Home Integrating machine learning and multi-omics analysis to develop an immune-derived multiple programmed cell death signature for predicting clinical outcomes in gastric cancer
Article Open Access

Integrating machine learning and multi-omics analysis to develop an immune-derived multiple programmed cell death signature for predicting clinical outcomes in gastric cancer

  • Chunhong Li ORCID logo EMAIL logo , Jiahua Hu , Mengqin Li , Xiao Fan and Yiming Mao EMAIL logo
Published/Copyright: September 3, 2024
Download Article (PDF)
Oncologie
From the journal Oncologie Volume 26 Issue 6

Abstract

Objectives

Metastasis of tumor cells is the leading reason for mortality among patients diagnosed with gastric cancer (GC). Emerging evidence indicated a strong correlation between programmed cell death (PCD) and the invasion and metastasis of tumor cells. Therefore, we aimed to develop a programmed cell death signature to assess the prognosis and therapeutic efficacy in GC patients.

Methods

Here, we collected 1911 PCD-related genes from 19 different PCD patterns, and developed an immune-derived multiple programmed cell death index (MPCDI) using the integrating machine learning and multi-omics analysis, and systematically dissected heterogeneity in GC patients. Subsequently, we divided GC patients into two categories, namely high-MPCDI group and low-MPCDI group, using the median MPCDI as the threshold. We performed a comprehensive analysis of the clinical characteristics, somatic mutations, immune infiltration, drug sensitivity, and immunotherapeutic efficacy of the two groups.

Results

Survival and immunotherapy response analyses indicated that the high-MPCDI patients experienced a poorer overall survival (p=0.018) and were more resistant to commonly used chemotherapeutic drugs but benefited from immunotherapy compared to the low-MPCDI patients. In addition, MPCDI was confirmed as a standalone risk factor for overall survival, and nomograms can provide a precise tool for the clinical diagnosis of GC patients.

Conclusions

Taken together, the MPCDI can serve as a robust clinical diagnostic classifier to guide medication administration and improve outcomes in GC patients.

Keywords: gastric cancer; multi-omics; immunotherapy; programmed cell death; machine learning

Introduction

Gastric cancer (GC) is a complex and multifactorial disease characterized by a multi-stage progression. It ranks among the most lethal cancers globally, as revealed by the Global Cancer Statistics 2020. The report highlighted that gastric cancer (GC) has a high mortality and incidence rate, placing fifth in incidence and fourth in mortality globally, while the incidence rate of GC is gradually decreasing in developed countries [1]. This trend can be attributed to the low rate of early GC diagnosis and treatment in China, which stands at less than 10 % [2], significantly lower than that of Japan and Korea. Consequently, a large number of patients are diagnosed in advanced stages of GC, missing crucial treatment opportunities.

The treatment of advanced GC patients is mainly based on traditional chemotherapy, with relatively limited efficacy [3]. With the emergence of the targeted drug trastuzumab, the clinical outcome of patients with advanced GC has been dramatically improved, but the number of patients who benefited is still limited 3], [4], [5], [6. Despite receiving comprehensive surgical intervention, the survival rate for advanced GC is still less than 30 % after 5 years, resulting in a significant burden on both the country and affected families [7, 8]. In the realm of GC, immunotherapy has recently witnessed remarkable advancements and when combined with chemotherapy, it presents a new ray of hope for patients with advanced GC [9, 10]. However, the immune microenvironment of advanced GC patients is complex and heterogeneous, and there are significant differences between different stages and types, before and after treatments, as well as progressive changes, so the progress of immunotherapy is still difficult. On the other hand, the 5-year survival rate of most early-stage GC patients can be more than 90 % after treatment, and even reach the curative effect, which is the key to enhancing the survival time for GC patients [11, 12]. Therefore, it is imperative to develop reliable clinical diagnostic tools for guiding the early diagnosis of GC and providing patients with personalized treatment strategies to improve their survival time and quality of life.

Programmed Cell Death (PCD) is a precise and orderly pattern of cell death in the organism determined by genetic factors, which is essential for maintaining tissue health and regulating disease progression [13]. Aberrant regulation of this pattern is influential in various illnesses including cancer, neurological diseases, and cardiovascular diseases 14], [15], [16], [17], [18. To date, about 19 PCD patterns have been discovered, which can be activated when cells are stimulated by various internal and external environmental factors 19], [20], [21], [22], [23], [24. In the past, the key initiators, effectors, and implementers of the above PCD patterns were thought to be independent of each other, but a growing body of evidence suggested that there are interactions between the different PCD patterns [25, 26]. There is now increasingly clear evidence that the PCD patterns of necroptosis, pyroptosis, and apoptosis are closely interconnected and can cross-regulate each other [27].

PANoptosis is a unique pattern of inflammatory PCD regulated by the PANoptosome, which recognizes the activation of receptors, binds inflammatory factors, and signals through receptors containing structural domains of death to initiate a highly interconnected cell death process 27], [28], [29. As a key intermediary molecule in the apoptosis and necrosis pathways, Caspase-8 is one of the earliest discovered bridges between different patterns of cell death [30]. It has the function of inducing and inhibiting cell death: it can induce apoptosis through death receptors, such as TNFR1 [31], and, at the same time, inhibit necroptosis through the kinases RIPK3 and MLKL 32], [33], [34], [35.

Different PCD patterns are inactive or resistant in cancer development, allowing cancer cells to proliferate unrestrictedly, which is mainly achieved by regulating the expression or the activity of proteins involved in various PCD patterns [36, 37]. Currently, one of the major challenges in cancer therapy is drug resistance due to the escape of cancer cells to various PCD patterns [36]. Based on the accumulation of current research results, focusing on the interactions between different PCD patterns may be a new direction for future cancer therapeutic research. In addition, when we do basic research in the field of cancer, we can quickly obtain important information related to the disease through bioinformatics analysis [38, 39]. Therefore, this study aimed to develop an immune-derived multiple programmed cell death index (MPCDI) to predict clinical outcomes and medication guidance in GC patients. According to the MPCDI scoring system, we discovered the novel molecular subtypes of GC patients and revealed the heterogeneity among them, which may provide new options for personalized treatment strategies for GC patients.

Materials and methods

Data source

The study’s overall design is presented in Supplementary Figure S1. As for the training cohort of the GC patients, we obtained RNA sequencing raw expression profiles and clinical data for 375 GC samples and 32 paraneoplastic samples from the TCGA public database (https://portal.gdc.cancer.gov/). As for the validation cohorts of the GC patients, we obtained two independent cohorts, GSE84433 (357 GC patients, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE84433) and GSE84437 (433 GC patients, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE84437), from the GEO public database. In addition, we collated and collected a set of PCD-related genes from the published literature 19], [20], [21, 24, 39], [40], [41], [42. Finally, we eliminated any duplicated genes and maintained a total of 1911 PCD-related genes selected for further in-depth bioinformatics analyses (Supplementary Table S1).

Tumor microenvironment analysis

To measure the status of tumor microenvironment in GC patients, we conducted immune microenvironment analysis using the “IOBR” package [43] in R software (v4.1.3) based on different analytical methods, including MCP-counter method [44], EPIC method [45], xCell method [46], CIBERSORT method [47], IPS method [48], quanTIseq method [49], ESTIMATE method [50], TIMER method [51], and single-sample gene set enrichment analysis (ssGSEA) method [52]. Furthermore, the immunotherapeutic response of GC samples was analyzed using the Tumor Immune Dysfunction and Rejection (TIDE) website (http://tide.dfci.harvard.edu/login/) [53].

Unsupervised clustering analysis

The ssGSEA method relies on 29 immune gene sets that encompass genes associated with various immune cell types, checkpoints, pathways, and functions. We used the ssGSEA method through the “IOBR” package to comprehensively evaluate the immune infiltration characteristics of each GC sample. Based on ssGSEA results of TCGA-STAD cohort, the consensus clustering analysis was conducted using the “ConsensusClusterPlus” R-package [54]. Through the assessment of 28 immune cell infiltrations derived from the ssGSEA algorithm, various clusters were discovered within the TCGA-STAD cohort. The outcomes of survival for these distinct immune clusters were graphically represented utilizing the “Survival” R-package.

WGCNA analysis

To discover PCD-related genes that have close ties to immune clusters, the WGCNA analysis was conducted using the “WGCNA” R-package [55] to pinpoint PCD modules most linked to immune clusters. Initially, to construct a scale-free co-expression network, we calculated an appropriate soft threshold β. Next, we converted the weighted adjacency matrix into a topological overlap matrix (TOM) and calculated the dissimilarity (dissTOM). To identify modules, we used the dynamic tree cutting approach. After analyzing 13 gene modules, we ended up with the blue gene module with the highest positive relevance with the immune cluster for further analysis. Subsequently, PCD-related genes with elevated GS and MM were classified as immune-related PCD-related genes (IRPCDRGs).

Development and validation of a multiple programmed cell death signature

To analyze the effect of these IRPCDRGs on the survival outcome of GC patients, we employed univariate Cox regression analysis. To minimize any possible negligence, the p-value was set to be less than 0.05. Moreover, we utilized three machine learning methods, specifically XGBoost [56], Random Forest [57] and Lasso regression to identify the common IRPCDRGs by capturing the common genes discovered by all three methods. Subsequent to this, stepwise regression analysis was carried out to identify the predictive indicators. The Patient’s MPCDI was computed using the subsequent equation: MPCDI=Coeff (gene 1) × Exp (gene 1) + Coeff (gene 2) × Exp (gene 2) + Coeff (gene n) × Exp (gene n). Herein, Exp (gene n) refers to the expression level of the gene, while the Coeff (gene n) represents the coefficient obtained from multivariate Cox analysis. Based on the median MPCDI value, GC patients were classified into groups of low- and high- MPCDI. We then conducted survival analyses using the “survival” package to explore survival outcomes in both groups and further validated them using the GSE84433 and GSE84437 cohorts. Univariate and multivariate Cox analyses were also used to assess whether MPCDI could be used as a self-contained predictor of GC patients.

