Molecular Subtypes Of Gastric Cancer Research Articles

Hierarchical learning of gastric cancer molecular subtypes by integrating multi‐modal DNA‐level omics data and clinical stratification

Hierarchical learning of gastric cancer molecular subtypes by integrating multi‐modal DNA‐level omics data and clinical stratification

Clinicopathological characteristics and prognosis of Epstein-Barr virus-associated gastric cancer.

Epstein-Barr virus (EBV)-associated gastric cancer (EBVaGC) is a distinct molecular subtype of gastric cancer (GC). At present, the clinical characteristics and prognostic implications of EBV infection and the potential clinical benefits of immune checkpoint blockade in GC remain to be clarified. Hence, this study was designed to analyze the clinical and pathological characteristics of GC patients with varying EBV infection states and compare their overall survival (OS). A retrospective study was performed on 1031 consecutive GC patients who underwent gastrectomy at the Affiliated Hospital of Xuzhou Medical University from February 2018 to November 2022. EBV-encoded RNA (EBER) in situ hybridization (ISH) was used for EBV assessment, and immunohistochemical staining was used for evaluation of human epidermal growth factor receptor 2 (HER2), programmed death ligand 1 (PD-L1), and Ki67 expression. EBVaGC was defined as tumors with EBV positivity. In addition, EBV-negative GC (EBVnGC) patients were matched with EBVaGC patients based on seven clinicopathological parameters (age, gender, anatomic subsite, tumor size, Lauren classification, degree of differentiation, and tumor-node-metastasis [TNM] stage). The correlations of clinical features with HER2, PD-L1, and Ki67 expression were evaluated statistically. The survival of patients was assessed through medical records, telephone, or WeChat communication, and prognostic analysis was performed using the logrank test as well as univariable and multivariable regression analysis. Out of 1031 GC patients tested, 35 (3.4%) were diagnosed with EBVaGC. Notably, the EBVaGC group exhibited a distinct predominance of males and younger patients, significantly higher Ki67 and PD-L1 expression levels, and a lower prevalence of pericancerous nerve invasion than the EBVnGC group (P < 0.01). In the 35 EBVaGC cases, Ki67 expression was negatively correlated with age (P < 0.05), suggesting that a younger onset age was associated with higher Ki67 expression. In addition, PD-L1 expression was correlated with the degree of differentiation, T-stage, and clinical stage of the patient. Furthermore, PD-L1 expression was elevated in tumors with lower differentiation or at later stages (P < 0.05). Using univariate analysis, Ki67, PD-L1, and clinical stage were identified as significant factors influencing the overall survival (OS) of EBVaGC patients (P < 0.05). Moreover, multivariate survival analysis revealed that clinical stage and Ki67 expression were independent risk factors for the OS of the patients (P < 0.05), and the three-year OS rate of EBVaGC patients was 64.2%. EBV-ISH is a practical and valuable method to identify EBVaGC. Owing to its unique etiological, pathological, and clinical characteristics, patients with EBVaGC might benefit from immune checkpoint blockade therapy.

Abstract P28: Comprehensive Molecular Phenotyping of ARID1A-deficient Gastric Cancer Reveals Pervasive Epigenomic Reprogramming and Therapeutic Opportunities

