnosed with early gastric cancer, endoscopic or surgical therapy is recommended. As a stomach-preserving technique, endoscopic resection has been most widely used in the treatment of early gastric cancer.^[82,83] According to the guidelines, the absolute indications for endoscopic mucosal resection/endoscopic submucosal dissection (ESD) are “UL0 cT1a differentiated-type carcinomas with a long diameter ≤2 cm”. Absolute indications for ESD are “UL0 cT1a differentiated-type carcinomas with a long diameter >2 cm; UL1 cT1a differentiated-type carcinomas with a long diameter ≤3 cm; and UL0 cT1a undifferentiated-type carcinomas with a long diameter ≤2 cm” (UL0 = no ulceration).^[84] Studies indicated that endoscopic treatment might lead to higher rates of recurrence compared with other therapies. Therefore, strict long-term postoperative surveillance is required after endoscopic treatment.^[85]

Clinical question 4.2: Can the indications for ESD for early gastric signet ring cell carcinoma be expanded?

Recommendation 4.2: We conditionally suggest that patients with early gastric signet ring cell carcinoma undergo ESD (weak recommendation, low certainty of evidence).

Currently, undifferentiated cancer measuring <2 cm is an expanded indication for ESD.^[86,87] However, owing to the high risk of lymph node metastasis and recurrence of undifferentiated cancer, especially signet ring cell cancer, whether patients with early signet ring cell cancer should be treated with ESD remains controversial.^[88] Current studies have shown that in undifferentiated cancer patients, ESD has similar efficacy and safety compared with surgery, especially in cases with expanded indications.^[88,89,90] However, the complete resection rate of lesions in the ESD group in previous studies was relatively low, and the recurrence rate was significantly higher than that in the operation group.^[89,90] Therefore, whether patients with early gastric signet ring cell carcinoma should receive ESD still requires careful consideration.

Clinical question 4.3: Can endoscopic full-thickness gastric resection, based on the concept of super-minimally invasive surgery, and the combination of ESD and laparoscopic lymph node dissection broaden the indications for endoscopic treatment for gastric cancer?

Recommendation 4.3: For patients with early gastric cancer that cannot be treated by ESD, combined ESD and laparoscopic lymph node dissection (LLND) is suggested (weak recommendation, low certainty of evidence).

Observational cohort studies indicated that combined ESD and LLND could avoid unnecessary gastrectomy in some early gastric cancer patients, enabling complete resection of the primary tumor and accurate histopathological assessment of the lymph node status. Five patients underwent combined ESD and LLND in a preliminary study. Histopathological examination revealed that all resected nodes were free of cancer cells in four patients; however, metastasis to the lymph nodes was confirmed in one patient who chose not to undergo additional surgery. During the follow-up, quality of life was restored in all patients, and no tumor recurrence was observed.^[91] In a single-center prospective study, 100 patients underwent sentinel node navigation surgery. For patients with a negative sentinel node biopsy during the intraoperative examination, treatment recommendation was made based on the location and type of cancer and the patient’s preference. Of 11 patients who underwent ESD, one patient also underwent distal gastrectomy because of tumor recurrence.^[92]

Clinical question 4.4: How do we manage complications due to ESD for early gastric cancer?

Recommendation 4.4.1: Routine second-look endoscopy (SLE) is not recommended to prevent bleeding after gastric ESD (strong recommendation, high certainty of evidence).

Post-ESD bleeding is difficult to predict and can be a potentially life-threatening complication. Although the administration of proton pump inhibitors and prophylactic coagulation after ESD were reportedly effective for preventing post-ESD bleeding,^{[93,94,95,96]} the incidence of bleeding after gastric ESD remains approximately 5%. Bleeding after ESD may result in serious events, such as hypovolemic shock; thus, prevention is important. Regarding the role of routine SLE in reducing the incidence of post-ESD bleeding, three RCTs reported that routine SLE did not reduce bleeding after gastric ESD in patients with an average risk of bleeding.^[97,98,99] Furthermore, a recent meta-analysis revealed that SLE after ESD did not reduce the risk of delayed post-gastric ESD bleeding.^[100]

Recommendation 4.4.2: Endoscopic closure methods are suggested to be the first choice for post-gastric ESD perforation. When panperitonitis develops, emergency surgery is indicated (weak recommendation, low certainty of evidence).

