Evaluation of neoadjuvant immunotherapy in resectable gastric/gastroesophageal junction tumors: a meta-analysis and systematic review

Background Neoadjuvant therapy for resectable gastric cancer/gastroesophageal junction tumors is progressing slowly. Although immunotherapy for advanced gastric cancer/gastroesophageal junction tumors has made great progress, the efficacy and safety of neoadjuvant immunotherapy for locally resectable gastric cancer/gastroesophageal junction tumors have not been clearly demonstrated. Here, we conducted a systematic review and meta-analysis to assess the efficacy and safety of neoadjuvant immunotherapy and advance the current research. Methods Original articles describing the safety and efficacy of neoadjuvant immunotherapy for resectable gastric cancer/gastroesophageal junction tumors published up until October 15, 2023 were retrieved from PubMed, Embase, the Cochrane Library, and other major databases. The odds ratios (OR) and 95% confidence intervals (CIs) were calculated for heterogeneity and subgroup analysis. Results A total of 1074 patients from 33 studies were included. The effectiveness of neoadjuvant immunotherapy was mainly evaluated using pathological complete remission (PCR), major pathological remission (MPR), and tumor regression grade (TRG). Among the included patients, 1015 underwent surgical treatment and 847 achieved R0 resection. Of the patients treated with neoadjuvant immunotherapy, 24% (95% CI: 19%–28%) achieved PCR and 49% (95% CI: 38%–61%) achieved MPR. Safety was assessed by a surgical resection rate of 0.89 (95% CI: 85%–93%), incidence of ≥ 3 treatment-related adverse events (TRAEs) of 28% (95% CI: 17%–40%), and incidence of ≥ 3 immune-related adverse events (irAEs) of 19% (95% CI: 11%–27%). Conclusion Neoadjuvant immunotherapy, especially neoadjuvant dual-immunotherapy combinations, is effective and safe for resectable gastric/gastroesophageal junction tumors in the short term. Nevertheless, further multicenter randomized trials are required to demonstrate which combination model is more beneficial. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=358752, identifier CRD42022358752.