Establishment of the nomogram

Combining clinical features of GC patients with MPCDI, the nomogram was established using multivariable Cox and stepwise regression analyses. Visualization of the nomogram and calibration plots was achieved through the use of the “rms” package. To evaluate the predictive performance, ROC analysis was conducted on GC samples using the “timeROC” package [58]. Depending on the clinical features, the relationship and stratification of the MPCDI were further analyzed.

Mutation profile analysis

Data on somatic mutations in GC patients were acquired from the TCGA publicly available databases utilizing the “TCGAbiolinks” package [59]. To describe the mutation profiles of GC samples, we utilized the “Maftools” package [60] to generate a waterfall diagram. Furthermore, we examined the genes that were most commonly mutated in the high- and low-MPCDI groups for potential mutual exclusion and cooperation.

Functional enrichment analysis

To investigate the fundamental biological processes, we conducted the Gene set enrichment analysis (GSEA), Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis in the two groups utilizing the R packages “clusterProfiler” [61], “enrichplot” and “GseaVis”. The criteria for selection were |NES|>1, along with the p<0.05.

Drug sensitivity analysis

To tailor the therapy, we utilized the “oncoPredict” package [62] to forecast the response to chemotherapy in GC patients based on varying MPCDI scores. Gene expression levels in tissue samples from patients were compared to those in cancer cell lines through the use of oncoPredict, and subsequently, the IC50 value of conventional chemotherapy drugs was calculated. To evaluate the variations in drug IC50 values between the two MPCDI groups, the Wilcoxon test was applied, establishing a significance threshold at p<0.05. To improve the accuracy of drug sensitivity assessments, we analyzed the gene expression profiles and drug sensitivity data of the three characteristic IRPCDRGs by accessing the CellMiner database (https://discover.nci.nih.gov/cellminer/home.do), GDSC and CTRP databases in the GSCALite platform (https://guolab.wchscu.cn/GSCA/#/), respectively [63].

Identification of the immunotherapy efficacy

We evaluated the likely response of GC patients to checkpoint immunotherapy using the Cancer Immunome Atlas (TCIA, https://tcia.at/home) [48]. Moreover, a thorough analysis was carried out to explore the relation between Immunophenoscores (IPS) scores for anti-PD-1 and anti-CTLA4 treatments in GC patients and MPCDI. Additionally, we obtained the IMvigor210 cohort from the “IMvigor210CoreBiologies” R-package [64], which includes important clinical details on the treatment of uroepithelial carcinoma with atezolizumab that is designed to target PD-L1 [65], and was further used to validate the clinical predictive value of the MPCDI.

Statistical analysis

Survival analyses in this study were calculated using the Kaplan–Meier method. Two or multiple group comparisons in the experiment were analyzed using the Wilcoxon test or the Kruskal–Wallis test for the corresponding statistics, respectively. Spearman’s correlation analysis was used to calculate correlations. When the p-value was equal to or less than 0.05, it was defined as a statistically significant difference analysis. All statistical analyses were conducted using R software (v4.1.3).

Results

Development and validation of immune infiltration consensus clusters

A consensus clustering analysis was conducted using 28 immune cell profiles assessed by the ssGSEA analysis, which initially classified all GC patients into k (k=2–9) clusters. The resilience of the cluster at k=2 was deemed excellent, as indicated by the cumulative distribution function (CDF) curve (Figure 1A–C). Supplementary Table S2 presents detailed baseline data for C1 and C2 patients. It was observed that C2 patients had a poorer prognosis compared to C1 patients (Figure 1D, p<0.001). Based on the ESITIMATE results, C2 patients exhibited higher stromal score, estimated score and immune score, while experiencing a decline in tumor purity (Figure 1E–H). The TIDE analysis also revealed elevated T cell rejection scores, TIDE scores, and T cell dysfunction scores, along with reduced MSI scores in C2 patients, indicating a heightened hidden capability for immune evasion and possibly lower efficacy of ICI therapies for C2 patients (Figure 1I–L). Furthermore, a notable variance in immune infiltration was observed between the two immune clusters, with a significantly greater degree of immune infiltration in C2 patients as compared to C1 patients (Figure 1M and N). To prevent potential biases in the analysis algorithms from affecting the results of the two immune clusters, we validated the ssGSEA results using six additional algorithms, which consistently found significantly higher levels of immune infiltration in C2 patients (Supplementary Figure S2).

Figure 1: Development and validation of immune infiltration consensus clusters. (A) Consensus clustering analysis of ssGSEA results within the TCGA-STAD cohort (k=2). (B) CDF curves representing the consensus matrix for each k, distinguished by different colors. (C) Changes in the area underneath the CDF curve for each k. (D) Kaplan–Meier survival curves illustrating overall survival (OS) for various clusters. (E–H) Scores for Stroma, Immune, ESTIMATE, and Tumor purity across both clusters. (I–L) Scores for TIDE, T-cell dysfunction, T-cell rejection, and microsatellite instability (MSI) across both clusters. (M) Distribution of infiltration among 28 subsets of immune cells across both clusters. (N) The abundance of 28 immune cell subsets’ infiltration was assessed via ssGSEA in both clusters.
Figure 1:

Development and validation of immune infiltration consensus clusters. (A) Consensus clustering analysis of ssGSEA results within the TCGA-STAD cohort (k=2). (B) CDF curves representing the consensus matrix for each k, distinguished by different colors. (C) Changes in the area underneath the CDF curve for each k. (D) Kaplan–Meier survival curves illustrating overall survival (OS) for various clusters. (E–H) Scores for Stroma, Immune, ESTIMATE, and Tumor purity across both clusters. (I–L) Scores for TIDE, T-cell dysfunction, T-cell rejection, and microsatellite instability (MSI) across both clusters. (M) Distribution of infiltration among 28 subsets of immune cells across both clusters. (N) The abundance of 28 immune cell subsets’ infiltration was assessed via ssGSEA in both clusters.

Identification of PCD modules derived from immune infiltration patterns

To determine PCD modules that exhibited a strong correlation with immune clusters, an analysis of the TCGA-STAD cohort was conducted utilizing WGCNA. An optimal soft threshold of 8 was selected to construct a scale-free topological network (Figure 2A). Subsequently, we identified 13 modules, each shown by a disparate color (Figure 2B–E). Correlations were also calculated between the modules and clinical parameters, which included immune clusters, events, time, pathological stage, age, gender, as well as TNM stage. Among the correlations observed, the highest correlation was found between the blue module and the immune clusters (Figure 2C). Modules’ representatives are termed as eigengenes, as shown in the heatmap (Figure 2D). Evaluation of the blue module showed a correlation coefficient of 0.95 between GS and MM, suggesting a high level of quality in constructing the PCD module (Figure 2F). The immune-related hub genes were identified by selecting 2266 genes in the blue module that comply with GS>0.5 and MM>0.6 (Figure 2G). By merging the WGCNA findings, 227 common IRPCDRGs were singled out for subsequent bioinformatics analysis (Figure 2G). According to the gene expression levels of these 227 immune-related IRPCDRGs, univariate Cox analysis identified nine prognostic IRPCDRGs (Figure 2H).

Figure 2: Identification of PCD modules derived from immune infiltration patterns. (A) Examination of network topology in relation to varying levels of soft-threshold power. The left side illustrates how soft-threshold power influences the scale-free topology fit index; conversely, the right side indicates its effect on average connectivity. (B) Dendrogram displaying clusters derived from the WGCNA analysis. (C) Analysis of correlations between module eigengenes and clinical characteristics. (D) The heatmap illustrated the adjacency of eigengenes among different modules. (E) Heatmap representing the network from the WGCNA analysis. (F) Scatter plot depicting the correlation between gene significance (GS) and module membership (MM) within the blue module. (G) Venn diagram illustrating the shared genes between the MEblue module and PCD-related genes. (H) Univariate Cox regression analysis identified nine prognostic IRPCDRGs within the TCGA-STAD cohort.
Figure 2:

Identification of PCD modules derived from immune infiltration patterns. (A) Examination of network topology in relation to varying levels of soft-threshold power. The left side illustrates how soft-threshold power influences the scale-free topology fit index; conversely, the right side indicates its effect on average connectivity. (B) Dendrogram displaying clusters derived from the WGCNA analysis. (C) Analysis of correlations between module eigengenes and clinical characteristics. (D) The heatmap illustrated the adjacency of eigengenes among different modules. (E) Heatmap representing the network from the WGCNA analysis. (F) Scatter plot depicting the correlation between gene significance (GS) and module membership (MM) within the blue module. (G) Venn diagram illustrating the shared genes between the MEblue module and PCD-related genes. (H) Univariate Cox regression analysis identified nine prognostic IRPCDRGs within the TCGA-STAD cohort.

