PD-1-targeted therapeutic strategies for advanced or recurrent gastric and gastroesophageal junction cancer: a systematic review and network meta-analysis

Introduction

Advanced gastric and gastroesophageal junction (GEJ) cancers remain among the most lethal and rapidly progressing malignancies worldwide. Owing to their often-asymptomatic onset, most cases are diagnosed at an advanced stage and are associated with poor survival outcomes, posing major therapeutic challenges (1). Globally, gastric cancer is the fifth most frequently diagnosed malignancy and the third leading cause of cancer-related mortality, accounting for over one million new cases and roughly 769,000 deaths each year (2). The burden of advanced gastric and GEJ cancers continues to increase, especially in populations with high exposure to established risk factors such as Helicobacter pylori infection, dietary patterns, and lifestyle behaviors. Their high morbidity and mortality impose not only severe personal and social consequences but also significant economic strain; the annual global cost of gastric cancer care exceeds 20 billion USD, underscoring its profound burden on healthcare systems (3).

Standard management of advanced gastric and GEJ cancers mainly includes surgery, systemic chemotherapy, and targeted therapy. However, their effectiveness in late-stage disease remains limited, and their toxicity often reduces patient quality of life (4). Chemotherapy provides only modest survival benefits and is frequently associated with considerable adverse effects that impair treatment tolerance. In recent years, immune checkpoint blockade targeting programmed cell death protein 1 (PD-1) has become an important therapeutic approach (5). PD-1 blockade enhances antitumor immunity by interrupting inhibitory checkpoint pathways, reactivating T-cell function, and promoting immune-mediated tumor elimination. Compared with conventional therapies, PD-1 inhibitors can induce more durable clinical responses and often have a more favorable safety profile, reducing the incidence of severe treatment-related toxicities (6). These therapeutic advantages form the basis for investigating PD-1-based regimens in advanced gastric and GEJ cancers.

Several clinical studies have evaluated PD-1 blockade in advanced gastric and GEJ cancers and have consistently demonstrated favorable therapeutic outcomes. The landmark CheckMate-649 trial confirmed the therapeutic benefit of adding nivolumab to chemotherapy, showing a significant survival improvement compared with standard regimens and establishing a new standard for first-line treatment (7). Similarly, the KEYNOTE-062 trial demonstrated that pembrolizumab monotherapy provided sustained clinical benefit and was associated with a more favorable safety profile in patients with PD-L1-positive advanced gastric or GEJ cancers, reinforcing the clinical applicability of immunotherapy in this setting (8). Concurrently, emerging evidence has examined PD-1 blockade combined with other immunomodulatory or targeted therapies, such as anti-angiogenic agents and lymphocyte activation gene 3 (LAG-3) inhibitors, to enhance synergistic antitumor effects (9). Despite these advances, the relative efficacy and safety of PD-1-based regimens remain uncertain because of variations in study design and the lack of direct comparative trials.

Given the inherent limitations of individual clinical trials and traditional meta-analyses, a comprehensive and methodologically robust synthesis of evidence is warranted. Network meta-analysis (NMA), which integrates both direct and indirect evidence across multiple treatment comparisons within a unified analytical model, provides a robust method to address this evidence gap (10). This method enables more reliable estimation of relative efficacy and safety, thereby supporting evidence-based treatment selection. Accordingly, this study sought to systematically evaluate and compare the efficacy and safety of different PD-1-based treatment regimens in patients with unresectable, advanced, or recurrent gastric and GEJ cancers. By addressing this critical evidence gap, the study seeks to inform optimal therapeutic strategies, improve patient outcomes, and alleviate the substantial global burden of these highly aggressive malignancies. We present this article in accordance with the PRISMA reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2512/rc) (11).

Methods

Data sources and literature search

A comprehensive search of PubMed, Embase, the Cochrane Central Register of Controlled Trials, and Web of Science was conducted from database inception to July 31, 2025. The search incorporated both Medical Subject Headings (MeSH) and free-text keywords related to “gastric cancer”, “PD-1”, “immunotherapy”, and “randomized controlled trial”, combined through the Boolean operators “AND” and “OR” (details provided in Table S1). The search was expanded by manually reviewing reference lists of included trials and recent systematic reviews (published within the past five years) to capture additional eligible studies. Two investigators independently screened all titles and abstracts, and the full texts of potentially relevant articles were assessed for eligibility. Any differences in study selection were resolved by discussion or, if needed, adjudicated by a third reviewer.

Study selection

Studies were considered eligible if they satisfied the following inclusion criteria:

• adult patients (≥18 years) with histologically verified gastric or gastroesophageal junction carcinoma that was unresectable, advanced, or recurrent;
• the intervention involved PD-1-based immunotherapy, delivered either as monotherapy or combined with chemotherapy, anti-angiogenic agents, or LAG-3 inhibitors plus chemotherapy;
• the comparator consisted of standard chemotherapy, placebo (with or without chemotherapy), or an alternative PD-1-based regimen that allowed direct or indirect head-to-head comparison within the NMA;
• at least one prespecified clinical endpoint was reported, such as overall survival (OS), progression-free survival (PFS), objective response, disease control, treatment-related adverse events (TRAEs), or severe toxicities (grade ≥3);
• the design was a randomized controlled trial (RCT); and
• the article was published in full-text English.