Delayed perforation is a rare but severe complication of ESD for early gastric cancer.^[101] The reported incidence of perforation after gastric ESD ranges from 0.04% to 0.7%,^{[102,103,104]} and almost half of the cases were managed by endoscopic methods; the remaining cases required emergency surgery.

Clinical question 4.5: Does additional surgery improve oncologic outcomes in early gastric cancer patients after noncurative resection with ESD?

Recommendation 4.5: Additional surgical treatment is suggested for early gastric cancer after noncurative resection with ESD. However, close follow-up or additional endoscopic treatment is suggested when a positive horizontal margin is the only noncurative resection factor, especially in elderly patients (weak recommendation, low certainty of evidence).

Current meta-analyses showed that the additional surgical group had longer 5-year overall survival (OS), disease-free survival, and disease-specific survival compared with the nonsurgical group (follow-up observation).^[105,106] Regarding close follow-up or endoscopic treatment in cases of a positive horizontal margin as the only noncurative resection factor, there were few additional endoscopic treatment cases, and all studies were retrospective cohort analyses; there is a lack of reviews and meta-analyses.

Clinical question 4.6: Is routine detection and eradication of H. pylori after early gastric cancer surgery recommended?

Recommendation 4.6: Routine detection and eradication of H. pylori after early gastric cancer surgery is recommended for the prevention of metachronous cancer (strong recommendation, high certainty of evidence).

Previous studies showed that the incidence of metachronous gastric cancer was significantly lower in the group that underwent detection and eradication of H. pylori after early gastric cancer surgery (including ESD) compared with the undetected or uneradicated group.^{[107,108,109,110]} Furthermore, there was a dose-response effect, with the benefit increasing with longer follow-up duration.^{[107,108,109,110]}

Clinical question 4.7: Is quality of life for patients with early gastric cancer after super-minimally invasive surgery better than that with traditional and minimally invasive surgery?

Recommendation 4.7: Quality of life for early gastric cancer patients after super-minimally invasive surgery is better than that with traditional and minimally invasive surgery (strong recommendation, moderate certainty of evidence).

For early gastric cancer, partial or whole-stomach gastrectomy and gastrointestinal reconstruction is necessary as surgical treatment through a laparoscopic (minimally invasive surgery) or open approach (traditional surgery). Gastrectomy is associated with considerable adverse events, prolonged hospitalization, high cost, delayed recovery of gut function, and poor quality of life.^{[111,112,113,114,115,116,117,118,119]} In contrast to minimally invasive and open surgery, super-minimally invasive surgery, which involves resecting the lesions while preserving anatomical integrity, is associated with enhanced recovery, minimal invasion, lower costs, and better quality of life.^{[113,114,115,116,117,118,119,120]} The principle of super-minimally invasive surgery is to cure disease with quick recovery, and is a new treatment mode in medicine. In super-minimally invasive surgery, there are four channels, namely, the natural lumen, submucosal tunnel, multicavity channel, and transmural channel, and the methods comprise endoscopic mucosal resection, ESD, endoscopic submucosal tunnel dissection, and various modified endoscopic resection methods.^[120]

Clinical question 4.8: What is the indication for surgical operation of early gastric cancer?

Recommendation 4.8: Radical gastrectomy is recommended for tumors that do not meet the endoscopic resection indications and with the possible presence of lymph node metastasis, as suggested by preoperative examinations, and patients with noncurative ESD resection of eCura C-2 (strong recommendation, low certainty of evidence).