Development of the programmed cell death signature

Three machine learning methods were then applied to detect the three most valuable IRPCDRGs among the nine prognostic IRPCDRGs (Figure 3A–G). The multivariate Cox regression analysis outcomes displayed the identification of three characteristic IRPCDRGs in the TCGA-STAD cohort. Subsequently, leveraging the gene expression levels and regression coefficients of the three characteristic IRPCDRGs, a formula was formulated to compute MPCDI: MPCDI=0.3412255 × Exp (ITGAV) + 0.2105629 × Exp (HCAR1) + 0.1472504 × Exp (WNT4). According to the MPCDI scoring system, the GC individuals were classified into two groups: one with low MPCDI and the other with high MPCDI. The clinical baseline information for each group is outlined in Supplementary Table S3. It was observed that the high-MPCDI patients experienced a poorer overall survival, but the low-MPCDI patients had a more positive overall survival in the TCGA-STAD cohort (Figure 3H). Survival analyses of GC patients in the GSE84433 and GSE84437 cohorts showed survival outcomes consistent with the TCGA-STAD cohort (Figure 3I–L). An MPCDI-based nomogram was constructed utilizing multivariate Cox analysis and stepwise regression in the TCGA-STAD cohort to predict overall survival time at 1, 3, and 5 years (Figure 3J). The calibration curves indicated that the predicted survival rates aligned closely with the actual rates at those time points (Figure 3N). The results of univariate and multivariate regression analyses in the TCGA-STAD, GSE84433 and GSE84437 cohorts supported the notion that MPCDI could be used as a self-contained predictor of GC patients (Figure 3K–M). The AUC values for 1, 3, and 5 years were 0.665, 0.615, and 0.614, respectively, in the TCGA-STAD cohort (Figure 4A). When it comes to the standalone validation cohorts, the AUC values for 1, 3, and 5 years display different results for two GEO cohorts (GSE84433 and GSE84437), with values of 0.505, 0.600, 0.564 and 0.516, 0.551, and 0.566, respectively (Figure 4B and C). Additionally, Figure 4D–F depict the MPCDI, the survival outcomes of GC patients, and the gene expression levels of three characteristic IRPCDRGs in various MPCDI groups. The high expression of three characteristic IRPCDRGs in the patients with high-MPCDI indicates their role as a risk factor, a finding that was further confirmed in the other two GEO cohorts.

Figure 3: Development of the programmed cell death signature. (A) Cross-validation method to select the best gene. (B) Lasso coefficient curve. (C) Number of trees and variable importance of Random Forest. (D) VIP genes in Random Forest. (E) The negatively biased logarithm associated with the Cox proportional hazards regression of the XGBoost technique plotted against the number of iterations. (F) Top 9 genes in XGBoost. (G) Venn plot showing the intersected valuable IRPCDRGs based on three machine learning algorithms. (H) Kaplan–Meier survival curves for the TCGA-STAD cohort. (I) Kaplan–Meier survival curves for the GSE84433 cohort. (J) Development of the nomogram deriving from the MPCDI alongside clinical characteristics (including age, gender, clinical stage, T stage, N stage, and M stage) in the TCGA-STAD cohort. (K, M) Univariate and multivariate regression analysis. (L) Survival curves generated using Kaplan-Meier for the GSE84437 cohort. (N) Calibration plot of the nomogram illustrating 1-year, 3-year, and 5-year OS in the TCGA-STAD cohort.
Figure 3:

Development of the programmed cell death signature. (A) Cross-validation method to select the best gene. (B) Lasso coefficient curve. (C) Number of trees and variable importance of Random Forest. (D) VIP genes in Random Forest. (E) The negatively biased logarithm associated with the Cox proportional hazards regression of the XGBoost technique plotted against the number of iterations. (F) Top 9 genes in XGBoost. (G) Venn plot showing the intersected valuable IRPCDRGs based on three machine learning algorithms. (H) Kaplan–Meier survival curves for the TCGA-STAD cohort. (I) Kaplan–Meier survival curves for the GSE84433 cohort. (J) Development of the nomogram deriving from the MPCDI alongside clinical characteristics (including age, gender, clinical stage, T stage, N stage, and M stage) in the TCGA-STAD cohort. (K, M) Univariate and multivariate regression analysis. (L) Survival curves generated using Kaplan-Meier for the GSE84437 cohort. (N) Calibration plot of the nomogram illustrating 1-year, 3-year, and 5-year OS in the TCGA-STAD cohort.

Figure 4: The prognostic value of MPCDI. (A) ROC curves illustrating the predictive capability of the MPCDI in 1, 3, and 5-year OS in the TCGA-STAD cohort. (B) ROC curves depicting the prediction efficacy of the MPCDI for 1, 3, and 5-year OS in the GSE84433 cohort. (C) ROC curves representing the prediction performance of the MPCDI for 1, 3, and 5-year OS in the GSE84437 cohort. (D) Heatmap displaying the distribution of MPCDI, patient survival data, and the expression profiles of the three characteristic IRPCDRGs that make up MPCDI in the TCGA-STAD cohort. (E) Heatmap illustrating the distribution of MPCDI, patient survival outcomes, and expression profiles of the three characteristic IRPCDRGs that constitute MPCDI in the GSE84433 cohort. (F) Heatmap of MPCDI distribution, patient survival, and expression profiles of the three characteristic IRPCDRGs that comprise MPCDI in the GSE84437 cohort.
Figure 4:

The prognostic value of MPCDI. (A) ROC curves illustrating the predictive capability of the MPCDI in 1, 3, and 5-year OS in the TCGA-STAD cohort. (B) ROC curves depicting the prediction efficacy of the MPCDI for 1, 3, and 5-year OS in the GSE84433 cohort. (C) ROC curves representing the prediction performance of the MPCDI for 1, 3, and 5-year OS in the GSE84437 cohort. (D) Heatmap displaying the distribution of MPCDI, patient survival data, and the expression profiles of the three characteristic IRPCDRGs that make up MPCDI in the TCGA-STAD cohort. (E) Heatmap illustrating the distribution of MPCDI, patient survival outcomes, and expression profiles of the three characteristic IRPCDRGs that constitute MPCDI in the GSE84433 cohort. (F) Heatmap of MPCDI distribution, patient survival, and expression profiles of the three characteristic IRPCDRGs that comprise MPCDI in the GSE84437 cohort.

Clinical correlation and stratified analysis of the MPCDI

To explore MPCDI’s predictive capacity when combined with diagnostic characteristics, we scrutinized the clinicopathological features of GC patients. An assessment was conducted on parameters such as gender, age, pathologic stage, TNM stage, and immune clusters, followed by a thorough stratified analysis. Our findings revealed a marked increase in MPCDI levels in deceased patients and those in stages IV and N3 (Figure 5A–H). Furthermore, survival outcomes demonstrated that the high-MPCDI patients with varying clinical attributes (e.g., male, age≥60 years, stage T4) exhibited considerably lower prognosis than the low-MPCDI patients (Figure 5I–P). These outcomes strongly underscore the pivotal role of MPCDI in predicting GC prognosis.

Figure 5: Clinical correlation and stratified analysis of the MPCDI. (A–H) Differences in MPCDI between groups with different clinical features. (I–P) OS KM curves for MPCDI in the two groups stratified by clinicopathologic factors.
Figure 5:

Clinical correlation and stratified analysis of the MPCDI. (A–H) Differences in MPCDI between groups with different clinical features. (I–P) OS KM curves for MPCDI in the two groups stratified by clinicopathologic factors.

Functional enrichment analysis

To uncover the molecular mechanism behind the effects of abnormal expression of PCD-related genes on varying clinical outcomes, we performed the functional enrichment analysis. The GO results of differentially expressed PCD-related genes (DEPCDGs) revealed a notable enrichment in collagen-containing extracellular matrix, external encapsulating structure, and hypertrophic pachymeningitis (HP) abnormal meningeal morphology (Figure 6A). The KEGG results of DEPCDGs revealed a notable enrichment in coagulation, epithelial mesenchymal transition, and ultraviolet (UV) response down-regulated genes (DN) (Figure 6B). Functional enrichment results revealed that these terms such as enzyme-linked receptor protein signaling pathway, collagen-containing extracellular matrix, apical junction, external encapsulating structure, coagulation, epithelial mesenchymal transition, and UV response DN were enriched in the high-MPCDI patients (Figure 6C and D). It is important that a notable enrichment of RNA transcription-related pathways was enriched in the high-MPCDI patients, containing RNA binding involved in RNA mediated gene silencing by inhibition of translation, posttranscriptional gene silencing, inner mitochondrial membrane protein complex, E2F targets, G2M checkpoint, and oxidative phosphorylation (Figure 6E and F).

Figure 6: Functional enrichment analysis. (A) GO functional enrichment analyses. (B) KEGG pathway enrichment analyses. (C) The GO functional enrichment analyses in the high-MPCDI patients. (D) The KEGG pathway enrichment analyses in the high-MPCDI patients. (E) The GO functional enrichment analyses of the low-MPCDI patients. (F) The KEGG pathway enrichment analyses of the low-MPCDI patients.
Figure 6:

Functional enrichment analysis. (A) GO functional enrichment analyses. (B) KEGG pathway enrichment analyses. (C) The GO functional enrichment analyses in the high-MPCDI patients. (D) The KEGG pathway enrichment analyses in the high-MPCDI patients. (E) The GO functional enrichment analyses of the low-MPCDI patients. (F) The KEGG pathway enrichment analyses of the low-MPCDI patients.

Comprehensive immune analysis

To assess the status of immunological microenvironment in GC patients, we adopted distinct means to compute the scores of stromal and immune cells. As per the ESITIMATE consequences, high-MPCDI patients exhibited an enhancement in immune score, stromal score, and estimated score, along with a reduction in tumor purity (Figure 7A–D). According to the ssGSEA results, there were significant variances in memory B cells, eosinophils, activated CD4 T cells, mast cells, immature dendritic cells, and plasmacytoid dendritic cells in the two groups (Figure 7E). Various software was utilized to analyze the relationship between MPCDI and diverse immune cell categories. Results demonstrated a significant positive association of MPCDI score with DCs, macrophages, and neutrophils. Additionally, a negative relevance was noted between the MPCDI and the proportion of Th1 cells, Th2 cells, and pro-B cells (Figure 7F). Furthermore, we observed that three characteristic IRPCDRGs within MPCDI existed with strong relevance to the tumor immune microenvironment. Specifically, ITGAV displayed positive correlations with natural killer cells, type 2 T helper cells, and effector memory CD4+ T cells. Conversely, HCAR1 demonstrated negative correlations with effector memory CD4+ T cells, activated CD4+ T cells, and activated CD8+ T cells. Lastly, WNT4 exhibited negative correlations with memory B cells, activated CD8+ T cells, and activated CD4+ T cells (Figure 7G–J). To prevent potential biases in the analysis algorithms from affecting the results of the two MPCDI groups, we validated the ssGSEA results using six additional methods, which consistently found significantly higher levels of immune infiltration in high-MPCDI patients (Supplementary Figure S3 and S4).