Abstract Objective Gastric cancer (GC) is a leading cause of cancer mortality, with ARID1A being the second most frequently mutated driver gene in GC. We sought to decipher ARID1A-specific GC regulatory networks and examine therapeutic vulnerabilities arising from ARID1A loss. Design Genomic profiling of GC patients including a Singapore cohort (&amp;gt;200 patients) was performed to derive mutational signatures of ARID1A inactivation across molecular subtypes. Single-cell transcriptomic profiles of ARID1A-mutated GCs were analyzed to examine tumor microenvironmental changes arising from ARID1A loss. Genome-wide ARID1A binding and chromatin profiles (H3K27ac, H3K4me3, H3K4me1, ATAC-seq) of gastric cell lines were generated to identify gastric-specific epigenetic landscapes regulated by ARID1A. Distinct cancer hallmarks of ARID1A-mutated GCs were converged at the genomic, single-cell, and epigenomic level, and targeted by pharmacological inhibition. Results We observed prevalent ARID1A inactivation across GC molecular subtypes, with distinct mutational signatures and linked to a NFKB-driven pro-inflammatory tumour microenvironment. ARID1A-depletion caused loss of H3K27ac activation signals at ARID1A-occupied distal enhancers, but unexpectedly gain of H3K27ac at ARID1A-occupied promoters in genes such as NFKB1 and NFKB2. Promoter activation in ARID1A-mutated GCs was associated with enhanced gene expression, increased BRD4 binding, and reduced HDAC1 and CTCF occupancy. Combined targeting of promoter activation and tumour inflammation via bromodomain and NFKB inhibitors confirmed therapeutic synergy specific to ARID1A-genomic status. Conclusion Our results suggest a therapeutic strategy for ARID1A-mutated GCs targeting both tumour-intrinsic (BRD4-assocatiated promoter activation) and extrinsic (NFKB immunomodulation) cancer phenotypes. Citation Format: Chang Xu, Kie Kyon Huang, Jia Hao Law, Joy Shijia Chua, Taotao Sheng, Natasha M. Flores, Melissa Pool Pizzi, Atsushi Okabe, Angie Lay Keng Tan, Feng Zhu, Vikrant Kumar, Xiaoyin Lu, Ana Morales Benitez, Benedict Shi Xiang Lian, Haoran Ma, Shamaine Wei Ting Ho, Kalpana Ramnarayanan, Chukwuemeka George Anene-Nzelu, Milad Razavi-Mohseni, Siti Aishah Binte Abdul Ghani, Su Ting Tay, Xuewen Ong, Ming Hui Lee, Yu Amanda Guo, Hassan Ashktorab, Duane Smoot, Shang Li, Anders Jacobsen Skanderup, Michael A. Beer, Roger Sik Yin Foo, Joel Shi Hao Wong, Kaushal Sanghvi, Wei Peng Yong, Raghav Sundar, Atsushi Kaneda, Shyam Prabhakar, Pawel Karol Mazur, Jaffer A. Ajani, Khay Guan Yeoh, Jimmy Bok-Yan So, Patrick Tan, Singapore Gastric Cancer Consortium. Comprehensive Molecular Phenotyping of ARID1A-deficient Gastric Cancer Reveals Pervasive Epigenomic Reprogramming and Therapeutic Opportunities [abstract]. In: Proceedings of Frontiers in Cancer Science; 2023 Nov 6-8; Singapore. Philadelphia (PA): AACR; Cancer Res 2024;84(8_Suppl):Abstract nr P28.

Identification of ALYREF in pan cancer as a novel cancer prognostic biomarker and potential regulatory mechanism in gastric cancer

ALYREF is considered as a specific mRNA m5C-binding protein which recognizes m5C sites in RNA and facilitates the export of RNA from the nucleus to the cytoplasm. Expressed in various tissues and highly involved in the transcriptional regulation, ALYREF has the potential to become a novel diagnostic marker and therapeutic target for cancer patients. However, few studies focused on its function during carcinogenesis and progress. In order to explore the role of ALYREF on tumorigenesis, TCGA and GTEx databases were used to investigate the relationship of ALYREF to pan-cancer. We found that ALYREF was highly expressed in majority of cancer types and that elevated expression level was positively associated with poor prognosis in many cancers. GO and KEGG analysis showed that ALYREF to be essential in regulating the cell cycle and gene mismatch repair in tumor progression. The correlation analysis of tumor heterogeneity indicated that ALYREF could be specially correlated to the tumor stemness in stomach adenocarcinoma (STAD). Furthermore, we investigate the potential function of ALYREF on gastric carcinogenesis. Prognostic analysis of different molecular subtypes of gastric cancer (GC) unfolded that high ALYREF expression leads to poor prognosis in certain subtypes of GC. Finally, enrichment analysis revealed that ALYREF-related genes possess the function of regulating cell cycle and apoptosis that cause further influences in GC tumor progression. For further verification, we knocked down the expression of ALYREF by siRNA in GC cell line AGS. Knockdown of ALYREF distinctly contributed to inhibition of GC cell proliferation. Moreover, it is observed that knocked-down of ALYREF induced AGS cells arrested in G1 phase and increased cell apoptosis. Our findings highlighted the essential function of ALYREF in tumorigenesis and revealed the specific contribution of ALYREF to gastric carcinogenesis through pan-cancer analysis and biological experiments.

GCclassifier: An R package for the prediction of molecular subtypes of gastric cancer

Gastric cancer (GC) is one of the most commonly diagnosed malignancies, threatening millions of lives worldwide each year. Importantly, GC is a heterogeneous disease, posing a significant challenge to the selection of patients for more optimized therapy. Over the last decades, extensive community effort has been spent on dissecting the heterogeneity of GC, leading to the identification of distinct molecular subtypes that are clinically relevant. However, so far, no tool is publicly available for GC subtype prediction, hindering the research into GC subtype-specific biological mechanisms, the design of novel targeted agents, and potential clinical applications. To address the unmet need, we developed an R package GCclassifier for predicting GC molecular subtypes based on gene expression profiles. To facilitate the use by non-bioinformaticians, we also provide an interactive, user-friendly web server implementing the major functionalities of GCclassifier. The predictive performance of GCclassifier was demonstrated using case studies on multiple independent datasets.

