Gastric cancer is the fourth leading cause of death from malignant tumors worldwide and the third main cause of cancer death in China (Cao et al., 2021; Navashenaq et al., 2021; Sun et al., 2021; Zeng and Jin, 2021). Despite the incidence of gastric cancer has reduced, gastric cancer related mortality has not changed (Sukri et al., 2020; Varon et al., 2021). Benefiting from advances in medical technology and the popularity of endoscopy, more and more gastric cancer is diagnosed at a relatively early stage. pT2N0M0 gastric cancer is defined as tumors infiltrating the muscularis propria without regional lymph node metastasis or distant metastasis based on the 8^th edition of the AJCC TNM staging system for gastric cancer (Amin et al., 2017; Brierley et al., 2017).

Surgery is the only potential chance of cure for gastric cancer, but a certain percentage of patients relapse after curative surgery, which leads to a poor prognosis. Adjuvant chemotherapy (ACT) or chemoradiotherapy has been demonstrated to be beneficial in numerous clinical trials worldwide (Macdonald et al., 2001; Sasako et al., 2011; Noh et al., 2014; Park et al., 2015). Nevertheless, these trials did not report whether patients with less advanced disease would benefit from adjuvant therapy. There are few studies on patients with pT2N0M0 gastric cancer.

Consensus guidelines provide disparate recommendations. Based on the National Comprehensive Cancer Network (NCCN) Guidelines (version 1.2021, Gastric Cancer), options for pT2N0M0 gastric cancer patients after D2 lymph node dissection include surveillance or ACT. Patients with poorly differentiated or high-grade cancer, lymphovascular invasion, neural invasion or aged <50 years are candidates for ACT (National and Comprehensive, 2020). Meanwhile, observation without adjuvant therapy after curative resection is recommended for stage I (including T2N0M0) gastric cancer according to the Japanese Gastric Cancer Treatment Guidelines 2018 (5^th edition) (Japanese Gastric Cancer A, 2020). ACT may decrease the risk of metastasis in high-risk pT2N0M0 patients, such as those aged <40 years or with high-grade or poorly differentiated tumor and nervous, lymphovascular invasion, based on the Chinese Society of Clinical Oncology (CSCO) Guidelines (version 1.2021, Gastric Cancer); however, it is unclear whether there is survival benefit of ACT for stage I gastric cancer ((Wang et al., 2021)).

Based on the 8th edition of the TNM staging system of gastric cancer, pT2N0M0 gastric cancer belongs to stage IB, which has good prognosis, with 5-year survival rate of approximately 80–90% after curative surgery ((He et al., 2018; Ji et al., 2018)). The recurrence rates for pT2N0M0 gastric cancer after resection range from 3 to 9% (Jin et al., 2015; Park et al., 2016). Considering the number of patients with stage I gastric cancer is increasing, several retrospective studies focused on the role of ACT in patients with pT2N0M0 gastric cancer and evaluated the high-risk factors of relapse and death; however, they reported diverse opinions on the effect of ACT on pT2N0M0 gastric cancer.

Since it was an open question whether ACT would benefit patients with pT2N0M0 gastric cancer, we aimed to determine the effect of ACT after curative resection in this study.

ACTS-GC trial (Sasako et al., 2011) demonstrated that patients with stage II/III gastric cancer could significantly benefit from adjuvant S-1. CLASSIC trial (Noh et al., 2014) also showed survival benefit of adjuvant XELOX for stage II/III gastric cancer patients. ARTIST II trial (Park et al., 2021) showed that adjuvant SOX was more effective than S-1 in patients with node positive, stage II/III gastric cancer. Exiting prospective randomized clinical trials demonstrating the benefit of ACT could not explain whether all gastric cancer patients (especially stage IB gastric cancer) would benefit from ACT. Although the prognosis of pT2N0M0 gastric cancer is relatively good in general, postoperative relapse still occurs in some patients with various recurrence sites.

In the current study, we found a good prognosis of pT2N0M0 gastric cancer, with the 5-year OS rate of 87.9% and 5-year DSS rate of 91.8%, similar to other studies ((In et al., 2016; Park et al., 2016)).

Some retrospective studies identified risk factors in stage I gastric cancer patients. The authors of a Korean study focusing on stage I gastric cancer reported that age, sex, stage IB, lymphatic vessel invasion, nerve invasion and a high serum carcinoembryonic antigen level, were independent prognostic factors (Caccialanza et al., 2016; Liu et al., 2016). A population-based study using the Surveillance, Epidemiology, and End Results (SEER) database demonstrated that older age, proximal tumor location, high tumor grade and large tumor size were independent factors of poor disease-related survival (Gold et al., 2013). Other studies found that several clinical pathological factors were significantly associated with a high risk of relapse and death in pT2N0M0 gastric cancer patients and suggested that patients with high-risk factors receive ACT. A further Chinese study identified the upper 1/3 of the stomach, large tumor diameter, perineural and lymphovascular invasion as independent risk factors associated with decreased OS rates (Wang et al., 2018). Another study also reported that lymphatic vessel and nerve invasion and tumor size were independent risk factors (Caccialanza et al., 2016; Liu et al., 2016).