A systematic review included 15 retrospective studies with 3737 patients in the endoscopic resection (ER) group and 4246 patients in the radical gastrectomy group. Although the 3- and 5-year survival rates were similar between the groups, the complication rate in the ER group was significantly lower than that in the radical gastrectomy group. ER is a good choice for patients with small early gastric cancer lesions without lymph node metastasis,^[112] especially in older patients with various medical comorbidities and in patients who cannot tolerate abdominal surgery or decline surgery. In contrast, radical gastrectomy is recommended when preoperative examination suggests the possible presence of lymph node metastasis. A systematic review included five studies and described in detail the efficacy and safety of ESD or surgery in the treatment of undifferentiated cancer. The incidence of metachronous cancer in the ESD group was significantly higher than that in the operation group. Accordingly, radical gastrectomy can be used for early gastric cancer. Furthermore, additional surgery is occasionally necessary when patients have undergone noncurative ESD. Regarding distal gastrectomy for stage cTIN0 gastric cancer, laparoscopic surgery has similar safety compared with open surgery, with no significant difference in the long-term prognosis, and can be used as a routine treatment option.^[121]

Clinical question 4.9: Which method is recommended for the localization of early gastric cancer?

Recommendation 4.9: Both preoperative and intraoperative gastroscopy are suggested for the localization of early gastric cancer (weak recommendation, very low certainty of evidence).

Recently, laparoscopic radical gastrectomy has developed rapidly in the field of gastric cancer treatment, and laparoscopic surgery has matured. When the gastric serosa is not involved, the specific location of the tumor is difficult to determine by observing the screen image during laparoscopic surgery. Therefore, the surgical resection scope of gastric cancer can be determined only by preoperative gastroscopic observation results or by experience and intuition. However, simple and effective methods for locating tumors in laparoscopic surgery for gastric cancer are lacking. Staining^[122] and metal clip marking^{[123,124,125,126]} can be used for preoperative gastroscopy, and endoscopy can be used for tumor localization, but there is no clear conclusion regarding the accuracy of these methods and which method is better. Therefore, both preoperative and intraoperative gastroscopy are recommended for the localization of early gastric cancer.

Clinical question 4.10: Is laparoscopic gastrectomy recommended for the treatment of early gastric cancer?

Recommendation 4.10.1: Laparoscopic surgery is recommended for distal gastrectomy in early gastric cancer (strong recommendation, high certainty of evidence).

Recommendation 4.10.2: Laparoscopic surgery is recommended for total gastrectomy in early gastric cancer (strong recommendation, high certainty of evidence).

Recommendation 4.10.3: Laparoscopic surgery is suggested for proximal gastrectomy and other functional sparing surgeries in early gastric cancer (weak recommendation, low certainty of evidence).

With developments in laparoscopic technology, laparoscopic gastrectomy has become widely used; however, surgical safety and OS after laparoscopic gastrectomy for early gastric cancer have not been confirmed. The evidence retrieval and evaluation team performed a systematic review. The systematic review included 37 studies^{[121,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163]} involving 12,172 patients. There were 23 studies of distal gastrectomy for early gastric cancer (6 RCTs, 2 prospective cohort studies, 15 retrospective studies). In the subgroup analysis by research type, six RCTs suggested a reduction in surgical complications with laparoscopic gastrectomy compared with open surgery (hazard ratio [HR]: 0.60, 95% CI: 0.46–0.76). Additionally, 17 nonRCTs suggested a reduction in surgical complications with laparoscopic gastrectomy compared with open surgery (HR: 0.60, 95% CI: 0.40–0.90). Regarding OS, studies of distal gastrectomy, including three RCTs, suggested no significant difference in OS between laparoscopic and open surgery (HR: 0.93, 95% CI: 0.80–1.08), and five retrospective studies suggested no significant difference in OS between laparoscopic and open surgery (HR: 0.99, 95% CI: 0.68–1.44).

