مقالات پذیرفته شده در نهمین کنگره بین المللی زیست پزشکی
AI-Driven Radiogenomics and microRNA Biomarkers for Predicting Radiation Therapy Response in Non-Small Cell Lung Cancer: A Comprehensive Review
AI-Driven Radiogenomics and microRNA Biomarkers for Predicting Radiation Therapy Response in Non-Small Cell Lung Cancer: A Comprehensive Review
Hossein Izi,1,*
1. Department of Medical Physics, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran
Introduction: Non-small cell lung cancer (NSCLC) accounts for the majority of lung cancer cases and remains a leading cause of cancer-related mortality worldwide. Radiation therapy (RT) is a cornerstone of treatment, particularly for patients with locally advanced or unresectable diseases. Despite advances in imaging, radiation delivery, and molecular profiling, therapeutic responses to RT remain highly variable due to tumour heterogeneity and complex biological interactions. In recent years, radiogenomics and microRNA (miRNA) profiling have emerged as promising tools for characterizing tumour behaviour and predicting RT outcomes. When integrated with artificial intelligence (AI), these modalities offer the potential to develop accurate, non-invasive predictive models that support personalized radiation strategies. This review critically examines the current landscape of AI-driven radiogenomics and miRNA biomarker research in NSCLC, highlighting their potential to refine RT decision-making and improve clinical outcomes.
Methods: A comprehensive literature search was conducted across PubMed, Scopus, and Web of Science databases through March 2025. Search terms included combinations of non-small cell lung cancer, radiogenomics, artificial intelligence, microRNA, and radiation therapy. Eligible studies included original research articles, clinical trials, and systematic reviews that explored the integration of radiomics features, molecular markers (e.g., EGFR, ALK, PD-L1), and miRNA profiles with AI-based predictive models. Selected studies were evaluated for methodological quality, clinical significance, and translational potential.
Results: Radiogenomics approaches have revealed significant associations between imaging-derived features, particularly from CT and PET, and molecular alterations such as EGFR mutations, ALK rearrangements, and PD-L1 expression. These correlations allow for the non-invasive characterization of tumour molecular profiles and may serve as surrogates for radio sensitivity or resistance. Concurrently, circulating miRNAs, especially miR-21 and miR-210, have been consistently implicated in mechanisms of radioresistance, including DNA damage repair, hypoxia adaptation, and regulation of apoptosis.
Artificial intelligence techniques, particularly deep learning and ensemble machine learning algorithms have shown considerable promise in modelling complex relationships within high-dimensional radiomics and molecular datasets. Integrative models that combine radiogenomics features with miRNA expression have demonstrated improved accuracy in predicting RT outcomes and stratifying patients by risk. These advances suggest a viable path toward precision-guided RT in NSCLC.
Conclusion: The integration of radiogenomics and miRNA biomarkers through AI-driven analytical platforms represents a significant advancement in the personalization of radiation therapy for NSCLC. This multi-modal strategy enhances the biological understanding of treatment response, supports the development of non-invasive predictive tools, and facilitates the implementation of individualized radiation plans. While early results are encouraging, clinical adoption will require further validation in prospective, multi-institutional trials, along with standardization of imaging protocols and molecular assays. Nonetheless, this emerging paradigm holds substantial promise in transforming radiation oncology from a generalized to a precision-based discipline.