Our study found that total gastrectomy and surgery alone were independent risk factors for survival. Other studies also found many other risk factors associated with a poor prognosis. The main reason for this inconsistency was study heterogeneity, with differences in race, surgical practice and initial prognosis.

A single-center study from the CLASSIC trial (Caccialanza et al., 2016; Liu et al., 2016) found a marked loss in body composition parameters (muscle, visceral fat and subcutaneous fat) significantly predicted short disease-free survival and OS among patients who underwent gastrectomy. Malnutrition was considered as poor prognostic factor in cancer patients (Caccialanza et al., 2016; Liu et al., 2016). Fujiya demonstrated that persistent postoperative malnutrition was frequently observed in patients who underwent total gastrectomy (Fujiya et al., 2018). These studies might explain why patients who received total gastrectomy had poor prognoses in our study, although we could not evaluate the nutrition index.

Despite a lack of prospective studies that explored the benefit of ACT in less advanced gastric cancer, there were some retrospective studies exploring the effect of ACT such patients. Based on the 8th edition of the TNM staging system of gastric cancer, stage IB gastric cancer includes pT1N1M0 and pT2N0M0. Wang used the SEER database to explore the difference between T1N1M0 and T2N0M0 and found that patients with T2N0M0 gastric cancer may not benefit from adjuvant treatment (Wang et al., 2019b). Recently, Jin et al.(Jin et al., 2021) found that pT2N0 gastric cancer patients with non-signet ring cell carcinoma, tumor size >3 cm and examined lymph nodes≤15 may be particularly appropriate candidates for ACT. In our study, there was no significant difference in OS between patients with signet ring cell carcinoma and patients with other histopathology type.

Since 1997, the retrieval of at least 15 lymph nodes has been recommended for adequate gastric cancer staging, and several studies have found that lymphadenectomy with <15 lymph nodes removed was an adverse independent prognostic factor for OS. A SEER study demonstrated that OS was dependent on the number of harvested lymph nodes; in patients with node-negative T1-2 gastric cancer, every additional 10 lymph nodes harvested increased the 5-year survival rate of 7.6% (Smith et al., 2005). Haejin found that their subgroup of T2N0M0 gastric cancer patients who underwent suboptimal lymphadenectomy benefitted from chemoradiotherapy rather than chemotherapy (Coburn et al., 2008; Du et al., 2011). Due to a lack of patients who received postoperative radiotherapy, the differences in radiotherapy and chemoradiotherapy roles could not be established in our study. Other studies failed to show the number of removed lymph nodes as an independent prognostic factor (Coburn et al., 2008; Du et al., 2011). One large population-based study demonstrated that surgery with adequate lymph node removing alone (≥15 lymph nodes) predicted better prognosis compared with adjuvant therapy in patients with stage I or node-negative gastric cancer (Dudeja et al., 2012). Our study found that the number of harvested lymph nodes was not associated with prognosis, which may be related to excluding patients with fewer than 15 harvested lymph nodes.

Several studies on patients with pT2 gastric cancer focused on the invasion depth. Some studies have showed that pT2 gastric cancer patients showing invasion into dMP had a relatively poor prognosis than those only invasion sMP (Zhang et al., 2016; Park et al., 2021), while others reported no significant difference in the prognosis between the two groups (Son et al., 2007; Nakamura et al., 2019). In our study, the difference in the 5-year OS rate between the sMP and dMP subgroups was not significant (89.6 vs. 86.1%, p = 0.327).