Studies were excluded if they fulfilled any of the following criteria:

• inclusion of mixed cancer populations without clearly separable data for gastric or GEJ cancer;
• neither the intervention nor the comparator involved PD-1-based immunotherapy, or both treatment arms used identical PD-1 regimens;
• inadequate or unclear description of treatment methods;
• absence of required outcome data or insufficient quantitative information for analysis despite at least four unsuccessful contact attempts to the corresponding authors within six weeks;
• nonrandomized design, including observational, single-arm, or case-report studies; or
• publication limited to conference abstracts, study protocols, or other gray literature.

Two reviewers independently screened all titles and abstracts, with full texts retrieved for studies that met preliminary inclusion criteria. Any discrepancies were resolved through discussion, and a third reviewer adjudicated unresolved cases.

Data collection and extraction

Reference management software (EndNote X9, Clarivate) was employed to compile search results and remove duplicate records. Two investigators independently extracted data using a standardized template. For each eligible study, key information was extracted, including publication details such as the first author and year of publication, study design and location, number of participants per arm, baseline characteristics (e.g., median age, sex distribution, disease setting or prior therapy, and HER2 status when available), details of the interventions (specific PD-1 inhibitor, any combination agents in the experimental arm, and the regimen administered to the control arm), and outcomes of interest, including efficacy and safety endpoints (see Table S2 for a summary of extracted data).

For time-to-event outcomes, including OS and PFS, hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) were extracted. When HRs were not directly reported, they were estimated from Kaplan-Meier curves using established statistical methods. Log (HR) values and their standard errors were calculated for inclusion in the NMA. The previous conversion of medians and interquartile ranges into approximate means and standard deviations was not applied to time-to-event outcomes (12).

When outcome data were not available in the article or its supplementary files, corresponding authors were contacted up to four times over a six-week period to request the missing information. Studies with unavailable or incomplete data were excluded from the quantitative synthesis for that specific endpoint.

In trials containing multiple intervention arms within the same treatment category (for example, two arms evaluating the same PD-1 inhibitor combined with different doses or schedules of another agent), data were pooled into a single intervention node to prevent duplication and maintain the assumption of comparability in the NMA.

Assessment of methodological quality

The methodological quality of each included RCT was independently evaluated by two reviewers using the revised Cochrane tool for assessing risk of bias (RoB 2) (13). This tool assesses potential bias across five key domains: (I) randomization process, including sequence generation and allocation concealment; (II) deviations from intended interventions, such as adherence and imbalance in co-interventions; (III) missing outcome data, addressing completeness of follow-up and management of attrition; (IV) outcome measurement, considering whether assessment might have been influenced by awareness of the assigned intervention or other systematic factors; and (V) the selection of reported findings, including potential selective reporting of outcomes.

Each domain was rated as having a “low risk”, “some concerns”, or “high risk” of bias, and an overall judgment was made in accordance with the RoB 2 framework. Disagreements were resolved through discussion, with a third reviewer consulted if consensus was not achieved.

Data classification and coding

For the NMA, PD-1-based therapeutic regimens from the included trials were systematically grouped into predefined intervention categories:

• PD-1 Mono, defined as single-agent PD-1 inhibitor therapy without combination;
• PD-1 + Chemo, referring to PD-1 inhibitor therapy combined with chemotherapy;
• PD-1/Angio, indicating regimens incorporating a PD-1 inhibitor with an anti-angiogenic agent, such as VEGF-targeted therapy;
• PD-1/LAG-3 + Chemo, describing triple combinations of a PD-1 inhibitor, a LAG-3 inhibitor, and chemotherapy;
• Chemo, representing standard chemotherapy without immunotherapy;
• Placebo + Chemo, denoting chemotherapy combined with placebo as a control; and
• Placebo, referring to placebo or best supportive care without active anticancer treatment.

This categorization enabled comparison among different PD-1-based therapeutic strategies rather than between individual agents.

Statistical analysis

All analyses were conducted in Stata (version 17.0; StataCorp LLC, College Station, TX, USA). A random-effects NMA within a frequentist framework was used to compare the efficacy and safety of various PD-1-based treatment regimens. A common between-study variance was assumed across comparisons under the random-effects model. Time-to-event outcomes (OS and PFS) were analyzed as HRs with corresponding 95% CIs. Dichotomous outcomes, including objective response rate (ORR), disease control rate (DCR), TRAEs, and severe events (grade ≥3), were reported as odds ratios (ORs) with corresponding 95% CIs.

Network geometry was illustrated through network plots to visualize direct and indirect comparisons among treatment regimens. Between-study heterogeneity was evaluated using the I² statistic, with thresholds of 25%, 50%, and 75% indicating low, moderate, and high heterogeneity, respectively. Global inconsistency was formally assessed using the design-by-treatment interaction model. When the global inconsistency test yielded a P value less than 0.05, local inconsistency was further examined using the node-splitting method. Results of the global inconsistency analysis are presented in Figure S1, and node-splitting results are provided in Table S3.

Treatment ranking was derived from the surface under the cumulative ranking curve (SUCRA), with higher scores reflecting superior efficacy or safety. To identify potential small-study effects, comparison-adjusted funnel plots were constructed and tested using Egger’s regression; statistical significance was defined as P<0.05 (14). Prediction interval plots were also generated to account for uncertainty beyond mean effect estimates. All tests were two-sided, and a P value below 0.05 was considered statistically significant.

Results

Study characteristics

The electronic search initially identified 1,160 records. After removing 472 duplicates, 688 unique citations were screened by title and abstract, and 619 were excluded. Full-text review of the remaining 69 articles yielded 20 RCTs, including 9928 patients with unresectable, advanced, or recurrent gastric or GEJ cancer (Figure 1) (7,8,15-32).

Figure 1 PRISMA Flow diagram of the search process for studies. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RCT, randomized controlled trial; WOS, Web of Science.

