Uveal melanoma (UM) represents an uncommon intraocular malignancy with high aggressiveness. Dysregulation of ferroptosis has been associated with UM progression. Dihydroartemisinin (DHA), a natural derivative of Artemisia annua, exhibits potent antitumor activity with a favorable safety profile, yet its role in ferroptosis regulation in UM remains unclear. Here, we showed that DHA significantly reduced the proliferation and invasiveness of UM cells-both primary and secondary-with effects intensifying over time and dose. Transcriptomic analysis indicated that DHA may exert antitumor effects by modulating the ferroptosis-related pathway, characterized by upregulating heme oxygenase-1 (HO-1) and downregulating the SLC7A11 (xCT)/GPX4 axis, leading to iron accumulation, increased ROS and lipid peroxidation, and mitochondrial dysfunction. Iron chelators and pancaspase inhibitors partially reverse these effects, whereas HO-1 inducers enhance them. Overall, our results suggest that DHA suppresses UM progression by inducing ferroptosis and mitochondrial dysfunction, while the HO-1 and xCT/GPX4 pathways may contribute to these effects. DHA may represent a potential therapeutic approach for UM, warranting further investigation.

Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy marked by substantial molecular heterogeneity and variable response to gemcitabine-based therapy. While KRAS mutations are nearly universal, the broader RTK-RAS and MAPK signaling architecture and its relationship to treatment response remain incompletely defined. We conducted an integrative clinical-genomic analysis of 184 PDAC tumors stratified by age at diagnosis and gemcitabine exposure, interrogating somatic alterations across curated RTK-RAS/MAPK gene sets. Conversational artificial intelligence agents (AI-HOPE-RTK-RAS and AI-HOPE-MAPK) enabled dynamic cohort construction and pathway-level analyses, with findings validated using standard statistical methods. In late-onset PDAC, ERBB2 and RET mutations were significantly enriched in gemcitabine-treated tumors. Early-onset cases demonstrated differential enrichment of CACNA2D family alterations in non-treated tumors and higher frequencies of FLNB and TP53 mutations in treated disease. Importantly, late-onset patients not treated with gemcitabine who lacked RTK-RAS or MAPK alterations exhibited significantly improved overall survival. These findings reveal age- and treatment-dependent pathway dependencies beyond canonical KRAS status and support a precision oncology framework in PDAC. Conversational AI facilitated rapid, multidimensional clinical-genomic integration to uncover clinically relevant signaling substructures.

Biomarkers that predict disease progression and treatment-free remission (TFR) would be of significant clinical value in chronic myeloid leukaemia (CML). We have previously shown that CIP2A levels at diagnosis can identify patients at increased risk of progression. One mechanism by which CIP2A acts is through upregulation of the anti-apoptotic gene BCL-XL. In this study, we evaluated BCL-XL mRNA expression as a diagnostic biomarker using samples from the SPIRIT2 and DESTINY clinical trials. In SPIRIT2, which compared imatinib and dasatinib as first-line therapies, high BCL-XL expression was associated with treatment failure, poor early molecular response, and lower rates of MR2 and MR3 achievement in patients treated with imatinib. In the DESTINY trial, which assessed treatment de-escalation and discontinuation, BCL-XL expression was significantly higher in patients who experienced molecular relapse compared to those achieving sustained TFR. Notably, increases in BCL-XL were detectable 6 to 8 months prior to molecular relapse, suggesting it may serve as an early biomarker of unsuccessful TFR. We now propose a clinical diagnostic toolkit combining CIP2A and BCL-XL biomarkers to stratify CML patients by the risk of disease progression and likelihood of achieving successful TFR.

Neurodegeneration and cancer are fundamentally distinct disorders: one signifies gradual neuronal loss while the latter signifies uncontrolled cell growth and survival. However, emerging evidence explores an inverse association between these conditions, suggesting that they do not arise from independent biological processes. Understanding the context-dependent behaviour of major pathways (for example, p53, PI3K/AKT/mTOR, Wnt, and immune-stress signaling) remains pivotal in elucidating the relationship between these two diseases. Pathways promoting early-life fitness, tissue repair, and tumor suppression in dividing cells can become detrimental later in life for post-mitotic neurons. Cross-species genomics studies reveal how evolution has repeatedly adapted these shared networks to balance cancer resistance with survival. Research on species exhibiting exceptional longevity and disease resistance, including naked mole rats and bowhead whales, shows that cancer resistance and longevity are not fixed traits but rather are controlled by precise regulatory mechanisms. In this review, we integrate insights from broad species genomics and multi-omic and single-cell studies to understand how evolutionarily conserved molecular crosstalks diverge at the interface of cancer and neurodegeneration.

