Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2024 Mar 8;103(10):e37314. doi: 10.1097/MD.0000000000037314

Machine learning-based cell death signature for predicting the prognosis and immunotherapy benefit in stomach adenocarcinoma

Fan Li a, Qian Feng a, Ran Tao a,*
PMCID: PMC10919539  PMID: 38457593

Abstract

Stomach adenocarcinoma (STAD) is a one of most common malignancies with high mortality-to-incidence ratio. Programmed cell death (PCD) exerts vital functions in the progression of cancer. The role of PCD-related genes (PRGs) in STAD are not fully clarified. Using TCGA, GSE15459, GSE26253, GSE62254 and GSE84437 datasets, PCD-related signature (PRS) was constructed with an integrative procedure including 10 machine learning methods. The role of PRS in predicting the immunotherapy benefits was evaluated by several predicting score and 3 immunotherapy datasets (GSE91061, GSE78220, and IMvigor210). The model developed by Lasso + CoxBoost algorithm having a highest average C-index of 0.66 was considered as the optimal PRS. As an independent risk factor for STAD patients, PRS had a good performance in predicting the overall survival rate of patients, with an AUC of 1-, 3-, and 5-year ROC curve being 0.771, 0.751 and 0.827 in TCGA cohort. High PRS score demonstrated a lower gene set score of some immune-activated cells and immune-activated activities. Patient with high PRS score had a higher TIDE score, higher immune escape score, lower PD1&CTLA4 immunophenoscore, lower TMB score, lower response rate and poor prognosis, indicating a less immunotherapy response. The IC50 value of some drugs correlated with chemotherapy and targeted therapy was higher in high PRS score group. Our investigation developed an optimal PRS in STAD and it acted as an indicator for predicting the prognosis, stratifying risk and guiding treatment for STAD patients.

Keywords: immunotherapy, machine learning, prognostic signature, programmed cell death, Stomach adenocarcinoma

1. Introduction

Gastric cancer is one of the most common malignancies and it has a high mortality-to-incidence ratio.[1] There are estimated 1089,103 new diagnosed gastric cancer cases and 768,793 gastric cancer-related deaths.[2] Stomach adenocarcinoma (STAD) is the most important subtype of gastric cancer.[3] STAD is a multifactorial and complex tumor with environmental and genetic factors contributing to its occurrence and development.[4] Though multidisciplinary approaches, including surgery, chemoradiotherapy, targeted therapy and immunotherapy, have been used for the treatment of STAD, the clinical outcomes of STAD patients are still poor.[5] Worse still, limited effective biomarkers for predicting the prognosis and therapy benefits of STAD patients have been put into clinical use. These grim data highlight the urgent need for markers to predict the prognosis and therapy benefits in STAD.

Cell death, including accident cell death and programmed cell death (PCD), exerts vital functions in the progression of cancer. PCD is a complex multi-step regulatory process involving multiple mechanisms.[6] As far as we know, PCD could be clustered into 15 patterns (Supplementary Table 1, http://links.lww.com/MD/L774).[6–8] Pyroptosis, gasdermin-mediated inflammatory cell death, plays a vital role in the progression and immunotherapy of cancer.[9,10] As a copper-related cell death, cuproptosis is correlated with tumor metastasis and therapy response.[11] Ferroptosis also has many biological functions in the development of cancer.[12] By regulating tumor progression, autophagy is linked to the prognosis of cancer patients.[13] At present, limited studies have been performed to fully explore the role of PCD-related genes (PRGs) in STAD.

After obtaining PRGs from previous studies or databases, we developed a PCD-related signature (PRS) using the data from the Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) and Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/gds). We then evaluated the role of PRS in predicting the prognosis, immune infiltration, and therapy benefits in STAD.

