A single biomarker is not adequate to identify patients with gastric cancer (GC) who have the potential to benefit from anti-PD-1/PD-L1 therapy, presumably owing to the complexity of the tumour microenvironment. The predictive value of tumour-infiltrating immune cells (TIICs) has not been definitively established with regard to their density and spatial organisation. Here, multiplex immunohistochemistry is used to quantify in situ biomarkers at sub-cellular resolution in 80 patients with GC. To predict the response to immunotherapy, we establish a multi-dimensional TIIC signature by considering the density of CD4⁺FoxP3⁻PD-L1⁺, CD8⁺PD-1⁻LAG3⁻, and CD68⁺STING⁺ cells and the spatial organisation of CD8⁺PD-1⁺LAG3⁻ T cells. The TIIC signature enables prediction of the response of patients with GC to anti-PD-1/PD-L1 immunotherapy and patient survival. Our findings demonstrate that a multi-dimensional TIIC signature may be relevant for the selection of patients who could benefit the most from anti-PD-1/PD-L1 immunotherapy.

Gastric cancer is the fifth most common cancer globally and is associated with significant morbidity and mortality. Despite its alarming prevalence, limited comparative evidence exists on its treatment efficacy and prognosis across diverse China populations. To address this, our study used a large-scale dataset from the National Cancer Information Database, including data from 220,304 patients from 53 leading hospitals across 27 provinces in China. From 2017 to 2023, early-stage (Stages I-II) gastric cancer diagnoses increased to 35.63% of all cancer cases. Our study evaluated the neoadjuvant treatment strategies, adjuvant post-operative therapy, first- and second-line management for progressive stages, alongside current gastric cancer treatment guidelines in China. Notably, immunotherapy accounted for 16.17% and 23.28% of first- and second-line treatments for late-stage gastric cancers, and 14.56% and 5.00% for neoadjuvant and adjuvant therapies, respectively. Analysis of survival rates revealed that the 1-, 2-, 3-, 4-, and 5-year survival rates were 74.07%, 54.89%, 44.21%, 37.97%, and 33.53%, respectively. The 5-year survival rates across stages I-IV were 85.07%, 49.34%, 35.56%, and 13.15%, respectively. These findings offer critical insights into the current state of gastric cancer treatment in China and can inform future initiatives to improve therapeutic outcomes for patients with gastric cancer.

Background: Gastric cancer patients with peritoneal metastasis (GCPM) experience a rapidly deteriorating clinical trajectory characterized by therapeutic resistance and dismal survival, particularly following the development of malignant ascites. However, the intricate dynamics within the peritoneal microenvironment (PME) during the treatment process remain largely unknown.Methods: Matched samples from primary tumours (PT), peritoneal metastases (PM), and paired pre‐treatment and post‐chemo/immunotherapy (anti‐PD‐1/PD‐L1) progression malignant ascites samples, were collected from 48 patients. These samples were subjected to single‐cell RNA sequencing (n = 30), multiplex immunofluorescence (n = 30), and spatial transcriptomics (n = 3). Furthermore, post hoc analyses of a phase 1 clinical trial (n = 20, NCT03710265) and an in‐house immunotherapy cohort (n = 499) were conducted to validate the findings.Results: Tracing the evolutionary trajectory of epithelial cells unveiled the terminally differentially MUC1+ cancer cells with a high epithelial‐to‐mesenchymal transition potential, and they demonstrated spatial proximity with fibroblasts and endothelial cells, correlating with poor prognosis. A significant expansion of macrophage infiltrates, which exhibited the highest proangiogenic activity, was observed in the ascites compared with PT and PM. Besides, higher C1Q+ macrophage infiltrates correlated with significantly lower GZMA+ T‐lymphocyte infiltrates in therapeutic failure cases, potentially mediated by the LGALS9‐CD45 and SPP1‐CD44 ligand–receptor interactions. In the chemoresistant group, intimate interactions between C1Q+ macrophages and fibroblasts through the complement activation pathway were found. In the group demonstrating immunoresistance, heightened TGF‐β production activity was detected in MUC1+ cancer cells, and they were skewed to interplay with C1Q+ macrophages through the GDF15‐TGF‐βR2 axis. Ultimately, post hoc analyses indicated that co‐targeting TGF‐β and PDL1 pathways may confer superior clinical benefits than sole anti‐PD‐1/PD‐L1 therapy for patients presenting with GCPM at the time of diagnosis.Conclusions: Our findings elucidated the cellular differentiation trajectories and crucial drug resistance features within PME, facilitating the exploration of effective targets for GCPM treatment.Highlights: MUC1+ cancer cells with a high epithelial‐to‐mesenchymal transition potential and exhibiting spatial proximity to fibroblasts and endothelial cells constitute the driving force of gastric cancer peritoneal metastasis (GCPM). Higher C1Q+ macrophage infiltrates correlated with significantly lower GZMA+ T‐lymphocyte infiltrates within the peritoneal microenvironment in therapeutic failure cases. Co‐targeting TGF‐β and PDL1 pathways may confer superior clinical benefits than sole anti‐PD‐1/PD‐L1 therapy for patients presenting with GCPM at diagnosis.