Three studies of total gastrectomy for early gastric cancer (one RCT and two retrospective studies) suggested that there was no significant difference in safety between laparoscopic and open total gastrectomy (HR: 1.09, 95% CI: 0.80–1.48). There was only one retrospective study of oncologic safety, which suggested no significant difference in OS between laparoscopic total gastrectomy and open total gastrectomy (HR: 0.96, 95% CI: 0.57–1.65).

Five studies, which focused on laparoscopic proximal gastrectomy and other functional sparing surgeries for early gastric cancer (all retrospective studies), reported no significant difference in surgical safety between laparoscopic gastric sparing surgery and open surgery (HR: 0.80, 95% CI: 0.51–1.25). There were no data on oncologic safety.

Clinical question 4.11: What is the recommended extent of gastrectomy for early gastric cancer?

Recommendation 4.11: For early gastric cancer, radical gastrectomy should fully ensure a safe margin distance from the tumor’s edge. For stage T1 tumors, the recommended resection margin is >2 cm from the tumor. The standard operations are distal gastrectomy and total gastrectomy. With a sufficient resection margin, functional-preserving gastrectomy can be selected based on the tumor location. Regarding upper-third lesions, if more than 50% of the stomach can be preserved, proximal gastrectomy can be performed. Regarding middle-third lesions, if the distal lesion is >4 cm from the pylorus, pylorus-preserving gastrectomy can be chosen (strong recommendation, low certainty of evidence).

Surgery plays an important role in the treatment of early gastric cancer. To ensure R0 resection, the resection margin should be >2 cm from the tumor for stage T1 tumors.^[164] The standard surgical procedures for gastric cancer are distal gastrectomy and total gastrectomy.^[165] Many prospective clinical trials have shown that laparoscopic radical gastrectomy has become a routine choice for gastrectomy.^[166,167] The prognosis of early gastric cancer is good; therefore, function-preserving gastrectomy can be selected based on the tumor location if the distance to the resection margin is sufficient. Proximal gastrectomy can be performed for upper gastric tumors if more than 50% of the stomach can be preserved.^[168] Some studies suggest that proximal gastrectomy with double-tract or double-flap technique reconstruction provides excellent antireflux efficacy.^[169,170] For tumors located in the middle third of the stomach, pylorus-preserving gastrectomy (PPG) can be performed if the distal lesion is >4 cm from the pylorus. A meta-analysis showed that the OS after PPG was not worse than that with distal gastrectomy, and PPG was associated with a lower incidence of postoperative weight loss and dumping syndrome.^[171] The KLASS-04 study is designed to evaluate the difference between PPG and distal gastrectomy, and the study will provide high-level evidence regarding PPG.^[172]

Clinical question 4.12: What is the recommended extent of gastrectomy for hereditary diffuse early gastric cancer?

Recommendation 4.12: Total gastrectomy is suggested for hereditary diffuse early gastric cancer (weak recommendation, low certainty of evidence).

Hereditary diffuse gastric cancer (HDGC) is caused by CDH1 gene mutation with autosomal dominant inheritance.^[173] People who carry the CDH1 gene mutation have a lifetime risk of gastric cancer of up to 80%. Importantly, HDGC is difficult to diagnose in the early stage. Most patients are diagnosed as having HDGC in the advanced stage, which carries a poor prognosis. Only approximately 10% of patients with advanced HDGC diagnosed before the age of 40 years can undergo radical resection. The prognosis of HDGC is poor; the 5-year survival rate is <30% even with radical surgery with neoadjuvant or adjuvant chemotherapy.^{[174,175,176,177]} Therefore, for people who carry the CDH1 gene mutation, prophylactic total gastrectomy is suggested to eliminate the risk of gastric cancer and improve survival rates. More than 80%^[178,179] of the patients who underwent prophylactic total gastrectomy were proven to have cancer on pathology. Most of the cancers were early gastric cancer, including advanced gastric cancer. Therefore, total gastrectomy is suggested for early gastric cancer that meets the diagnostic criteria of HDGC. Total gastrectomy and Roux-en-Y digestive tract reconstruction is suggested, and the proximal resection margin should include the distal esophagus with squamous epithelium to ensure no residual gastric mucosa. Owing to the small number of these patients, the level of evidence provided by the relevant literature is low. However, total gastrectomy is still suggested for patients with early gastric cancer meeting the diagnostic criteria of HDGC.