The included studies were published between 2018 and 2024, with a median publication year of 2022. Sample sizes ranged from 27 to 1,581 participants, with a median of 395. The mean age of participants across trials varied from 55.1 to 67.0 years, with a median of 61.5. HER2 status was reported in 14 trials: 11 enrolled only HER2-negative patients, 1 enrolled exclusively HER2-positive patients, and 2 included both subgroups.

Regarding treatment strategies, 12 studies investigated PD-1 monotherapy, 9 evaluated PD-1 combined with chemotherapy, 2 examined PD-1 plus LAG-3 blockade with chemotherapy, and 1 assessed PD-1 combined with anti-angiogenic therapy. Control groups included chemotherapy alone in 10 studies, placebo plus chemotherapy in 5, and placebo alone in 5. A detailed summary of trial characteristics is presented in Table S2.

Results of the NMA

OS

Seventeen RCTs including 9,191 patients provided HRs for OS and were included in the NMA. The network configuration and distribution of direct comparisons are presented in Figure 2.

Figure 2 Network plot comparing efficacy and safety outcomes: (A) OS; (B) PFS; (C) ORR; (D) DCR; (E) ≥3 TRAEs. DCR, disease control rate; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; TRAEs, treatment-related adverse events.

According to the SUCRA rankings (Table 1), PD-1 plus chemotherapy (PD-1 + Chemo) demonstrated the highest probability of reducing the hazard of death (90.0%), followed by PD-1 monotherapy (87.5%) and chemotherapy alone (52.4%), whereas placebo ranked lowest (8.2%). However, treatment effects were interpreted primarily on the basis of HRs and their corresponding CIs rather than ranking probabilities alone.

Compared with placebo, PD-1 + Chemo significantly reduced the hazard of death (HR, 0.59; 95% CI: 0.48 to 0.73). PD-1 monotherapy was also associated with a significantly lower mortality risk (HR, 0.60; 95% CI: 0.52 to 0.69). Chemotherapy (HR, 0.73; 95% CI: 0.61 to 0.88) and placebo plus chemotherapy (HR, 0.77; 95% CI: 0.62 to 0.96) similarly demonstrated significant reductions in the hazard of death relative to placebo.

When compared with placebo plus chemotherapy, PD-1 + Chemo (HR, 0.77; 95% CI: 0.71 to 0.84) and PD-1 monotherapy (HR, 0.78; 95% CI: 0.66 to 0.93) were associated with significantly lower mortality risk. In addition, PD-1 + Chemo (HR, 0.81; 95% CI: 0.73 to 0.90) and PD-1 monotherapy (HR, 0.82; 95% CI: 0.72 to 0.93) significantly reduced the hazard of death compared with chemotherapy alone. It should be noted that certain treatment nodes were informed by a limited number of trials, and therefore the stability of their ranking positions should be interpreted with caution.

PFS

Seventeen trials including 9,191 patients contributed PFS data to the NMA. The network geometry is illustrated in Figure 2B.

Based on SUCRA rankings (Table 2), PD-1 + Chemo showed the highest probability of reducing the hazard of disease progression or death (93.2%), followed by PD-1 combined with anti-angiogenic therapy (PD-1 + Angio) (71.2%) and chemotherapy alone (64.0%), whereas placebo ranked lowest (1.6%). As with OS, interpretation was grounded primarily in the magnitude and precision of the HRs.

Compared with placebo, PD-1 + Chemo significantly reduced the hazard of progression or death (HR, 0.40; 95% CI: 0.27 to 0.58). PD-1 + Angio (HR, 0.45; 95% CI: 0.21 to 0.96), chemotherapy (HR, 0.51; 95% CI: 0.36 to 0.71), placebo plus chemotherapy (HR, 0.55; 95% CI: 0.37 to 0.83), and PD-1 monotherapy (HR, 0.60; 95% CI: 0.46 to 0.78) were all associated with significantly lower progression risk relative to placebo.

Furthermore, PD-1 + Chemo significantly reduced the hazard of progression compared with PD-1 monotherapy (HR, 0.67; 95% CI: 0.50 to 0.88), placebo plus chemotherapy (HR, 0.72; 95% CI: 0.58 to 0.88), and chemotherapy alone (HR, 0.78; 95% CI: 0.63 to 0.97).

ORR

All 20 RCTs involving 10,194 patients reported ORR. The network structure is illustrated in Figure 2C.

SUCRA rankings (Figure 3A) indicated that PD-1 + Chemo had the highest probability of improving ORR (85.8%), followed by PD-1 + LAG-3 + Chemo (84.0%) and Placebo + Chemo (71.0%). Placebo ranked lowest (0.1%). Nevertheless, conclusions were based primarily on ORs and CIs rather than on ranking hierarchy alone.

Figure 3 SUCRA probability ranking plot of efficacy and safety outcomes: (A) ORR; (B) DCR; (C) ≥3 TRAEs. DCR, disease control rate; ORR, objective response rate; PD-1, programmed cell death protein 1; SUCRA, surface under the cumulative ranking curve; TRAEs, treatment-related adverse events.

As shown in Table 3, all active regimens significantly increased ORR compared with placebo, including PD-1 + Chemo (OR, 55.02; 95% CI: 13.22 to 228.96), PD-1 + LAG-3 + Chemo (OR, 58.18; 95% CI: 10.80 to 313.47), Placebo + Chemo (OR, 44.44; 95% CI: 10.00 to 197.51), Chemo (OR, 24.82; 95% CI: 6.26 to 98.51), PD-1 + Angio (OR, 16.84; 95% CI: 1.84 to 154.46), and PD-1 monotherapy (OR, 20.71; 95% CI: 5.78 to 74.22).

