Prehospital Detection of Large Vessel Occlusion and Intracerebral Hemorrhage Using a Dual-Biomarker Point-of-Care Test

BACKGROUND: Timely identification of stroke subtype is critical for triage and treatment. Existing prehospital stroke scales have limited diagnostic accuracy, especially in differentiating large vessel occlusion (LVO) from intracerebral hemorrhage (ICH). The LVOne point-of-care test measures d -dimer and GFAP (glial fibrillary acidic protein), biomarkers associated with thrombotic and hemorrhagic stroke, enabling biologically informed triage. METHODS: We validated a new version of the LVOne assay using plasma samples from 210 suspected stroke patients presenting within 6 hours of symptom onset, enrolled in the TIME (Testing of Identification Markers for Stroke) prospective observational cohort. This article reports a retrospective (secondary) laboratory analysis of these prospectively collected samples and clinical data. Samples were collected before neuroimaging or thrombolysis. LVOne version 1 combined d -dimer positivity, GFAP negativity, and Field Assessment Stroke Triage for Emergency Destination scoring for LVO detection only; version 2 uses improved GFAP measurement and a clinical decision rule enabling simultaneous diagnosis of LVO and ICH. Diagnostic performance (sensitivity, specificity, positive predictive value, negative predictive value) was calculated for each version. Receiver operating characteristic analyses and biomarker correlations with Field Assessment Stroke Triage for Emergency Destination were also performed. RESULTS: Both LVOne versions achieved high accuracy for LVO detection (sensitivity, 0.75; specificity, 0.92; negative predictive value, 0.97). Version 2 improved ICH specificity (0.99 versus 0.90) and positive predictive value (0.83 versus 0.23) versus version 1. Biomarker levels correlated with Field Assessment Stroke Triage for Emergency Destination scores but with limited explanatory power (GFAP R 2 =0.077; d -dimer R 2 =0.047), suggesting added information beyond clinical severity scales. In a specimen-matrix comparison (n=25), qualitative GFAP/ d -dimer classification was concordant across plasma, venous whole blood, and capillary whole blood. CONCLUSIONS: The updated LVOne assay enables simultaneous biomarker-based identification of both LVO and ICH in suspected stroke patients. Its rapid, low-complexity format supports prehospital use and may improve triage, reduce time to treatment, and optimize outcomes. Prospective field validation is warranted.