BackgroundWhile neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and platelet-to-lymphocyte ratio (PLR) have been associated with acute ischemic stroke (AIS) outcomes, their differential predictive value across etiological subtypes (TOAST classification) in thrombolysis-treated patients remains underexplored.MethodsIn this retrospective cohort study, we analyzed 381 AIS patients receiving intravenous thrombolysis. Hematological indices were calculated from pre-thrombolysis. Using multivariable logistic regression adjusted for age, NIHSS, and comorbidities, we assessed associations between baseline ratios and 90-day unfavorable outcomes (mRS 3–6). Receiver operating characteristic (ROC) analysis was used to determine optimal cutoffs stratified by TOAST subtypes.ResultsA total of 381 patients were included in the study. NLR showed superior predictive performance: large-artery atherosclerosis: AUC = 0.702 (aOR = 1.35, 95%CI = 1.14–1.61, p = 0.001), small-artery occlusion: AUC = 0.750 (aOR = 1.51, 95%CI = 1.08–2.10, p = 0.015), cardioembolic stroke: AUC = 0.679 (aOR = 1.82, 95%CI = 1.07–3.10, p = 0.028). LMR showed predictive value only in large-artery atherosclerosis (AUC = 0.632, p = 0.004). Optimal NLR cutoffs: 3.19 (large-artery), 3.94 (small-artery), 3.17 (cardioembolic stroke).ConclusionNLR emerged as a robust, subtype-specific predictor of post-thrombolysis outcomes, particularly in atherosclerotic stroke variants. These findings supported NLR’s clinical utility for risk stratification in thrombolysis-eligible AIS patients.