2. Materials and methods

2.1. Datasets sources and identification of differentially expressed genes

Supplementary Table 1, http://links.lww.com/MD/L774 showed the 15 PCD patterns-related genes from MSigDB (http://software.broadinstitute.org/gsea/msigdb/index.jsp), Kyoto Encyclopedia of Genes and Genomes, review articles, and manual collection of gene sets from Gene cards website (https://www.genecards.org/).[14,15] mRNA data of STAD were collated from TCGA (n = 325), GSE15459(n = 175), GSE26253(n = 432), GSE62254 (n = 297) and GSE84437 (n = 423) datasets. Three immunotherapy datasets, including GSE91061 (n = 98), GSE78220 (N = 28), and IMvigor210 (n = 298), were used to evaluate the role of PRS in predicting the immunotherapy benefits. Compared with normal tissues, the differentially expressed genes (DEGs) in STAD were identified with “limma” packages with |Log2FC| value > 1.5 and P value < .05 as threshold value.

2.2. Integrative machine learning algorithms constructed an optimal PRS

After performing univariate cox analysis for screening out potential prognostic biomarkers from these DEGs in STAD, we submitted these biomarkers into integrative analysis procedure containing 10 machine learning methods for constructing a stable PRS. These 10 machine learning methods are random survival forest, elastic network, Lasso, Ridge, stepwise Cox, CoxBoost, partial least squares regression for Cox, supervised principal components, generalized boosted regression modeling, and survival support vector machine. The PRS was constructed with following 4 steps: The prediction model of TCGA data set was fitted with 101 algorithms combinations; All algorithm combinations were performed in GEO cohorts; C-index was calculated across all cohorts. The process of the R scripts (https://github.com/Zaoqu-Liu/IRLS) were following with previous studies.[16,17] Based on the constitutive genes and corresponding coefficients, we then calculated the PRS score (risk score) of STAD cases. With the best cutoff determined by “surv_cutpoint” function within the R package “survminer,” we clustered STAD cases into high and low-risk score groups.

2.3. The performance of PRS in predicting the prognosis of STAD patients

The survival curves of different risk score groups were drawn using Kaplan–Meier survival method. The draw of ROC curve and C-index was determined with the “survivalROC” package. The risk factors for the OS rate of STAD patients were identified by univariate and multivariate Cox analyses. Using PRS score and risk factors, we then developed a predictive nomogram using “nomogramEx” package, with which we could predict the prognosis of STAD patients.

2.4. Immune infiltration analysis

The immune score and ESTIMATE score of each STAD case were determined by ESTIMATE method.[18] After being systematically benchmarked and found to have unique properties and strengths, 7 methods (TIMER, xCell, MCP-counter, CIBERSORT, CIBERSORT-ABS, EPIC, and quanTIseq) were applied to evaluate the level of immune cells in STAD.[19] Relative expression of human leukocyte antigen (HLA)-related genes and immune checkpoints in different risk score groups were visualized with “ggpubr” or “ggplot” R package. Immune cells and immune functions-related gene set score of STAD cases were evaluated by ssGSEA method with “GSVA” package.

2.5. Therapy benefits analysis

Tumor immune dysfunction and exclusion score from TIDE (http://tide.dfci.harvard.edu/), immunophenoscore from The Cancer Immunome Atlas (https://tcia.at/home), immune escape score from previous study[20] and tumor mutation burden (TMB) score from TCGA were used to evaluate the performance of PRS in predicting the immunotherapy response of STAD patients. Higher TIDE score, higher tumor escape score, lower immunophenoscore and lower TMB score indicate a more likelihood of immune escape and worse effectiveness of immunotherapy treatment. The IC50 of drugs in each STAD case was evaluated by “oncoPredict R” package with the data from Genomics of Drug Sensitivity in Cancer (https://www.cancerrxgene.org/). A higher IC50 value indicated lower sensitivity.

3. Results

3.1. Identification of potential prognostic biomarkers

Compared with normal tissues, we identified 7813 DEGs in STAD tissues with |Log fold change| ≥ 1.5 and P value < .05 as the cutoff. (Fig. 1A). And a total of 1071 genes were PRGs among these DEGs (Fig. 1B). As shown in Figure 1C, we then identified 62 genes that were significantly linked to the overall survival rate of STAD (P < .05).

Figure 1.

Figure 1.

Open in a new tab

Potential biomarkers among programmed cell death-related genes in STAD. (A) Differentially expressed genes in STAD. (B) The overlap between differentially expressed genes and programmed cell death-related genes. (C) Potential biomarkers identified using Univariate cox analysis. STAD = stomach adenocarcinoma.

