A Hybrid Dependable Deep Feature Extraction and Ensemble-Based Machine Learning Approach for Breast Cancer Detection

IEEE Access, Год журнала: 2023, Номер 11, С. 87694 - 87708

Опубликована: Янв. 1, 2023

Breast cancer is a prevalent and life-threatening disease that requires effective detection diagnosis methods to improve patient outcomes. Deep learning (DL) machine (ML) techniques have emerged as powerful tools in breast detection, offering benefits such improved accuracy efficiency. However, existing scalability performance limitations, emphasizing the need for further research. In this paper, we propose hybrid dependable approach combines power of DL using pre-trained ResNet50V2 model ensemble-based ML methods. The integration enables learn extract hidden patterns from complex images, while algorithms contribute interpretability generalization capabilities. We conducted extensive experiments histopathology image-based publicly available Invasive Ductal Carcinoma (IDC) dataset comprising samples different sizes. results obtained our rigorous provide compelling evidence model's robustness high performance. achieved higher rate 95%, precision 94.86%, recall 94.32%, F1 score 94.57% compared state-of-the-art models. also identified Light Boosting Classifier (LGB) most suitable conjunction with architecture. research offer significant contributions through an innovative approach, comprehensive analysis, assessment. Moreover, it has potential assist medical professionals making informed decisions, improving care, enhancing outcomes patients.

Язык: Английский

---

A review of cancer data fusion methods based on deep learning

Information Fusion, Год журнала: 2024, Номер 108, С. 102361 - 102361

Опубликована: Март 20, 2024

Язык: Английский

---

Carboxylesterase Activatable Molecular Probe for Personalized Treatment Guidance by Analyte‐Induced Molecular Transformation

Angewandte Chemie International Edition, Год журнала: 2024, Номер 63(31)

Опубликована: Май 10, 2024

Abstract Accurate visualization of tumor microenvironment is great significance for personalized medicine. Here, we develop a near‐infrared (NIR) fluorescence/photoacoustic (FL/PA) dual‐mode molecular probe (denoted as NIR−CE) distinguishing tumors based on carboxylesterase (CE) level by an analyte‐induced transformation (AIMT) strategy. The recognition moiety CE activity the acetyl unit NIR−CE, generating pre‐product, NIR−CE−OH, which undergoes spontaneous hydrogen atom exchange between nitrogen atoms in indole group and phenol hydroxyl group, eventually transforming into NIR−CE−H. In cellular experiments vivo blind studies, human hepatoma cells with high were successfully distinguished both NIR FL PA imaging. Our findings provide new imaging strategy treatment guidance.

Язык: Английский

---

Refining breast cancer biomarker discovery and drug targeting through an advanced data-driven approach

BMC Bioinformatics, Год журнала: 2024, Номер 25(1)

Опубликована: Янв. 22, 2024

Abstract Breast cancer remains a major public health challenge worldwide. The identification of accurate biomarkers is critical for the early detection and effective treatment breast cancer. This study utilizes an integrative machine learning approach to analyze gene expression data superior biomarker drug target discovery. Gene datasets, obtained from GEO database, were merged post-preprocessing. From dataset, differential analysis between normal samples revealed 164 differentially expressed genes. Meanwhile, separate dataset 350 Additionally, BGWO_SA_Ens algorithm, integrating binary grey wolf optimization simulated annealing with ensemble classifier, was employed on datasets identify predictive genes including TOP2A, AKR1C3, EZH2, MMP1, EDNRB, S100B, SPP1. over 10,000 genes, identified 1404 in (F1 score: 0.981, PR-AUC: 0.998, ROC-AUC: 0.995) 1710 GSE45827 0.965, 0.986, 0.972). intersection DEGs selected 35 that consistently significant across methods. Enrichment analyses uncovered involvement these key pathways such as AMPK, Adipocytokine, PPAR signaling. Protein-protein interaction network highlighted subnetworks central nodes. Finally, drug-gene investigation connections anticancer drugs. Collectively, workflow robust signature cancer, illuminated their biological roles, interactions therapeutic associations, underscored potential computational approaches discovery precision oncology.

Язык: Английский

---

Innovative applications of artificial intelligence in zoonotic disease management

Science in One Health, Год журнала: 2023, Номер 2, С. 100045 - 100045

Опубликована: Янв. 1, 2023

Zoonotic diseases, transmitted between humans and animals, pose a substantial threat to global public health. In recent years, artificial intelligence (AI) has emerged as transformative tool in the fight against diseases. This comprehensive review discusses innovative applications of AI management zoonotic including disease prediction, early diagnosis, drug development, future prospects. AI-driven predictive models leverage extensive datasets predict outbreaks transmission patterns, thereby facilitating proactive health responses. Early diagnosis benefits from AI-powered diagnostic tools that expedite pathogen identification containment. Furthermore, technologies have accelerated discovery by identifying potential targets optimizing candidate drugs. addresses these advancements, while also examining promising control. We emphasize pivotal role revolutionizing our approach managing diseases highlight its safeguard both animals on scale.