Figure 7: Comprehensive immune analysis. (A–D) Stroma score (A), Immune score (B), ESITIMATE score (C), and Tumor purity (D) in the two MPCDI groups. (E) ssGSEA score of immune cell infiltration in the two MPCDI groups. (F) Correlation between immune scores and immune cells assessed by different software. (G) The correlation between each key molecule and each TME infiltration cell type. (H) Correlation analysis between TME infiltrated cells and ITGAV. (I) Correlation analysis between TME infiltrated cells and HCAR1. (J) Correlation analysis between TME infiltrated cells and WNT4. (ns, no significance; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).
Figure 7:

Comprehensive immune analysis. (A–D) Stroma score (A), Immune score (B), ESITIMATE score (C), and Tumor purity (D) in the two MPCDI groups. (E) ssGSEA score of immune cell infiltration in the two MPCDI groups. (F) Correlation between immune scores and immune cells assessed by different software. (G) The correlation between each key molecule and each TME infiltration cell type. (H) Correlation analysis between TME infiltrated cells and ITGAV. (I) Correlation analysis between TME infiltrated cells and HCAR1. (J) Correlation analysis between TME infiltrated cells and WNT4. (ns, no significance; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).

Comprehensive gene mutation profiles

To examine the variances in gene mutation profiles of GC patients, the mutation patterns were assessed within the two MPCDI groups, identifying differences in the genomic mutation landscape of MUC16, TTN, and TP53 between the two groups (Figure 8A and B). Evaluation of the commutative exclusion and coordination of genomic mutation revealed synergistic impacts for most genes. Notably, In the low-MPCDI patients, significant mutually exclusive effects were observed for mutations in the TP53-ARID1A gene. Conversely, within the high-MPCDI group, mutations in the CSMD1-SPTA1 genes demonstrated clear cooperative relationships, while in the low-MPCDI patients, the ARID1A-PIK3CA genomic mutations exhibited significant cooperation (Figure 8C and D). Notably, the predominant missense mutations were concentrated on the three characteristic IRPCDRGs (Figure 8E). Lollipop plots of mutation types, locations and corresponding amino acid changes of CSMD1 and SPTA1 in the high-MPCDI group showed that they were mainly missense mutations with an even distribution (Figure 8F and G). Lollipop plots of mutation types, locations and corresponding amino acid changes of TP53, ARID1A and PIK3CA in the low-MPCDI group showed that they were mainly missense mutations with an even distribution (Figure 8H–J).

Figure 8: Comprehensive gene mutation profiles. (A, B) Plots depicting waterfalls of somatic mutations in the high-MPCDI patients (A) and the low-MPCDI patients (B). (C, D) Heatmaps illustrating mutually exclusive and synergistic mutated genes in the high-MPCDI patients (C) and the low-MPCDI patients (D). (E) Mutation rates of three characteristic IRPCDRGs constructed for MPCDI. (F, G) Lollipop plots of mutation types, locations and corresponding amino acid changes in GSMD1 (F) and SPTA1 (G) in the high-MPCDI group. (H–J). Lollipop plots of mutation types, locations and corresponding amino acid changes of TP53 (H), ARID1A (I) and PIK3CA (J) in the low-MPCDI group.
Figure 8:

Comprehensive gene mutation profiles. (A, B) Plots depicting waterfalls of somatic mutations in the high-MPCDI patients (A) and the low-MPCDI patients (B). (C, D) Heatmaps illustrating mutually exclusive and synergistic mutated genes in the high-MPCDI patients (C) and the low-MPCDI patients (D). (E) Mutation rates of three characteristic IRPCDRGs constructed for MPCDI. (F, G) Lollipop plots of mutation types, locations and corresponding amino acid changes in GSMD1 (F) and SPTA1 (G) in the high-MPCDI group. (H–J). Lollipop plots of mutation types, locations and corresponding amino acid changes of TP53 (H), ARID1A (I) and PIK3CA (J) in the low-MPCDI group.

Comprehensive drug sensitivity analyses

To further investigate the clinical importance of PCD-related genes in the precise chemotherapy of GC patients, an assessment was conducted on the effectiveness of conventional chemotherapy drugs within various groups. The results from the IC50 analysis were promising: the low-MPCDI patients displayed improved effectiveness with conventional chemotherapy drugs compared to the high-MPCDI patients, indicating that the MPCDI scoring system can offer guidance in the chemotherapy of GC patients (Figure 9A). To further study the connection between the levels of the three characteristic IRPCDRGs and conventional chemotherapy drugs, an examination was carried out using the CellMiner publicly accessible database. Correlation histograms were generated for the top five conventional chemotherapy drugs exhibiting the most remarkable associations between the gene and drug sensitivity. Of significance, heightened expression of HCAR1 was found to increase resistance to the majority of drugs, while elevated expression of ITGAV enhanced sensitivity to most drugs (Figure 9B). Within the GDSC publicly accessible database, expression of ITGAV demonstrated a positive correlation with multiple conventional chemotherapy drug sensitivities, whereas expression of WNT4 exhibited a negative correlation with various drug sensitivities (Figure 9C). Notably, we obtained similar results when using the CTRP publicly accessible database for drug sensitivity analysis (Figure 9D).

Figure 9: Analysis of drug sensitivity among the two MPCDI categories. (A) Variations in the reaction to commonly used chemotherapeutic drugs between groups with high and low MPCDI. (B) Relationship between the constructed MPCDI genes and their sensitivity to drugs. (C) The association of GDSC drug sensitivity with the mRNA expression levels of the three characteristic IRPCDRGs within the constructed MPCDI. (D) The association of CTRP drug sensitivity with the mRNA expression levels of the three characteristic IRPCDRGs within the constructed MPCDI.
Figure 9:

Analysis of drug sensitivity among the two MPCDI categories. (A) Variations in the reaction to commonly used chemotherapeutic drugs between groups with high and low MPCDI. (B) Relationship between the constructed MPCDI genes and their sensitivity to drugs. (C) The association of GDSC drug sensitivity with the mRNA expression levels of the three characteristic IRPCDRGs within the constructed MPCDI. (D) The association of CTRP drug sensitivity with the mRNA expression levels of the three characteristic IRPCDRGs within the constructed MPCDI.

Prediction of immunotherapy efficacy

Patients with various malignancies have benefited from clinical immunotherapy using immune checkpoint inhibitors. TIDE analysis results showed that in low-MPCDI patients, T cell dysfunction score, T cell rejection score, and TIDE score increased, while MSI score decreased, indicating that the immune escape ability was enhanced, which may lead to the reduced therapeutic effect of ICI (Figure 10A–D). Conversely, findings from the IPS analysis revealed that low-MPCDI patients derived significant benefits from ICI therapy (Figure 10E–H). Based on the score of immune-related pathways by ssGSEA, the high-MPCDI patients showed markedly higher activity in a variety of pathways containing APC co-inhibition, cytolytic activity, type II IFN response, inflammation-promoting, T cell co-inhibition, APC co-stimulation, and MHC class I pathways (Figure 10I). It was observed that, except for the ITGAV gene, three characteristic IRPCDRGs were significantly negatively related to the largest number of immune checkpoints (Figure 10J). There was a remarkable increase in the expression levels of the largest number of immune checkpoints in the high-MPCDI patients, including NRP1, VTCN1, TNFSF15, CD276, TNFSF18 and CD160, compared to the low-MPCDI patients (Figure 10K). In the IMvigor210 cohort, MPCDI scores were higher in immunotherapy-responsive patients (PD/SD) than in nonresponsive patients (CR/PR) (Figure 10L).

Figure 10: The value of the MPCDI in immunotherapy responses. (A) TIDE score, (B) T-cell dysfunction score, (C) T-cell rejection score and (D) microsatellite instability (MSI) score in the two groups. (E–H) Comparisons of the IPS in the two MPCDI groups. (I) Immune-related pathways’ activity showing a significant difference between the high-MPCDI group and the low-MPCDI group. (J) MPCDI gene expression was associated with common immune checkpoints. (K) Expression levels of immune checkpoints in high- and low-MPCDI groups. (L) The IMvigor 210 database analyzed MPCDI in the responding and non-responding groups to immunotherapy. (Ns, no significance; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).
Figure 10:

The value of the MPCDI in immunotherapy responses. (A) TIDE score, (B) T-cell dysfunction score, (C) T-cell rejection score and (D) microsatellite instability (MSI) score in the two groups. (E–H) Comparisons of the IPS in the two MPCDI groups. (I) Immune-related pathways’ activity showing a significant difference between the high-MPCDI group and the low-MPCDI group. (J) MPCDI gene expression was associated with common immune checkpoints. (K) Expression levels of immune checkpoints in high- and low-MPCDI groups. (L) The IMvigor 210 database analyzed MPCDI in the responding and non-responding groups to immunotherapy. (Ns, no significance; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).

Discussion

For all we know, this research marks the initial effort to provide an all-encompassing profile of nineteen PCD interaction patterns in GC patients. We developed an immune-derived MPCDI in the TCGA-STAD cohort and further validated their superior property in two external GEO testing cohorts (GSE84433 and GSE84437). An MPCDI-based nomogram was constructed, and the results showed superior performance. Finally, we determined whether the MPCDI was associated with clinical characterization parameters, clinical prognosis, immune subtypes, somatic mutations, immune infiltration, immune checkpoints, immunotherapeutic reaction, and drug susceptibility in GC patients. In brief, we discovered novel molecular subtypes of GC patients and revealed heterogeneity between subtypes, which may bring new promise for personalized immunotherapy combined with chemotherapy strategies for GC patients.

