Toward AI-Powered Neurovascular Intervention: From Imaging to XR-Robotic Convergence

Abstract

Stroke remains a leading cause of mortality and long-term disability worldwide, where rapid diagnosis and timely intervention are critical for improving outcomes. Neurovascular imaging modalities, including magnetic resonance angiography, computed tomography angiography, magnetic resonance imaging, and computed tomography, remain central for detecting stenosis, aneurysms, occlusions, and arteriovenous malformations. This review synthesizes recent advances in artificial intelligence-augmented stroke care, spanning the continuum from diagnosis to intervention. We first examine artificial intelligence applications in diagnosis, including classification, object detection, and segmentation models that automate lesion localization and subtype identification across multimodal imaging. Advanced architectures, such as convolutional neural networks, transformers, and large language models, are assessed for their potential in multimodal stroke analysis. Beyond diagnosis, we discuss emerging artificial intelligence-driven planning frameworks, extended reality simulators for training and intraoperative guidance, and robotic platforms that enable precise catheter navigation, force sensing, and telerobotic operation. While most systems remain at preclinical or feasibility stages, their integration illustrates a roadmap toward intelligent, multimodal platforms for stroke care. We also highlight key translational and ethical challenges, including regulatory and policy considerations, which must be addressed for safe adoption. Together, these developments point toward a future of precision-driven and globally accessible neurovascular intervention.