Язык: Английский

---

Artificial intelligence and personalized medicine: transforming patient care

Elsevier eBooks, Год журнала: 2024, Номер unknown, С. 131 - 142

Опубликована: Янв. 1, 2024

Язык: Английский

---

Applications of machine learning and deep learning in medical diagnosis

Elsevier eBooks, Год журнала: 2025, Номер unknown, С. 47 - 82

Опубликована: Янв. 1, 2025

Язык: Английский

---

Multiomics Integration for Identifying Treatment Targets, Drug Development, and Diagnostic Designs in PAH

Advances in Pulmonary Hypertension, Год журнала: 2025, Номер 23(2), С. 33 - 42

Опубликована: Янв. 1, 2025

Unraveling the complexities of pulmonary arterial hypertension (PAH) is challenging due to its multifaceted nature, encompassing molecular, cellular, tissue, and organ-level alterations. The advent omics technologies, including genomics, ­epigenomics, transcriptomics, metabolomics, proteomics, has generated a vast array public nonpublic datasets from both humans model organisms, opening new avenues for understanding PAH. However, insights provided by individual into molecular mechanisms PAH are inherently limited. In response, efforts increasing develop integrative approaches designed synthesize multidimensional data cohesive dynamics this review, we discuss various strategies integrating multiomic illustrate their application in research. We explore challenges encountered profound potential leveraging comprehensive insight as well identification novel therapeutic targets biomarkers specific Furthermore, seek elucidate process rationale behind conducting studies PAH, raising critical questions about feasibility future prospects integration unraveling disease.

Язык: Английский

---

Comprehensive Bioinformatics Analysis of Glycosylation-Related Genes and Potential Therapeutic Targets in Colorectal Cancer

International Journal of Molecular Sciences, Год журнала: 2025, Номер 26(4), С. 1648 - 1648

Опубликована: Фев. 14, 2025

Colorectal cancer (CRC) is a leading cause of cancer-related deaths worldwide, characterized by high incidence and poor survival rates. Glycosylation, fundamental post-translational modification, influences protein stability, signaling, tumor progression, with aberrations implicated in immune evasion metastasis. This study investigates the role glycosylation-related genes (Glycosylation-RGs) CRC using machine learning bioinformatics. Data from The Cancer Genome Atlas (TCGA) Molecular Signatures Database (MSigDB) were analyzed to identify 67 differentially expressed Glycosylation-RGs. These used classify patients into two subgroups distinct outcomes, highlighting their prognostic value. Weighted gene coexpression network analysis (WGCNA) revealed key modules associated traits, including pathways like glycan biosynthesis PI3K-Akt signaling. A machine-learning-based model demonstrated strong predictive performance, stratifying high- low-risk groups significant differences. Additionally, correlations between risk scores cell infiltration, providing insights microenvironment. Drug sensitivity identified potential therapeutic agents, Trametinib, SCH772984, Oxaliplatin, showing differential efficacy groups. findings enhance our understanding glycosylation CRC, identifying it as critical factor disease progression promising target for future strategies.

Язык: Английский

---

Más allá del laboratorio: biomarcadores para transformar la detección y el tratamiento del cáncer con un enfoque social

LATAM Revista Latinoamericana de Ciencias Sociales y Humanidades, Год журнала: 2025, Номер 6(1)

Опубликована: Март 3, 2025

La capacidad de medir o evaluar el pronóstico una enfermedad y, en consecuencia, orientar tratamiento es posible gracias al desarrollo indicadores biológicos, denominados manera general biomarcadores. Las empresas del área la salud utilizan ampliamente este tipo biomoléculas para exposición, eficacia y seguridad los fármacos, así como mejorar diseño ensayos clínicos selección pacientes. Los biomarcadores también ayudan a dosificación determinar cuándo acelerar un fármaco, por lo que representan interés las biotecnológicas. En trabajo, revisamos resumimos progreso logrado era postgenómica, con enfoque aquellos relacionados cáncer. Además, exponemos diversas tecnologías, Cell-SELEX sistema CRISPR-Cas, hacen más rentable identificación estos Durante análisis información, observamos cómo algunos tipos cáncer tienen menor incidencia cuentan número estudios desarrollados.

---