Regarding the therapy regimen, monotherapy and dual-drug therapy showed no significant difference. ACTS-GC trial and CLASSIC trial demonstrated that ACT with S-1 or XELOX was safe. In our study, the main grade 3 or 4 adverse events were neutropenia, peripheral neuropathy in the dual-drug group and anemia, anorexia in the monotherapy group. According to the ACTS-GC trial ((Sakuramoto et al., 2007)), the most common adverse events of grade 3 or grade 4 were anorexia (6.0%), nausea (3.7%), and diarrhea (3.1%) in the S-1 group. According to the CLASSIC trial ((Bang et al., 2012)), the main grade 3 or 4 adverse events were neutropenia (22%), thrombocytopenia (8%), nausea (8%), and vomiting (7%) in the XELOX group. According the ARTIST II study ((Nagaraja et al., 2019; Li et al., 2020; Low et al., 2021; Nakazawa et al., 2021)), the most common adverse events of grade 3 or 4 were peripheral neuropathy (12%), anemia (8%) and anorexia (4%) in the SOX group. The common dose-limiting toxicity of oxaliplatin is peripheral neuropathy, which affects 90% patients ((Kweekel et al., 2005)). The incidence of peripheral neuropathy is considered to be related to the prolonged use of oxaliplatin ((Baek et al., 2010)). Thus, we recommend monotherapy to prevent toxicity and discomfort. However, other studies, which aim to explore the role of ACT in stage IB gastric cancer, failed to analyse the difference between monotherapy and dual-drug therapy.

According to the CSCO guidelines, patients with pT2N0M0 gastric cancer with high-risk factors (age <40 years or with high-grade or poorly differentiated tumor and nervous, lymphovascular invasion) are recommended to receive ACT to reduce the risk of recurrence. Then, we divided patients with pT2N0M0 gastric cancer into two subgroups (with high-risk factors and without high-risk factors) and evaluated whether the effect of postoperative therapy was diverse. ACT indeed provided survival benefits to patients with high-risk factors, while patients without high-risk factors also benefitted from ACT, which was inconsistent with the CSCO guidelines. Regarding the therapy regimen, monotherapy and dual-drug therapy showed no significant difference; thus, considering possible adverse events, we recommend monotherapy regardless of high-risk factors. There were no studies exploring the role of ACT stratified by high-risk factors based on the CSCO guidelines.

Although patients with gastric cancer received ACT after radical gastrectomy, some patients still experienced relapse. Timely detection of recurrence, as well as identification of patients at high risk of relapse after surgery or completion of adjuvant therapy are major challenges in the treatment of gastric cancer. Drug resistance is the major factor of treatment failure and relapse and numerous studies aim to investigate the mechanisms of drug resistance ((Nagaraja et al., 2019; Li et al., 2020; Low et al., 2021; Nakazawa et al., 2021)). Over the past few decades, predictive biomarkers have received increasing attention in diagnosis, treatment, and prognosis of gastric cancer. Studies have found many predictive biomarkers for the precision treatment of gastric cancer (Petrillo and Smyth, 2020). In 2014, The Cancer Genome Atlas (Cancer Genome Atlas Resea, 2014) proposed a molecular classification of gastric cancer into 4 subtypes: chromosomal instability, Epstein-Barr virus positive, genomically stable and microsatellite instability (MSI). An et al.(An et al., 2012) found that in stage II/III gastric cancer, patients with microsatellite stable and MSI-low type significantly benefited from 5-FU-based ACT, while patients with MSI-high type did not benefit from 5-FU-based ACT. Findings from the MAGIC trial (Smyth et al., 2017) showed that mismatch repair deficiency (dMMR) and MSI-high were associated with good prognosis in patients treated with surgery alone, whereas in gastric cancer patients treated with perioperative chemotherapy, dMMR and MSI-high were associated with worse prognosis. Post hoc analysis of CLASSIC trial (Choi et al., 2019) showed that MSI-high was independent prognostic factor and ACT significantly improved disease-free survival in MSS group while no benefit was found in the MSI-high group. MSI status could be used for precision treatment of gastric cancer in the future.

A prospective randomized trial comparing surgery alone with ACT in stage IB gastric cancer patients with at least one risk factor for recurrence (male sex, age>65 years, perineural and lymphovascular invasion) is now ongoing (ClinicalTrials.gov identifier NCT01917552), and this large-scale prospective trial is expected to compensate for previous research shortcomings and yield satisfactory results. Although the trial is based on the 6th edition of the AJCC staging system, it also includes pT2N0M0 gastric cancer based on the 8th edition of the AJCC staging system.

Nevertheless, there are several potential limitations in this study. The number of patients with pT2N0M0 gastric cancer was relatively small since it was a single-center study, the resultant effects may have been underestimated, and the results should be interpreted with caution. In addition, this was a retrospective study, and there were likely patient and tumor baseline characteristic imbalances between the treatment groups. Finally, the role of radiotherapy was not analysed due to a lack of patients who received postoperative radiotherapy. Therefore, the conclusions of this study need to be verified by prospective study with a large sample size.

ACT was accompanied with higher 5-year OS and DSS rates of patients with pT2N0M0 gastric cancer. Patients with pT2N0M0 gastric cancer, regardless of high-risk factors based on the CSCO guidelines, might be considered candidates for ACT. In regard to the therapy regimen, monotherapy might be the optimal choice, considering the adverse events.