It is widely recognized that metastasis in cancer is the primary reason for cancer-related fatalities in humans. Growing evidence suggests that the molecular mechanisms of PCD are significantly linked to the progression and spread of different types of tumors [66, 67]. Programmed cell death factor 10 (PDCD10) is an important pro-oncogenic factor that is essential for mesenchymal transition in hepatocellular carcinoma by maintaining low cell adhesion. It has been found that PDCD10 can affect the Hippo signaling pathway through PP2AC-mediated YAP activation, which in turn promotes epithelial mesenchymal transition in hepatocellular carcinoma. knockdown of PDCD10 can inhibit the interaction between PP2AC and YAP, which in turn affects the activation of YAP [66]. There is evidence that tumor cell infiltration and metastasis can be regulated by inducing endothelial cell programmed death through death receptor 6 (DR6), suggesting that the endothelial DR6-mediated necroptosis signaling pathway can become a target for anti-metastasis therapy [67]. In this study, we developed a prognostic model, named MPCDI, containing three characteristic IRPCDRGs (WNT4, ITGAV, and HCAR1), and found that it predicted clinical prognosis in GC patients.

The Wnt signaling pathway is a complex and diverse signaling pathway, its function is most commonly found in embryonic development and cancer, and can be divided into classical and non-canonical pathways [68, 69]. Among them, non-canonical signaling pathways that are not dependent on β-catenin include WNT4, WNT5A, WNT6 and WNT11 [70]. The study found that serum WNT4 levels in colorectal cancer patients were significantly elevated, but the serum level of WNT4 decreased significantly after tumor resection. It can promote epithelial-to-mesenchymal transition (EMT) and activate fibroblasts by activating the WNT4/β-catenin pathway both in vitro and in vivo, and can induce angiogenesis by the WNT4/β-catenin/Ang2 pathway [71]. Experimental evidence suggests that low concentrations of 3,3′-diindolylmethane can induce secretion of Wnt4 and activation of β-catenin signaling, which promotes disease progression in GC patients [72]. Integrin αV (ITGAV, CD51) is a member of the integrin family, which is closely related to the development, advancement, and poor clinical prognosis of many malignant tumors [73]. It was found that CD51 can generate a functional intracellular domain (CD51-ICD) through transmembrane cleavage by γ-secretase, and the cleaved CD51-ICD promotes hepatocellular carcinoma metastasis and invasion by promoting the transcription of OXPHOS-related genes [73]. A recent study found that the process of CD51 cleavage by γ-secretase establishes an association between nerve infiltration, CD51 and γ-secretase, providing a target for clinical treatment of nerve infiltration in colorectal cancer [74]. HCAR1 is a G-protein-coupled receptor for L-lactic acid that is expressed in different cellular populations, including brain cells, adipocytes, various cancer cells, and the retina [75, 76]. It has been found that lactic acid can act as a metabolic regulator of breast cancer cells through HCAR1 [76]. Additionally, HCAR1 exhibited significant expression levels in tissues affected by ovarian cancer [77].

In this study, a consensus clustering analysis was conducted on 28 immune cell infiltrations using ssGSEA analysis to categorize patients with GC into two immune clusters, C1 and C2. It was observed that patients in the C2 had increased stromal, immune, and estimated scores, while tumor purity decreased compared to those in C1. Thus, C1 was labeled as “immune-cold” tumors and C2 as “immune-hot” tumors. By creating new immune clusters in GC patients, an immune-derived MPCDI was developed to assess the risk of death. It was observed that the high-MPCDI patients experienced a poorer overall survival, but the low-MPCDI patients had a more positive overall survival in the TCGA-STAD cohort. Survival analyses of GC patients in the GSE84433 and GSE84437 cohorts showed survival outcomes consistent with the TCGA-STAD cohort. Furthermore, the AUC values at 1, 3, and 10 years were 0.665, 0.615, and 0.614, respectively, in the TCGA-STAD cohort. In the external validation cohorts, the AUC values at 1, 3, and 5 years in the GSE84433 and GSE84437 cohorts were 0.505, 0.600, 0.564 and 0.516, 0.551, 0.566, respectively. These findings indicated that MPCDI has superior predictive power for medium- and long-term survival outcomes in GC patients.

In recent times, exciting breakthroughs have been made in immunotherapy for advanced gastric cancer, with immune checkpoint inhibitors contributing to anti-tumor effects in first-line, second-line, and subsequent treatments, yielding particularly encouraging efficacy in first-line therapy 78], [79], [80], [81. Some patients can obtain long-term clinical benefits from immunotherapy, but most patients inevitably develop primary or secondary drug resistance leading to disease progression [81]. In this study, the high-MPCDI group showed markedly higher activity in a variety of pathways based on the score of immune-related pathways by ssGSEA, containing APC co-inhibition, cytolytic activity, type II IFN response, inflammation-promoting, T cell co-inhibition, APC co-stimulation, and MHC class I pathways. It is exciting to observe that the low-MPCDI patients displayed improved effectiveness with conventional chemotherapy drugs compared to the high-MPCDI patients, indicating that the MPCDI scoring system can offer guidance in the chemotherapy of GC patients. Interestingly, the IPS results manifested that the low-MPCDI patients observably benefited from immunotherapy, suggesting that MPCDI can be used to guide strategies for personalized immunotherapy combined with chemotherapy in GC patients.

Although the MPCDI clinical diagnostic classifier showed excellent performance in both the TCGA-STAD cohort and the externally independent GEO testing cohorts, there are still some limitations. Prospective studies with large clinical cohorts, sufficiently detailed multicenter clinical follow-up data for further validation, and high-quality ex vivo and in vivo molecular experimental evidence (such as PCR, western blot and animal experiments) are still needed to explore the feasibility of using this clinical diagnostic classifier to predict the clinical outcome of GC patients and to guide the combination of clinical immunotherapy with chemotherapy. In addition, the prognostic models developed in the current study still need to be cross-validated in multi-centers and large cohorts in order to provide practical experience for clinical application.

Conclusions

The main findings of this study are shown in Figure 11. We systematically dissected 19 PCD patterns and formulated an MPCDI for GC patients, which provides new indicators for predicting clinical prognosis and immunotherapeutic interventions in GC patients. Besides, we researched the relationship between MPCDI and clinical characterization parameters, clinical prognosis, immune subtypes, somatic mutations, immune infiltration, immune checkpoints, immunotherapeutic responses, and drug sensitivities, which provides a basis for future mechanistic studies.

Figure 11: The main findings of this study. To the best of our knowledge, this study is the first to comprehensively profile nineteen PCD patterns in gastric cancer (GC) patients. In this research, we created an immune-derived multiple programmed cell death index (MPCDI) utilizing a machine learning framework, which can serve as a robust clinical diagnostic classifier to guide medication administration and improve outcomes in GC patients. A nomogram model including MPCDI, age, gender, clinical stage, T stage, N stage, and M stage was established through multivariate Cox and stepwise regression analyses in within TCGA-STAD cohort to predict OS at 1, 3, and 5 years, with the findings demonstrating exceptional performance. The results from the IC50 analysis were promising: the low-MPCDI patients displayed improved effectiveness with conventional chemotherapy drugs compared to the high-MPCDI patients, indicating that the MPCDI scoring system can offer guidance in the chemotherapy of GC patients.
Figure 11:

The main findings of this study. To the best of our knowledge, this study is the first to comprehensively profile nineteen PCD patterns in gastric cancer (GC) patients. In this research, we created an immune-derived multiple programmed cell death index (MPCDI) utilizing a machine learning framework, which can serve as a robust clinical diagnostic classifier to guide medication administration and improve outcomes in GC patients. A nomogram model including MPCDI, age, gender, clinical stage, T stage, N stage, and M stage was established through multivariate Cox and stepwise regression analyses in within TCGA-STAD cohort to predict OS at 1, 3, and 5 years, with the findings demonstrating exceptional performance. The results from the IC50 analysis were promising: the low-MPCDI patients displayed improved effectiveness with conventional chemotherapy drugs compared to the high-MPCDI patients, indicating that the MPCDI scoring system can offer guidance in the chemotherapy of GC patients.

Corresponding authors: Chunhong Li, Central Laboratory, The Second Affiliated Hospital of Guilin Medical University, Guilin 541199, Guangxi, China; and Guangxi Health Commission Key Laboratory of Glucose and Lipid Metabolism Disorders, The Second Affiliated Hospital of Guilin Medical University, Guilin 541199, Guangxi, China, E-mail: ; and Yiming Mao, Department of Thoracic Surgery, Suzhou Kowloon Hospital, Shanghai Jiao Tong University School of Medicine, Suzhou 215028, China, E-mail:

Acknowledgements

We thank Dr. Jianming Zeng (University of Macau), and all the members of his bioinformatics team, biotrainee, for generously sharing their experience and codes.

  1. Research ethics: This study does not contain any studies with human participants or animals performed by any of the authors.

  2. Informed consent: Not applicable.

  3. Author contributions: Li CH, Fan X and Mao YM designed the study and supervised the completion, Li CH, Hu JH, Li MQ, Fan X and Mao YM contributed to data collection and analysis, Li CH wrote the manuscript, Li CH, Fan X and Mao YM reviewed the background and edited the manuscript. All the authors approved the final version of the manuscript.

  4. Competing interests: The authors declare that they have no competing interests.

  5. Research funding: The authors received no financial support for the research authorship and/or publication of this article.