Genetic landscape and PD-L1 expression in Epstein–Barr virus-associated gastric cancer according to the histological pattern

Epstein–Barr virus (EBV)-associated gastric cancer (EBVaGC) is a distinct molecular subtype of gastric cancer. This study aims to investigate genomic and clinicopathological characteristics of EBVaGC according to the histological pattern. We retrospectively collected 18 specimens of surgically resected EBVaGCs. Whole-exome sequencing was performed for all cases. Moreover, PD-L1 expression and tumor-infiltrating lymphocyte (TIL) percentage were investigated. Among 18 EBVaGCs, 10 cases were of intestinal histology, 3 were of poorly cohesive histology, and the remaining 5 were of gastric carcinoma with lymphoid stroma histology. Whole-exome sequencing revealed that EBVaGCs with intestinal histology harbored pathogenic mutations known to frequently occur in tubular or papillary adenocarcinoma, including TP53, KRAS, FBXW7, MUC6, ERBB2, CTNNB1, and ERBB2 amplifications. One patient with poorly cohesive carcinoma histology harbored a CDH1 mutation. Patients with EBVaGCs with intestinal or poorly cohesive carcinoma histology frequently harbored driver mutations other than PIK3CA, whereas those with EBVaGCs with gastric carcinoma with lymphoid stroma histology lacked other driver mutations. Moreover, the histological pattern of EBVaGCs was significantly associated with the levels of TILs (P = 0.005) and combined positive score (P = 0.027). In conclusion, patients with EBVaGCs with different histological patterns exhibited distinct genetic alteration, PD-L1 expression, and degree of TILs.

Effect of hypoxic status on survival and therapeutic responses in gastric cancer.

e16003 Background: As a significant factor of tumor microenvironment (TME), hypoxia is closely related to cancer metastasis, immune escape and drug resistance. The potential prognosis of hypoxia, as well as its influence on the TME, and therapeutic response have not yet been comprehensively studied in gastric cancer (GC). Here, we developed a prognosis risk assessment model (hypoxia_pred model) based on genes related to hypoxia, and further explored the exact role an impact of hypoxia in GC. Methods: Clinical information and gene expression profile were obtained from Gene Expression Omnibus (GEO) and the Cancer Genome Atlas (TCGA). 375 patients from TCGA-STAD cohort were used as training set. A 21-hypoxic genes expression dataset from TCGA-STAD was performed an unsupervised clustering analysis for different molecular subtypes of GC. Genes associated prognosis were screened by univariate Cox analysis with a cutoff of pvalue＜0.05. The prognostic risk model based on 9-genes transcriptome data of both hypoxia and prognosis related genes was constructed by Least Absolute Shrinkage and Selection Operator regression analysis. Another two gene expression datasets including 409 patients from GEO (GSE62254 and GSE26901) were employed for verifying the forecast performance of this model. In addition, we used ESTATE and single sample Gene Set Enrichment Analysis (ssGSEA) to explore tumor immune infiltration and potential molecular mechanism of hypoxia in GC. We also evaluated the influence of hypoxia on the prognosis, targeted therapy and immunotherapy of GC patients by survival analysis, targeted drug sensitivity analysis and the assessments of the Cancer Immunome Atlas database. Results: We found hypoxia_pred model is a valuable predictor for the survival of GC in validation cohort (AUC &gt; 0.7). AUC for 1- and 3-year overall survival of GC patients in the TCGA cohort were 0.71 and 0.73. In GC, it was supposed that the aggravation of hypoxia was accompanied by the increase of immune rejection phenotype and the infiltration of immunosuppressive cells. In the validation set, patients predicted high risk by the model have worse immunotherapeutic response and prognosis, and may profit from drugs that inhibit cell cycle signaling pathways. In addition, the outcomes by hypoxia_pred for changes in TME, efficacy of immunotherapy and drug sensitivity were quite hypoxia_pred. targeted drug sensitivity, immune microenvironment, biological function, immunotherapeutic response, and immune cell infiltration were found to be significantly different in those molecular subtypes and risk groups. Conclusions: The model based on hypoxic environment provides theoretical foundation for predicting tumor prognosis and clinical treatment, and may improve the prognosis of GC patients by recommending personalized immunotherapy.

