Gastric cancer ranks among the most prevalent cancers globally, standing as the fourth most common cause of cancer-related deaths in 2020, with an estimated 769,000 fatalities (1). The incidence of gastric cancer exhibits regional variation, with a higher prevalence observed in East Asia, particularly in China. In 2022, it ranked as the fifth most common cancer in China, with estimated incidence and mortality rates of 10.5% and 12.4%, respectively (2). Regrettably, alongside the high incidence of this disease, the majority of patients are diagnosed at an advanced stage upon detection. This is often due to the early clinical symptoms of gastric cancer being mild and easily overlooked. In patients with advanced or metastatic gastric cancer (AGC/MGC), conventional systemic chemotherapy remains the predominant treatment option. Commonly used regimens typically revolve around fluorouracil-based and platinum-based treatments such as SOX (oxaliplatin + S-1), XELOX (oxaliplatin + capecitabine), FOLFOX (calcium folinate + fluorouracil + oxaliplatin) and so on. However, chemotherapy shows limited effect, with a median survival of 11–12 months (3). For human epidermal growth factor receptor 2 (HER2)-positive patients, the recommended first-line treatment option for advanced cancers is the combination of the anti-HER2 antibody trastuzumab with chemotherapy (4, 5). Nonetheless, patient benefit remains constrained, underscoring the imperative to investigate further treatment alternatives.

Immune checkpoint inhibitors (ICIs), including antibodies against programmed cell death protein-1 (PD-1) or its ligand (PD-L1), have been rapidly developed over the past few years and are now established treatments for chemotherapy-refractory gastric cancers (cancers that progress after two or more lines of chemotherapy) (6, 7). Previous clinical studies have confirmed that some gastric cancer patients may benefit from PD-1 inhibitor therapy. Regarding first-line treatments for AGC/MGC, the KEYNOTE-062 trial, the first global randomized phase III trial comparing the efficacy and safety of pembrolizumab, a humanized anti-PD-1 monoclonal antibody, with chemotherapeutic agents in AGC/MGC, demonstrated that pembrolizumab alone was not inferior to chemotherapy [median overall survival (mOS): 10.6 vs. 11.1 months, hazard ratio (HR): 0.91, 95% confidence interval (CI): 0.69–1.18]. Moreover, pembrolizumab even extended OS in patients with a combined positive score (CPS) of 10 or greater (mOS: 17.4 vs. 10.8 months, HR: 0.69, 95% CI: 0.49–0.97). However, pembrolizumab in combination with chemotherapy did not result in an improvement in OS or progression-free survival (PFS) compared to chemotherapy, either in patients with CPS ≥1 or CPS ≥10 (8). In another phase III trial, KEYNOTE-859, pembrolizumab in combination with chemotherapy prolonged patients’ OS and PFS compared to chemotherapy alone (9). For nivolumab, another clinically utilized PD-1 inhibitor, the combination of nivolumab with a chemotherapeutic agent notably enhanced PFS across all CPS subgroups in both the CheckMate 649 and ATTRACTION-4 trials. Additionally, in the CheckMate 649 trial, overall survival was prolonged. However, there was no significant difference between the two treatment modalities in the ATTRACTION-4 trial (10, 11). Regarding second- and third-line therapy, two trials conducted globally and in Asia showed favorable activity and a manageable safety profile for PD-1 inhibitors in patients with refractory advanced gastric cancer or gastroesophageal junction cancer (GEJC) who had received at least two prior therapies (12, 13). In contrast, another phase III trial, JAVELIN Gastric 300, which explored the efficacy of avelumab, a complete human IgG1 monoclonal antibody, as maintenance therapy in patients with AGC/GEJC, led to the conclusion that avelumab did not improve OS or PFS compared to chemotherapy as third-line treatment (14).

Most of the above trials have demonstrated the effectiveness of PD-1 inhibitors in AGC/MGC, but the treatment response varies among the studies [objective response rate (ORR): 20%-65.1%] (8–10, 15). Findings on the benefits of OS, PFS, and ORR were also inconsistent. The clinical application of PD-1 inhibitor-based therapies is increasingly prevalent; however, there is limited real-world evidence regarding the use of PD-1 inhibitors as monotherapy or in combination treatments. This study endeavors to investigate the effectiveness of PD-1 inhibitor monotherapy or combination therapy in patients with metastatic gastric cancer through analysis of real-world clinical data. Furthermore, we conducted an additional analysis to explore the relationship between clinical characteristics and treatment efficacy, aiming to identify potential biomarkers for predicting the effectiveness of anti-PD-1 therapy.