Clinical question 4.13: What is the recommended region (distal or total) for gastrectomy for multiple early gastric cancer (MEGC) that is unsuitable for endoscopic treatment?

Recommendation 4.13: Subtotal gastrectomy is suggested for MEGC that is unsuitable for endoscopic treatment (weak recommendation, very low certainty of evidence).

In 2020, Lee et al^[180] performed a meta-analysis that included 25 cohort studies involving 3058 early gastric cancer patients. The results showed that the short-term outcome of radical subtotal gastrectomy was superior to that of radical total gastrectomy. Regarding the long-term outcomes, a meta-analysis that included two cohort studies involving 131 MEGC patients^[181,182] reported that the 5-year OS after subtotal gastrectomy was comparable to that of total gastrectomy (OR = 1.40, 95% CI: 0.21–9.23).

Regarding short-term outcomes after subtotal vs. total gastrectomy, no direct evidence supports either approach in patients with MEGC. One meta-analysis included 25 cohort studies involving 3058 early gastric cancer patients. Compared with total gastrectomy, subtotal gastrectomy significantly reduced the operation time (weighted mean difference (WMD) = –17.89, 95% CI: –29.64 to –6.13) and intraoperative blood loss (WMD = –35.38, 95% CI: –61.27 to –9.48). Furthermore, no significant differences were observed between subtotal gastrectomy and total gastrectomy regarding hospitalization duration (WMD = –0.94, 95% CI: –2.73–0.85) or postoperative complications (OR = 0.84, 95% CI: 0.55–1.30).^[180]

Clinical question 4.14: What is the recommended extent of lymph node dissection for early gastric cancer?

Recommendation 4.14: D2 lymphadenectomy is recommended for cT1N+ tumors, and D1/1+ lymphadenectomy is recommended for cT1N0 tumors (strong recommendation, low certainty of evidence).

Lymph node metastasis has an important impact on the prognosis of patients with early gastric cancer.^[183] Studies have shown that approximately 5% of patients with mucosal cancer and 20% of patients with submucosal cancer have lymph node metastasis.^[184,185] However, the rational extent of lymphadenectomy remains controversial. The risk of lymph node metastasis in early gastric cancer is much lower than that in advanced forms.^[186] Invasion of the submucosa, poor differentiation, tumor size (diameter >2 cm), and ulceration are risk factors for lymph node metastasis in early gastric cancer.^{[105,187,188,189]} In early gastric cancer that is unsuitable for endoscopic treatment (differentiated-type adenocarcinoma, UL0, T1a, diameter ≤2 cm), the Japanese Gastric Cancer Association guidelines suggest D1 or D1+ lymphadenectomy in cases with clinically-negative nodes. The 5-year disease-specific survival of D1/D1+ gastrectomy for early gastric cancer is 96%–98%,^{[159,190,191]} which is similar to survival after D2 gastrectomy.^{[121,192,193]} D2 lymphadenectomy should be performed for patients with cT1N+ tumors and whenever the possibility of nodal involvement cannot be dismissed.^[87,194]

Currently, no consensus on the extent of lymphadenectomy for early gastric cancer is available. A systematic review that included 21 retrospective studies (4789 patients) showed that in patients with early gastric cancer who required salvage surgery after endoscopic surgery, those who underwent additional gastrectomy had longer 5-year OS (HR = 0.34; P <0.001) and 5-year disease-free survival (HR = 0.52; P = 0.001). Lymph node metastasis was associated with larger tumor size (>3 cm) (OR = 1.73; P <0.001), elevated tumor type (OR = 1.60; P = 0.035), deep tumor invasion (submucosal grade >1) (OR = 2.68; P <0.001), lymphatic invasion (OR = 4.65; P <0.001), and positive vertical margin (OR = 2.30; P <0.001).^[105]

Clinical question 4.15: What are the indications for surgical resection of early gastric cancer with noncurative ESD?

