Biomedical subjects

Caiwen Ou

Publications and source records attributed to Caiwen Ou.

3 recordsLinked to original sources

PubMed2026

Advancing cancer detection and treatment using longitudinal routine clinical data.

Cancer management remains fragmented across its continuum, from late-stage diagnosis and salvage therapies to non-personalized surveillance. Here, we present Oncoformer, a unified multimodal transformer model trained on the China Oncology Multimodal Prediction and Surveillance Study (COMPASS) cohort (3.67 million individuals, 17.7 million clinical visits) and validated on independent external cohorts, including the UK Biobank. Oncoformer integrates longitudinal electronic health records with chest X-ray imaging to address multiple clinical tasks: pan-cancer diagnosis (area under the receiver operating characteristic curve [AUROC] = 0.956), future cancer prediction up to 1 year before diagnosis (AUROC = 0.869), tumor stage inference (mean AUROC > 0.90), patient-specific treatment-response forecasting, and recurrence-free survival stratification across ten cancer types (all p < 0.01). Staging predictions were independently validated against postoperative pathological endpoints and shown to converge on core cancer genomic pathways. By translating routine clinical data into a dynamic view of cancer evolution, Oncoformer provides a framework for risk-informed cancer prediction and treatment stratification using routine clinical data.

Humans↗

PubMed2026

A TIGIT nanotrapping-guided STING-activatable immunometabolic strategy overcomes innate immune silence and T cell exhaustion in breast cancer.

Breast cancer exhibits a profoundly immunosuppressive tumor microenvironment (TME), where innate immune silence prevents antigen sensing and persistent T cell exhaustion limits effector responses, rendering most immunotherapies ineffective. Clinical profiling of 1093 The Cancer Genome Atlas (TCGA) cases identified a glucose-fueled glutathione (GSH)-glutathione peroxidase 4 (GPX4)-dihydrolipoamide S-acetyltransferase (DLAT) axis as a dominant metabolic shield that suppresses oxidative stress, and thereby enforces both stimulator of interferon genes (STING) silence and CD8+ T cell exclusion. To dismantle this barrier, we developed an immunometabolic nanotherapy, GOx/ES-CO-LDH@TIGIT-Nanotrap (TNT). In acidic tumors, proton-driven layered double hydroxide (LDH) disassembly releases glucose oxidase (GOx) and extremely small cuprous oxide (ES-CO). GOx depletes glucose and nicotinamide adenine dinucleotide phosphate (NADPH) to induce disulfidptosis, while ES-CO releases cuprous ions (Cu+) that trigger cuproptosis via binding to lipoylated mitochondrial proteins. Their mutual biochemical amplification produces a cycloacclerated disulfidptosis-cuproptosis cascade that collapses the GSH-GPX4-DLAT axis and restores STING activation. Meanwhile, the macrophage-derived T cell immunoreceptor with Ig and ITIM domains (TIGIT) Nanotrap sequesters CD155 to prevent T cell suppression. Together, this coordinated innate reactivation and adaptive rescue converts immune-cold tumors into STING-inflamed and T cell responsive lesions.

Female↗

PubMed2025

Association of genetically proxied cancer-targeted drugs with cardiovascular diseases through Mendelian randomization analysis.

BACKGROUND: Cancer-targeted therapies are progressively pivotal in oncological care. Observational studies underscore the emergence of cancer therapy-related cardiovascular toxicity (CTR-CVT), impacting patient outcomes. We aimed to investigate the causal relationship between different types of cancer-targeted therapies and cardiovascular disease (CVD) outcomes through a two-sample Mendelian randomization (MR) study. METHODS: This genome-wide association study was conducted using a two-sample Mendelian randomization framework. Genetic instruments for drug target gene expression were extracted from the eQTLGen consortium (31684 individuals, 37 cohorts). Genome-wide association study (GWAS) summary statistics for 19 cardiovascular diseases were derived from the FinnGen database. Primary analysis was carried out using the summary-data-based MR (SMR) method, with sensitivity analysis for validation. Colocalization analysis identifies shared causal variants between exposure eQTLs and CVD-associated single-nucleotide polymorphisms (SNPs). RESULTS: Among the 39 drug target genes, 8 were identified with detectable cis-eQTLs and were subsequently validated through positive control analysis for further investigation. In the SMR and sensitivity analyses, genetically proxied VEGFA inhibition showed significantly strong association with stroke (odds ratio [OR] = 1.17, 95% confidence interval [CI] = 1.09-1.26, p = 1.33 × 10- 5). Additionally, the inhibition of FGFR1, FLT1, and MAP2K2 exhibited suggestive association with corresponding cardiovascular disease outcomes. Nevertheless, only VEGFA expression and stroke shared a causal variant (93.6%), whereas FGFR1, MAP2K2, and FLT1 did not share causal variants with corresponding cardiovascular diseases in the colocalization analysis. CONCLUSIONS: This genetic association study revealed evidence supporting the genetic association between the use of VEGFA inhibitors and increased stroke risk, highlighting the need for enhanced pharmacovigilance. These findings underscore the delicate balance between cardiovascular toxicity risk and the benefits of cancer-targeted therapy.

Humans↗