  6. Data availability: The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.

References

1. Sung, H, Ferlay, J, Siegel, RL, Laversanne, M, Soerjomataram, I, Jemal, A, et al.. Global cancer statistics 2020: globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2021;71:209–49. https://doi.org/10.3322/caac.21660.Search in Google Scholar PubMed

2. Hatamian, S, Etesam, S, Mazidimoradi, A, Momenimovahed, Z, Salehiniya, H. The barriers and facilitators of gastric cancer screening: a systematic review. J Gastrointest Cancer 2021;52:839–45. https://doi.org/10.1007/s12029-021-00652-8.Search in Google Scholar PubMed

3. Wagner, AD, Syn, NL, Moehler, M, Grothe, W, Yong, WP, Tai, BC, et al.. Chemotherapy for advanced gastric cancer. Cochrane Db Syst Rev 2017;8. https://doi.org/10.1002/14651858.cd004064.pub4.Search in Google Scholar PubMed PubMed Central

4. Janjigian, YY, Kawazoe, A, Yañez, P, Li, N, Lonardi, S, Kolesnik, O, et al.. The KEYNOTE-811 trial of dual PD-1 and HER2 blockade in HER2-positive gastric cancer. Nature 2021;600:727–30. https://doi.org/10.1038/s41586-021-04161-3.Search in Google Scholar PubMed PubMed Central

5. Wang, J, Huang, Q, Hu, X, Zhang, S, Jiang, Y, Yao, G, et al.. Disrupting circadian rhythm via the PER1-HK2 axis reverses trastuzumab resistance in gastric cancer. Cancer Res 2022;82:1503–17. https://doi.org/10.1158/0008-5472.can-21-1820.Search in Google Scholar PubMed PubMed Central

6. Shitara, K, Baba, E, Fujitani, K, Oki, E, Fujii, S, Yamaguchi, K. Discovery and development of trastuzumab deruxtecan and safety management for patients with HER2-positive gastric cancer. Gastric Cancer 2021;24:780–9. https://doi.org/10.1007/s10120-021-01196-3.Search in Google Scholar PubMed PubMed Central

7. Lu, J, Huang, CM, Zheng, CH, Li, P, Xie, JW, Wang, JB, et al.. Consideration of tumor size improves the accuracy of TNM predictions in patients with gastric cancer after curative gastrectomy. Surg Oncol 2013;22:167–71. https://doi.org/10.1016/j.suronc.2013.05.002.Search in Google Scholar PubMed

8. Huang, ZN, Ma, Y, Chen, QY, Zheng, CH, Li, P, Xie, JW, et al.. Potential survival benefits of open over laparoscopic radical gastrectomy for gastric cancer patients beyond three years after surgery: result from multicenter in-depth analysis based on propensity matching. Surg Endosc 2022;36:1456–65. https://doi.org/10.1007/s00464-021-08430-0.Search in Google Scholar PubMed PubMed Central

9. Dai, P, Rao, X, Zhang, X, Qiu, E, Wu, G, Lin, Y, et al.. Case report: complete remission of a patient with metastatic gastric cancer treated with nivolumab combined with chemotherapy after palliative surgery. Front Immunol 2022;13. https://doi.org/10.3389/fimmu.2022.908558.Search in Google Scholar PubMed PubMed Central

10. Zhu, Z, Dai, PL, Han, S, Qiu, E, Wang, Y, Li, Z. Complete remission in a patient with metastatic gastric cancer receiving tislelizumab combined with chemotherapy: a case report. Front Oncol 2023;13. https://doi.org/10.3389/fonc.2023.1147636.Search in Google Scholar PubMed PubMed Central

11. Cristescu, R, Lee, J, Nebozhyn, M, Kim, KM, Ting, JC, Wong, SS, et al.. Molecular analysis of gastric cancer identifies subtypes associated with distinct clinical outcomes. Nat Med 2015;21:449–56. https://doi.org/10.1038/nm.3850.Search in Google Scholar PubMed

12. Williams, ED, Gao, D, Redfern, A, Thompson, EW. Controversies around epithelial-mesenchymal plasticity in cancer metastasis. Nat Rev Cancer 2019;19:716–32. https://doi.org/10.1038/s41568-019-0213-x.Search in Google Scholar PubMed PubMed Central

13. Tower, J. Programmed cell death in aging. Ageing Res Rev 2015;23:90–100. https://doi.org/10.1016/j.arr.2015.04.002.Search in Google Scholar PubMed PubMed Central

14. Tong, X, Tang, R, Xiao, M, Xu, J, Wang, W, Zhang, B, et al.. Targeting cell death pathways for cancer therapy: recent developments in necroptosis, pyroptosis, ferroptosis, and cuproptosis research. J Hematol Oncol 2022;15:174. https://doi.org/10.1186/s13045-022-01392-3.Search in Google Scholar PubMed PubMed Central

15. Liu, J, Hong, M, Li, Y, Chen, D, Wu, Y, Hu, Y. Programmed cell death tunes tumor immunity. Front Immunol 2022;13. https://doi.org/10.3389/fimmu.2022.847345.Search in Google Scholar PubMed PubMed Central

16. Moujalled, D, Strasser, A, Liddell, JR. Molecular mechanisms of cell death in neurological diseases. Cell Death Differ 2021;28:2029–44. https://doi.org/10.1038/s41418-021-00814-y.Search in Google Scholar PubMed PubMed Central

17. Del, RDP, Amgalan, D, Linkermann, A, Liu, Q, Kitsis, RN. Fundamental mechanisms of regulated cell death and implications for heart disease. Physiol Rev 2019;99:1765–817. https://doi.org/10.1152/physrev.00022.2018.Search in Google Scholar PubMed PubMed Central

18. Dong, Y, Chen, H, Gao, J, Liu, Y, Li, J, Wang, J. Molecular machinery and interplay of apoptosis and autophagy in coronary heart disease. J Mol Cell Cardiol 2019;136:27–41. https://doi.org/10.1016/j.yjmcc.2019.09.001.Search in Google Scholar PubMed

19. Galluzzi, L, Vitale, I, Aaronson, SA, Abrams, JM, Adam, D, Agostinis, P, et al.. Molecular mechanisms of cell death: recommendations of the nomenclature committee on cell death 2018. Cell Death Differ 2018;25:486–541. https://doi.org/10.1038/s41418-017-0012-4.Search in Google Scholar PubMed PubMed Central

20. Tang, D, Kang, R, Berghe, TV, Vandenabeele, P, Kroemer, G. The molecular machinery of regulated cell death. Cell Res 2019;29:347–64. https://doi.org/10.1038/s41422-019-0164-5.Search in Google Scholar PubMed PubMed Central

21. Liu, X, Nie, L, Zhang, Y, Yan, Y, Wang, C, Colic, M, et al.. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol 2023;25:404–14. https://doi.org/10.1038/s41556-023-01091-2.Search in Google Scholar PubMed PubMed Central

22. Zou, Y, Xie, J, Zheng, S, Liu, W, Tang, Y, Tian, W, et al.. Leveraging diverse cell-death patterns to predict the prognosis and drug sensitivity of triple-negative breast cancer patients after surgery. Int J Surg 2022;107. https://doi.org/10.1016/j.ijsu.2022.106936.Search in Google Scholar PubMed

23. Wei, Q, Jiang, X, Miao, X, Zhang, Y, Chen, F, Zhang, P. Molecular subtypes of lung adenocarcinoma patients for prognosis and therapeutic response prediction with machine learning on 13 programmed cell death patterns. J Cancer Res Clin Oncol 2023;149:11351–68. https://doi.org/10.1007/s00432-023-05000-w.Search in Google Scholar PubMed

24. Qin, H, Abulaiti, A, Maimaiti, A, Abulaiti, Z, Fan, G, Aili, Y, et al.. Integrated machine learning survival framework develops a prognostic model based on inter-crosstalk definition of mitochondrial function and cell death patterns in a large multicenter cohort for lower-grade glioma. J Transl Med 2023;21:588. https://doi.org/10.1186/s12967-023-04468-x.Search in Google Scholar PubMed PubMed Central

25. Chen, X, Kang, R, Kroemer, G, Tang, D. Broadening horizons: the role of ferroptosis in cancer. Nat Rev Clin Oncol 2021;18:280–96. https://doi.org/10.1038/s41571-020-00462-0.Search in Google Scholar PubMed

26. Christgen, S, Tweedell, RE, Kanneganti, TD. Programming inflammatory cell death for therapy. Pharmacol Ther 2022;232. https://doi.org/10.1016/j.pharmthera.2021.108010.Search in Google Scholar PubMed PubMed Central

27. Bertheloot, D, Latz, E, Franklin, BS. Necroptosis, pyroptosis and apoptosis: an intricate game of cell death. Cell Mol Immunol 2021;18:1106–21. https://doi.org/10.1038/s41423-020-00630-3.Search in Google Scholar PubMed PubMed Central

28. Samir, P, Malireddi, RKS, Kanneganti, TD. The PANoptosome: a deadly protein complex driving pyroptosis, apoptosis, and necroptosis (PANoptosis). Front Cell Infect Microbiol 2020;10:238. https://doi.org/10.3389/fcimb.2020.00238.Search in Google Scholar PubMed PubMed Central

29. Pandian, N, Kanneganti, TD. PANoptosis: a unique innate immune inflammatory cell death modality. J Immunol 2022;209:1625–33. https://doi.org/10.4049/jimmunol.2200508.Search in Google Scholar PubMed PubMed Central

30. Jiang, M, Qi, L, Li, L, Wu, Y, Song, D, Li, Y. Caspase-8: a key protein of cross-talk signal way in “PANoptosis” in cancer. Int J Cancer 2021;149:1408–20. https://doi.org/10.1002/ijc.33698.Search in Google Scholar PubMed

31. Varfolomeev, EE, Schuchmann, M, Luria, V, Chiannilkulchai, N, Beckmann, JS, Mett, IL, et al.. Targeted disruption of the mouse Caspase 8 gene ablates cell death induction by the TNF receptors, Fas/Apo1, and DR3 and is lethal prenatally. Immunity 1998;9:267–76. https://doi.org/10.1016/s1074-7613(00)80609-3.Search in Google Scholar PubMed