3.2. Machine learning developed a prognostic PRS

These 62 genes were submitted into a machine learning-based integrative procedure for developing a PRS. A total of 101 kinds prediction models fitted via the LOOCV framework in TCGA cohort, and the C-index of each model across all GEO cohorts was further calculated (Fig. 2A). As Figure 2A showed the C-index of each prediction model in all cohorts, the PRS developed by Lasso + CoxBoost algorithm having a highest average C-index of 0.66 was considered as the optimal PRS (Fig. 2A). In prediction model developed by Lasso + CoxBoost algorithm, the PRS was constructed with 16 genes and the PRS score (risk score) of STAD cases were calculated using following the formula: risk score = 0.142 × MATN3exp + 0.089 × SERPINE1exp + 0.035 × GRPexp + 0.094 × LOXexp + 0.328 × CXCR4exp + 0.130 × ZFP36exp + (−0.168) × CDC37exp + 0.039 × ANXA5exp + (−0.060) × FAM111Aexp + 0.159 × NT5Eexp + (−0.067) × CD82exp + (−0.080) × BCL2L14exp + (−0.326) × CTLA4exp + (−0.022) × TNFAIP2exp + (−0.084) × PTPN6exp + (−0.129) × DAB2IP. We then clustered STAD cases into high and low PRS score group. In survival analysis, we found a better clinical outcome in STAD patients with low PRS score among TCGA, GSE15459, GSE26253, GSE62254 and GSE84437 datasets, with 1-, 3-, and 5-year AUCs of 0. 771, 0.752, and 0.827 in TCGA cohort; 0.646, 0.666, and 0.667 in GSE15459 cohort; 0.625, 0.665, and 0.684 in GSE26253 cohort; 0.667, 0.678, and 0.677 in GSE62254 cohort, 0.636, 0.634, and 0.667 in GSE84437 cohort, respectively (Fig. 2B–F, all P < .05).

Figure 2.

Figure 2.

Open in a new tab

Construction of PRS with integrative machine learning algorithms. (A) The C-index of 101 kinds prognostic models constructed with 10 machine learning algorithms in TCGA and GEO datasets. (B–F) The survival curves of different PRS score groups and their corresponding ROC curves in TCGA, GSE15459, GSE26253, GSE62254 and GSE84437 cohort. GEO = gene expression omnibus, PRS = PCD-related signature, TCGA = the Cancer Genome Atlas.

3.3. The performance of PRS in evaluating the prognosis of STAD patients

For comparing the performance of PRS and other clinical characters in evaluating the prognosis of STAD cases, we also calculated the C-index of PRS and these clinical characters. The data revealed a higher C-index of PRS than that of clinical characters, including age, gender, tumor grade and clinical stage, in TCGA and GEO cohorts (Fig. 3A). The results of univariate and multivariate cox regression analysis were shown in Figure 3B and C, demonstrating PRS as an independent risk factor for the prognosis of STAD cases in TCGA and GEO datasets (all P < .05). As shown in Figure 3D, we then developed a nomogram for predicting the clinical outcome of STAD patients using risk score and stage. The calibration plots suggested excellent agreement between the nomogram prediction and actual observation in terms of the 1, 3- and 5-year survival rates in the TCGA cohort (Fig. 3E). All the data may suggest a powerful and stable performance of PRS in predicting the clinical outcome of STAD cases.

Figure 3.

Figure 3.

Open in a new tab

The performance of PRS in evaluating the prognosis of STAD patients. (A) The C-index of PRS and clinical characters in predicting the prognosis of STAD patients in TCGA and GEO datasets. (B–C) Risk factors in STAD identified by univariate and multivariate cox regression analysis. (D–E) Predictive nomogram and calibration evaluating the overall survival rate of STAD patients. GEO = gene expression omnibus, PRS = PCD-related signature, STAD = stomach adenocarcinoma, TCGA = the Cancer Genome Atlas.

