RBIO-02. Integrative multi-omics profiling of radiation necrosis reveals pathogenic mechanisms and therapeutic targets

Abstract BACKGROUND Radiation necrosis (RN) is a debilitating, delayed complication of radiotherapy characterized by progressive neurological decline and limited treatments. Despite its clinical significance, the molecular mechanisms underlying RN remain poorly defined, hampering the development of targeted therapies. This study employed an integrative multi-omics approach to uncover conserved molecular signatures and identify potential therapeutic targets in RN. METHODS RN tissue samples were collected from patients with diverse primary brain lesions (gliomas, metastases, AVMs). Histologically confirmed RN and matched non-necrotic tissues underwent RNA sequencing (RNA-Seq) and mass spectrometry-based proteomics (MS). Graph theory and partial least squares (PLS) modeling were used for network characterization and multi-omics integration. Complementary in-vivo studies involved whole-brain irradiation of mice, followed by a latency period and cognitive assessment using Morris water maze. Mouse brains were then analyzed (RNA-Seq & MS). RESULTS Unsupervised clustering revealed a distinct molecular profile in RN tissue, independent of primary lesion. A total of 678 differentially expressed genes (DEGs) and 153 differentially abundant proteins (DAPs) were identified. RN tissue showed upregulation of pro-inflammatory genes, complement activation, extracellular matrix remodeling, and hypoxia-associated pathways. Integrated analysis identified 14 concordantly upregulated DEG and DAP, (including C3, APOD, IGHG2), as well as 6 DEG/DAP pairs with discordant expression patterns. Network analysis revealed non-random hub connectivity (including STAT3, FN1, SPP1, TIMP1)—genes implicated in inflammation and tissue remodeling. PLS-based multi-omics integration identified a robust RN-specific molecular signature that strongly correlated with radiographic cerebral edema (R²=0.8,P<0.001). Survival analysis demonstrated that patients with higher dysregulation of network-core DEGs/DAPs had a 5.1-fold increased mortality risk(P<0.001). In-vivo, irradiated mice exhibited significant impairments in spatial memory tasks(P<0.05), and proteomic profiling revealed marked changes in protein expression. CONCLUSION This comprehensive multi-omics analysis defines the molecular architecture of RN, identifies central regulatory nodes, and provides a foundation for biomarker-guided therapeutic development.