Clinical molecular subtyping reveals intrinsic mesenchymal reprogramming in gastric cancer cells

The mesenchymal cancer phenotype is known to be clinically related to treatment resistance and a poor prognosis. We identified gene signature-based molecular subtypes of gastric cancer (GC, n = 547) based on transcriptome data and validated their prognostic and predictive utility in multiple external cohorts. We subsequently examined their associations with tumor microenvironment (TME) features by employing cellular deconvolution methods and sequencing isolated GC populations. We further performed spatial transcriptomics analysis and immunohistochemistry, demonstrating the presence of GC cells in a partial epithelial-mesenchymal transition state. We performed network and pharmacogenomic database analyses to identify TGF-β signaling as a driver pathway and, thus, a therapeutic target. We further validated its expression in tumor cells in preclinical models and a single-cell dataset. Finally, we demonstrated that inhibition of TGF-β signaling negated mesenchymal/stem-like behavior and therapy resistance in GC cell lines and mouse xenograft models. In summary, we show that the mesenchymal GC phenotype could be driven by epithelial cancer cell-intrinsic TGF-β signaling and propose therapeutic strategies based on targeting the tumor-intrinsic mesenchymal reprogramming of medically intractable GC.

Abstract 4863: PMCA inhibition induces cytotoxic calcium stress and suppression of Wnt signaling in refractory molecular subtype of gastric cancer

Abstract Background: Gastric cancer (GC) is the fourth leading cause of cancer-related deaths worldwide, but its genetic heterogeneity retards the development of novel therapeutics. Recently, tumor molecular classification identified clinically refractory SEM (i.e., Stem-like, EMT and Mesenchymal) subtype GC which exhibited over-expression of the plasma membrane Ca2+ ATPase (PMCA) whose role is circumventing calcium overload causing ER stress and apoptosis. Here, we showed that PMCA4 is associated with Wnt signaling activity by regulating Secreted Frizzled-Related Protein 2 (SFRP2) expression and demonstrated that PMCA4 inhibition might achieve dual mechanism-specific anti-cancer effect by inducing calcium-mediated apoptosis and restraining oncogenic Wnt activity in SEM subtype GC. Materials and Methods: GC cell lines and patient-derived organoids (PDOs) were categorized either SEM-subtype (SNU484, SNU668, SK4, MKN1; GA077T, GA265T, GA333T, GA352T) or non-SEM subtype (MKN45, NCIN87, SNU16, SNU601; GA200T, GA247T, GA215T) based on transcriptome signatures or subtype-specific biomarker genes. PMCA4 expression was confirmed using western blot and its calcium clearance function was examined using Fluo-4 AM fluorescent Ca2+ dye. Biological and molecular characteristics of GC cell lines and PDOs upon PMCA4 inhibition were investigated using RT-qPCR, western blot, and TOPflash/FOPflash reporter assay. The target engagement of novel PMCA inhibitor PM058 (bis-indole maleimide derivative) was evaluated by cellular thermal shift assay (CETSA), and the efficacy was determined by CCK8 cell viability assay for cell lines and by Cell Titer-Glo for PDOs.Results: PMCA4 was highly expressed in SEM subtype GC cell lines and PDOs. The siRNA-mediated PMCA4 knockdown (KD) or pharmaceutical inhibition of PMCA4 suppressed the calcium clearance, the cell viability, and decreased cell invasion/migration. TCF/LEF-responsive reporter assay revealed that pharmaceutical inhibition of PMCA4 suppressed Wnt/beta-catenin signaling activity, proven by decrease in beta-catenin and Wnt target gene expression. PMCA inhibition was mechanistically associated with the increased expression of SFRP2, a Wnt antagonist. Importantly, PM058, a PMCA inhibitor, treatment resulted in highly efficacious anti-cancer effect in PDOs at nano molar range, which correlated with PMCA4 protein expression.Discussion and Conclusion: Our results provide the proof of concept that PMCA4 can be a promising therapeutic target for refractory subtype cancer. The proposed mechanism of PMCA4 inhibition is calcium overload - induced apoptosis and interfering with Wnt signaling activity via SFRP2 in GC. Importantly, we demonstrated a novel potent inhibitor PM058 that blocks PMCA function in clinically intractable SEM subtype GC cancer. Citation Format: Saeli Ban, Jungmin Kim, Jonghwan Bae, Ki Cheong Park, Suji Lee, Moon Hwan Kim, Jae-Ho Cheong. PMCA inhibition induces cytotoxic calcium stress and suppression of Wnt signaling in refractory molecular subtype of gastric cancer. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4863.

