World Model for AI Autonomous Navigation in Mechanical Thrombectomy

2025-09-29

AI Summary

Endovascular autonomous navigation during mechanical thrombectomy faces dual challenges: high anatomical heterogeneity across patients and poor generalization of long-horizon decision-making. This work introduces, for the first time, the world-model-based TD-MPC² algorithm to this domain, enabling learnable, transferable dynamic environment representation and achieving multi-task end-to-end navigation across vasculatures from 10 real patients. In simulation, TD-MPC² achieves a 65% average task success rate—significantly outperforming the model-free Soft Actor-Critic (SAC) baseline (37%)—while also reducing path length substantially. The results empirically validate the efficacy of world models in capturing complex, non-stationary vascular dynamics, establishing a scalable, model-based foundation for clinically deployable autonomous navigation systems.

Technology Category

• Planning, Routing, and Scheduling: Model-Based Reasoning
• Multiagent Systems: Adversarial Agents
• Intelligent Robots: State Estimation

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

Autonomous navigation for mechanical thrombectomy (MT) remains a critical challenge due to the complexity of vascular anatomy and the need for precise, real-time decision-making. Reinforcement learning (RL)-based approaches have demonstrated potential in automating endovascular navigation, but current methods often struggle with generalization across multiple patient vasculatures and long-horizon tasks. We propose a world model for autonomous endovascular navigation using TD-MPC2, a model-based RL algorithm. We trained a single RL agent across multiple endovascular navigation tasks in ten real patient vasculatures, comparing performance against the state-of-the-art Soft Actor-Critic (SAC) method. Results indicate that TD-MPC2 significantly outperforms SAC in multi-task learning, achieving a 65% mean success rate compared to SAC's 37%, with notable improvements in path ratio. TD-MPC2 exhibited increased procedure times, suggesting a trade-off between success rate and execution speed. These findings highlight the potential of world models for improving autonomous endovascular navigation and lay the foundation for future research in generalizable AI-driven robotic interventions.

Problem

• Autonomous navigation challenges in mechanical thrombectomy procedures
• Generalization issues across patient vasculatures in reinforcement learning
• Long-horizon task performance limitations in endovascular navigation

Innovation

• World model using TD-MPC2 for autonomous navigation
• Multi-task training across ten patient vasculatures
• Model-based RL outperforms SAC in success rates