PD-1 + Chemo was also superior to Chemo alone (OR, 2.22; 95% CI: 1.28 to 3.82) and PD-1 monotherapy (OR, 2.66; 95% CI: 1.41 to 5.02).

DCR

Sixteen RCTs including 7,961 patients reported DCR. The network geometry is shown in Figure 2D.

According to SUCRA rankings (Figure 3B), PD-1 + LAG-3 + Chemo ranked highest (97.2%), followed by PD-1 + Chemo (84.4%) and Placebo + Chemo (63.5%). Placebo ranked lowest (2.9%). Given that some treatment nodes were supported by relatively few trials, these rankings should be interpreted in conjunction with effect estimates and CIs.

As summarized in Table 4, PD-1 + LAG-3 + Chemo significantly improved DCR compared with Chemo (OR, 2.37; 95% CI: 1.15 to 4.86), PD-1 monotherapy (OR, 3.90; 95% CI: 1.84 to 8.24), PD-1 + Angio (OR, 4.68; 95% CI: 1.04 to 21.05), and placebo (OR, 9.44; 95% CI: 4.10 to 21.71).

PD-1 + Chemo was superior to Placebo + Chemo (OR, 1.30; 95% CI: 1.04 to 1.63), Chemo (OR, 1.63; 95% CI: 1.15 to 2.30), PD-1 monotherapy (OR, 2.67; 95% CI: 1.77 to 4.03), and placebo (OR, 6.47; 95% CI: 3.75 to 11.19).

Severe (grade ≥3) TRAEs

Sixteen RCTs involving 7,856 patients reported grade ≥3 TRAEs. The network configuration is shown in Figure 2E and Table 5.

SUCRA rankings (Figure 3C) indicated that placebo was the safest regimen (99.9%), followed by PD-1 monotherapy (82.8%) and PD-1 + Angio (58.0%). PD-1 + Chemo ranked lowest (10.6%). Safety interpretation focused primarily on ORs and their precision rather than on ranking positions alone.

Placebo was associated with significantly fewer grade ≥3 TRAEs than PD-1 monotherapy (OR, 0.40; 95% CI: 0.22 to 0.73), PD-1 + Angio (OR, 0.11; 95% CI: 0.02 to 0.52), Chemo (OR, 0.07; 95% CI: 0.03 to 0.14), Placebo + Chemo (OR, 0.05; 95% CI: 0.02 to 0.11), PD-1 + LAG-3 + Chemo (OR, 0.04; 95% CI: 0.02 to 0.10), and PD-1 + Chemo (OR, 0.04; 95% CI: 0.02 to 0.08).

PD-1 monotherapy was also associated with significantly lower toxicity compared with Chemo (OR, 0.17; 95% CI: 0.12 to 0.24), Placebo + Chemo (OR, 0.13; 95% CI: 0.08 to 0.22), PD-1 + LAG-3 + Chemo (OR, 0.10; 95% CI: 0.05 to 0.20), and PD-1 + Chemo (OR, 0.10; 95% CI: 0.07 to 0.15). Chemo was associated with fewer severe TRAEs than PD-1 + Chemo (OR, 0.60; 95% CI: 0.43 to 0.83).

Risk of bias and publication bias

Among the 20 RCTs included in the analysis, 10 were assessed as having an overall low risk of bias, 9 were judged to have some concerns, and 1 was rated as high risk of bias. In the domain of the randomization process, 14 trials were classified as low risk, 5 were assessed as having some concerns, and 1 was considered high risk. Regarding deviations from intended interventions, 17 trials were evaluated as low risk, 2 presented some concerns, and 1 was rated as high risk. For missing outcome data, 14 trials were judged to be at low risk, 5 had some concerns, and 1 was classified as high risk. In the domain of outcome measurement, 19 trials were assessed as low risk and 1 was considered to have some concerns. All 20 trials were judged to have low risk of bias for selective reporting of results. Detailed assessments for each domain are presented in Table S4.

Potential publication bias was evaluated using funnel plots, which are provided in Figure S2. Visual inspection of the plots demonstrated varying degrees of symmetry around the vertical axis. In particular, Figures S1,S2 exhibited a certain degree of asymmetry, suggesting the possibility of publication bias or small study effects. Egger’s regression test indicated statistically significant evidence of publication bias for ORR, with P<0.05. For all other outcomes, Egger’s test yielded P values greater than 0.05, suggesting no statistically significant evidence of publication bias in the overall analyses.

Discussion

This NMA incorporated 20 RCTs involving 9928 patients with unresectable advanced or recurrent gastric and GEJ cancer. It provides a comprehensive comparison of the efficacy and safety of various PD-1-based therapeutic strategies in this population. The findings of this analysis can be summarized in three main points. First, among all evaluated regimens, PD-1 combined with chemotherapy showed the most consistent and favorable results across all major efficacy outcomes, including OS, PFS, and tumor response indicators. Importantly, this conclusion is based primarily on the magnitude and precision of HRs and ORs rather than on SUCRA ranking probabilities alone. PD-1 combined with chemotherapy demonstrated statistically significant reductions in mortality and progression risk across multiple comparisons, with CIs that excluded the null value. While SUCRA values supported this hierarchical pattern, they were interpreted as complementary evidence rather than determinative criteria. Second, PD-1 monotherapy achieved meaningful improvement in survival but demonstrated less consistent effects on disease control and tumor response compared with combination regimens. These results suggest that while single-agent immunotherapy can provide durable benefit for selected patients, it may not achieve the same depth or breadth of tumor suppression as combination approaches that engage multiple therapeutic pathways. Third, in terms of safety, PD-1 monotherapy and non-chemotherapy combinations, particularly those including anti-angiogenic agents, were associated with the lowest rates of severe treatment-related toxicity, whereas PD-1 combined with chemotherapy showed the highest toxicity burden. These observations emphasize the importance of balancing treatment efficacy and tolerability, especially in patients with limited functional reserve. For such patients, PD-1 monotherapy or non-cytotoxic combinations may offer safer yet clinically meaningful options. These findings offer a robust, evidence-based foundation for refining PD-1-oriented therapeutic strategies in advanced gastric and GEJ cancers. They highlight the importance of tailoring treatment intensity to individual patient capacity and clinical objectives, ensuring an optimal balance between efficacy and safety in practice.