Ochratoxin A (OTA), a prevalent food contaminant, is closely linked to the development of various cancers, including clear cell renal cell carcinoma (ccRCC). However, the potential mechanisms remain to be explored. In this study, we employed network toxicology, machine learning, and molecular docking techniques to systematically investigate the potential molecular mechanisms underlying OTA-associated ccRCC. We normalized transcriptional data from two Gene Expression Omnibus (GEO) datasets and analyzed it using differential expression analysis and weighted gene co-expression network analysis (WGCNA), identifying 3224 ccRCC-associated target genes. These were intersected with 232 predicted OTA target genes, yielding a total of 56 overlapping targets. The results of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses indicated that these targets were primarily enriched in critical biological processes, including extracellular matrix remodeling, immune microenvironment regulation, signaling pathway transduction, cellular metabolism, and protein homeostasis. Machine learning analysis identified "glmBoost + RF" (a sequential combination of feature selection and classifier) as the optimal model, from which nine key genes were extracted. SHapley Additive exPlanations (SHAP) analysis revealed five core genes (IGFBP3, ITGA5, PYGL, SLC22A8, LTB4R), with IGFBP3 and ITGA5 serving as the principal driver genes of the model. Validation of the model's diagnostic efficacy and single-cell transcriptome analysis indicated that the core genes exhibited significant differential expression patterns, cell-type-specific expression characteristics, and high independent diagnostic efficacy. Molecular docking analyses predicted stable interactions between OTA and the core target proteins. These findings suggest potential molecular links between OTA exposure and ccRCC, providing a foundation for hypothesis generation and future experimental validation.

Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive malignancy with a five-year survival rate of approximately 13%. Patients with PDAC also have an elevated incidence of venous thromboembolism (VTE), despite prophylactic anticoagulation. Thus, there is an urgent need for therapeutic strategies that target tumor progression and hypercoagulability. Protein S (PS), a physiological anticoagulant encoded by the PROS1 gene, has recently been shown to inhibit PDAC growth in preclinical models. To further examine the physiological relevance of intratumoral PS in PDAC, we performed a meta-analysis of four independent PDAC patient cohorts obtained from cBioPortal. Patients were stratified based on low versus high intratumoral PROS1 expression based on below- and above-average mean expression, overall survival, and gene expression of select pro-growth genes, implementing a fixed-effects model. High intratumoral PROS1 expression was associated with a 41.8% reduction in the risk of death compared with low PROS1 expression (pooled hazard ratio = 0.581) within 30 months of diagnosis from survival data in three cohorts. Elevated PROS1 expression correlated with marked downregulation of key genes implicated in PDAC invasion and metastasis, including MMP2 and SNAI2, in all four cohorts. Collectively, these findings suggest that PROS1 is a potential prognostic biomarker and molecular regulator in PDAC and thus support further investigation into the dual role of PS in tumor progression.

Loss of epithelial polarity is a critical driver of tumor progression; however, how core polarity regulators interface with oncogenic signaling pathways in hepatocellular carcinoma (HCC) remains incompletely defined. LLGL scribble cell polarity complex component 1 (LLGL1) is an evolutionarily conserved polarity protein with well-established tumor-suppressive roles in multiple epithelial malignancies. Nevertheless, how LLGL1 loss shapes oncogenic signaling outputs and cellular phenotypes in HCC remains unclear. In this study, we investigated the consequences of LLGL1 knockout (KO) in epithelial-like Huh-7 HCC cells. LLGL1 loss resulted in enhanced proliferative capacity and increased clonogenic potential, accompanied by altered cell-cycle distribution characterized by reduced G1-phase and increased S-phase fractions (p < 0.001). At the signaling level, LLGL1 KO cells displayed potentiated EGFR-driven RAS/MAPK pathway activation, with increased EGFR phosphorylation, enhanced downstream RAF1-MEK-ERK-RSK signaling, elevated EGFR abundance, and selective modulation of RAF1 protein levels. Functionally, LLGL1 loss markedly enhanced migratory and invasive behavior (p < 0.0001). Despite increased motility, LLGL1 KO cells exhibited remodeling of epithelial-mesenchymal transition (EMT)-associated markers without evidence of a classical EMT program. Collectively, these findings position LLGL1 loss as a central factor associated with altered MAPK signaling, EMT marker remodeling, and tumor-promoting cellular phenotypes in HCC.

