ReFRM3D: A Radiomics-enhanced Fused Residual Multiparametric 3D Network with Multi-Scale Feature Fusion for Glioma Characterization

2025-12-27

AI Summary

To address the high inter-scanner variability in multi-parametric MRI data of gliomas and the inefficiency of jointly modeling segmentation and classification, this paper proposes an end-to-end 3D joint segmentation-classification framework. The method builds upon a radiomics-enhanced fused residual 3D network, incorporating multi-scale feature fusion, hybrid upsampling, and dilated residual skip connections, alongside a radiomics classifier leveraging multiple quantitative tumor biomarkers. It tightly integrates semantic image features and quantitative radiomic representations within a 3D U-Net backbone. Evaluated on the BraTS2019–2021 datasets, the framework achieves mean Dice scores of 94.0%, 92.6%, and 93.2% for the whole tumor (WT), enhancing tumor (ET), and tumor core (TC) subregions—surpassing state-of-the-art methods. This demonstrates synergistic optimization of high-precision tumor segmentation and histopathological subtype classification.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Gliomas are among the most aggressive cancers, characterized by high mortality rates and complex diagnostic processes. Existing studies on glioma diagnosis and classification often describe issues such as high variability in imaging data, inadequate optimization of computational resources, and inefficient segmentation and classification of gliomas. To address these challenges, we propose novel techniques utilizing multi-parametric MRI data to enhance tumor segmentation and classification efficiency. Our work introduces the first-ever radiomics-enhanced fused residual multiparametric 3D network (ReFRM3D) for brain tumor characterization, which is based on a 3D U-Net architecture and features multi-scale feature fusion, hybrid upsampling, and an extended residual skip mechanism. Additionally, we propose a multi-feature tumor marker-based classifier that leverages radiomic features extracted from the segmented regions. Experimental results demonstrate significant improvements in segmentation performance across the BraTS2019, BraTS2020, and BraTS2021 datasets, achieving high Dice Similarity Coefficients (DSC) of 94.04%, 92.68%, and 93.64% for whole tumor (WT), enhancing tumor (ET), and tumor core (TC) respectively in BraTS2019; 94.09%, 92.91%, and 93.84% in BraTS2020; and 93.70%, 90.36%, and 92.13% in BraTS2021.

Problem

• Enhancing glioma segmentation and classification using multi-parametric MRI data
• Addressing high variability and inefficiency in glioma imaging analysis
• Improving computational resource optimization for brain tumor characterization

Innovation

• Radiomics-enhanced fused residual multiparametric 3D network for glioma characterization
• Multi-scale feature fusion and hybrid upsampling in 3D U-Net architecture
• Multi-feature tumor marker classifier using radiomic features from segmentation