Introduction
Gastric cancer (GC), representing the fifth most common malignancy and one of the third most common cancer-related causes of death, is responsible for approximately 33% of all cancerrelated deaths worldwide, and the highest mortality and incidence rates are found in East Asia (1).The junction of the esophagus and stomach (termed the gastroesophageal junction [GEJ]) is the area of transition between the esophageal squamous epithelium and the pancreatic gland columnar epithelium.Adenocarcinoma of the esophagogastric junction (AEG) is a tumor type with different biological behavior and clinical features from squamous cell carcinoma and gastric adenocarcinoma.AEG is separated into three types based on the distance from the tumor center to the GEJa classification first proposed by Siewert in 1999 (2)(3)(4)(5).The majority of these tumors are histologically adenocarcinomas.The main treatment option for early and locally advanced tumors is surgical resection.Nevertheless, there is only a 10%-33% 5-year overall survival (OS) rate for patients treated with surgery alone (6)(7)(8).Therefore, it is a serious challenge to treat these patients appropriately and improve their survival rates.
In gastric and gastroesophageal junction cancers, neoadjuvant therapy is a well-established practice for reducing tumor burden, assessing tumor response preoperatively, and improving OS (9).While the landmark phase III MAGIC trial established perioperative ECF/ECX chemotherapy for resectable G/GEJ cancers as the standard of care period (10), the recently published CROSS trial established neoadjuvant radiotherapy as a valid treatment option for esophageal and GEJ tumors (7,11,12).To date, three completed randomized trials have directly compared neoadjuvant radiotherapy (NACRT) with neoadjuvant chemotherapy (NAC) and found that NACRT increased pathological complete remission rates and margin-negative resection rates without increasing OS (13)(14)(15).Moreover, a previous meta-analysis highlighted that the incidence of > 3 treatment-related adverse events (TRAEs) in the NAC group was as high as 25.7% in patients with resectable gastric cancer (16).It was also noted that the treatment-related complications were similar between the NAC and NACRT groups, while the postoperative complications were more severe in the NACRT group (15).As a result, there is promise that a new, more effective, neoadjuvant regimen with reduced toxicity will improve clinical outcomes in patients with G/GEJ tumors without increasing the incidence of adverse events.
As medicine continues to advance, immunotherapy has begun to gain approval in the clinical setting, thus changing the landscape of oncology treatment, with satisfactory results observed for the treatment of melanoma and non-small-cell lung cancer (17,18).Immune checkpoint inhibitors, mainly consisting of programmed cell death protein 1 (PD-1) and programed cell death ligand 1 (PD-L1), have revolutionized the treatment of malignant tumors.PD-1 is expressed by activated lymphocytes and is expressed in combination with ligands, including PD-L1.By blocking the immune response and promoting immune escape, it further promotes the development of various malignancies and disease progression (19,20).A recent study demonstrated that, compared to adjuvant immunotherapy, neoadjuvant immunotherapy showed significantly high therapeutic efficacy in eradicating metastases in a preclinical mouse model through systemic expansion of tumorspecific CD8 + T cells in peripheral blood and organs (21).Based on these findings, it is possible that neoadjuvant PD-1 blockers activate effective systemic immunity and consequently obliterate residual micrometastases after surgical resection of the primary tumor.Moreover, conventional chemotherapy has been shown to potentiate tumor antigenicity, which interferes with suppressive immune pathways and increases effector T cell responses (22).The efficacy of PD-1 receptor blockers can be enhanced by combining them with appropriate chemotherapeutic agents, especially for less immunogenic tumors with poor chemotherapy sensitivity (23,24).
A multi-study meta-analysis will provide a more promising alternative to several neoadjuvant treatment strategies and enhance self-confidence regarding future clinical trials for neoadjuvant immunotherapy.The purpose of this meta-analysis, based on existing data, is to provide evidence for the efficacy and safety of neoadjuvant immunotherapy for resectable G/GEJ tumors and to offer options for further treatment of locally advanced G/GEJ tumors with better survival benefits in the future.

Materials and methods
This systematic review and meta-analysis was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.The trial protocol can b e fou nd in PR OSPER O ( registration number: CRD42022358752.

Search strategy and study selection
Using PubMed, Embase, and the Cochrane Library, we performed a comprehensive search for articles on neoadjuvant immunotherapy for resectable G/GEJ tumors published in English up to October 15, 2023.We also conducted a retrospective search for the latest unpublished data on clinical trials of neoadjuvant immunotherapy for resectable G/GEJ tumors conducted at international oncology congresses, such as ASCO and ESMO, up to October 15, 2023.These medical databases were searched for terms such as "gastric cancer/gastroesophageal junction tumor," "neoadjuvant therapy," and "immunotherapy" (which includes all currently known ICIs).The full reference list of all searched texts was filtered to further identify potentially relevant studies.

Selection criteria and data extraction
Publications that met the following criteria were selected: 1. Publications that reported tissue-confirmed resectable G/GEJ tumors; 2. Clinical trials currently applied in clinical practice or registered; 3. Reports containing comprehensive scenarios, patient information, and a minimum critical clinical outcome with respect to PCR, MPR, TRG, TRAEs, irAEs, surgical complication rates, surgical resection rates, and operative delay rates.Publications that met the following criteria were excluded: 1. Presence of inoperable or advanced metastatic disease; 2. At no point did the study focus on MPR, PCR, TRG, TRAEs, irAEs, or surgical resection rates; 3. Enrollment of fewer than ten patients; 4. A lack of available effective data to assess the effectiveness and safety of neoadjuvant immunotherapy in combination with chemotherapy; 5.The existence of duplicate publications.Two investigators (JCW and ZXZ) independently selected a list of the retrieved publications.Following a review by a senior researcher (YHZ), any disagreements were resolved by discussion and consensus between the two reviewers, which was followed by a search of the full text to assess its eligibility.We searched and browsed each of the citations of the included studies to ensure that no relevant studies were missing.

