Integration of Artificial Intelligence in Predicting Radiotherapy Outcomes for Glioblastoma Multiforme (GBM)

Glioblastoma multiforme (GBM) is an aggressive and lethal brain tumor with limited treatment options and a poor prognosis. Radiotherapy is a cornerstone of GBM management, yet predicting treatment response remains a challenge due to tumor heterogeneity. Artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), has emerged as a promising tool for enhancing predictive accuracy and personalizing treatment strategies. This systematic review evaluates the current advancements in AI-driven models for predicting radiotherapy outcomes in GBM patients. A comprehensive search of PubMed, Scopus, and Web of Science identified 35 relevant studies employing various AI methodologies, including ML, DL, and hybrid approaches. The results indicate that convolutional neural networks (CNNs) and hybrid AI models incorporating radiomics and genetic biomarkers achieved the highest predictive performance, with accuracy rates ranging from 75% to 92% and area under the curve (AUC) values up to 0.91. Despite these advancements, challenges such as data heterogeneity, small sample sizes, and model interpretability remain significant barriers to clinical implementation. Future research should focus on large-scale multicenter collaborations, the integration of multi-omics data, and the development of explainable AI (XAI) models to enhance transparency and clinical applicability. This systematic review aims to: 1. Comprehensively evaluate the performance characteristics of various AI models in predicting key radiotherapy outcomes for GBM, including overall survival (OS), progression-free survival (PFS), patterns of failure, and treatment-related toxicities such as radiation necrosis 2. Assess the incremental value of integrating multiple data modalities (e.g., structural and functional imaging, molecular biomarkers, dosimetry) in predictive model performance 3. Critically examine methodological considerations in AI model development and validation specific to GBM radiotherapy applications 4. Identify current barriers to clinical implementation and propose pathways for translation of these technologies into routine neuro-oncology practice