Screening of Differentially Expressed Genes Based on the ACRG Molecular Subtypes of Gastric Cancer and the Significance and Mechanism of AGTR1 Gene Expression

The Asian Cancer Research Group (ACRG) classification is a molecular classification established based on the tissues of gastric cancer (GC) patients in Asia. Patients with different ACRG subtypes differ significantly with regard to treatment response and prognosis, which indicates that the ACRG molecular classification is more valuable than the traditional pathological classification. However, the specific differentially expressed genes (DEGs) and the value of the ACRG molecular subtypes of GC have not been studied in depth. Through the analysis of the GEO database, the DEGs in GC tissues of different ACRG molecular subtypes were investigated. The expression and mechanism of the screened angiotensin II receptor type 1 (AGTR1) gene were bioinformatically analyzed and experimentally verified. The role of AGTR1 in GC cells was mainly investigated using CCK-8, wound-healing, transwell invasion assays, qRT-PCR, and Western blotting. The bioinformatics results showed the presence of multiple DEGs in GC tissues with different ACRG molecular subtypes. Certain DEGs in GC tissues of different ACRG molecular subtypes have prognostic significance. AGTR1 levels in tumor tissues were significantly higher than in paired paracancerous tissues. The prognosis of GC patients with high expression of AGTR1 was poor (p < 0.05). The AGTR1 gene in GC samples was associated with the expression of immune pathways and immune checkpoint genes. After modifying AGTR1 expression in cell lines, cells' proliferation, invasion, and migration abilities and the expression of related genes changed. There were significant DEGs in GC tissues with different ACGR molecular types, among which the increased expression of AGTR1 was a molecular feature of MSS/EMT type gastric cancer. Further study found that AGTR1 was closely related to tumor immune infiltration and invasion and may be a new therapeutic target gene for gastric cancer.

Assessing Epstein–Barr virus in gastric cancer: clinicopathological features and prognostic implications

BackgroundEpstein–Barr virus (EBV)-associated gastric cancer (EBVaGC) was a unique molecular subtype of gastric cancer (GC). However, the clinicopathological characteristics and prognostic role of EBV infection remains unclear. We aimed to evaluate the clinicopathological features of EBVaGC and its role on prognosis.MethodsEBV-encoded RNA (EBER) in situ hybridization method was used to evaluate the EBV status in GC. The serum tumor markers AFP, CEA, CA19-9 and CA125 of patients were detected before treatment. HER2 expression and microsatellite instability (MSI) status was evaluated according to established criteria. The relationship between EBV infection and clinicopathological factors as well as its role on prognosis were investigated.Results420 patients were enrolled in the study and of 53 patients (12.62%) were identified as EBVaGC. EBVaGC was more common in males (p = 0.001) and related to early T stage (p = 0.045), early TNM stage (p = 0.001) and lower level of serum CEA (p = 0.039). No association could be found between EBV infection and HER2 expression, MSI status and other factors (p all > 0.05). Kaplan–Meier analysis revealed that both the overall survival and disease-free survival of EBVaGC patients were similar to that of EBV-negative GC (EBVnGC) patients (p = 0.309 and p = 0.264, respectively).ConclusionEBVaGC was more common in males and in patients with the early T stage and TNM stage as well as patients with lower serum CEA level. Difference in overall survival and disease-free survival between EBVaGC and EBVnGC patients cannot be detected.

Cancer functional states-based molecular subtypes of gastric cancer

BackgroundThe treatment of gastric cancer remains a challenge.MethodsWe divided gastric cancer into three subtypes based on 14 cancer functional states. We investigated differences between subtypes through multi-omics data, especially at the single-cell level, which allowed us to analyze differences from the perspective of each type of cell rather than the whole.ResultsThe cluster 1 is characterized by high levels of tumor progression-related cancer functional status, worst survival outcomes, low metabolic level, high infiltration of immunosuppressive cells, high copy number variations (CNV), and low tumor mutational burden (TMB). The cluster 2 is characterized by low levels of tumor progression-related cancer functional status, favorable prognosis, moderate metabolic level, low immune cell infiltration, high CNV, and moderate TMB. Then, the cluster 3 is characterized by the high level of all cancer functional status, high metabolic level, low CNV, high TMB, high infiltration of immune cells with high cytotoxicity, and better response to immunotherapy. We also established a prognostic model based on cancer functional status and validated its robustness.ConclusionsCollectively, our study identified gastric cancer subtypes and provided new insights into the clinical treatment of gastric cancer.