Data abstraction
In the present meta-analysis, both investigators extracted the data separately.The following information was recorded for each study: first author, year of publication, clinical trial, NCT code, ICI drug, sample size, median age, MPR, PCR, TRG0-3, CR, PR, SD, ORR, DCR, TRAEs, irAEs, surgical complication rate, and surgical resection rate.In this meta-analysis, as most studies that met the inclusion criteria were conference abstracts, some epidemiological data were incomplete, such as the male to female ratio and median age.Studies for which data could not be extracted to calculate the key clinical outcomes described above from the articles, or for which data were discussed but not presented as raw data at international meetings, could not be included.Each study was reviewed several times to ensure that there were no missing or mislabeled data.Any differences regarding inclusion were addressed by discussion or by a third-party investigator who decided whether to incorporate the study.In the case of incomplete literature, the original authors were contacted for additional information wherever possible.

Statistical analysis
The meta-analysis was primarily undertaken using Review Manager version 5.4 (RevMan; (Cochrane Collaboration), which is a professional software provided by the Cochrane Collaboration (25).As most of the included studies were single-arm clinical trials with PCR and MPR representing the predominant outcome indicators, the research team used noncomparative binary data for meta-analysis in RevMan software.P-values and standard errors (SE(p)) were calculated by the following formula: p = ln(odds) = ln (X/(n-x)).SE(p) = SE(ln(odds)) = √1/X+1 (n-x).The dominance odds ratio (OR) and 95% confidence interval (CI) were used as efficacy evaluation indicators.For heterogeneity, the c 2 test and I 2 test were used.The included studies were taken out sequentially for sensitivity analysis, which revealed that the combined results were not significantly affected by each individual trial.A random-effects model was used where heterogeneity was significant; the alternative was to use a fixed-effects model.P-values < 0.05 were considered statistically significant.Higgins I 2 statistic < 50% was considered low heterogeneity, while > 50% of the statistic was considered high heterogeneity.Subgroup analysis was conducted to pinpoint the source of heterogeneity and factors related to clinical outcomes.RevMan 5.4 software and Stata/SE 15.0 software were used for statistical analysis of the data.

Assessments of publication bias and study quality
The quality of the included studies was assessed using the risk of bias assessment tools recommended by the Cochrane Handbook 5.1.0,including: (1) random assignment method; (2) allocation concealment; (3) whether participants and investigators were blinded; (4) whether efficacy was evaluated using blinded methods; (5) completeness of outcome information; (6) selective reporting of results; and (7) other biases.A qualitative assessment was independently performed by two investigators, while disagreements were decided by discussion between two or thirdparty investigators.The Begg test was used to test for possible publication bias in clinical studies.

Characteristics of the included studies
As a result of the literature study, we identified 3,564 potentially relevant papers.After removing duplicates, there remained 2,692 papers to be analyzed.We selected 191 articles for extensive analysis by screening their titles and abstracts.After full-text screening, 33 papers fulfilled the inclusion criteria and were included in our systematic review.The main characteristics of the included studies are summarized in Table 1.We recorded the total number of screened, selected, and excluded studies in a prismatic flow diagram (Figure 1).

Quality assessment of the studies
To assess the efficacy and safety of neoadjuvant immunotherapy in 33 clinical studies, we performed a Begg test using Stata software to identify possible publication bias.The results revealed no significant publication bias because most of the data were from uncontrolled cohort clinical trials and the graphs showed a symmetrical distribution (Supplementary Figure 2).

R0 Resection rate
The R0 resection rate is another important index for estimating the efficiency of neoadjuvant therapy.Nine of the 26 studies achieved a 100% R0 resection rate, and the remaining 17 studies had an individual OR < 1, with a combined OR of 96% (95% CI: 95%-98%, p < 0.0001; Figure 3E).A fixed-effects model was used given that the heterogeneity (P = 0.06, I 2 = 37%) was not significant.