Oral squamous cell carcinoma (OSCC) remains a major global health burden due to aggressive invasion, early metastasis, therapeutic resistance, and poor long-term survival. Beyond tumor-intrinsic genetic and epigenetic alterations, accumulating evidence highlights the critical role of the tumor microenvironment in shaping OSCC progression and clinical outcomes. Cancer-associated fibroblasts (CAFs) and immune cells orchestrate tumor initiation, immune evasion, and recurrence through extracellular matrix remodeling, cytokine signaling, angiogenesis, and metabolic and redox regulation. Key oncogenic pathways, including EGFR/PI3K/AKT/mTOR, TGF-β, Wnt, and Notch, integrate with non-coding RNA networks to reinforce stemness, epithelial-mesenchymal transition, and therapy resistance. Moreover, PD-1/PD-L1-mediated immune escape, CAF-driven biomechanical remodeling, and metabolic reprogramming such as aerobic glycolysis and lipid metabolism contribute to OSCC heterogeneity. This review synthesizes current insights into OSCC across genomic, epigenetic, metabolic, and microenvironmental dimensions, emphasizing CAF biology, immune landscape reprogramming, and non-coding RNA regulation. We further discuss emerging biomarkers, liquid biopsy approaches, and targeted therapeutic strategies, providing a system-level framework for biomarker-guided stratification and precision combination therapies in OSCC.

Programmed death-ligand 1 (PD-L1) expression is routinely used to guide immune checkpoint inhibitor (ICI) therapy in advanced non-small cell lung cancer (NSCLC), yet clinical benefit remains heterogeneous even among PD-L1-high tumors. Liquid biopsy based on cell-free DNA (cfDNA) enables minimally invasive, real-time monitoring of tumor evolution. We report four cases of metastatic lung adenocarcinoma treated with atezolizumab, integrating longitudinal whole-exome sequencing (WES) of cfDNA with radiological assessment. Four patients with PD-L1-positive (≥60%) metastatic NSCLC received atezolizumab monotherapy. Serial cfDNA samples (1-3 per patient) were analyzed by high-depth WES. Distinct molecular trajectories paralleled divergent clinical outcomes. One patient achieved a complete molecular response, characterized by progressive clearance of KRAS, ATM, and NF1 mutant clones, which was concordant with radiological remission. A second patient showed an initial molecular response, followed by clonal rebound of TP53, NF1, and NOTCH2 mutant populations and the emergence of PTEN and KIF1A variants, suggesting clinical progression. Two patients exhibited primary resistance despite high PD-L1 expression, with persistent or expanding clones and early subclonal diversification; in one case, new EGFR and BRAF alterations emerged under treatment pressure. Notably, switching to platinum-based chemotherapy in a non-responder induced a measurable molecular response, highlighting discordance between PD-L1 status and immunotherapy efficacy. Longitudinal cfDNA WES captured dynamic clonal remodeling under immunotherapy and anticipated radiological outcomes. These findings underscore the clinical necessity of integrating dynamic molecular monitoring by liquid biopsy to overcome the limitations of static PD-L1 assessment, refine therapeutic stratification, and identify early resistance mechanisms in advanced NSCLC.