The combination of PD-1 inhibition and chemotherapy constitutes a rational and increasingly implemented approach for managing unresectable advanced or recurrent gastric and GEJ cancers. This approach integrates immune checkpoint blockade—which reinvigorates cytotoxic T-cell activity by inhibiting the PD-1/PD-L1 axis—with cytotoxic chemotherapy that induces immunogenic cell death and modulates the tumor microenvironment to enhance antigen presentation and T-cell infiltration (33). In our NMA, PD-1 + chemotherapy consistently emerged as the most effective strategy across all major efficacy endpoints, including OS, PFS, ORR, DCR. This uniform superiority across both survival- and response-related outcomes highlights the synergistic potential of combining cytotoxic and immunotherapeutic mechanisms in this clinical context. Although SUCRA rankings placed this regimen at the top across efficacy endpoints, the interpretation was anchored in effect estimates and CIs. Clinically, these endpoints—particularly OS and PFS—are vital metrics for assessing therapeutic value in patients with advanced gastric and GEJ cancers, who often present with aggressive disease, limited treatment options, and high symptom burden. ORR and DCR, while secondary, offer crucial insights into tumor responsiveness and disease stabilization, especially in palliative settings where durable control can meaningfully impact quality of life (34). The observed superiority of PD-1 + chemotherapy across all four metrics suggests not only an extension of life expectancy but also enhanced immediate tumor control—both of which are highly relevant to this population. Our results align with and build upon prior phase III evidence. In the CheckMate-649 trial, the addition of nivolumab to chemotherapy significantly enhanced OS and PFS compared with chemotherapy alone in the first-line treatment of HER2-negative advanced gastric and GEJ cancers (22). Our analysis, incorporating a broader range of trials and treatment regimens, further reinforces the central role of PD-1 plus chemotherapy as a first-line option and uniquely quantifies its relative superiority among competing strategies. However, while previous studies typically evaluate PD-1 combinations in head-to-head comparisons against standard chemotherapy, our network framework allows for simultaneous indirect comparisons across all major PD-1-based regimens, providing a more nuanced therapeutic hierarchy.

The superior efficacy of PD-1 combined with chemotherapy across multiple clinical endpoints can be explained by several complementary biological mechanisms. Chemotherapy reduces immunosuppressive cell populations, including myeloid-derived suppressor cells and regulatory T cells, thereby enhancing the immunogenic potential of the tumor microenvironment (35). It also triggers immunogenic tumor cell death, leading to the release of tumor antigens that drive dendritic cell activation and enhance antigen presentation to effector T cells (36). This process broadens the repertoire of tumor-specific T cells and amplifies the antitumor response initiated by PD-1 blockade. Furthermore, chemotherapy can upregulate PD-L1 expression on tumor cells, increasing their susceptibility to immune checkpoint inhibition (37). These mechanisms provide a biologically coherent explanation for the superior outcomes observed with PD-1 plus chemotherapy, supporting its role as a mechanistically synergistic and clinically effective regimen for patients with advanced gastric and gastroesophageal junction cancers.

While PD-1 plus chemotherapy demonstrated the most consistent efficacy across all endpoints, the findings for PD-1 monotherapy revealed a distinct response pattern, warranting further exploration. In our NMA, PD-1 monotherapy ranked second only to PD-1 + chemotherapy in OS, yet its relative performance was substantially lower in PFS, ORR, and DCR, with SUCRA values of 47.0%, 38.7%, and 47.1%, respectively. This divergence suggests that PD-1 monotherapy may offer survival prolongation without achieving the same magnitude of tumor shrinkage or disease stabilization observed with combination strategies. These findings align with but also extend insights from prior clinical trials. For example, the KEYNOTE-061 trial, which evaluated pembrolizumab monotherapy versus paclitaxel in previously treated gastric cancer, failed to demonstrate a significant improvement in PFS or ORR, despite showing a favorable trend in OS among patients with high PD-L1 expression (CPS ≥10) (30). Similarly, in the KEYNOTE-062 trial, pembrolizumab monotherapy yielded non-inferior OS compared to chemotherapy in PD-L1-positive patients but was associated with lower ORR and shorter median PFS (8). Our findings are consistent with this pattern and highlight, from a network-wide perspective, the therapeutic dichotomy between survival benefit and tumor response when PD-1 inhibitors are used as monotherapy.