Incidence of grade ≥ 3 TRAEs
Adverse events caused by ICIs are defined as TRAEs and are evaluated by the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 5 (61), which is the key metric for evaluating neoadjuvant immunotherapy safety.A total of 117 cases of grade ≥ 3 TRAEs were reported in 14 clinical studies.Neoadjuvant immunotherapy was supported when the overall individual OR and combined analysis showed an individual OR < 1 and a combined OR of 28% (95% CI: 17%-40%); the difference was statistically significant (p < 0.0001; Figure 4A).A random-effects model was used due to the significant heterogeneity (p < 0.0001, I 2 = 91%).

Surgical resection rate
As a percentage of patients with resected tumors versus those awaiting resection, the surgical resection rate is also an important indicator of the safety of neoadjuvant immunotherapy.Nineteen of the 33 studies had a 100% surgical resection rate.The combined OR of neoadjuvant immunotherapy was 89% (95% CI: 85%-93%; p < Frontiers in Immunology frontiersin.org0.0001) using a random-effects model (P = 0.009, I 2 = 54%; Figure 4B).

Incidence of grade ≥ 3 irAEs
Adverse events associated with treatment with immune checkpoint inhibitors (ICIs) were defined as irAEs.A total of 50 cases of grade ≥ 3 irAEs were reported in eight clinical studies.Neoadjuvant immunotherapy was supported when the combined OR of grade ≥ 3 irAE with resectable G/GEJ was 19% (95% CI: 11%-27%); the difference was statistically significant (P = 0.002; Figure 4C).A random-effects model was used due to the significant heterogeneity (p < 0.0001, I 2 = 69%).

Drug toxicity
Drug toxicity is the other important factor used to evaluate the safety of neoadjuvant immunotherapy.In the current analysis, two (26, 32) studies mentioned grade G3/4 toxicity, including stomatitis, nausea, vomiting, diarrhea, hypothyroidism, arthralgia, neutropenia, and pneumonia, but no instances of postponement of surgery or death due to drug toxicity.

Sensitivity analysis
The criteria for re-examining the search, selection, and merging of studies did not reduce the heterogeneity.A sensitivity analysis was performed by removing the included studies in order, with the aim to confirm that the conjunction results were not significantly impacted by an individual trial.Of the 33 studies analyzing individual studies of PCR incidence, the two studies by Hirotaka Hasegawa (47) and Jia-lin Tang (41) contributed the most to the heterogeneity, although the weights given in these two studies were not the greatest.Heterogeneity was significantly reduced after excluding the two studies (P = 0.02; I 2 = 39%), and the remaining 31 combined trials still significantly demonstrated the safety of neoadjuvant immunotherapy in combination with chemotherapy (OR = 25%; 95% CI: 22%-27; p < 0.0001; Supplementary Figure 3).The most significant cause of heterogeneity in TRG3 surgical resection rates was the study by Song Li et al. (55).After removing this study, the OR for combined TRG3 in the remaining ten trials was 19 (95% CI: 15%-23%; I 2 = 29%, p < 0.0001; Supplementary Figure 4).The OR for the surgical resection rate in the other 14 trials combined was 92 (95% CI: 89%-94%; I 2 = 33%, p < 0.0001; Supplementary Figure 5).Moreover, Jiang Z's (43) study was the main reason for the heterogeneity of irAEs.The remaining eight trials combined with an OR of 21% (95% CI: 16%-27%; I 2 = 20%; Supplementary Figure 6) remained supportive of neoadjuvant immunotherapy after exclusion (P = 0.27).In conclusion, sensitivity analyses of the study results continue to confirm the efficacy and safety of neoadjuvant immunotherapy along with chemotherapy.