Sensitivity towards HDAC inhibition is associated with RTK/MAPK pathway activation in gastric cancer

Gastric cancer ranks the fifth most common and third leading cause of cancer‐related deaths worldwide. Alterations in the RTK/MAPK, WNT, cell adhesion, TP53, TGFβ, NOTCH, and NFκB signaling pathways could be identified as main oncogenic drivers. A combination of altered pathways can be associated with molecular subtypes of gastric cancer. In order to generate model systems to study the impact of different pathway alterations in a defined genetic background, we generated three murine organoid models: a RAS‐activated (KrasG12D, Tp53R172H), a WNT‐activated (Apcfl/fl, Tp53R172H), and a diffuse (Cdh1fl/fl, Apcfl/fl) model. These organoid models were morphologically and phenotypically diverse, differed in proteome expression signatures and possessed individual drug sensitivities. A differential vulnerability to RTK/MAPK pathway interference based on the different mitogenic drivers and according to the level of dependence on the pathway could be uncovered. Furthermore, an association between RTK/MAPK pathway activity and susceptibility to HDAC inhibition was observed. This finding was further validated in patient‐derived organoids from gastric adenocarcinoma, thus identifying a novel treatment approach for RTK/MAPK pathway altered gastric cancer patients.

A Novel Microbiome Signature in Gastric Cancer: A Two Independent Cohort Retrospective Analysis.

The microbiome is hypothesized to have a significant impact on cancer development. In gastric cancer (GC), Helicobacter pylori is an established class I carcinogen. However, additional organisms in the intratumoral microbiome play an important role in GC pathogenesis and progression. In this study, we characterize the full spectrum of the microbes present within GC and identify distinctions among molecular subtypes. A microbiome bioinformatics pipeline that is generalizable across multiple next-generation sequencing platforms was developed. Microbial profiles for alpha diversity and enrichment were generated for 2 large, demographically distinct cohorts: (1) internal Memorial Sloan Kettering Cancer Center (MSKCC) and (2) The Cancer Genome Atlas (TCGA) cohorts. A total of 520 GC samples were compared with select tumor-adjacent nonmalignant samples. Microbiome differences among the GC molecular subtypes were identified. Compared with nonmalignant samples, GC had significantly decreased microbial diversity in both MSKCC and TCGA cohorts ( P <0.05). Helicobacter , Lactobacillus , Streptococcus , Prevotella , and Bacteroides were significantly more enriched in GC samples when compared with nonmalignant tissue ( P <0.05). Microsatellite instability-high GC had distinct microbial enrichment compared with other GC molecular subtypes. Distinct patterns of microbial diversity and species enrichment were identified in patients with GC. Given the varied spectrum of disease progression and treatment response of GC, understanding unique microbial signatures will provide the landscape to explore key microbial targets for therapy.

DEMoS: a deep learning-based ensemble approach for predicting the molecular subtypes of gastric adenocarcinomas from histopathological images.

The molecular subtyping of gastric cancer (adenocarcinoma) into four main subtypes based on integrated multiomics profiles, as proposed by The Cancer Genome Atlas (TCGA) initiative, represents an effective strategy for patient stratification. However, this approach requires the use of multiple technological platforms, and is quite expensive and time-consuming to perform. A computational approach that uses histopathological image data to infer molecular subtypes could be a practical, cost- and time-efficient complementary tool for prognostic and clinical management purposes. Here, we propose a deep learning ensemble approach (called DEMoS) capable of predicting the four recognized molecular subtypes of gastric cancer directly from histopathological images. DEMoS achieved tile-level area under the receiver-operating characteristic curve (AUROC) values of 0.785, 0.668, 0.762 and 0.811 for the prediction of these four subtypes of gastric cancer [i.e. (i) Epstein-Barr (EBV)-infected, (ii) microsatellite instability (MSI), (iii) genomically stable (GS) and (iv) chromosomally unstable tumors (CIN)] using an independent test dataset, respectively. At the patient-level, it achieved AUROC values of 0.897, 0.764, 0.890 and 0.898, respectively. Thus, these four subtypes are well-predicted by DEMoS. Benchmarking experiments further suggest that DEMoS is able to achieve an improved classification performance for image-based subtyping and prevent model overfitting. This study highlights the feasibility of using a deep learning ensemble-based method to rapidly and reliably subtype gastric cancer (adenocarcinoma) solely using features from histopathological images. All whole slide images used in this study was collected from the TCGA database. This study builds upon our previously published HEAL framework, with related documentation and tutorials available at http://heal.erc.monash.edu.au. The source code and related models are freely accessible at https://github.com/Docurdt/DEMoS.git. Supplementary data are available at Bioinformatics online.

Characterization of Immune-Related Molecular Subtypes and a Prognostic Signature Correlating With the Response to Immunotherapy in Patients With Gastric Cancer.