Objective: This study was to determine whether peripheral blood biomarkers including neutrophil‑lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), and systemic immune inflammation index (SII) could predict early response to cetuximab; moreover, the prognostic ability of those biomarkers on progression free survival (PFS) and overall survival (OS) of metastatic colorectal cancer (mCRC) patients with wild-type (WT) RAS was also investigated. Methods: mCRC patients with WT RAS treated with cetuximab plus chemotherapy were retrospectively analyzed, and early response was evaluated according to RECIST 1.1 after three or four treatment cycles. In prior to chemotherapy, hematologic data and clinic-pathological parameters were collected. The associations between pre-treatment inflammatory biomarkers and early response, and the prognostic value of those biomarkers were analyzed. A total of 102 patients were enrolled and divided into low or high NLR, PLR, and SII groups, respectively. Results: The early response rate was significantly higher in the low NLR (p<0.001), low PLR (p=0.045), and low SII (p=0.011), respectively. In multivariate analyses, primary tumor resection (hazard ratio (HR) 0.411, p<0.001), carcino-embryonicantigen ≤5 ng/mL (HR 0.406, p<0.001), early treatment response (HR 0.322, p<0.001), and low NLR (HR 0.665, p=0.031) were independent factors of longer PFS. Primary tumor resection (HR 0.488, p=0.003) and early response (HR 0.392, p<0.001) were independent factors of longer OS. Further analysis showed that patients with early response, even in the high groups, can achieve better PFS and OS than non-responders. Conclusion: Pre-treatment inflammatory biomarkers, especially NLR were predictors of benefit from cetuximab-combined therapy in mCRC patients. They were also predictors of significantly longer PFS and OS of early responders compared to non-responders.

With the rapid advancement of artificial intelligence (AI) technologies, including deep learning, large language models, and neural networks, these methodologies are increasingly being developed and integrated into cancer research. Gastrointestinal tumors are characterized by complexity and heterogeneity, posing significant challenges for early detection, diagnostic accuracy, and the development of personalized treatment strategies. The application of AI in digestive oncology has demonstrated its transformative potential. AI not only alleviates the diagnostic burden on clinicians, but it improves tumor screening sensitivity, specificity, and accuracy. Additionally, AI aids the detection of biomarkers such as microsatellite instability and mismatch repair, supports intraoperative assessments of tumor invasion depth, predicts treatment responses, and facilitates the design of personalized treatment plans to potentially significantly enhance patient outcomes. Moreover, the integration of AI with multiomics analyses and imaging technologies has led to substantial advancements in foundational research on the tumor microenvironment. This review highlights the progress of AI in gastrointestinal oncology over the past 5 years with focus on early tumor screening, diagnosis, molecular marker identification, treatment planning, and prognosis predictions. We also explored the potential of AI to enhance medical imaging analyses to aid tumor detection and characterization as well as its role in automating and refining histopathological assessments.