Exploratory subgroup analysis
In the subgroup analysis, randomized clustering was applied to investigate potential branches of heterogeneity.We performed subgroup analysis to identify possible associations between neoadjuvant immunotherapy and different treatment regimens.The study was divided into nI, nII, nICT, nICRT and nAICT +apatinib groups.Heterogeneity was not significantly reduced because of the exclusion of MPRs as well as grade ≥ 3 TRAEs from the literature on a case-by-case basis in the sensitivity analysis.Because PCR is an important indicator of our concern, we performed subgroup analyses for all three indicators.

Discussion
Regional variations exist for treating locally progressive G/GEJ cancers.In Europe, the emphasis is on perioperative chemotherapy, whereas in North America, simultaneous postoperative radiotherapy is advocated, and in Asia, D2 curative surgery in conjunction with complementary postoperative chemotherapy is favored, with surgery continuing to be a definitive treatment (62).The standard treatments for locally progressive G/GEJ tumors are both NAC and NACRT.PCR rates are high with nCRT, but it is unclear whether such high PCR rates translate into long-term survival benefits.For the first-line treatment of advanced gastric cancer, immunotherapy has been broadly applied and extensively studied in the circumoperative phase.It is possible that compound chemotherapy and immunotherapy have concerted effects and produce better antitumor effects.At present, there remains controversy regarding the effectiveness and safety of neoadjuvant immunotherapy for resectable G/GEJ tumors; however, our study strongly supports the effectiveness and safety of neoadjuvant immunotherapy for resectable G/GEJ tumors.More importantly, not only is PCR a potent marker of excellent performance of neoadjuvant therapy in the short term but also a relatively accurate predictor with respect to gastric cancer recurrence, metastasis, and patient survival following treatment with neoadjuvant therapy (63).In our meta-analysis, the mean PCR for neoadjuvant immunotherapy was 26%, which was slightly higher than 23% in the CROSS study (11).Nine of the 33 studies had PCRs > 30%, the highest of which was the study by Pietrantonio F (49). and colleagues, with a PCR of 60%.Surprisingly, the mean MPR was 50.4% in the 15 included clinical studies, with the highest being the clinical trial by Pietrantonio F. et al. (49), with an MPR of 80%.Six studies had MPRs > 60%.PD-L1 expression is a potential biomarker for anti-PD -1/PD-L1 therapy.However, its predictive value in GC is unclear.Furthermore, it was found that microsatellite instability (MSI)/ mismatch repair (MMR) levels and tumor mutation burden (TMB) may be effective markers for screening patients for potential benefit from immunotherapy (64,65).Ten (28,35,36,38,44,47,(50)(51)(52)58) of the 33 included studies were found to indicate a higher incidence of PCR with MPR after neoadjuvant immunotherapy in patients with detectable CPS ≥ 1 or dMMR than in patients with CPS < 1 or pMMR.This result indicates that CPS ≥ 1 or dMMR may serve as a biomarker of prognosis in patients with resectable G/ GEJ.With these promising and encouraging outcomes, we have provided abundant evidence to support the effectiveness of neoadjuvant immunotherapy.
With regard to the safety of surgery, an average R0 resection rate of 94.5% was achieved with the combination of neoadjuvant immunotherapy and chemotherapy, much higher than the 82%-85% observed in the CROSS study (11) and 84% in the FLOT4-AIO study (66).Again, this suggests a promising and attractive effect of neoadjuvant immunotherapy.Nevertheless, it is difficult to elucidate the benefit of neoadjuvant immunotherapy in prolonging survival because of the short follow-up period and the fact that complete survival data based on RCTs have not been published.However, there is promise that future OS and DFS/PFS data will provide insight into the impact and benefit of the combination of neoadjuvant immunotherapy with chemotherapy on survival over the long term.