3.4. Dissection of PRS-based tumor immune microenvironment

Figure 4A showed the correlation atlas between PRS score and immune cells in STAD. PRS score showed negative correlation with immune activated cells (CD8+ T cells, CD4+ T cells) and positive correlation with immune suppressed cells (cancer-related fibroblast) (Fig. 4B–D, P < .05). As shown in Figure 4E, a higher level of CD8+ T cells, mast cells, NK cells, and TIL in STAD patients with low PRS score. Moreover, the data suggested a higher gene set score of APC_costimulation, cytolytic activity, inflammation-promoting, and T cell co-stimulation in low PRS score group (Fig. 4F). As shown in Figures 4G to I, high PRS score group had a lower stromal score, immune score and ESTIMAE score in STAD (all P < .05).

Figure 4.

Figure 4.

Open in a new tab

PRS-based immune infiltration atlas in STAD. (A) Association atlas between PRS and immune cells in STAD using 7 state-of-the-art algorithms for analysis. (B–D) The association between PRS score and CD8 + T cell, CD4 + T cell and fibroblast. The relative level of immune cells (E), immune related functions (F), immune score (G), stroma score (H) and ESTIMAE score (I) in different PRS score group. *P < .05, **P < .01, ***P < .001. PRS = PCD-related signature, STAD = stomach adenocarcinoma.

3.5. PRS as an indicator for predicting therapy benefits in STAD

STAD patients with lower PRS score had a higher PD1&CTLA4 immunophenoscore and TMB score (Fig. 5A–B, all P < .05). We also found a lower immune escape, lower TIDE score, lower T cell exclusion and dysfunction score in low PRS score group (Fig. 5C–D, all P < .05). As shown in Figure 5E and F, STAD patients with low PRS score also had a higher expression of immune checkpoints and HLA-related genes (all P < .05). Higher immunophenoscore demonstrated higher potential for receiving immunotherapy benefits.[21] Higher TIDE score suggested a less likelihood of immune escape and a better response to immunotherapy.[22,23] High HLA-related gene and immune checkpoints expression suggested wider range of antigen presentation, increasing the likelihood of immunotherapy benefits.[24] Thus, STAD patients with low PRS scores may have a better immunotherapy benefit. We also verified our results using 3 immunotherapy datasets, including GSE91061, GSE78220 and IMivigor210. In cutaneous melanoma patients receiving anti-PD-L1 therapy (GSE91061), non-responders had a higher PRS score versus responders (Fig. 5G, P < .01). Moreover, patients with low PRS score experienced a better clinical outcome (P = .016). We also found a higher response rate in patients low PRS score (P < .05). Interestingly, we obtained similar results in GSE78220 cohort and IMigor210 cohort (Fig. 5H–I). We also explored the IC50 value of STAD cases as the significant role of chemotherapy and targeted therapy in the therapy of STAD. As shown in Figure 6A and B, the result suggested a low IC50 value of 5-Fluorouracil, Camptothecin, Gemcitabine, Docetaxel, Cyclophosphamide, Nilotinib, Crizotinib, Erlotinib, Gefitinib, Lapatinib in STAD patients with high PRS score (all P < .05), indicating that STAD patients with low PRS score had better sensitivity to chemotherapy and targeted therapy.

Figure 5.

Figure 5.

Open in a new tab

PRS acted as an indicator for predicting the immunotherapy benefits in STAD. The level of PD1&CTLA4 immunophenoscore (A), TMB score (B), immune escape (C), TIDE score (D), HLA-related genes (E) and immune checkpoints (F) in different PRS score of STAD patients. The immunotherapy response and overall rate in patients with high and low PRS score in GSE91061 (G), GSE78220 (H) datasets and IMvigor210 (I). *P < .05, **P < .01, ***P < .001. PRS = PCD-related signature, STAD = stomach adenocarcinoma.

Figure 6.

Figure 6.

Open in a new tab

Drug sensitivity analysis. High PRS score indicated a high IC50 value of common drugs in chemotherapy (A) and targeted therapy (B). PRS = PCD-related signature.