Pulmonary large-cell neuroendocrine carcinoma (LCNEC) is a rare lung malignancy characterized by an aggressive clinical course and an unfavorable prognosis. Next-generation sequencing (NGS) has revealed that LCNECs exhibit molecular features resembling either small-cell lung carcinoma (SCLC-like LCNEC) or non-small cell lung carcinoma (NSCLC-like LCNEC). This study aimed to characterize the incidence of actionable gene variants in a retrospective cohort of LCNEC patients using a targeted NGS approach. Microscopic diagnosis was established according to the 2021 World Health Organization (WHO) classification using a standard immunohistochemical (IHC) panel. In total, 216 LCNEC tumor samples were analyzed for molecular variants in 17 genes using the RNA-based Archer FusionPlex Lung NGS assay (Integrated DNA Technologies, USA) and the MiSeq platform (Illumina, USA)-an algorithm utilized for routine NSCLC diagnosis. Overall, 46 variants were identified in 46/216 (21.3%) tumor samples, with 28/216 (13%) LCNECs harboring at least one actionable molecular variant potentially targetable by registered or investigational agents. KRAS variants (5%; including G12C at 2%) and PIK3CA variants (5%) were the most prevalent, followed by RET single-nucleotide variants (3%), uncommon EGFR variants (1%), and BRAF class II and III variants (<1%). Notably, no classical EGFR exon 18-21 mutations nor ALK, FGFR1/2/3, or ROS1 alterations (mutations or fusions) were detected, despite the technical capability of the assay to identify such variants. A novel in-frame gene fusion (TMEM79::NTRK1) was identified in a single tumor sample (0.5%). Our results confirm that LCNECs harbor potentially targetable alterations in KRAS, PIK3CA, RET, BRAF, and NTRK1, albeit at lower frequencies than those typically observed in NSCLC.

Cancer immunotherapy has recently become an essential approach for treating cancer, showing considerable promise as a substitute for surgery, radiation therapy, and conventional chemotherapy. It primarily aims to boost the host's natural defense system to combat cancer malignancies by utilizing components of immune checkpoint blockades (ICBs), mainly programmed cell death protein 1 (PD-1) and cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4), along with elements of adoptive cellular therapies (ACTs) like Chimeric Antigen Receptor (CAR) therapy, T Cell Receptor (TCR) therapy and Tumor-Infiltrating Lymphocyte (TIL) therapy. However, cancer cells tend to undermine the effectiveness of cancer immunotherapeutic strategies by employing one or more immune evasion mechanisms. This review briefly highlights how key mechanisms of cancer immune evasion confer resistance to immunotherapy and how the Clustered Regularly Interspaced Short Palindromic Repeats/Cas9 (CRISPR)/Cas9 systems, as gene-editing tools, are poised to enhance cancer immunotherapy for treating challenging cancers. We emphasize that (CRISPR/Cas9) systems can be used to explore and positively alter the genes of the immune system, boosting the effectiveness of cancer immunotherapy by editing immune checkpoints, TILs, and CAR-T cells, and disrupting genes, facilitating tumors' evasion of the immune system. Furthermore, we highlight the growing interest in emerging base editor technology to engineer natural killer (NK) cells to overcome NK-cell-based immunotherapy challenges, particularly human leukocyte antigens (HLA)-mediated limitations, and to engineer CAR-T cells for improved immunotherapy outcomes.

Targeted therapies are reshaping oncology by enabling treatment selection based on actionable molecular alterations, improving precision, and reducing unnecessary toxicity. This review provides an up-to-date overview of current targeted treatment modalities and the medicinal chemistry principles that support their discovery and optimization. We synthesize evidence on small-molecule and biologic strategies spanning receptor and non-receptor kinases and their major signaling axes (PI3K-AKT-mTOR and RAS-RAF-MEK-ERK), apoptosis regulation (BCL-2 family), DNA repair via poly(ADP-ribose) polymerase (PARP) inhibition, and epigenetic or metabolic targets including histone deacetylases (HDACs), bromodomain and extra-terminal proteins (BET), and mutant isocitrate dehydrogenases (IDH1/2). Across these areas, we summarize recurrent resistance mechanisms and the rationale for combination or sequential approaches. Biologic targeted therapy is discussed in parallel, including immune checkpoint blockade, antibody-drug conjugates, bispecific antibodies (BsAb), and cell therapies such as chimeric antigen receptor T cells, with emphasis on biomarker-guided patient stratification. Finally, we outline emerging directions beyond canonical nodes, including modulation of the p53-MDM2/MDM4 axis, ferroptosis control through AIFM2/FSP1, and innate immune pathways such as CD47-SIRPa and the stimulator of interferon genes (STING). Overall, the field is shifting from single-target inhibition toward integrated strategies that combine precise molecular targeting with an understanding of signaling network dynamics, resistance evolution, and therapeutic vulnerabilities.

