Abstract Background Clinical-Diffusion mismatch (CDM; NIHSS score ≥8 & DWI lesion volume ≤25 mL) and Perfusion-Diffusion mismatch (PDM; difference >20% between initial DWI and MTT lesion volumes) have been proposed as surrogates for ischemic brains that are at risk of infarction. However, their utility to improve the selection of patients for thrombolytic treatment remains controversial. Our aim was to identify molecular biomarkers that can be used with neuroimaging to facilitate the selection of ischemic stroke patients for systemic thrombolysis. Methods We prospectively studied 595 patients with ischemic stroke within 12 h of the stroke onset. A total of 184 patients received thrombolytic treatment according to the SITS-MOST criteria. DWI and MTT volumes were measured at admission. The main outcome variable was good functional outcome at 3 months (modified Rankin scale <3). Serum levels of glutamate (Glu), IL-10, TNF-α, IL-6, NSE, and active MMP-9 also were determined at admission. Results Patients treated with t-PA who presented with PDM had higher IL-10 levels at admission (p < 0.0001). In contrast, patients with CDM had higher levels of IL-10 (p < 0.0001) as well as Glu and TNF-α (all p < 0.05) and lower levels of NSE and active MMP-9 (all p < 0.0001). IL-10 ≥ 30 pg/mL predicts good functional outcome at 3 months with a specificity of 88% and a sensitibity of 86%. IL-10 levels ≥30 pg/mL independently in both patients with PDM (OR, 18.9) and CDM (OR, 7.5), after an adjustment for covariates. Conclusions Serum levels of IL-10 facilitate the selection of ischemic stroke patients with CDM and PDM for systemic thrombolysis.