Gastric cancer (GC) is a disease characterized by high molecular and phenotypic heterogeneity and represents a leading cause of cancer-related death worldwide. The tumor immune microenvironment (TIME) affects the response to immunotherapy and the prognosis of patients with GC. Explorations of the TIME in GC and characterization of molecular subtypes might enhance personalized treatment and facilitate clinical decision-making. In this study, two molecular subtypes were defined through unsupervised consensus clustering based on immune-related dysregulated genes. Then, patients with different molecular subtypes of GC were shown to have distinct differences in sensitivity to immune checkpoint blockers (ICBs). The immune-related prognostic signature was established utilizing least absolute shrinkage and selection operator (LASSO)-Cox regression analysis. Three independent external cohorts and the IMvigor210 cohort were introduced to validate the robustness of IPRS. scRNA-seq data of GC samples were used to decipher the underlying mechanisms of how IPRS contributes to the TIME. GC biospecimens were collected for RT-qPCR to further validate our findings. In summary, we characterized the abnormal TIME of GC and constructed a reliable immune-related prognostic signature correlating with the response to immunotherapy. This study may provide new strategies for developing individualized treatments for patients with GC.

The promoter aberrant methylation status of TMEM130 is associated with gastric cancer.

Gastric cancer (GC) is a malignant tumor that seriously affects human health and Epstein-Barr virus (EBV)-associated gastric cancer (EBVaGC) is a molecular subtype of GC. This study aims to determine the relationship between the methylation status of the TMEM130 gene and GC, and to explore the influence of EBV infection. qRT-PCR was conducted to investigate the transcriptional expression of TMEM130 in GC. BSP and MSP assays were used to detect the methylation level of the TMEM130 promoter. The cell migration ability was detected by Transwell and western blot after transfection of TMEM130 plasmids in GC cells. The transcriptional expression of TMEM130 decreased in GC with hypermethylation of the promoter region. The DNA methyltransferase inhibitor could increase the mRNA expression of TMEM130. Moreover, hypermethylation of the TMEM130 promoter in GC tissues was associated with EBV infection. Overexpression of TMEM130 in GC cell lines suppresses cell migration ability. This study was the first to research the expression and function of TMEM130 and found that TMEM130 gene hypermethylation might contribute to GC migration and EBV infection as a cause of hypermethylation of the TMEM130 gene. TMEM130 is a promising biomarker for the diagnosis of GC.

A 107-gene Nanostring assay effectively characterizes complex multiomic gastric cancer molecular classification in a translational patient-derived organoid model.

4049 Background: Multi-omics profiling of gastric cancer (GC) has produced numerous molecular classification systems. However, widespread clinical implementation and testing of molecular subtypes are currently limited. Here, we develop, validate and implement a custom Nanostring assay capable of allocating GC molecular subtypes to clinical specimens in a translational patient-derived organoid model. Methods: Using publicly available whole-transcriptome data, machine learning models were developed to predict GC molecular subtypes from 376 Cancer Genome Atlas (TCGA) and 1797 Tumour Microenvironment Score (TME) patients. Models were generated using feature selection with 10-fold nested cross-validation. GC biopsies from 10 local patients were preserved in paraffin (tumour) and established as an organoid culture (organoid). Gene expression was measured using Nanostring. The allocation of molecular subtypes was internally and externally validated using gold-standard reference features in public databases comprising 2202 GC patients and 10 tumour-organoid pairs, respectively. We evaluated the concordance of tumour-organoid molecular subtypes and explored the correlation between subtype scores and in-vitro chemotherapy response. Results: Classification models for TCGA (57 genes) and TME (50 genes) predicted subtypes with an accuracy ± standard deviation of 89.46% ± 0.04 and 89.33% ± 0.02, respectively. Subtype assignment of microsatellite instability (MSI) in reference to capillary electrophoresis was found to have 99.3% [95% CI 97.4-99.9, n = 277] internal and 100% [95% CI 83.2-100, n = 20] external accuracy. In reference to Epstein-Barr Virus (EBV) in-situ hybridization, EBV type internal and external accuracy was 98.7% [95% CI 97.4-99.5, n = 552] and 100% [95% CI 83.2-100, n = 20], respectively. TCGA Genomically Stable (GS) scores followed a previously characterized enrichment of diffuse-type histology compared to intestinal-type in internal and external cohorts (Dunn’s Test, p &lt; 0.0001 and p &lt; 0.05, n = 1471 and n = 15, respectively). Statistically similar subtype scores (Paired Wilcoxon, p &gt; 0.05) were found for tumour-organoid pairs. Discordance occurred in three tumour-organoid pairs. In-vitro Drug Sensitivity Score was not statistically efficacious in any molecular subtype, but Pearson correlation identified increasing efficacy with increasing EBV and MSI scores. Conclusions: Patient-derived organoids generally recapitulate the molecular subtype of parent tumours; however, in specific cases, subtype discordance occurs. Although additional external validation is required, our 107 gene assay effectively captures multi-omics classification systems in GC and allows future inquiry into the prognostic and therapeutic implications of these molecular subtypes.

