Predicting neuroimaging eligibility for extended-window endovascular thrombectomy.

Area of Science:

• Neurology
• Stroke Medicine
• Medical Imaging

Background:

• Endovascular thrombectomy (EVT) and tissue plasminogen activator (tPA) are effective for acute ischemic stroke within the initial treatment window.
• Extended treatment windows for stroke may offer benefits, but eligibility often requires advanced imaging like CT or MR perfusion.
• Many hospitals lack advanced imaging or EVT capabilities, necessitating patient transfers for optimal care.

Purpose of the Study:

• To develop risk indices using clinical variables to identify stroke patients eligible for extended-window EVT or tPA.
• To assist in transfer decisions for patients requiring specialized stroke interventions.

Main Methods:

• Retrospective analysis of stroke patients with concurrent CT angiography (CTA) and perfusion imaging.
• Evaluation of three potential outcomes indicating benefit from transfer: LVO and target mismatch (TM) (5-23h LKN), TM with LVO (5-15h LKN), and TM (4.5-12h LKN).
• Development of multivariable models using backward stepping and assessed with C statistics.

Main Results:

• Clinical predictors including NIHSS, ASPECTS, and age accurately predicted salvageable penumbra (C statistics 71%-86%).
• Prediction accuracy improved to 85%-86% for patients with confirmed LVO or less stringent TM cutoffs.
• A 3-point risk score demonstrated high probabilities (80%-94%) for positive outcomes across different scenarios.

Conclusions:

• Clinical variables can effectively predict salvageable brain tissue in stroke patients, comparable to perfusion imaging.
• Developed risk indices offer a valuable tool for transfer decisions in extended-window stroke treatment, especially where advanced imaging is limited.
• Validation and refinement of these predictive models using larger patient registries are recommended.