32. Oberst, A, Dillon, CP, Weinlich, R, McCormick, LL, Fitzgerald, P, Pop, C, et al.. Catalytic activity of the caspase-8-FLIP(L) complex inhibits RIPK3-dependent necrosis. Nature 2011;471:363–7. https://doi.org/10.1038/nature09852.Search in Google Scholar PubMed PubMed Central

33. Kaiser, WJ, Upton, JW, Long, AB, Livingston-Rosanoff, D, Daley-Bauer, LP, Hakem, R, et al.. RIP3 mediates the embryonic lethality of caspase-8-deficient mice. Nature 2011;471:368–72. https://doi.org/10.1038/nature09857.Search in Google Scholar PubMed PubMed Central

34. Alvarez-Diaz, S, Dillon, CP, Lalaoui, N, Tanzer, MC, Rodriguez, DA, Lin, A, et al.. The pseudokinase MLKL and the kinase RIPK3 have distinct roles in autoimmune disease caused by loss of death-receptor-induced apoptosis. Immunity 2016;45:513–26. https://doi.org/10.1016/j.immuni.2016.07.016.Search in Google Scholar PubMed PubMed Central

35. Newton, K, Wickliffe, KE, Maltzman, A, Dugger, DL, Reja, R, Zhang, Y, et al.. Activity of caspase-8 determines plasticity between cell death pathways. Nature 2019;575:679–82. https://doi.org/10.1038/s41586-019-1752-8.Search in Google Scholar PubMed

36. Kitada, S, Andersen, J, Akar, S, Zapata, JM, Takayama, S, Krajewski, S, et al.. Expression of apoptosis-regulating proteins in chronic lymphocytic leukemia: correlations with in vitro and in vivo chemoresponses. Blood 1998;91:3379–89. https://doi.org/10.1182/blood.v91.9.3379.Search in Google Scholar

37. Saraswathy, SD, Mirunalini, A, Karthikeyan, K, Premkumar, K. BH3 mimetic peptides: an effective strategy to complement anticancer therapy. Curr Protein Pept Sci 2023;24:853–64. https://doi.org/10.2174/1389203724666230822100131.Search in Google Scholar PubMed

38. Xie, J, Chen, L, Sun, Q, Li, H, Wei, W, Wu, D, et al.. An immune subtype-related prognostic signature of hepatocellular carcinoma based on single-cell sequencing analysis. Aging 2022;14:3276–92. https://doi.org/10.18632/aging.204012.Search in Google Scholar PubMed PubMed Central

39. Li, C, Hu, J, Jiang, X, Tan, H, Mao, Y. Identification and validation of an immune-derived multiple programmed cell death index for predicting clinical outcomes, molecular subtyping, and drug sensitivity in lung adenocarcinoma. Clin Transl Oncol 2024;26:2274–95 [Online ahead of print]. https://doi.org/10.1007/s12094-024-03439-y.Search in Google Scholar PubMed

40. Li, Z, Li, Y, Liu, L, Zhang, C, Li, X. Multiple programmed cell death patterns and immune landscapes in bladder cancer: evidence based on machine learning and multi-cohorts. Environ Toxicol 2024;39:1780–801. https://doi.org/10.1002/tox.24066.Search in Google Scholar PubMed

41. Zhou, N, Yuan, X, Du, Q, Zhang, Z, Shi, X, Bao, J, et al.. FerrDb V2: update of the manually curated database of ferroptosis regulators and ferroptosis-disease associations. Nucleic Acids Res 2023;51:D571–82. https://doi.org/10.1093/nar/gkac935.Search in Google Scholar PubMed PubMed Central

42. Li, C, Qin, W, Hu, J, Lin, J, Mao, Y. A machine learning computational framework develops a multiple programmed cell death index for improving clinical outcomes in bladder cancer. Biochem Genet 2024;14 [Online ahead of print]. https://doi.org/10.1007/s10528-024-10683-y.Search in Google Scholar PubMed

43. Zeng, D, Ye, Z, Shen, R, Yu, G, Wu, J, Xiong, Y, et al.. IOBR: multi-omics immuno-oncology biological research to decode tumor microenvironment and signatures. Front Immunol 2021;12. https://doi.org/10.3389/fimmu.2021.687975.Search in Google Scholar PubMed PubMed Central

44. Becht, E, Giraldo, NA, Lacroix, L, Buttard, B, Elarouci, N, Petitprez, F, et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol 2016; 17:218–249 [Erratum in 2016;17]. https://doi.org/10.1186/s13059-016-1070-5.Search in Google Scholar PubMed PubMed Central

45. Racle, J, de Jonge, K, Baumgaertner, P, Speiser, DE, Gfeller, D. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. Elife 2017;6:e26476. https://doi.org/10.7554/elife.26476.Search in Google Scholar

46. Aran, D, Hu, Z, Butte, AJ. XCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol 2017;18:220. https://doi.org/10.1186/s13059-017-1349-1.Search in Google Scholar PubMed PubMed Central

47. Newman, AM, Liu, CL, Green, MR, Gentles, AJ, Feng, W, Xu, Y, et al.. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods 2015;12:453–7. https://doi.org/10.1038/nmeth.3337.Search in Google Scholar PubMed PubMed Central

48. Charoentong, P, Finotello, F, Angelova, M, Mayer, C, Efremova, M, Rieder, D, et al.. Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade. Cell Rep 2017;18:248–62. https://doi.org/10.1016/j.celrep.2016.12.019.Search in Google Scholar PubMed

49. Finotello, F, Mayer, C, Plattner, C, Laschober, G, Rieder, D, Hackl, H, et al.. Molecular and pharmacological modulators of the tumor immune contexture revealed by deconvolution of RNA-seq data. Genome Med 2019;11:34. https://doi.org/10.1186/s13073-019-0638-6.Search in Google Scholar PubMed PubMed Central

50. Yoshihara, K, Shahmoradgoli, M, Martínez, E, Vegesna, R, Kim, H, Torres-Garcia, W, et al.. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612. https://doi.org/10.1038/ncomms3612.Search in Google Scholar PubMed PubMed Central

51. Li, B, Severson, E, Pignon, JC, Zhao, H, Li, T, Novak, J, et al.. Comprehensive analyses of tumor immunity: implications for cancer immunotherapy. Genome Biol 2016;17:174. https://doi.org/10.1186/s13059-016-1028-7.Search in Google Scholar PubMed PubMed Central

52. Barbie, DA, Tamayo, P, Boehm, JS, Kim, SY, Moody, SE, Dunn, IF, et al.. Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature 2009;462:108–12. https://doi.org/10.1038/nature08460.Search in Google Scholar PubMed PubMed Central

53. Jiang, P, Gu, S, Pan, D, Fu, J, Sahu, A, Hu, X, et al.. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018;24:1550–8. https://doi.org/10.1038/s41591-018-0136-1.Search in Google Scholar PubMed PubMed Central

54. Wilkerson, MD, Hayes, DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics 2010;26:1572–3. https://doi.org/10.1093/bioinformatics/btq170.Search in Google Scholar PubMed PubMed Central

55. Langfelder, P, Horvath, S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinf 2008;9:559. https://doi.org/10.1186/1471-2105-9-559.Search in Google Scholar PubMed PubMed Central

56. López, V, Fernández, A, García, S, Palade, V, Herrera, F. An insight into classification with imbalanced data: empirical results and current trends on using data intrinsic characteristics. Inf Sci 2013;250:113–41. https://doi.org/10.1016/j.ins.2013.07.007.Search in Google Scholar

57. Petralia, F, Wang, P, Yang, J, Tu, Z. Integrative random forest for gene regulatory network inference. Bioinformatics 2015;31:i197–205. https://doi.org/10.1093/bioinformatics/btv268.Search in Google Scholar PubMed PubMed Central

58. Blanche, P, Dartigues, JF, Jacqmin-Gadda, H. Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med 2013;32:5381–97. https://doi.org/10.1002/sim.5958.Search in Google Scholar PubMed

59. Colaprico, A, Silva, TC, Olsen, C, Garofano, L, Cava, C, Garolini, D, et al.. TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res 2016;44:e71. https://doi.org/10.1093/nar/gkv1507.Search in Google Scholar PubMed PubMed Central

60. Mayakonda, A, Lin, D-C, Assenov, Y, Plass, C, Koeffler, HPJG. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res 2018;28:1747–56. https://doi.org/10.1101/gr.239244.118.Search in Google Scholar PubMed PubMed Central

61. Wu, T, Hu, E, Xu, S, Chen, M, Guo, P, Dai, Z, et al.. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation (Camb) 2021;2. https://doi.org/10.1016/j.xinn.2021.100141.Search in Google Scholar PubMed PubMed Central

62. Maeser, D, Gruener, RF, Huang, RS. oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief Bioinform 2021;22. https://doi.org/10.1093/bib/bbab260.Search in Google Scholar PubMed PubMed Central

63. Liu, CJ, Hu, FF, Xia, MX, Han, L, Zhang, Q, Guo, AY. GSCALite: a web server for gene set cancer analysis. Bioinformatics 2018;34:3771–2. https://doi.org/10.1093/bioinformatics/bty411.Search in Google Scholar PubMed

64. Mariathasan, S, Turley, SJ, Nickles, D, Castiglioni, A, Yuen, K, Wang, Y, et al.. TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 2018;554:544–8. https://doi.org/10.1038/nature25501.Search in Google Scholar PubMed PubMed Central

65. Rosenberg, JE, Hoffman-Censits, J, Powles, T, Van Der Heijden, MS, Balar, AV, Necchi, A, et al.. Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial. Lancet 2016;387:1909–20. https://doi.org/10.1016/s0140-6736(16)00561-4.Search in Google Scholar PubMed PubMed Central