The observed discrepancy between improved OS and relatively poor PFS and response rates in PD-1 monotherapy likely reflects the unique immunodynamic properties of immune checkpoint blockade. Unlike cytotoxic therapies, PD-1 inhibitors rely on T-cell-mediated immune activation, which may result in delayed antitumor effects—a phenomenon often described as “pseudoprogression” or delayed clinical response. Consequently, initial radiographic assessments may underestimate true therapeutic activity, particularly when relying on conventional RECIST criteria (38). Moreover, survival benefit from PD-1 monotherapy may be driven by a subset of immunologically “hot” tumors—characterized by high PD-L1 expression, microsatellite instability (MSI-H), or high tumor mutational burden (TMB)—in which durable responses are more likely (39). In contrast, tumors with low immunogenicity may not respond rapidly or at all, leading to inferior short-term disease control metrics such as PFS or ORR, despite the potential for eventual long-term benefit. Additionally, the discrepancy may reflect patient selection in monotherapy trials, which often include individuals with better performance status or lower disease burden—factors associated with longer survival independent of immediate treatment effect. Therefore, while PD-1 monotherapy may be appropriate for selected patients who are unable to tolerate chemotherapy or exhibit predictive biomarkers of response, its limited capacity to deliver broad and consistent tumor control must be carefully considered in clinical decision-making.

Treatment safety is a critical component in the management of advanced gastric and GEJ cancers, where many patients face impaired functional status, comorbidities, and limited treatment tolerance. In this context, understanding the toxicity profiles of PD-1-based regimens are essential for informed therapeutic decision-making. Our analysis demonstrated that PD-1 monotherapy and combinations excluding chemotherapy—such as PD-1 plus anti-angiogenic therapy—were associated with the most favorable safety profiles. In contrast, PD-1 combined with chemotherapy consistently showed the highest incidence of severe TRAEs, indicating a substantially greater toxicity burden relative to other regimens. These findings align with prior clinical trials in which PD-1 monotherapy was associated with fewer high-grade toxicities compared to chemotherapy-based regimens. While immunotherapy-related adverse events, such as immune-mediated endocrinopathies or colitis, can occur with PD-1 blockade, they are generally manageable and less frequent when used as monotherapy (40). The addition of chemotherapy, however, introduces cumulative toxicities that include hematologic suppression, gastrointestinal distress, and neuropathy, which may be amplified by the immunomodulatory effects of PD-1 inhibitors.

From a mechanistic perspective, chemotherapy exerts systemic cytotoxic effects that not only compromise normal tissue integrity but may also potentiate immune activation when administered concurrently with PD-1 inhibitors. This dual insult—direct cytotoxicity and checkpoint inhibition—can lead to overlapping toxicities and heightened immune-related inflammation (41). Moreover, combination regimens typically require more intensive treatment schedules, increasing drug exposure and cumulative toxicity risk. In contrast, PD-1 monotherapy primarily relies on the host immune response to mediate antitumor activity, and its toxicity profile is more predictable and immunologically confined. Taken together, these findings reinforce the importance of tailoring treatment based not only on efficacy but also on safety considerations. For patients with limited physiological reserve or contraindications to chemotherapy, PD-1 monotherapy or other non-cytotoxic combinations may represent a more tolerable yet still effective option. Conversely, in patients with robust performance status and high tumor burden, the superior efficacy of PD-1 plus chemotherapy may justify the increased risk of toxicity, provided that adverse events are closely monitored and managed. This nuanced understanding of safety differentials among PD-1-based regimens is critical to optimizing individualized treatment strategies in advanced gastric and GEJ cancers.

There are several methodological strengths in this study. First, to our knowledge, it is the most comprehensive NMA to date assessing the efficacy and safety of multiple PD-1–based treatment strategies in patients with unresectable advanced or recurrent gastric and gastroesophageal junction cancers. By integrating both direct and indirect evidence from RCTs, it provides a robust hierarchical comparison of therapeutic regimens even in the absence of direct head-to-head studies. Second, the analysis adopted a frequentist network meta-analytic framework with SUCRA-based ranking, which enabled structured comparison of clinical benefit and toxicity across treatments. However, SUCRA values were interpreted as supportive measures of relative ranking rather than as independent determinants of superiority. Primary interpretation was grounded in effect estimates, corresponding CIs, and their clinical relevance. However, several limitations should be acknowledged. First, the validity of any NMA depends on the assumption of transitivity across included trials. Variations in baseline characteristics, biomarker profiles such as PD-L1 expression and HER2 status, and geographic diversity may contribute to heterogeneity and influence treatment estimates. In addition, certain treatment nodes were informed by a limited number of trials, which may affect the stability of ranking positions and warrants cautious interpretation of SUCRA-based hierarchies. Second, some studies lacked complete outcome reporting, particularly for safety endpoints, which may have introduced selective reporting bias. Although data imputation was avoided and authors were contacted for clarification, some missing data could not be retrieved. Third, subgroup-level information, including tumor PD-L1 expression, microsatellite instability, and prior treatment exposure, was not consistently reported, limiting further stratified analyses that could refine treatment recommendations. Despite these limitations, the present study offers clinically meaningful insights into the comparative benefit-risk profiles of PD-1 based regimens and provides a valuable framework for evidence-based treatment selection in advanced gastric and gastroesophageal junction cancers. Future prospective and real-world studies incorporating biomarker-guided stratification and patient-centered outcomes are needed to validate and extend these findings.