Progressive gastric cancer has a poor prognosis and current treatment options are limited. With the rise of immunotherapy in recent years, PD-1 inhibitors such as nivolumab and pembrolizumab in combination with chemotherapy also have been recommended for treatment of metastatic gastric cancer (7, 16, 17). However, most previous conclusions were drawn from clinical trials, and there is considerable variation in patient outcomes, with ORR ranging from 20% to 65.1% (9, 10, 15, 18). Hence, it is imperative to investigate the real-world benefits that gastric cancer patients derive from first-line treatment with immune checkpoint inhibitors. To the best of our knowledge, this study represents a largest single-center real-world analysis among Chinese patients with MGC undergoing first-line treatment with anti-PD-1 therapy. Our findings suggest that first-line treatment with PD-1 inhibitors yields significant efficacy, particularly show that the median OS was 22.0 months (95% CI: 16.204–22.796), and the median PFS was 9.0 months (95% CI: 7.184–10.816) with an ORR of 30.5% and a DCR of 89.5%. Furthermore, compared to chemotherapy regimens alone in previous studies (OS:6.2–14 months, PFS: 4.3–12.1 months), the efficacy of chemotherapy combined with PD-1 inhibitors was indeed superior (3). Our patients obtained a longer OS.

It appears that patients may potentially benefit more from combination therapy with chemotherapy plus trastuzumab compared to PD-1 monotherapy or combination with chemotherapy alone. After a follow-up period of 28.0 months in our study, patients treated with PD-1 inhibitors exhibit a mOS of 21 months, while those receiving the combination of PD-1 inhibitors and trastuzumab remain alive, and thus have not yet reached their mOS. For HER2-positive patients, the current recommended standard treatment entails combining chemotherapy with trastuzumab. Recent studies have shown that adding PD-1 inhibitors to the combination of chemotherapy and trastuzumab can enhance efficacy in HER2-positive patients (19–21). Our study further supports this evidence. Nevertheless, due to the limited number of HER2-positive cases, further analysis and expansion of this patient subgroup are warranted.

Although anti-PD-1 treatment has improved patient prognosis compared to early studies, a considerable portion of patients still do not benefit from it, and there is also an increase of the risk of severe immune-related adverse effects. Therefore, researchers continue to explore and identify biomarkers to select the most suitable candidates for immunotherapy. Previous experiments have explored various biomarkers for predicting the efficacy of immunotherapy, including two gastric cancer genomic subtypes identified by The Cancer Genome Atlas (TCGA) studies: EBV infection status and MSI status (22). EBV infection can upregulate PD-L1 expression (23, 24), while DNA replication defects caused by dMMR/MSI-H result in the accumulation of mutations and the expression of new antigens, which may serve as potential targets for immune cells (25). Clinical trial findings suggest that patients with EBV infection or dMMR/MSI-H tumors may be more sensitive to immunotherapy (26–29). However, some experiments also demonstrate that these two biomarkers may not predict the response to immunotherapy in gastric cancer effectively (30–32). In this study, the ORR for two biomarkers EBV-positive and dMMR/MSI-H were 66.7% and 40%, respectively, which align with previous research findings (28, 29, 33). However, there were no significant statistical differences in OS and PFS between the subgroups. This lack of significance may be attributed to the low proportion of patients testing positive for these indicators. Consequently, the small sample size of these patients may have limited statistical power, underscoring the necessity to expand the sample size.

Currently, these biomarkers have limited clinical utility due to their uncertain predictive value, low positivity rates, and high cost. Therefore, we are striving to identify a more effective and convenient biomarker to predict the efficacy of immunotherapy. CEA is a widely used tumor marker in gastric cancer, playing a significant role in disease diagnosis and prognosis (34, 35). Additionally, research has demonstrated ability of CEA to predict the efficacy of PD-1 inhibitors, in non-small cell lung cancer. CEA values have been identified as predictors of patient responsiveness to immune checkpoint inhibitors (36, 37). In the current study, patients with CEA levels above the normal level had a mOS of 15.0 months, while those with normal CEA levels had a mOS of 27.0 months. Similarly, the mPFS was 6.0 months for patients with CEA levels above the standard level and 9.0 months for those with normal level, with statistically significant differences observed. It can be inferred that patients with CEA baseline levels below the standard threshold may derive greater benefits from immunotherapy. In the future, CEA levels might be utilized to predict patients’ response to anti-PD-1 therapy.

This study also has some limitations. Firstly, our study is a retrospective analysis, which inevitably introduces biases. Secondly, despite being the largest single-center, retrospective study to date, the sample size remains insufficient, for instance, efficacy comparisons among various PD-1 inhibitors were not feasible. Thirdly, due to the retrospective nature of the study, certain molecular markers in patients were not assessed, such as PD-L1 expression and tumor mutation burden (TMB), thus preventing analysis of these markers’ predictive value for treatment efficacy.

In conclusion, our study illustrates the effectiveness of PD-1 inhibitor-based therapy for patients with metastatic gastric cancer. Additionally, we found that carcinoembryonic antigen shows promise as a predictor of immunotherapy efficacy.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Biomedical Research Ethics Committee of West China Hospital of Sichuan University. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because studies utilizing medical records were obtained from previous clinical consultations.

YD: Writing – original draft, Writing – review & editing. JL: Writing – review & editing. SZ: Writing – review & editing. FB: Writing – review & editing.