3.6. Difference in functional enrichment in different PRS score group

Low PRS score group had a lower gene sets score involved in hypoxia, E2F targets, angiogenesis, coagulation, DNA repair, G2M checkpoints, glycolysis, IL2-STAT5 signaling, mTORC1 signaling, and NOTCH signaling in STAD (Fig. 7A, all P < .05).In GSEA analysis, high PRS score was significantly correlated with NOTCH signaling pathway, pathways in cancer and DNA replication while low PRS score was significantly correlated with chemokine signaling, ribosome, T cell receptor signaling (Fig. 7B–C).

Figure 7.

Figure 7.

Open in a new tab

PRS-based function analysis in STAD. (A) The relative gene set score correlated with cancer hallmarks in different PRS score group in STAD. (B–C) GSEA analysis revealing the functional enrichment in different PRS score group in STAD. GSEA = gene set enrichment analyses, PRS = PCD-related signature, STAD = stomach adenocarcinoma.

4. Discussion

The current investigation developed a PRS for STAD based on 10 integrative machine learning methods. The predicted model developed by Lasso + CoxBoost algorithm had a highest average C-index and was considered as the optimal PRS. The PRS acted as an independent risk factor for STAD patients and it had a good performance in predicting the overall survival rate of patients. We also found that PRS act as indicator for predicting immunotherapy benefits and low PRS score indicated a better immunotherapy benefit.

A total of 16 genes, including MATN3, SERPINE1, GRP, LOX, CXCR4, ZFP36, CDC37, ANXA5, FAM111A, NT5E, CD82, BCL2L14, CTLA4, TNFAIP2, PTPN6, and DAB2IP, were used to construct the PRS. Previous study also suggested MATN3 as a prognostic biomarker for gastric cancer.[25] High LOX resulted in dismal clinical outcome of STAD by promoting M2 macrophage polarization.[26] CXCR7 was involved in the proliferation and migration of STAD patients.[27] miR-362-3p suppressed the migration and invasion in STAD by regulating CD82 expression.[28] Downregulation of DAB2IP lead to tumor progression and cisplatin resistance in STAD.[29]

In our study, the data suggested PRS as an independent risk factor for STAD patients and PRS had a good performance in predicting the overall survival rate of patients. Programmed cell death related signature also acted as biomarker in some cancers. In triple-negative breast cancer, PRS could predict the prognosis and drug sensitivity of patients.[6] Immunogenic PRS was correlated with the prognosis of lower-grade glioma patients.[30] PRS also acted as a biomarker for the prognosis in lung adenocarcinoma.[31]

Increasing evidences suggested the vital role of immunotherapy in the therapeutic strategy in STAD.[32,33] In our study, the data found that STAD patients with low PRS score had a lower TIDE score, lower immune escape score, higher PD1&CTLA4 immunophenoscore, higher TMB score, higher response rate and better prognosis. Actually, TMB and MSI were biomarkers for predicting immunotherapy benefits and high TMB/MSI score were correlated with better immunotherapy response.[34,35] Low TIDE score indicated a less likelihood of immune escape.[22] Thus, PRS may act as an indicator for predicting immunotherapy benefits in STAD and low IRS score indicated a better immunotherapy benefit.

Low PRS score group had a lower gene sets score involved in hypoxia, E2F targets, angiogenesis, coagulation, DNA repair, G2M checkpoints, glycolysis, IL2-STAT5 signaling, mTORC1 signaling, and NOTCH signaling in STAD. Previous study suggested angiogenesis as primary promotor for tumor progression and metastasis in STAD.[36] EMT signaling was also involved in the development and progression of STAD.[37] As the vital role of glycolysis in the survival and metastasis of STAD, targeting glycolysis was suggested as a potential strategy for STAD patients.[38] These data suggested that PRS may mediate the progression of STAD via these cancer related hallmarks.

5. Conclusion

Our investigation developed an optimal PRS in STAD and it acted a indicator for predicting the prognosis, stratifying risk and guiding treatment for STAD patients.

Author contributions

Investigation: Fan Li.

Methodology: Qian Feng.

Project administration: Ran Tao.

Validation: Ran Tao.

Writing – original draft: Fan Li.

Writing – review & editing: Qian Feng, Ran Tao.