Conclusions

This NMA shows that PD-1 combined with chemotherapy provides the most effective and balanced therapeutic option for advanced or recurrent gastric and gastroesophageal junction cancers. PD-1 monotherapy offers survival benefit but limited tumor control, whereas non-cytotoxic combinations achieve better safety profiles. These results support PD-1 plus chemotherapy as the preferred regimen for patients who can tolerate higher toxicity, and suggest PD-1 monotherapy or targeted combinations as safer, individualized alternatives. Future biomarker-driven trials are needed to refine patient selection and guide precision immunotherapy in this population.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2512/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2512/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-aw-2512/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Smyth EC, Nilsson M, Grabsch HI, et al. Gastric cancer. Lancet 2020;396:635-48. [Crossref] [PubMed]
2. Lin JL, Lin JX, Lin GT, et al. Global incidence and mortality trends of gastric cancer and predicted mortality of gastric cancer by 2035. BMC Public Health 2024;24:1763. [Crossref] [PubMed]
3. Song Y, Liu X, Cheng W, et al. The global, regional and national burden of stomach cancer and its attributable risk factors from 1990 to 2019. Sci Rep 2022;12:11542. [Crossref] [PubMed]
4. Guan WL, He Y, Xu RH. Gastric cancer treatment: recent progress and future perspectives. J Hematol Oncol 2023;16:57. [Crossref] [PubMed]
5. Yuan H, Bao M, Chen M, et al. Advances in Immunotherapy and Targeted Therapy for Gastric Cancer: A Comprehensive Review. Br J Hosp Med (Lond) 2025;86:1-24. [Crossref] [PubMed]
6. Lin X, Kang K, Chen P, et al. Regulatory mechanisms of PD-1/PD-L1 in cancers. Mol Cancer 2024;23:108. [Crossref] [PubMed]
7. Janjigian YY, Shitara K, Moehler M, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone for advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma (CheckMate 649): a randomised, open-label, phase 3 trial. Lancet 2021;398:27-40. [Crossref] [PubMed]
8. Shitara K, Van Cutsem E, Bang YJ, et al. Efficacy and Safety of Pembrolizumab or Pembrolizumab Plus Chemotherapy vs Chemotherapy Alone for Patients With First-line, Advanced Gastric Cancer: The KEYNOTE-062 Phase 3 Randomized Clinical Trial. JAMA Oncol 2020;6:1571-80. [Crossref] [PubMed]
9. Qiu X, Yu Z, Lu X, et al. PD-1 and LAG-3 dual blockade: emerging mechanisms and potential therapeutic prospects in cancer. Cancer Biol Med 2024;21:970-6. [Crossref] [PubMed]
10. Leucht S, Chaimani A, Cipriani AS, et al. Network meta-analyses should be the highest level of evidence in treatment guidelines. Eur Arch Psychiatry Clin Neurosci 2016;266:477-80. [Crossref] [PubMed]
11. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [Crossref] [PubMed]
12. Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Hoboken: J Wiley-Blackwell; 2008.
13. Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [Crossref] [PubMed]
14. Chaimani A, Higgins JP, Mavridis D, et al. Graphical tools for network meta-analysis in STATA. PLoS One 2013;8:e76654. [Crossref] [PubMed]
15. Boku N, Satoh T, Ryu MH, et al. Nivolumab in previously treated advanced gastric cancer (ATTRACTION-2): 3-year update and outcome of treatment beyond progression with nivolumab. Gastric Cancer 2021;24:946-58. [Crossref] [PubMed]
16. Chao J, Fuchs CS, Shitara K, et al. Assessment of Pembrolizumab Therapy for the Treatment of Microsatellite Instability-High Gastric or Gastroesophageal Junction Cancer Among Patients in the KEYNOTE-059, KEYNOTE-061, and KEYNOTE-062 Clinical Trials. JAMA Oncol 2021;7:895-902. [Crossref] [PubMed]
17. Chen LT, Satoh T, Ryu MH, et al. A phase 3 study of nivolumab in previously treated advanced gastric or gastroesophageal junction cancer (ATTRACTION-2): 2-year update data. Gastric Cancer 2020;23:510-9. [Crossref] [PubMed]
18. Chung HC, Kang YK, Chen Z, et al. Pembrolizumab versus paclitaxel for previously treated advanced gastric or gastroesophageal junction cancer (KEYNOTE-063): A randomized, open-label, phase 3 trial in Asian patients. Cancer 2022;128:995-1003. [Crossref] [PubMed]
19. Fuchs CS, Özgüroğlu M, Bang YJ, et al. Pembrolizumab versus paclitaxel for previously treated PD-L1-positive advanced gastric or gastroesophageal junction cancer: 2-year update of the randomized phase 3 KEYNOTE-061 trial. Gastric Cancer 2022;25:197-206. [Crossref] [PubMed]
20. Hegewisch-Becker S, Mendez G, Chao J, et al. First-Line Nivolumab and Relatlimab Plus Chemotherapy for Gastric or Gastroesophageal Junction Adenocarcinoma: The Phase II RELATIVITY-060 Study. J Clin Oncol 2024;42:2080-93. [Crossref] [PubMed]
21. Janjigian YY, Kawazoe A, Yañez P, et al. The KEYNOTE-811 trial of dual PD-1 and HER2 blockade in HER2-positive gastric cancer. Nature 2021;600:727-30. [Crossref] [PubMed]
22. Kang YK, Chen LT, Ryu MH, et al. Nivolumab plus chemotherapy versus placebo plus chemotherapy in patients with HER2-negative, untreated, unresectable advanced or recurrent gastric or gastro-oesophageal junction cancer (ATTRACTION-4): a randomised, multicentre, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2022;23:234-47. [Crossref] [PubMed]