Atractylodes lancea (AL) has been shown to be a promising candidate for the treatment of cholangiocarcinoma (CCA). The study explored the potential of atractylodin (AT) and β-eudesmol (BE) to chemosensitize the effects of standard chemotherapeutics in CCA. The cytotoxicities of AT and BE on CL6, HuCCT1, and HuH28 when used in combination with 5-fluorouracil (5FU), gemcitabine (GEM), and cisplatin (Cis) were assessed by MTT assay. The modulatory effects of both compounds on mRNA expression of the reuptake and efflux transporters were determined by real-time PCR. The FIC (Fractional Inhibitory Concentration) indices indicated synergistic interactions (AT-5FU in all cell lines and BE-5FU in HuH28) and antagonistic interactions (BE-Cis in all cell lines and AT-Cis or AT-GEM in HuCCT1). The synergistic interactions observed with the AT-5FU and BE-5FU combinations were well correlated with the significant upregulation of the mRNA expression of the reuptake transporter genes hENT1 (2.64-fold) and hOCT3 (5.02-fold) and the significant downregulation of the mRNA expression of the efflux transporter gene ABCC2 (0.33-fold). AT and BE, when purified or present as significant components in AL, may benefit CCA treatment when used as adjunct therapy to standard chemotherapeutic drugs, particularly 5FU. The mechanism of synergistic activity may, at least in part, involve modulation of transporter gene expression and activity.

Cancer cells resort to metabolic reprogramming to sustain proliferation. Lung cancer has one of the highest mortality rates of all types of cancer. An important factor in its high mortality rate is its tumors' ability to undergo significant metabolic reprogramming. Phytochemicals can counteract this altered metabolism and exhibit anticancer properties. Coleus hadiensis, a plant used in traditional medicine, has shown such potential. This study evaluated the in vitro cytotoxic activity of its methanolic extract and its effects on the metabolism of HTB-177 lung cancer cells. Qualitative and quantitative phytochemical analysis of this extract was performed to characterize its main constituents. Lung cancer cells were treated with different extract concentrations to evaluate their response to the extract. Cytotoxicity was determined using an MTT assay, and metabolites were analyzed through 1H-NMR spectroscopy combined with multivariate statistical analysis. Transcriptomic profiling was also conducted to assess gene expression changes in metabolic pathways. Three main phenolic compounds were identified in the extract. The HPLC profile revealed peaks corresponding to gallic acid (GA), ferulic acid (FA), and rosmarinic acid (RA). The extract exhibited cytotoxic activity with an IC50 of 192.85 µg/mL. Metabolic alterations were observed mainly in glycolysis, the Krebs cycle, and lipid metabolism-key pathways for tumor growth. Transcriptomic data revealed altered metabolism-related genes. The upregulation of ME1 correlated with the observed increase in pyruvate levels, while the downregulation of ALDH7A1 and ASRGL1 was linked to altered amino acid catabolism. Furthermore, transcriptomic data revealed the upregulation of the pro-apoptotic gene HRK. These results indicate that the methanolic extract of C. hadiensis possesses cytotoxic activity against lung cancer cells by modulating central metabolic routes and gene expression linked to cancer cell survival and proliferation.

Limited-stage small cell lung cancer (LS-SCLC) is a highly aggressive malignancy with a poor prognosis. For a long time, the standard first-line treatment has been Etoposide plus Platinum-based chemotherapy combined with concurrent thoracic radiotherapy; however, most patients still face recurrence and drug resistance. This article systematically reviews major recent advances in the field, aiming to illustrate the new paradigm shifting from conventional chemoradiotherapy toward precision-integrated therapy. Radiotherapy techniques continue to be optimized, and the value of surgery has been reaffirmed in rigorously selected early-stage patients; immunotherapy has achieved a milestone breakthrough, with the ADRIATIC trial demonstrating that consolidation therapy with Durvalumab after chemoradiotherapy significantly extends median overall survival (55.9 vs 33.4 months); meanwhile, targeted therapies based on molecular subtyping also show promise. This paper clarifies the optimized directions of multimodal treatment strategies, emphasizes the guiding significance of molecular subtyping and biomarkers in individualized therapy, with the aim of providing reference for clinical practice, offering a framework for future research, and ultimately improving the survival outcomes of LS-SCLC patients as a whole.
.

