Abstract Number: Esoc2026a1828 Super High-Frequency Multi-Omics in Acute Ischemic Stroke

Background and aims Stroke biology changes hour-by-hour, but biomarker studies typically sample blood daily at best - missing time-ordered signals. We tested whether hourly plasma multi-omics can uncover molecular trajectories aligned with clinical evolution and identify single-timepoint biomarkers. Methods We performed hourly plasma sampling from 1-48 hours after onset in two patients with acute large-vessel occlusion stroke for proteomics, metabolomics, and peptidomics. We used dimensionality reduction to capture temporal structure and an autoencoder to derive trajectory clusters with pathway enrichment. We quantified tissue contributions by mapping proteins/peptides to the Human Protein Atlas. Validation was conducted in an independent cohort (N=502, PROMISE). Results Hourly plasma multi-omics revealed time-ordered change and non-monotonic trajectory clusters (Fig. A-B), underscoring the need for hourly sampling for biomarkers with unknown kinetics. Despite similar baseline severity and reperfusion treatment, patient #1 exhibited molecular changes indicative of rising platelet/NETosis-related thromboinflammation and catabolic stress, whereas patient #2 displayed attenuated responses (Fig. C-E). These differences coincided with greater infarct-progression (final-infarct-volume: 146 vs 17ml), hemorrhagic transformation, and worse 90-day outcome in patient #1 (mRS 5 vs 3). Cystathionine (an intermediate to glutathione redox buffering) rose steadily in patient #1 but declined early in patient #2. In PROMISE (N=502), cystathionine correlated with infarct volume (ρ=0.19, p=2.3×10-5) and predicted worse 90-day mRS (OR=1.70 per SD, p=1.5×10-12). Conclusions Hourly plasma multi-omics maps clinically aligned, time-resolved stroke biology and nominates single-timepoint biomarkers. Scaling to larger cohorts will link trajectory clusters to tissue injury and outcomes, enabling early molecular stratification. Conflict of interest Ali Rezaei: nothing to disclose Figure 1 - belongs to Conclusions