In particular, the results of the safety analysis suggest that it is possible to proceed with neoadjuvant immunotherapy assertively.The average incidence of grade ≥ 3 TRAEs in our meta-analysis was 29.1%, showing good tolerability.As seen in the subgroup analysis, the nICRT group greatly increased the incidence of grade ≥ 3 TRAEs.The majority of ICIs have also been previously assessed in full initial clinical studies and have been used globally for the treatment of extensive advanced tumors and consequently carry significant insight into the recognition and care of undesired events, providing further evidence that TRAE is an effective treatment.In terms of surgical resection rates, neoadjuvant immunotherapy averaged 94.3%.Further, the mean incidence of grade ≥ 3 irAEs was only 19.8%, again indicating good tolerability.Five studies in the included literature mentioned postoperative complications, with 13 cases of grade ≥ 3, including two anastomotic fistulas (47), six infections, and five cardiac causes (50).Considering these consequences together, the safety of neoadjuvant immunotherapy is deemed acceptable.In the nII group, neoadjuvant immunotherapy was associated with slightly lower grade ≥ 3 TRAEs as well as grade ≥ 3 irAEs, which may indicate a synergistic effect of neoadjuvant immunization along with immunotherapy, which can achieve higher PCR and MPR and does not enhance the incidence of AEs.Of course, further clinical studies are needed to verify the feasibility of this regimen (67).ICI subgroups based on ICI type could not be analyzed because the choice of chemotherapy regimen was not identical in any of the included studies, and there was no evidence that different ICIs contributed differently to the efficacy and safety of neoadjuvant immunotherapy.Therefore, ICIs are not currently the preferred choice for neoadjuvant immunotherapy; instead, neoadjuvant immunotherapy drug selection is dependent on the individual patient and the clinical situation.Of course, additional clinical trial data are needed to support this conclusion.This meta-analysis has several limitations.First, the majority of the included clinical studies have not reached the terminal point.Therefore, there are few clinical studies without a comprehensive regimen and available data.Additionally, as most data came from conference abstracts, these studies were not formally published in these cases, which may affect the assessment of bias and influence publication bias.However, publication bias due to article type is acceptable because leakage plots for assessing publication bias are distributed symmetrically.In addition to the metrics noted in this paper, other metrics could be used to estimate efficacy and safety, such as CR, PR, DCR, SD, PFS, OS, and time to surgery.However, we could not use these metrics because of the lack of relevant data.Another major limitation is that some trials have small sample sizes and there are so few RCTs, for which biases may result.As a result, there is a need for larger sample sizes and more RCTs for further validation in polycenter studies.
Despite the excellent results in terms of removal rates, some concerns remain.First, regarding the frequency of cycles of neoadjuvant therapy, it is unclear whether an increase in the number of cycles will improve treatment efficacy, produce better MPR rates and PCR rates, and reduce both toxicity and side effects.Second, it remains to be determined whether the sequencing of chemotherapeutic agents with immune reagents will do a better job of enhancing metrics such as PCR rates.Third, future studies are required to establish which combination of immune drugs and which treatment modality achieves maximal PCR/MPR.To conclude, a high postoperative PCR rate does not directly imply a high survival rate.However, as most of the included studies had short follow-up periods, conclusive results could not be obtained.
To summarize, our meta-analysis of the effectiveness and safety of neoadjuvant immunotherapy for resectable G/GEJ tumors suggests that the extensive use of neoadjuvant immunotherapy is clinically supported.However, because most clinical trials have not yet met their endpoints, it is important to examine the long-term outcomes and toxicity to confirm this conclusion.It is comforting that more neoadjuvant immunotherapy studies are underway, which may confirm the above conclusions in the future.

Conclusion
Neoadjuvant immunotherapy, especially neoadjuvant dualimmunotherapy combinations, is effective and safe for resectable gastric/gastroesophageal junction tumors in the short term.Nevertheless, further multicenter randomized trials are required to demonstrate which combination model is more beneficial.

FIGURE 1 Preferred
FIGURE 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram of the study selection.

TABLE 1
Main characteristics of included studies.

TABLE 2
Main characteristics of included studies.