23. Kato K, Satoh T, Muro K, et al. A subanalysis of Japanese patients in a randomized, double-blind, placebo-controlled, phase 3 trial of nivolumab for patients with advanced gastric or gastro-esophageal junction cancer refractory to, or intolerant of, at least two previous chemotherapy regimens (ONO-4538-12, ATTRACTION-2). Gastric Cancer 2019;22:344-54. [Crossref] [PubMed]
24. Kim JH, Ryu MH, Park YS, et al. Predictive biomarkers for the efficacy of nivolumab as ≥ 3(rd)-line therapy in patients with advanced gastric cancer: a subset analysis of ATTRACTION-2 phase III trial. BMC Cancer 2022;22:378. [Crossref] [PubMed]
25. Liu T, Bai Y, Lin X, et al. First-line nivolumab plus chemotherapy vs chemotherapy in patients with advanced gastric, gastroesophageal junction and esophageal adenocarcinoma: CheckMate 649 Chinese subgroup analysis. Int J Cancer 2023;152:749-60. [Crossref] [PubMed]
26. Qiu MZ, Oh DY, Kato K, et al. Tislelizumab plus chemotherapy versus placebo plus chemotherapy as first line treatment for advanced gastric or gastro-oesophageal junction adenocarcinoma: RATIONALE-305 randomised, double blind, phase 3 trial. BMJ 2024;385:e078876. [Crossref] [PubMed]
27. Rha SY, Oh DY, Yañez P, et al. Pembrolizumab plus chemotherapy versus placebo plus chemotherapy for HER2-negative advanced gastric cancer (KEYNOTE-859): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol 2023;24:1181-95. [Crossref] [PubMed]
28. Satake H, Lee KW, Chung HC, et al. Pembrolizumab or pembrolizumab plus chemotherapy versus standard of care chemotherapy in patients with advanced gastric or gastroesophageal junction adenocarcinoma: Asian subgroup analysis of KEYNOTE-062. Jpn J Clin Oncol 2023;53:221-9. [Crossref] [PubMed]
29. Satoh T, Kang YK, Chao Y, et al. Exploratory subgroup analysis of patients with prior trastuzumab use in the ATTRACTION-2 trial: a randomized phase III clinical trial investigating the efficacy and safety of nivolumab in patients with advanced gastric/gastroesophageal junction cancer. Gastric Cancer 2020;23:143-53. [Crossref] [PubMed]
30. Shitara K, Özgüroğlu M, Bang YJ, et al. Pembrolizumab versus paclitaxel for previously treated, advanced gastric or gastro-oesophageal junction cancer (KEYNOTE-061): a randomised, open-label, controlled, phase 3 trial. Lancet 2018;392:123-33. [Crossref] [PubMed]
31. Wei Q, Xu X, Li J, et al. Apatinib Plus Toripalimab (Anti-PD1 Therapy) as Second-Line Therapy in Patients With Advanced Gastric or Esophagogastric Junction Cancer: Results From a Randomized, Open-Label Phase II Study. Oncologist 2024;29:364-e578. [Crossref] [PubMed]
32. Xu J, Jiang H, Pan Y, et al. Sintilimab Plus Chemotherapy for Unresectable Gastric or Gastroesophageal Junction Cancer: The ORIENT-16 Randomized Clinical Trial. JAMA 2023;330:2064-74. [Crossref] [PubMed]
33. Pu W, Li S, Zhang J, et al. The efficacy and safety of PD-1/PD-L1 inhibitors in combination with chemotherapy as a first-line treatment for unresectable, locally advanced, HER2-negative gastric or gastroesophageal junction cancer: a meta-analysis of randomized controlled trials. Front Immunol 2025;16:1566939. [Crossref] [PubMed]
34. Janjigian YY, Ajani JA, Moehler M, et al. First-Line Nivolumab Plus Chemotherapy for Advanced Gastric, Gastroesophageal Junction, and Esophageal Adenocarcinoma: 3-Year Follow-Up of the Phase III CheckMate 649 Trial. J Clin Oncol 2024;42:2012-20. [Crossref] [PubMed]
35. Jiang R, Cheng X, Li P, et al. Plasma circulating tumor DNA unveils the efficacy of PD-1 inhibitors and chemotherapy in advanced gastric cancer. Sci Rep 2024;14:14027. [Crossref] [PubMed]
36. Jou E, Chaudhury N, Nasim F. Novel therapeutic strategies targeting myeloid-derived suppressor cell immunosuppressive mechanisms for cancer treatment. Explor Target Antitumor Ther 2024;5:187-207. [Crossref] [PubMed]
37. Liu P, Zhao L, Kroemer G, et al. Conventional type 1 dendritic cells (cDC1) in cancer immunity. Biol Direct 2023;18:71. [Crossref] [PubMed]
38. Park HJ, Kim KW, Pyo J, et al. Incidence of Pseudoprogression during Immune Checkpoint Inhibitor Therapy for Solid Tumors: A Systematic Review and Meta-Analysis. Radiology 2020;297:87-96. [Crossref] [PubMed]
39. Cristescu R, Aurora-Garg D, Albright A, et al. Tumor mutational burden predicts the efficacy of pembrolizumab monotherapy: a pan-tumor retrospective analysis of participants with advanced solid tumors. J Immunother Cancer 2022;10:e003091. [Crossref] [PubMed]
40. Li ZQ, Yan HC, Gu JJ, et al. Comparative efficacy and safety of PD-1/PD-L1 Inhibitors versus platinum-based chemotherapy for the first-line treatment of advanced non-small cell lung cancer: a meta analysis of randomized controlled trials. Pharmacol Res 2020;160:105194. [Crossref] [PubMed]
41. Yao W, Zhao X, Gong Y, et al. Impact of the combined timing of PD-1/PD-L1 inhibitors and chemotherapy on the outcomes in patients with refractory lung cancer. ESMO Open 2021;6:100094. [Crossref] [PubMed]