66. Sun, B, Zhong, FJ, Xu, C, Li, YM, Zhao, YR, Cao, MM, et al.. Programmed cell death 10 promotes metastasis and epithelial-mesenchymal transition of hepatocellular carcinoma via PP2Ac-mediated YAP activation. Cell Death Dis 2021;12:849. https://doi.org/10.1038/s41419-021-04139-z.Search in Google Scholar PubMed PubMed Central

67. Strilic, B, Yang, L, Albarrán-Juárez, J, Wachsmuth, L, Han, K, Müller, UC, et al.. Tumour-cell-induced endothelial cell necroptosis via death receptor 6 promotes metastasis. Nature 2016;536:215–8. https://doi.org/10.1038/nature19076.Search in Google Scholar PubMed

68. Duchartre, Y, Kim, YM, Kahn, M. The Wnt signaling pathway in cancer. Crit Rev Oncol Hematol 2016;99:141–9. https://doi.org/10.1016/j.critrevonc.2015.12.005.Search in Google Scholar PubMed PubMed Central

69. Zhou, Y, Xu, J, Luo, H, Meng, X, Chen, M, Zhu, D. Wnt signaling pathway in cancer immunotherapy. Cancer Lett 2022;525:84–96. https://doi.org/10.1016/j.canlet.2021.10.034.Search in Google Scholar PubMed

70. Jang, S, Cho, HH, Park, JS, Jeong, HS. Non-canonical Wnt mediated neurogenic differentiation of human bone marrow-derived mesenchymal stem cells. Neurosci Lett 2017;660:68–73. https://doi.org/10.1016/j.neulet.2017.09.023.Search in Google Scholar PubMed

71. Yang, D, Li, Q, Shang, R, Yao, L, Wu, L, Zhang, M, et al.. WNT4 secreted by tumor tissues promotes tumor progression in colorectal cancer by activation of the Wnt/β-catenin signalling pathway. J Exp Clin Cancer Res 2020;39:251. https://doi.org/10.1186/s13046-020-01774-w.Search in Google Scholar PubMed PubMed Central

72. Zhu, Y, Zhang, B, Gong, A, Fu, H, Zhang, X, Shi, H, et al.. Anti-cancer drug 3,3’-diindolylmethane activates Wnt4 signaling to enhance gastric cancer cell stemness and tumorigenesis. Oncotarget 2016;7:16311–24. https://doi.org/10.18632/oncotarget.7684.Search in Google Scholar PubMed PubMed Central

73. Cai, J, Wang, J, Jiang, C, Ye, L, He, X, Huang, J, et al.. Combined inhibition of surface CD51 and γ-secretase-mediated CD51 cleavage improves therapeutic efficacy in experimental metastatic hepatocellular carcinoma. J Hepatol 2023;79:1418–34. https://doi.org/10.1016/j.jhep.2023.08.007.Search in Google Scholar PubMed

74. Huang, T, Lin, Y, Chen, J, Hu, J, Chen, H, Zhang, Y, et al.. CD51 intracellular domain promotes cancer cell neurotropism through interacting with transcription factor NR4A3 in colorectal cancer. Cancers 2023;15:2623. https://doi.org/10.3390/cancers15092623.Search in Google Scholar PubMed PubMed Central

75. Morland, C, Lauritzen, KH, Puchades, M, Holm-Hansen, S, Andersson, K, Gjedde, A, et al.. The lactate receptor, G-protein-coupled receptor 81/hydroxycarboxylic acid receptor 1: expression and action in brain. J Neurosci Res 2015;93:1045–55. https://doi.org/10.1002/jnr.23593.Search in Google Scholar PubMed

76. Jin, L, Guo, Y, Chen, J, Wen, Z, Jiang, Y, Qian, J. Lactate receptor HCAR1 regulates cell growth, metastasis and maintenance of cancer-specific energy metabolism in breast cancer cells. Mol Med Rep 2022;26:268. https://doi.org/10.3892/mmr.2022.12784.Search in Google Scholar PubMed PubMed Central

77. Kerslake, R, Panfilov, S, Mustafa, N, Hall, M, Kyrou, I, Randeva, HS, et al.. Elevated circulating lactate levels and widespread expression of its cognate receptor, hydroxycarboxylic acid receptor 1 (HCAR1). Ovarian Cancer J Clin Med 2022;12:217. https://doi.org/10.3390/jcm12010217.Search in Google Scholar PubMed PubMed Central

78. Ma, X, Jia, S, Wang, G, Liang, M, Guo, T, Du, H, et al.. TRIM28 promotes the escape of gastric cancer cells from immune surveillance by increasing PD-L1 abundance. Signal Transduct Target Ther 2023;8:246. https://doi.org/10.1038/s41392-023-01450-3.Search in Google Scholar PubMed PubMed Central

79. Li, S, Yu, W, Xie, F, Luo, H, Liu, Z, Lv, W, et al.. Neoadjuvant therapy with immune checkpoint blockade, antiangiogenesis, and chemotherapy for locally advanced gastric cancer. Nat Commun 2023;14:8. https://doi.org/10.1038/s41467-022-35431-x.Search in Google Scholar PubMed PubMed Central

80. Kono, K, Nakajima, S, Mimura, K. Current status of immune checkpoint inhibitors for gastric cancer. Gastric Cancer 2020;23:565–78. https://doi.org/10.1007/s10120-020-01090-4.Search in Google Scholar PubMed

81. Högner, A, Moehler, M. Immunotherapy in gastric cancer. Curr Oncol 2022;29:1559–74. https://doi.org/10.3390/curroncol29030131.Search in Google Scholar PubMed PubMed Central

Supplementary Material

This article contains supplementary material (https://doi.org/10.1515/oncologie-2024-0284).

Received: 2024-06-07
Accepted: 2024-08-19
Published Online: 2024-09-03

© 2024 the author(s), published by De Gruyter on behalf of Tech Science Press (TSP)

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Frontmatter
  2. Review Articles
  3. Recent advances in organelle-specific autophagy in melanoma
  4. Photodynamic therapy in glioma cell culture
  5. Targeting tumor microenvironments with gold nanoparticles for enhanced photothermal therapy
  6. Mechanistic insights into traditional Chinese medicine for digestive tract cancers: implications for gastric, hepatic, esophageal, intestinal, and pancreatic tumors
  7. Effectiveness of supervised combined aerobic and resistance exercise in fatigue of prostate cancer survivors under androgen deprivation therapy: a systematic review and meta-analysis
  8. Anti-cancer effects of hyperbaric oxygen therapy in mice: a meta-analysis
  9. Research Articles
  10. Y27632 induces tongue squamous cell carcinoma cell apoptosis through MAPK-ERK/JNK signal
  11. FTX promotes esophageal cancer progression and desensitizes esophageal cancer cells to ionizing radiation by microRNA-99a/b-3p/WEE1/ERCC1 axis
  12. Exosomal microRNA-21-5p from gastric cancer cells promotes angiogenesis by targeting LEMD3 in human endothelial cells
  13. Sex differences in prognosis of primary bone cancer: a propensity score-matched study
  14. Real-world analysis of the incidence and risk factors of pneumonitis in non-small cell lung cancer patients treated with combined thoracic radiotherapy and immunotherapy
  15. Integrating machine learning and multi-omics analysis to develop an immune-derived multiple programmed cell death signature for predicting clinical outcomes in gastric cancer
  16. Comprehensive analysis of NOTCH pathway with tumor environment in pancreatic adenocarcinoma
  17. Comprehensive bioinformatics analysis of lncRNA regulation and screening for pathogenic genes in NF2-related schwannomatosis
  18. Short Commentary
  19. The role of perioperative treatment in radiation-associated soft tissue sarcomas
https://doi.org/10.1515/oncologie-2024-0284
Keywords for this article
gastric cancer; multi-omics; immunotherapy; programmed cell death; machine learning
Creative Commons
BY 4.0

Articles in the same Issue

  1. Frontmatter
  2. Review Articles
  3. Recent advances in organelle-specific autophagy in melanoma
  4. Photodynamic therapy in glioma cell culture
  5. Targeting tumor microenvironments with gold nanoparticles for enhanced photothermal therapy
  6. Mechanistic insights into traditional Chinese medicine for digestive tract cancers: implications for gastric, hepatic, esophageal, intestinal, and pancreatic tumors
  7. Effectiveness of supervised combined aerobic and resistance exercise in fatigue of prostate cancer survivors under androgen deprivation therapy: a systematic review and meta-analysis
  8. Anti-cancer effects of hyperbaric oxygen therapy in mice: a meta-analysis
  9. Research Articles
  10. Y27632 induces tongue squamous cell carcinoma cell apoptosis through MAPK-ERK/JNK signal
  11. FTX promotes esophageal cancer progression and desensitizes esophageal cancer cells to ionizing radiation by microRNA-99a/b-3p/WEE1/ERCC1 axis
  12. Exosomal microRNA-21-5p from gastric cancer cells promotes angiogenesis by targeting LEMD3 in human endothelial cells
  13. Sex differences in prognosis of primary bone cancer: a propensity score-matched study
  14. Real-world analysis of the incidence and risk factors of pneumonitis in non-small cell lung cancer patients treated with combined thoracic radiotherapy and immunotherapy
  15. Integrating machine learning and multi-omics analysis to develop an immune-derived multiple programmed cell death signature for predicting clinical outcomes in gastric cancer
  16. Comprehensive analysis of NOTCH pathway with tumor environment in pancreatic adenocarcinoma
  17. Comprehensive bioinformatics analysis of lncRNA regulation and screening for pathogenic genes in NF2-related schwannomatosis
  18. Short Commentary
  19. The role of perioperative treatment in radiation-associated soft tissue sarcomas
Sign up now to receive a 20% welcome discount
Institutional Access
How does access work?
Have an idea on how to improve our website?
Please write us.
© 2025 De Gruyter Brill
Downloaded on 26.10.2025 from https://www.degruyterbrill.com/document/doi/10.1515/oncologie-2024-0284/html
Scroll to top button