A Position Statement on Endovascular Models and Effectiveness Metrics for Mechanical Thrombectomy Navigation, on Behalf of the Stakeholder Taskforce for Artificial Intelligence-Assisted Robotic Thrombectomy (START)

Clinical Snapshot

Key Findings

• P Value: Not applicable - consensus statement

• Effect Size: Not applicable - consensus statement

• Primary Outcome: Consensus framework establishing four essential testbed environments (in silico, in vitro, ex vivo, in vivo) with distinct validation roles

• Nnt Or Sensitivity: Two macroclasses of effectiveness metrics identified: technical navigation metrics for early stages, clinical outcome metrics for in vivo stages

• Confidence Interval: Not applicable - consensus statement

Clinical Application

Framework provides roadmap for development but requires significant technological advancement and regulatory approval before clinical implementation Highly relevant for Australia's geographic challenges in stroke care delivery; aligns with National Stroke Foundation guidelines emphasizing timely access; would require TGA approval pathway development for robotic systems Stroke patients requiring mechanical thrombectomy, particularly those in geographically remote areas with limited access to specialized neurointerventional services

Abstract

Although we are making progress in overcoming infectious diseases and cancer, one of the major medical challenges of the mid-21st century will be the increasing prevalence of stroke. Occlusions in large vessels are especially debilitating, yet effective treatment-needed within hours to achieve best outcomes-remains limited because of geographic accessibility. One solution for improving timely access to mechanical thrombectomy in geographically diverse populations is the widespread deployment of robotic surgical systems. Artificial intelligence assistance may enable the safe and effective upskilling of operators in this emerging therapeutic delivery approach. Our aim was to establish consensus frameworks for developing and validating artificial intelligence-assisted robots for thrombectomy. Objectives included standardizing effectiveness metrics and defining reference testbeds across in silico, in vitro, ex vivo, and in vivo environments. To achieve this, we convened experts in neurointervention, robotics, data science, health economics, policy, statistics, and patient advocacy. Consensus was built through an incubator day, a Delphi process, and a final position statement. We identified that the 4 essential testbed environments each had distinct validation roles. Realism requirements vary: simpler testbeds should include realistic vessel anatomy compatible with guidewire and catheter use, whereas standard testbeds should incorporate deformable vessels. More advanced testbeds should include blood flow, pulsatility, and disease features, such as atheromatous plaques. There are 2 macroclasses of effectiveness metrics: one for in silico, in vitro, and ex vivo stages focusing on technical navigation (eg, path-following error), and another for in vivo stages, focused on clinical outcomes (eg, modified treatment in cerebral infarction scores). Patient safety is central, and not a barrier, to this technology's development. One requisite patient safety task needed now is to correlate in vitro measurements to in vivo complications.

References

1. 1.Robertshaw, H., Barnes, A., Blakelock, P., Blanc, R., Crossley, R., Fahrig, R., Hassan, A. E., Jackson, B., Karstensen, L., Kaur, N., Kowarschik, M., Lynch, J., Mathis-Ullrich, F., Meglan, D., Pereira, V. M., Ourak, M., Pantano, M., Sadati, S. M. H., Taylor-Gee, A., Vercauteren, T., White, P., Granados, A., & Booth, T. C. (2026). A Position Statement on Endovascular Models and Effectiveness Metrics for Mechanical Thrombectomy Navigation, on Behalf of the Stakeholder Taskforce for Artificial Intelligence-Assisted Robotic Thrombectomy (START). Journal of the American Heart Association. https://doi.org/10.1161/JAHA.125.044931