Non-small cell lung cancer (NSCLC) is characterized by high incidence and mortality, with a low five-year survival rate. Lactate metabolism plays a central role in the metabolic reprogramming of NSCLC. Beyond serving as the end-product of glycolysis, lactate accumulates in the tumor microenvironment (TME), contributing to acidification, and can also enter the tricarboxylic acid cycle to participate in energy metabolism. Moreover, the G protein-coupled receptor 81 (GPR81)/phosphoinositide 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) signaling axis induces the expression of immune checkpoint molecules, such as programmed death-ligand 1 and cytotoxic T lymphocyte-associated protein 4 (CTLA-4), thereby suppressing the functions of T lymphocytes and natural killer cells and establishing an immunosuppressive microenvironment. Lactate further promotes epithelial-mesenchymal transition and tumor metastasis, and drives NSCLC chemoresistance and relapse via histone lactylation. Clinical studies indicate that enhanced lactate metabolism is associated with NSCLC progression and chemotherapy resistance, while targeting lactate metabolism in combination with immunotherapy exerts synergistic antitumor effects. Therefore, comprehensive inhibition of lactate metabolism together with enhancement of antitumor immunity may improve the efficacy of precision therapy in NSCLC.
.

Pyrroline-5-carboxylate reductase 1 (PYCR1) is a key enzyme in the proline biosynthesis pathway and has garnered widespread attention in the field of tumor research in recent years. Studies have shown that PYCR1 is abnormally expressed in a variety of malignant tumors. It plays a significant role in tumorigenesis and development by participating in metabolic reprogramming of tumor cells, regulating key signaling pathways, influencing the tumor microenvironment, and mediating immune evasion, among other mechanisms. This article systematically reviews the biological structure and functions of PYCR1, its expression characteristics in various tumors, and the underlying molecular mechanisms. The focus is on exploring its role in promoting the occurrence and development of different tumors, as well as its involvement in mediating chemotherapy resistance. Additionally, the research progress and clinical application prospects of PYCR1 as a potential therapeutic target are analyzed, providing a theoretical basis and research directions for the development of novel anti-tumor strategies.
.

Non-small cell lung cancer (NSCLC) is associated with a high rate of postoperative recurrence, and conventional tumor-node-metastasis (TNM) staging does not fully reflect its biological heterogeneity. Hypoxia-inducible factor-1alpha (HIF-1α) plays a critical role in tumor progression and remodeling of the tumor immune microenvironment. However, the spatial distribution of HIF-1α and its prognostic significance in the context of different lymph node metastatic states remain unclear. This study aimed to investigate the densities of HIF-1α-expressing tumor cells, CD4+ T cells, and CD8+ T cells in the primary tumors of NSCLC patients, and to assess their associations with lymph node metastasis and postoperative recurrence.

The clinical utility of integrated proteogenomic biomarkers for predicting chemotherapy response in triple-negative breast cancer remains underexplored. We prospectively analyzed paired baseline and post-treatment tumor samples from 50 patients with stage II-III TNBC treated with anthracycline- and taxane-based neoadjuvant chemotherapy, integrating whole-exome sequencing, RNA sequencing, global proteomics, and phosphoproteomics.

Breast cancer derived extracellular vesicles (EVs) mediate tumor progression through surface protein-dependent intercellular communication; however, their molecular heterogeneity remains poorly characterized. In this study, we employed a proximity-dependent barcoding assay (PBA) together with patient-derived organoid (PDO) models and identified CDCP1 as a key driver of EV-mediated oncogenesis. PBA-based surface proteomics revealed CDCP1 as the most upregulated protein in breast cancer-derived EVs compared with EVs from normal tissues. Clinical validation confirmed elevated CDCP1 expression in tumor tissues and matched EVs. PDOs generated from fresh clinical specimens recapitulated CDCP1 expression levels of the parental tumors and secreted CDCP1-enriched EVs. Functional experiments showed that CDCP1-knockdown EVs suppressed PDO proliferation and sensitized tumors to chemotherapy. Mechanistically, CDCP1-positive EVs promoted macrophage polarization toward an M2 phenotype, accompanied by upregulation of IL-10 and TGF-β and CCL22. Multiplex immunofluorescence confirmed that CDCP1-high tumors exhibited increased co-localization of CD68⁺ and CD163⁺ macrophages. These results establish CDCP1 as a master regulator of EV driven breast cancer progression, linking surface proteome remodeling to chemo-resistance and immunosuppressive microenvironment reprogramming. The integration of single-EV profiling and PDO modeling establishes a translational framework for targeting CDCP1 as a promising therapeutic target and a candidate biomarker for future liquid biopsy development in aggressive breast cancer subtypes.