bims-dinmec 2026-09-06 papers

Issue of 2026–09–06
three papers selected by
Lorena Ancona, Humanitas Research

Mil Med Res. 2026 ;13(1): 100066

GastroAlert: a multimodal epigenetic blood test for early detection of gastric cancer using cell-free DNA.

Yong-Jie Xu, Jian-Feng Xu, Chang-Fa Xia, Yu-Jie Wu, Dian-Ge Li, Tie-Yan Liu, Huai-Ping Cui, Jin Zhang, Liang Shang, Wan-Qing Chen.

Background: Gastric cancer remains a leading cause of cancer-related mortality worldwide, frequently diagnosed at advanced stages due to the limitations of current diagnostic approaches. While cell-free DNA (cfDNA)-based epigenetic profiling has emerged as a promising avenue for early cancer detection, comprehensive investigations into the epigenetic landscape of cfDNA in gastric cancer remain scarce. To address this gap, we aimed to develop and validate a multimodal epigenetic blood test for the non-invasive detection of gastric cancer.
Methods: We developed GastroAlert, a multimodal epigenetic blood test that integrates cfDNA methylation, nucleosome footprinting, and fragmentation features into a single diagnostic approach for gastric cancer detection. A custom-designed capture probe panel was utilized to target gastric cancer-associated differentially methylated CpG sites and functionally relevant genes. Targeted enzymatic methylation sequencing was then performed on a discovery dataset consisting of 136 participants, including 53 gastric cancer patients, 55 individuals with benign gastric conditions, and 28 healthy controls. Subsequently, the diagnostic performance of GastroAlert was rigorously validated in an independent external dataset of 149 participants, comprising 79 gastric cancer cases and 70 non-cancer controls.
Results: In cross-validation of the discovery dataset, GastroAlert achieved an area under the receiver operating characteristic curve (AUC) of 0.950 [95% confidence interval (CI) 0.874-0.995], with observed AUCs for conventional protein biomarkers ranging from 0.507 to 0.687. Notably, the locked GastroAlert model exhibited robust and reproducible performance in the independent validation dataset, yielding an AUC of 0.965 (95% CI 0.940-0.989) for distinguishing gastric cancer patients from non-cancer controls. For the clinically critical subset of early-stage gastric cancer cases, the model still maintained high diagnostic efficacy with an AUC of 0.921 (95% CI 0.862-0.980) in the discovery dataset and 0.948 (95% CI 0.910-0.985) in the validation dataset. Among all 79 gastric cancer cases in the validation dataset, the model attained an overall sensitivity of 0.898 (95% CI 0.813-0.948) at a specificity of 0.900 in the validation dataset.
Conclusions: The findings demonstrate that multimodal epigenetic analysis of cfDNA provides a robust, non-invasive strategy for early gastric cancer detection, with potential clinical utility to improve patient outcomes.

Keywords: Cell-free DNA; Early detection; Epigenetic biomarkers; Gastric cancer

DOI: https://doi.org/10.1016/j.mmr.2026.100066

Brief Bioinform. 2026 Sep 01. pii: bbag455. [Epub ahead of print]27(5):

DNAmBERT: a transformer-based model for non-invasive cancer diagnosis using DNA sequence and methylation data.

Maryam Yassi, Mark Ezegbogu, Euan J Rodger, Peter Stockwell, Aniruddha Chatterjee, Matthew Parry.

DNA methylation alterations are early and stable hallmarks of cancer and represent promising biomarkers for non-invasive detection using circulating cell-free DNA (cfDNA). However, current computational approaches often model DNA sequence and methylation features separately and struggle to capture complex read-level methylation architecture in heterogeneous, low-signal liquid biopsy data. Here, we present DNAmBERT, a Transformer-based deep learning framework designed to jointly model DNA sequence context and read-level methylation haplotype structure from cfDNA methylation sequencing data. DNAmBERT integrates k-mer-encoded DNA sequences with methylation haplotype tokens using a unified representation and masked language modelling objective, enabling context-aware learning of sequence-epigenetic dependencies through self-attention. We evaluated DNAmBERT across multiple cfDNA methylation platforms (RRBS, cfRRBS, and cfMethyl-seq) and cancer types, including colorectal cancer, lung adenocarcinoma and hepatocellular carcinoma. In binary classification tasks, the model achieved high performance across platforms (AUC up to 0.99-1.00) and outperformed conventional machine learning and existing deep learning approaches. Aggregation of read-level predictions enabled quantitative tumour probability estimation at the sample level. Beyond binary detection, DNAmBERT supported multi-cancer and stage-aware classification, including early-stage disease, with multiclass AUC values up to 0.99. The framework further demonstrated effective cross-cancer transfer learning, maintaining robust performance under limited data availability. These results indicate that integrated sequence-haplotype representation learning provides an accurate and scalable approach for cfDNA-based multi-cancer detection.

Keywords: DNA methylation; circulating cell-free DNA (cfDNA); multi-cancer classification; non-invasive cancer detection; transfer learning; transformer models

DOI: https://doi.org/10.1093/bib/bbag455

Front Biosci (Landmark Ed). 2026 Aug 17. 31(8): 50970

Aging and Reproductive Cancers: An Integrative View on Cell-Free DNA and Transposable Elements.

Monia Cecati, Salvatore Vaiasicca, Laura Cianfruglia, Stefania Fumarola, Roberto Campagna, Francesca Marchegiani, Maurizio Cardelli.

Aging is one of the strongest risk factors for cancer, and its impact is particularly evident in malignancies of the reproductive system. Ovarian, endometrial, cervical, vulvar, prostate, and penile cancers are mainly diagnosed in older adults and often show different clinical and biological features compared with the same tumors in younger patients. Aging is associated with hormonal changes, immune decline, epigenetic alterations, and accumulation of DNA damage, all of which contribute to cancer development and progression. At the same time, many older patients have frailty and multiple comorbidities, which can limit the use of screening programs and invasive diagnostic procedures. This often leads to delayed diagnosis and worse outcomes. Cell-free DNA (cfDNA) is a minimally invasive biomarker that can be obtained from blood samples and provides molecular information on both tumor and host tissues. Circulating DNA reflects tumor-specific alterations but is also influenced by aging-related changes in DNA release, fragmentation, and methylation. For this reason, aging must be considered when cfDNA-based biomarkers are applied in clinical practice. In this review, we describe how aging influences the biology of reproductive system cancers and how these processes are mirrored in cfDNA profiles. We focus on the clinical use of cfDNA for cancer detection and monitoring in older and fragile patients. Special attention is given to repetitive elements in cfDNA, which are strongly affected by aging and tumor-related epigenetic changes and can be detected with high sensitivity even when the tumor fraction is low. We propose an integrative mechanistic framework in which age-related epigenetic and genomic changes influence both tumor biology and cfDNA composition, with transposable elements acting as a central link between aging and cancer.

Keywords: aging; cancer biomarkers; cell-free DNA; early detection; elderly patients; frailty; liquid biopsy; reproductive cancers; transposable elements

DOI: https://doi.org/10.31083/FBL50970