Tumour burden score (TBS) shows promise for colorectal cancer liver metastasis (CRLM) prognostication, yet conflicting evidence has stalled its adoption. Our study evaluated the association between TBS and long-term outcomes after hepatectomy using a diagnostic meta-analysis workflow. We searched seven databases through September 30, 2025, for retrospective cohorts supplying hazard ratios for TBS-based survival. Random-effects synthesis with Hartung–Knapp adjustment was used; heterogeneity was explored via leave-one-out analyses. Data integrity was verified with subgroup analysis, sensitivity analyses and trim-and-fill publication-bias assessment; GRADE rated certainty. Thirteen cohorts (N = 5,100) were pooled. High TBS was linked to shorter overall survival (OS) (HR = 1.57; 95
Bladder cancer (BC) is a prevalent malignancy with high metastatic and recurrent potential. Filamin C (FLNC) has been reported implicated in various cancers, but its role in BC remains unclear. Bioinformatics analysis using The gene expression profiling interactive analysis (GEPIA) and The cancer genome atlas (TCGA) databases assessed FLNC expression, prognosis, and multivariate Cox regression in bladder urothelial carcinoma (BLCA). Human protein atlas (HPA) database provided immunohistochemical images. In vitro, CCK-8, wound healing, Transwell, and western blotting (WB) assays evaluated BC cell proliferation, migration, invasion, epithelial‑mesenchymal transition (EMT), and MAPK/ERK signaling. Co‑immunoprecipitation and immunofluorescence examined FLNC‑MEK1/2 interaction and co‑localization. Rescue experiments used the MEK agonist C16‑PAF or inhibitor U0126. In vivo, the effects of FLNC on BC progression and lung metastasis were verified through nude mouse xenograft tumor models and lung metastasis models, with WB and immunohistochemistry detecting EMT and MAPK/ERK pathway proteins. FLNC was upregulated in different disease stages of BLCA and was associated with poor prognosis. In vitro, FLNC significantly promoted the proliferation, migration, and invasion abilities of BC cells and facilitated the EMT process. In vivo, FLNC promoted tumor growth, lung metastasis, and EMT. Mechanistically, FLNC directly interacted with MEK1/2 and co‑localized in the cytoplasm, activating the MAPK/ERK pathway. Rescue experiments showed that a MEK agonist reversed the effects of FLNC knockdown, while a MEK inhibitor reversed the effects of FLNC overexpression. FLNC drives BC progression and metastasis via MAPK/ERK activation and EMT induction.
The BUB1 mitotic checkpoint serine/threonine kinase B (BUB1B), one of the core proteins of the spindle checkpoint, has been linked to a poor prognosis in many tumors. Our purpose was to explore the clinical value and biological function of BUB1B in lung adenocarcinoma (LUAD). TCGA Level 3 RSEM-normalized RNA-seq data, GTEx standardized expression data, TIMER (for purity-corrected immune infiltration estimation), and single-sample gene set enrichment analysis (ssGSEA) (using the GSVA R package with method = "ssgsea") were used to analyze BUB1B expression, prognosis, immune infiltration in LUAD. The role of BUB1B in H1650 and A549 cells was investigated through knocking-down BUB1B or overexpression BUB1B. BUB1B expression was significantly increased in LUAD tissues. High levels of BUB1B indicated a worse outcome. BUB1B expression was closely linked to tumor size and TMN stage, which was an independent risk factor for the prognosis of LUAD. BUB1B was associated with 20 types of TIICs as determined by ssGSEA analysis of 28 immune cell signatures. In TIMER database, BUB1B showed a positive correlation with neutrophil infiltration, but negatively correlated with immune infiltration levels of B cells and CD4 + T cells. These findings were cross-validated across TIMER, ssGSEA, and GEPIA platforms. Low dendritic cell infiltration, low B cell infiltration, and high expression of BUB1B were closely linked to lower cumulative survival in patients with LUAD. BUB1B overexpression augmented the proliferation, invasion and migration of A549 and H1650 cells, reduced the apoptosis, and promoted the EMT process. Whereas, BUB1B knockdown showed the opposite results. Additionally, BUB1B overexpression promoted macrophage recruitment by regulating CCL2 and p-P65 expression, as confirmed by in vitro Transwell chemotaxis assays using THP-1-derived macrophages. BUB1B may be a critical factor in regulating immune function and may serve as a novel clinical prognostic marker of LUAD patients.
Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer (BC) characterized by the absence of estrogen receptor (ER), progesterone receptor (PR), and HER2 expression, and is one of the most challenging cancers to treat. Non-coding RNAs have been identified as potential biomarkers for various diseases, including cancer. Identifying these biomarkers may facilitate early diagnosis and improve treatment strategies. Therefore, this study aimed to evaluate the non-coding RNAs NEAT1 and miR-506-3p expression levels in TNBC patients. Formalin-fixed, paraffin-embedded (FFPE) tumor tissues and paired adjacent non-tumor tissues were obtained from 35 TNBC patients. Total RNA was extracted, and the expression levels of NEAT1 and miR-506-3p were quantified by qPCR using GAPDH and U6 as internal controls, respectively. Relative expression was analyzed by the 2−ΔΔCt method. Statistical analyses included comparison of gene expression fold changes, assessment of associations with clinicopathological features, Spearman correlation analysis, and receiver-operating characteristic (ROC) curve analysis. NEAT1 was significantly upregulated (≈eightfold, P < 0.0001) and miR-506-3p was markedly downregulated (≈fivefold, P < 0.0001) in tumor tissues compared to adjacent non-tumor samples. Elevated NEAT1 expression was significantly associated with the younger age group, necrosis, calcification status, larger tumor size, and vascular invasion, whereas no significant associations were observed with BMI, tumor grade, or lymph-node involvement (P > 0.05). MiR-506-3p expression showed no significant clinicopathological differences in subgroup analysis. A significant inverse correlation was observed between NEAT1 and miR-506-3p levels (rs = −0.33, P < 0.05). ROC curve analysis showed promising discriminatory performance for NEAT1 in the present cohort (AUC = 0.9563, 95
Adjuvant imatinib is the standard postoperative treatment for patients with intermediate- and high-risk gastrointestinal stromal tumors (GISTs). However, uncertainty remains regarding the optimal duration of therapy and its long-term effects on recurrence-free survival (RFS) and overall survival (OS). This study aimed to evaluate the efficacy and safety of adjuvant imatinib following complete surgical resection of GISTs and to assess the influence of treatment duration on clinical outcomes. A systematic review and meta-analysis were conducted in accordance with PRISMA guidelines. PubMed, Scopus, and the Cochrane Library were searched from inception to March 2025. Studies evaluating adjuvant imatinib after complete GIST resection were included. The primary outcome was recurrence-based survival (RFS/DFS/RFI), while secondary outcomes included OS and treatment-related adverse events. Randomized and non-randomized studies were synthesized, and pooled hazard ratios (HRs) were calculated using random-effects models. Sixteen studies involving 3531 patients met the inclusion criteria, and eight studies provided data suitable for quantitative synthesis. Adjuvant imatinib significantly reduced the risk of recurrence or death compared with surgery alone (pooled HR = 0.49, 95
BACKGROUND:Non-small cell lung cancer (NSCLC) is a particularly aggressive subtype of lung cancer characterized by early metastasis and poor prognosis. Lung adenocarcinoma (LUAD) represents the most prevalent histological subtype within NSCLC. The development of brain metastases in LUAD is frequently associated with a severely unfavorable outcome. While extrachromosomal circular DNA (eccDNA) has been implicated in various tumors, its specific role in brain metastasis related to LUAD remains largely unexplored. METHODS:EccDNA associated with brain metastasis in LUAD profiles was collected through sequencing 20 cerebrospinal fluid (CSF) samples from both control subjects and LUAD patients with brain metastases, focusing on those with Epidermal Growth Factor Receptor (EGFR) mutations, rare mutations, or no mutations at all. The genomic characteristics of the eccDNAs were examined across groups utilizing a circular map. Differentially expressed eccDNAs were subjected to analysis and functional annotation by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), employing eccDNA-related genes. RESULTS:The distribution of eccDNAs within the genome is intricately associated with gene density. We compared the genomic features of eccDNAs between the subtypes of control and brain metastatic LUAD. Differentially expressed eccDNAs in CSF were identified, and subsequent GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses demonstrated their functional relevance in each group. CONCLUSIONS:Our findings offer initial insights into the features of CSF-enriched eccDNAs across LUAD subtypes with varying pathogenic mechanisms, emphasizing their potential as diagnostic and prognostic indicators for LUAD.