Supplementary Material

medi-103-e37314-s001.xlsx (64.9KB, xlsx)

Abbreviations:

GEO
gene expression omnibus
GSEA
gene set enrichment analyses
PCD
programmed cell death
PRS
PCD-related signature
STAD
stomach adenocarcinoma
TCGA
the Cancer Genome Atlas

Our study is not involved in the use of human subjects and animals. No ethical approval and informed consents are required.

All data generated or analyzed during this study are included in this published article [and its supplementary information files].

Supplemental Digital Content is available for this article.

The authors have no conflicts of interest to disclose.

This study was supported by Natural Science Foundation of Jiangxi Provincial Department of Science and Technology (20224BAB216059), Science and Technology Research Project of Jiangxi Provincial Department of Education (GJJ2200116, GJJ2200173), Science and Technology Program of Jiangxi Provincial Administration of traditional Chinese Medicine (2022A342), and Science and Technology Plan Project of Jiangxi Provincial Health Commission (202410229).

How to cite this article: Li F, Feng Q, Tao R. Machine learning-based cell death signature for predicting the prognosis and immunotherapy benefit in stomach adenocarcinoma. Medicine 2024;103:10(e37314).

Contributor Information

Fan Li, Email: 742901772@qq.com.

Qian Feng, Email: fengqian870804@163.com.