Recommendation 4.15.1: Surgical resection with lymphadenectomy is an option (others are repeat ESD or careful follow-up) for eCura C1 patients based on the status of the primary tumor(s) (conditional recommendation, low certainty of evidence).

Recommendation 4.15.2: Surgical resection with lymphadenectomy is recommended for eCura C2 patients (strong recommendation, low certainty of evidence).

Surgical resection with lymphadenectomy is the primary remedial measure for noncurative ESD in early gastric cancer. The eCura scoring system was developed in 2017 to categorize the curative degree of ESD.^[195] In the Japanese gastric cancer treatment guidelines (v5, 2018),^[196] the eCura system is used to evaluate the curability of ESD and determine additional treatment options. For patients with eCura C2, additional surgical treatment is recommended, but for eCura C1, additional surgery should be comprehensive, and the tumor status and ESD resection status should be considered. However, all of the current evidence regarding noncurative ESD is based on retrospective studies with different evaluation classifications for curative degree. Although there is low certainty of evidence, the Japanese eCura system is the best-evaluated system for patients who have undergone ESD.

A published meta-analysis, which included 10 studies from 2010 to 2018, found that additional surgery improved both OS and disease-specific survival compared with simple follow-up after noncurative ESD.^[106]

Clinical question 4.16: What is the recommended extent of gastrectomy for early gastric cancer with noncurative ESD?

Recommendation 4.16: The recommended extent of gastrectomy for early gastric cancer patients with noncurative ESD is the same as that for early gastric cancer (strong recommendation, low certainty of evidence).

Few published studies have compared the extent of gastrectomy for early gastric cancer patients with noncurative ESD. In reality, early gastric cancer patients with noncurative ESD are still considered to have “early gastric cancer”, and the extent of gastrectomy should not be influenced by previous ESD. Therefore, the extent of gastrectomy for early gastric cancer with or without noncurative ESD is the same as that for early gastric cancer and is consistent with the guidelines.^[197] Therefore, we emphasize that the extent of gastrectomy for patients with noncurative ESD should be based on the treatment guidelines for early gastric cancer.

Clinical question 4.17: What is the optimal timing for additional surgery in patients with early gastric cancer following noncurative ESD?

Recommendation 4.17: The suggested timing of additional surgery for early gastric cancer patients after noncurative ESD is within 3 months after ESD (weak recommendation, very low certainty of evidence).

There is no clear conclusion regarding the best time for additional surgery after noncurative ESD for early gastric cancer. Studies have shown that ESD-induced ulcers are usually in the healing or scar formation stage 4–8 weeks after surgery. Inflammation and ulcers caused by ESD lead to inflammation and edema of local tissues in the gastric wall and swelling of lymph nodes around the lesion. The inflammatory response reaches its peak 1–2 weeks postoperatively and begins to resolve after 1 month.^[198,199] With additional surgery performed less than 4 weeks after ESD, tissue separation and lymph node dissection become more difficult, the operation time is prolonged, and bleeding may increase. The recurrence rate and survival are not significantly affected by additional surgery performed >4 weeks postoperatively.

Few studies have focused on the optimal timing of additional surgery after noncurative ESD for early gastric cancer. In 2014, Kim et al^[200] first suggested that 1 month after ESD was the best time for additional surgery. The study reviewed data for 154 patients with early gastric cancer who underwent additional surgery after ESD. The patients were divided into two groups: additional surgery ≤29 days after ESD and >29 days after ESD. The ≤29-day surgery group had longer operation times and more intraoperative blood loss compared with the >29-day group. In 2019, the research team in the study expanded the sample size to 302 patients and extended the follow-up time to 41.98 ± 21.23 months, which again showed that the delayed surgery group had shorter operation times and postoperative hospital stays.^[201] The 5-year survival rate (99%) was higher in the later surgery group than that of the early surgery group (92%), but the difference was not statistically significant. Another retrospective study included 83 patients who were also divided into a ≤29-day additional surgery group and >29-day additional surgery group using the cutoff of 29 days after ESD. Compared with the early surgery group, the later surgery group had less intraoperative blood loss (P = 0.011). No significant differences for operation time, postoperative hospital stay, or postoperative complication rates were observed between the groups.

Academic supervisors

Enqiang Linghu (Chinese People’s Liberation Army General Hospital), Jiafu Ji (Peking University Cancer Hospital and Institute)

Expert committee (in alphabetical order by surname)

Yu Bai (Changhai Hospital, Navy Medical University), Ningli Chai (The First Medical Center of PLA General Hospital), Shiyao Chen (Zhongshan Hospital, Fudan University), Min Dai (Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College), Weiguo Dong (Renmin Hospital of Wuhan University), Shiyu Du (China-Japan Friendship Hospital), Jiagang Han (Beijing Chaoyang Hospital, Capital Medical University), Bing Hu (West China Hospital, Sichuan University), Jiankun Hu (West China Hospital, Sichuan University), Changming Huang (Fujian Medical University Union Hospital), Yonghui Huang (Peking University Third Hospital), Jiafu Ji (Peking University Cancer Hospital and Institute), Jingnan Li (Peking Union Medical College Hospital), Peng Li (Beijing Friendship Hospital, Capital Medical University), Xinxiang Li (Fudan University Shanghai Cancer Center), Xun Li (The First Hospital of Lanzhou University), Yanqing Li (Qilu Hospital, Shandong University), Ziyu Li (Peking University Cancer Hospital and Institute), Pin Liang (The First Affiliated Hospital of Dalian Medical University), Enqiang Linghu (The First Medical Center of PLA General Hospital), Fenglin Liu (Zhongshan Hospital, Fudan University), Jinsong Liu (Wuhan Union Hospital of Tongji Medical College), Mei Liu (Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology), Yun Lu (Affiliated Hospital of Qingdao University), Qingfeng Luo (Beijing Hospital), Suxia Luo (The Affiliated Cancer Hospital of Zhengzhou University), Nonghua Lyu (The First Affiliated Hospital of Nanchang University), Fandong Meng (Beijing Friendship Hospital, Capital Medical University), Yongzhan Nie (Xijing Hospital), Zhanlong Shen (Peking University People’s Hospital), Yongquan Shi (Xijing Hospital of Digestive Diseases, Fourth Military Medical University), Yantao Tian (Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College), Junping Wang (Shanxi Provincial People’s Hospital), Yi Wang (Peking University People’s Hospital), Zhenchang Wang (Beijing Friendship Hospital, Capital Medical University), Zhenjun Wang (Beijing Chaoyang Hospital, Capital Medical University), Zhenning Wang (The First Affiliated Hospital of China Medical University), Bin Wu (Peking Union Medical College Hospital), Hongwei Yao (Beijing Friendship Hospital, Capital Medical University), Lu Zang (Ruijin Hospital, Shanghai Jiaotong University), Guoxin Zhang (The First Affiliated Hospital of Nanjing Medical University), Shutian Zhang (Beijing Friendship Hospital, Capital Medical University), Zhongtao Zhang (Beijing Friendship Hospital, Capital Medical University)

Methodology group

Shiyao Chen (Zhongshan Hospital, Fudan University), Yuelun Zhang (Peking Union Medical College Hospital)