Programmed death-ligand 1 (PD-L1) expression assessed by the combined positive score (CPS) is required for eligibility to first-line pembrolizumab-based therapy in metastatic triple-negative breast cancer (mTNBC). However, the predictive value of CPS beyond treatment eligibility, particularly for disease control, response kinetics and response durability in real-world practice, remains uncertain. This study evaluated the association between CPS analyzed as a continuous variable and clinical outcomes in a Polish real-world mTNBC population. This multicenter retrospective study included patients with PD-L1-positive (CPS ≥ 10) mTNBC treated with first-line pembrolizumab plus chemotherapy across 13 oncology centers in Poland (2022–2025). CPS was assessed locally and primarily analyzed as a continuous variable with exploratory categorical analyses performed for descriptive and visualization purposes. Primary endpoints included objective response rate (ORR), disease control rate (DCR), and progression as best response. Secondary endpoints were time to best response (TTBR) and duration of response (DoR). Seventy-two patients were eligible for analysis. Median age was 57 years (IQR 48–67) and median CPS was 20 (IQR 15–50). After a median follow-up of 11.6 months, ORR was 48.6
Reliable prognostic markers are needed to improve risk stratification in patients with non-small cell lung cancer (NSCLC) receiving immune checkpoint inhibitors. We evaluated the prognostic value of the CHAM score, a composite marker integrating systemic inflammation, hematologic parameters, and metastatic burden, in nivolumab-treated NSCLC patients. To our knowledge, its prognostic relevance in this setting has not been previously investigated. This retrospective cohort study included 91 patients with histologically confirmed NSCLC who received nivolumab between 2020 and 2026. The CHAM score was derived from C-reactive protein, albumin, hemoglobin levels, and metastatic extent. Overall and progression-free survival were analyzed using the Kaplan–Meier method and compared with the log-rank test. Multivariable Cox proportional hazards regression evaluated the prognostic impact of the CHAM score, and receiver operating characteristic curve analysis determined the optimal cut-off. During follow-up, 59 deaths were observed. Patients with a CHAM score ≥ 4 had shorter overall survival than those with a CHAM score < 4, with median survival times of 6.7 and 18.7 months, respectively (log-rank p < 0.001). Progression-free survival was also reduced in the high-risk group, with median values of 2.6 versus 11.2 months (log-rank p < 0.001). In multivariable Cox regression analyses, a high CHAM score remained independently associated with worse overall and progression-free survival (p < 0.05 for both). Receiver operating characteristic analysis identified a CHAM score of 4 as the optimal cut-off. The CHAM score may provide practical prognostic stratification in nivolumab-treated NSCLC patients; however, external validation is required.
The widespread adoption of immune checkpoint inhibitors has significantly improved outcomes across multiple solid tumors but has also increased pressure on hospital-based oncology services, particularly oncology day hospitals. Subcutaneous (SC) formulations of immunotherapy offer a clinically validated alternative to intravenous (IV) administration and may facilitate more efficient, patient-centered models of cancer care delivery. However, the organizational, economic, and governance implications of this transition remain insufficiently explored. We conducted a mixed-methods analysis integrating four components: (1) a structured narrative review of clinical evidence regarding SC immune checkpoint inhibitors, including pharmacokinetics, efficacy, safety, and patient-reported outcomes; (2) an economic evaluation including cost-effectiveness and return-on-investment analyses; (3) benchmarking of national and international decentralized oncology care models; and (4) cross-sectional surveys of healthcare professionals (n = 84) and patients receiving immunotherapy (n = 76). Based on these findings, a hub-and-spoke care delivery framework supported by digital health tools was developed. Across tumor types, SC immunotherapy demonstrated pharmacokinetic bioequivalence to IV formulations, with comparable efficacy and safety profiles. Treatment time was reduced by approximately 85–90
