Clinical Significance of the Hyperintense Reperfusion Marker and Ocular Gadolinium Leakage After Thrombectomy: A Retrospective ETIS Study.

BACKGROUND AND

OBJECTIVES

Despite successful recanalization after mechanical thrombectomy for acute ischemic stroke, a substantial proportion of patients experience early neurologic deterioration or infarct progression. Early blood-brain barrier disruption, visualized on post-contrast fluid-attenuated inversion recovery (pcFLAIR) imaging as hyperintense acute reperfusion marker (HARM) or gadolinium leakage in ocular structures (GLOS), has been proposed as a marker of reperfusion injury, but its determinants and prognostic significance in thrombectomy cohorts remain uncertain. We aimed to evaluate whether HARM and GLOS are associated with poor clinical outcomes after mechanical thrombectomy. Secondary objectives included their association with poor radiologic outcomes and the identification of baseline and procedural predictors.

METHODS

We retrospectively analyzed data from the French Endovascular Treatment in Ischemic Stroke (ETIS) registry (2010-2021), including consecutive patients admitted to Versailles and Foch Hospital. All patients underwent mechanical thrombectomy at Foch Hospital and pcFLAIR imaging within 4-48 hours postgadolinium administration. Independent predictors of HARM and GLOS were identified through multivariable logistic regression, and their associations with early neurologic deterioration, poor 3-month functional outcome, infarct growth, and hemorrhagic transformation (HT) outcomes were evaluated.

RESULTS

Among 229 patients (mean age 69.7 ± 14.9 years, 52.0% female), HARM was observed in 48.9% and GLOS in 41.9%. Overall, 60.3% of patients had either HARM or GLOS, including 29.3% with both markers, 19.7% with isolated HARM, and 11.4% with isolated GLOS. HARM was independently associated with early neurologic deterioration (adjusted odds ratio [aOR] 3.45; 95% CI 1.33-8.95) and infarct growth (aOR 2.46; 95% CI 1.22-4.95), but not with HT or 3-month functional outcome. GLOS was not independently associated with clinical or radiologic outcomes. Older age, higher creatinine, and a shorter interval between the first and second FLAIR were common predictors of both HARM and GLOS, whereas single-pass recanalization was an independent predictor of HARM only.

DISCUSSION: In this large thrombectomy cohort, HARM was independently associated with early neurologic deterioration and infarct progression after thrombectomy, supporting its role as an imaging marker of early blood-brain barrier disruption. GLOS, although frequent, was not independently associated with clinical or radiologic outcomes. Prospective studies with early and serial imaging are required to clarify the prognostic relevance of HARM.

TRIAL REGISTRATION INFORMATION: NCT03776877 (ETIS registry).