183: MOLECULAR CHARACTERIZATION OF ESOPHAGEAL ADENOCARCINOMAS (EACS) AND ITS PROGNOSTIC IMPACT

Abstract Background and aim Esophageal adenocarcinomas (EACs) have an increasing incidence and a poor prognosis. While gastric cancer molecular subtypes are well characterized with 4 categories: EBV+, Microsatellite instability (MSI), Genomically stable (GS), Chromosomal instability (CIN), there is less data available on EACs. Often compared to stomach, recent literature only describes CIN subtype in EACs, while our clinical experience suggests greater heterogeneity. The aims of our study were (i) to characterize EACs from a molecular perspective based on a large clinico-biological database and (ii) compare our clinico-pathological data with molecular results to analyze their prognostic impact. Methods 516 patients underwent a transthoracic esophagectomy at the University Hospital of Lille. All HES staining slides from surgical specimen was reviewed, followed by a tissue microarrays analysis. Molecular subtypes were determined by immunohistochemistry (IHC) (MMR proteins for the deficient MMR (dMMR) and MSI-PCR for the dMMR/MSI subtype, E-cadherin for the GS subtype, HER2, C-MET and KRAS for the CIN subtype), and in situ hybridization (EBV+). Results In our series, 290 EACs are located on the esogastric junction (Siewert I and II), and 226 are above. Our current analysis highlighted one patient EBV+ and 26 dMMR status (9%). Interestingly, we observed discordances between MSI-PCR and IHC results in a third of cases (n = 8), suggesting possible modifications induced by pre-operative treatment. Our results about CIN and GS subtypes are consistent with literature with a loss of E-cadherin expression in 22 tumors (15%). An overexpression of HER2, C-MET and KRAS was noticed in 12 (8%), 41 (27%) and 13 (9%) tumors, respectively. We observed a trend toward better survival in the dMMR/MSI group comparing to GS subtype (90% vs 40% at 60 months, P = 0.07). Further analyses of the prognostic impact are in progress. Conclusion Our study highlights the presence of biomarkers in EACs, defining at least 4 molecular subtypes, partly overlapping to those already described in gastric cancer. Interestingly, finding of EBV+ and MSI tumors was unexpected and never described elsewhere in previous genomic characterization. The prognostic impact of these 4 subtypes underlined the necessity of routine molecular characterization to adapt the therapeutic strategy.

Molecular subtyping of gastric cancer according to ACRG using immunohistochemistry – Correlation with clinical parameters

BackgroundGastric cancer (GC) is a very heterogenous disease necessitating further stratification for prognostic and therapeutic aspects. Based on the recommendation of The Asisan Cancer Research Group (ACRG) recently established four molecular subtypes (MSI, MSS/EMT, MSS/TP53+, MSS/TP53-) which require molecular expression analysis. The technology required for comprehensive molecular analysis is expensive and not applicable for routine diagnostics. Thus, in this study we established a classification system utilizing immunohistochemistry and morphology-based analyses as surrogate markers in order to reproduce the ACRG molecular subtypes of gastric cancer. To clarify the clinical relevance of the novel classification system, we performed a correlation with established clinical parameters. MethodsThe study cohort consisted of 189 patients with GC (UICC III and IV). Using immunohistochemistry, the following markers were analysed: MLH1, MSH2, MSH6, PMS2 (as a surrogate for microsatellite status), p53, SOX9. We assessed tumor budding as a surrogate for EMT to distinguish between MSS/EMT and MSS/non-EMT groups. ResultsImmunohistochemical and morphologic subtyping classified cases as follows: 10% MSI, 35% MSS/EMT, 16% MSS/TP53 + and 39% MSS/TP53–. Subtypes significantly correlated with the Lauren classification, tumor stage, venous invasion and SOX9 expression (p < .05). There was no significant correlation between molecular subtype and lymph node growth pattern. ConclusionWe propose a simple algorithm for molecular subtyping of GC using universally available immunohistochemistry, which correlates with clinical parameters and is cost-effective and applicable in diagnostic routine. This classification might prospectively help to determine patient prognosis, optimize patient care and homogenize patient cohorts for clinical trials.