Sex-related disparities in non-small cell lung cancer (NSCLC) remain incompletely characterized, particularly in low- and middle-income countries. We aimed to evaluate sex-based differences across the entire lung cancer care continuum, from smoking exposure and diagnostic pathways to treatment access and survival outcomes, distinguishing biological susceptibility from healthcare system inequities. We conducted a retrospective cohort study of 1031 consecutive patients with histologically confirmed NSCLC managed at two tertiary oncology centers in Tunisia between 2013 and 2023. Smoking exposure, body mass index, age and stage at diagnosis, diagnostic and treatment intervals, access to molecular testing, treatment patterns, and overall survival were compared between women and men. Multivariable regression and interaction analyses were performed. Women represented 20.7
Melanoma brain metastases treated with Gamma Knife radiosurgery show heterogeneous local control trajectories. Accurate prediction of time to local control loss could support risk stratification and follow-up planning using routinely available clinical and treatment variables. A retrospective dataset was constructed from a melanoma brain metastasis radiosurgery cohort. The endpoint was time to local control loss with right censoring for lesions maintaining local control at last follow-up. Clinical, demographic, and treatment features were included, focusing on age, sex, race, pre-treatment KPS, systemic therapy, number of metastases, lesion location, eloquence, tumor volume, and margin dose. A Random Survival Forest model was trained using one-hot encoding for categorical variables. Performance was assessed with five-fold stratified cross-validation at the lesion-level using Harrell C index. Feature importance was estimated with permutation importance. A total of 884 lesions were analyzed, including 198 local control loss events. The Random Survival Forest achieved high discrimination with a mean C index of 0.919 ± 0.020 across folds. The most influential predictors were age at treatment, margin dose, pre-treatment KPS, tumor volume, and isodose line. Additional contributions were observed from anatomic location and therapy category. A Random Survival Forest survival model accurately predicted time to local control loss in melanoma brain metastases using routinely collected variables, with strong discrimination and transparent feature importance. This approach enables individualized risk estimation and time-based predictions that can be integrated into clinical decision support and follow-up strategies.
In patients with endometrial cancer (EC) with metastases at diagnosis, palliative treatment is the only therapy considered and to our knowledge there is no literature on the evolution of these patients following external-beam-irradiation (EBRT) and image-guided brachytherapy (IGBT). We present the clinical outcomes of 11 FIGO 2009 stage-IVB EC patients treated with IGBT ± EBRT. From July 2009 to February 2018, 11 stage IVB patients were treated with IGBT ± EBRT in 4 European centres. The treatments and the outcomes of these patients were analysed. Statistics: Kaplan-Meier and descriptive analysis were used for analysing overall survival (OS). The median age was 63 years (46–79) and the median follow-up was 39 months (8–128). The sites of metastases were reported in 8/11 patients: the lung in 4 patients, peritoneal and liver metastasis in one, pleural metastasis in one, bone in one and inguinal lymph nodes in one. Seven patients underwent chemotherapy (5 achieving complete response (CR)). Two patients received IGBT alone and 9 EBRT + IGBT. Three developed uterine relapse, 3 lymph node relapse and all died. Four had distant metastases and only one was alive after treatment. The OS was 69.3
Stereotactic body radiotherapy (SBRT) has entered daily clinical practice in the management of oligometastatic disease. Similarly to the evidence in support of spinal metastases, the use of SBRT has been recently reported also for the treatment of non-spinal bone metastases (NSBM). This is a single-institutional experience of oligometastatic patients treated with SBRT for NSBM. Oligometastases were defined according to the recent ESTRO/EORTC consensus. Inclusion criteria were as follows: ECOG PS ≤ 2, written informed consent, up to 5 lesions to be treated at the same time, and treatment with radiotherapy schedules applying a minimum of 6 Gy per fraction. The primary endpoint of the study was local control (LC); acute and late toxicity, distant progression-free survival (DPFS), time-to-next systemic treatment (TNST), and overall survival (OS) were secondary endpoints. Toxicity was assessed according to CTCAE criteria v5.0. Survival estimates were performed using the Kaplan–Meier method, uni- and multi-variate analyses were carried out to identify any potentially significant correlation. A total of 74 bone oligometastases in 52 patients were treated in our institution between February 2020 and December 2024. All patients received SBRT with Helical Tomotherapy for a median total dose of 33.7 Gy (range, 24–35 Gy) delivered in 3–5 fractions. In 51.9
To validate the predictive value of Total Tumor Load (TTL), based on Cytokeratin 19 (CK19) mRNA copy number measured by OSNA in sentinel lymph nodes (SLNs), and to establish an optimal cut-off for identifying HR + /HER2– breast cancer (BC) patients at high risk of ≥ 4 metastatic lymph nodes (LNs), eligible for adjuvant abemaciclib therapy. A retrospective, international, multicenter study pooling data from six studies conducted between 2000 and 2016 in adult BC patients with SLN biopsies analyzed by OSNA. Variables were TTL (CK19 mRNA copies/µL) in SLNs, number of LNs removed, and number of metastatic ones. We constructed receiver operating characteristics (ROC) curves to determine optimal TTL values and multivariate logistic regression models to identify relevant prognostic factors for predicting ≥ 4 LN metastases. We included 3804 patients with HR + /HER2- tumors of any histological type, complete axillary LN dissection, and > 4 LNs removed; 608 (16.0
BACKGROUND:Effective management of hereditary cancers requires the accurate detection of both germline and somatic pathogenic variants. Conventional approaches, such as liquid biopsy and short-read sequencing, are widely used but may have limitations in resolving structural variants (SVs), providing phasing information, and capturing complex genomic alterations. Oxford Nanopore Technologies (ONT) long-read sequencing enables improved characterization of SVs and phasing, providing additional insights for germline variant analysis. AIM:In this study, we implemented a multi-modal workflow in 50 cancer patients, integrating plasma-based liquid biopsy, short-read next-generation sequencing (NGS) and ONT-based germline testing. RESULTS:Somatic mutations were identified in 72% (36/50) of patients, primarily involving TP53, PIK3CA, EGFR, and KRAS. In parallel, germline sequencing revealed pathogenic or likely pathogenic variants in 60% (30/50) of patients, including alterations in BRCA2, MSH2, and CHEK2. Importantly, 10% (5/50) of patients harbored clinically actionable germline variants that were not detected by liquid biopsy, whereas 22% (11/50) showed somatic-only alterations. Notably, 18% (9/50) of patients had no detectable pathogenic variants in either analysis, potentially reflecting biological factors such as low circulating tumor DNA (ctDNA) levels or limitations in panel coverage. CONCLUSION:These findings highlight the complementary and non-redundant roles of somatic liquid biopsy and germline analyses. Rather than indicating diagnostic equivalence, the results show that each approach captures distinct but clinically relevant genomic information. The inclusion of ONT long-read sequencing may improve the characterization of inherited variants, particularly structural variants and phasing and may support more comprehensive risk assessment and therapeutic planning.
BACKGROUND:Protein arginine methyltransferase 5 (PRMT5) is an arginine methyltransferase that can methylate both histone and non-histone proteins. It has multiple roles, most notably in the development of cancers, including lung adenocarcinoma (LUAD). However, its prognostic value and immune involvement in LUAD remain to be elucidated. METHODS:PRMT5 expression in LUAD was analyzed using TIMER and UALCAN, validated by Western blot and immunohistochemical (IHC) assays. Knockdown experiments assessed its effect on proliferation. GEPIA was used for survival analysis. GSCA examined correlations with copy number variants and methylation. TIMER and TISIDB evaluated immune infiltration and chemokine profiles. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed. RESULTS:TIMER and UALCAN revealed high PRMT5 expression in various tumor types, including LUAD, as confirmed by Western blot and IHC. PRMT5 knockdown significantly reduced cell proliferation, indicating a tumor promoting role. GEPIA analysis identified PRMT5 as a risk factor for poor overall survival in LUAD. PRMT5 expression showed positive association with copy number variants but negative with methylation from GSCA. Further analyses showed that PRMT5 expression was negatively correlated with most immunocytes and chemokines in LUAD from TISIDB. These results indicate PRMT5 contributes significantly to LUAD through immune infiltration. Finally, KEGG analysis showed negative associations of PRMT5 expression with immune and inflammation-related pathways. CONCLUSION:These findings demonstrate that PRMT5 has potential as a prognostic biomarker and as an immunomodulatory target in LUAD.
BACKGROUND:Regulated necrosis, a type of programmed cell death, is a major factor in cancer development. The function of regulated necrosis-related genes (RNRGs) in pancreatic cancer (PC) remains unclear. This study aimed to determine the heterogeneity associated with RNRGs in PC. METHODS:Single-cell RNA-seq was downloaded from the Gene Expression Omnibus database. Transcriptome profiling, somatic mutations, somatic copy-number alterations, and the clinical data of PC samples were downloaded from The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) databases. RESULTS:At single-cell resolution, we discovered that the regulated necrosis status of non-malignant cells, rather than that of malignant cells, was a significant contributor. Notably, immune cells with immunosuppressive functions displayed higher regulated necrotic activity among distinct immune cells and clusters with discernible regulated necrosis traits. PCs were divided into two prognostic groups using signatures associated with regulated necrosis. Cluster 1 showed low regulated necrosis activity and a higher survival rate, whereas cluster 2 showed significant regulated necrosis activity and a lower survival rate. The clustering of TCGA cohorts mirrored the variety of intertumoral regulated necrosis, while ICGC cohorts further confirmed the intricacy. Finally, we discovered that the candidate gene for regulated necrotic activity in PC may be A2ML1. CONCLUSIONS:Using a classifier that integrate single-cell and bulk RNA-seq, we identified the heterogeneity of RNRGs in PC. These results may deepen our understanding of RNRGs in PC, offer new perspectives for physicians to forecast prognoses, and help create more efficient and customized therapeutic approaches in the future.
Trastuzumab deruxtecan (T-DXd) has demonstrated clinical activity in HER2-positive advanced gastric and gastroesophageal junction (GEJ) adenocarcinoma following trastuzumab failure. However, variability across studies has limited a comprehensive assessment of its efficacy and safety. A systematic review and meta-analysis were conducted following PRISMA guidelines. PubMed, Scopus, and Cochrane databases were searched through September 2025. Prospective studies evaluating T-DXd in HER2-positive gastric or GEJ adenocarcinoma were included. Outcomes were pooled using a random-effects model and reported as proportions with 95
The need for accurate, cost-effective, non-invasive screening tests for colorectal cancer (CRC) persists. Repetitive genomic elements like Long Interspersed Nuclear Element-1 (LINE-1), due to their high copy number and cancer-associated alterations, present a promising but unexplored target in stool DNA for CRC detection. In this cross-sectional study, stool DNA was extracted from 38 patients with histologically confirmed CRC and 39 healthy controls. The concentration of human DNA was quantified via quantitative PCR targeting two short LINE-1 amplicons (55-bp and 60-bp). Diagnostic performance was evaluated using ROC curve analysis to determine sensitivity, specificity, and predictive values. The concentration of stool DNA, measured by LINE-1 qPCR cycle threshold (Ct) values, was significantly higher in CRC patients compared to controls (p < 0.001 for both amplicons). The 60-bp LINE-1 amplicon demonstrated superior diagnostic accuracy with an area under the ROC curve (AUC) of 0.92 (95
BACKGROUND:Esophageal cancer (EC) is one of the most lethal malignancies worldwide, characterized by insidious early symptoms, late clinical presentation, and poor overall survival (OS). DNA methylation, a key epigenetic modulator, plays a pivotal role in tumorigenesis and represents a promising biomarker for early detection and prognosis due to its high specificity, technical stability, and non-invasive accessibility. OBJECTIVE:This study aimed to systematically identify diagnostic and prognostic DNA methylation signatures in EC by integrating methylome, transcriptome, and clinical data, and to evaluate their potential clinical utility for early diagnosis and prognosis prediction. METHODS:Differentially expressed genes (DEGs) and differentially methylated sites (DMSs) were identified using the R limma package. Univariate Cox regression analysis was applied to screen DEGs and DMSs significantly associated with OS (p < 0.05). An interaction network was constructed, and hub methylation sites were prioritized based on degree centrality. Multivariable logistic regression was employed to develop a diagnostic signature. Least absolute shrinkage and selection operator (LASSO), random forest, and multivariable Cox regression were sequentially applied to build a prognostic signature. Predictive performance was evaluated using receiver operating characteristic (ROC) curves. RESULTS:A total of 7,683 DEGs (adjusted p < 0.05, |log2fold-change (FC)|> 1) and 29,122 DMSs (adjusted p < 0.001) were identified. Among these, 891 DMSs and 270 DEGs were significantly associated with OS (p < 0.05), and the top 30 hub DNA methylation sites were prioritized. A 2-CpG diagnostic model (cg03850256 and cg11394785) achieved an area under the curve (AUC) of 0.994, demonstrating excellent diagnostic accuracy. A 4-CpG prognostic signature (cg05768047, cg22724943, cg08363794, and cg10405610) showed robust predictive performance for survival outcomes (AUC: 0.821). CONCLUSION:We constructed and validated accurate methylation-based classifiers for early diagnosis and risk stratification of EC patients. These findings provide novel insights and candidate biomarkers for precise management of EC and lay the foundation for developing DNA methylation-based screening and prognostic tools.