References

  • [1].Wong MCS, Huang J, Chan PSF, et al. Global incidence and mortality of gastric cancer, 1980-2018. JAMA Netw Open. 2021;4:e2118457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Sung H, Ferlay J, Siegel RL, 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. [DOI] [PubMed] [Google Scholar]
  • [3].Luo L, Li L, Liu L, et al. A necroptosis-related lncRNA-based signature to predict prognosis and probe molecular characteristics of stomach adenocarcinoma. Front Genet. 2022;13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Machlowska J, Baj J, Sitarz M, et al. Gastric cancer: epidemiology, risk factors, classification, genomic characteristics and treatment strategies. Int J Mol Sci. 2020;21:4012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].He Y, Zhang H, Zhang Y, et al. Comprehensive characterization of transforming growth factor beta receptor 1 in stomach adenocarcinoma identifies a prognostic signature for predicting clinical outcomes and immune infiltrates. Int J Gen Med. 2022;15:3375–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Zou Y, Xie J, Zheng S, 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;107:106936. [DOI] [PubMed] [Google Scholar]
  • [7].Müller GJ, Hasseldam H, Rasmussen RS, et al. Dexamethasone enhances necrosis-like neuronal death in ischemic rat hippocampus involving μ-calpain activation. Exp Neurol. 2014;261:711–9. [DOI] [PubMed] [Google Scholar]
  • [8].Liu J, Hong M, Li Y, et al. Programmed cell death tunes tumor immunity. Front Immunol. 2022;13:847345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Yang F, Bettadapura SN, Smeltzer MS, et al. Pyroptosis and pyroptosis-inducing cancer drugs. Acta Pharmacol Sin. 2022;43:2462–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Lin W, Chen Y, Wu B, et al. Identification of the pyroptosis-related prognostic gene signature and the associated regulation axis in lung adenocarcinoma. Cell Death Discovery. 2021;7:161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Li C, Xiao Y, Cao H, et al. Cuproptosis regulates microenvironment and affects prognosis in prostate cancer. Biol Trace Elem Res. 2023;202 [DOI] [PubMed] [Google Scholar]
  • [12].Wang H, Cheng Y, Mao C, et al. Emerging mechanisms and targeted therapy of ferroptosis in cancer. Mol Ther. 2021;29:2185–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Klionsky DJ, Petroni G, Amaravadi RK, et al. Autophagy in major human diseases. EMBO J. 2021;40:e108863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Tang D, Kang R, Berghe TV, et al. The molecular machinery of regulated cell death. Cell Res. 2019;29:347–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Galluzzi L, Vitale I, Aaronson SA, et al. Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death Differ. 2018;25:486–541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Liu Z, Liu L, Weng S, et al. Machine learning-based integration develops an immune-derived lncRNA signature for improving outcomes in colorectal cancer. Nat Commun. 2022;13:816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Li Z, Guo M, Lin W, et al. Machine learning-based integration develops a macrophage-related index for predicting prognosis and immunotherapy response in lung adenocarcinoma. Arch Med Res. 2023;54:102897. [DOI] [PubMed] [Google Scholar]
  • [18].Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013;4:2612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Li T, Fu J, Zeng Z, et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 2020;48:W509–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Sun Y, Wu L, Zhong Y, et al. Single-cell landscape of the ecosystem in early-relapse hepatocellular carcinoma. Cell. 2021;184:404–421.e16. [DOI] [PubMed] [Google Scholar]
  • [21].Palmeri M, Mehnert J, Silk AW, et al. Real-world application of tumor mutational burden-high (TMB-high) and microsatellite instability (MSI) confirms their utility as immunotherapy biomarkers. ESMO Open. 2022;7:100336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Fu J, Li K, Zhang W, et al. Large-scale public data reuse to model immunotherapy response and resistance. Genome Med. 2020;12:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Lin A, Zhang J, Luo P. Crosstalk between the MSI status and tumor microenvironment in colorectal cancer. Front Immunol. 2020;11:2039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Lin A, Yan WH. HLA-G/ILTs targeted solid cancer immunotherapy: opportunities and challenges. Front Immunol. 2021;12:698677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Wang P, Xiao WS, Li YH, et al. Identification of MATN3 as a novel prognostic biomarker for gastric cancer through comprehensive TCGA and GEO data mining. Dis Markers. 2021;2021:1769635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Nai A, Zeng H, Wu Q, et al. lncRNA/miR-29c-mediated high expression of LOX can influence the immune status and chemosensitivity and can forecast the poor prognosis of gastric cancer. Front Cell Dev Biol. 2021;9:760470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Li N, Xu H, Ou Y, et al. LPS-induced CXCR7 expression promotes gastric cancer proliferation and migration via the TLR4/MD-2 pathway. Diagn Pathol. 2019;14:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Zhang QH, Yao YL, Wu XY, et al. Anti-miR-362-3p inhibits migration and invasion of human gastric cancer cells by its target CD82. Dig Dis Sci. 2015;60:1967–76. [DOI] [PubMed] [Google Scholar]
  • [29].Wang G, Wang X, Han M, et al. Loss of DAB2IP contributes to cell proliferation and cisplatin resistance in gastric cancer. OncoTargets Ther. 2021;14:979–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Cai J, Hu Y, Ye Z, et al. Immunogenic cell death-related risk signature predicts prognosis and characterizes the tumour microenvironment in lower-grade glioma. Front Immunol. 2022;13:1011757. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Ding D, Wang L, Zhang Y, et al. Machine learning developed a programmed cell death signature for predicting prognosis and immunotherapy benefits in lung adenocarcinoma. Transl Oncol. 2023;38:101784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [32].Ruiz-Cordero R, Devine WP. Targeted therapy and checkpoint immunotherapy in lung cancer. Surg Pathol Clin. 2020;13:17–33. [DOI] [PubMed] [Google Scholar]
  • [33].Högner A, Moehler M. Immunotherapy in gastric cancer. Curr Oncol. 2022;29:1559–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Liu L, Bai X, Wang J, et al. Combination of TMB and CNA stratifies prognostic and predictive responses to immunotherapy across metastatic cancer. Clin Cancer Res. 2019;25:7413–23. [DOI] [PubMed] [Google Scholar]
  • [35].Seo I, Lee HW, Byun SJ, et al. Neoadjuvant chemoradiation alters biomarkers of anticancer immunotherapy responses in locally advanced rectal cancer. J ImmunoTher Cancer. 2021;9:e001610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Xu P, Liu S, Song S, et al. Identification and validation of a novel angiogenesis-related gene signature for predicting prognosis in gastric adenocarcinoma. Front Oncol. 2022;12:965102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Baj J, Korona-Głowniak I, Forma A, et al. Mechanisms of the epithelial-mesenchymal transition and tumor microenvironment in helicobacter pylori-induced gastric cancer. Cells. 2020;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Zheng S, Li H, Li Y, et al. The emerging role of glycolysis and immune evasion in gastric cancer. Cancer Cell Int. 2023;23:317. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

medi-103-e37314-s001.xlsx (64.9KB, xlsx)

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES