Predictors and a Nomogram for Parenchymal Hematoma After Endovascular Thrombectomy in Large Core Ischemic Stroke

We sought to establish a nomogram with internal validation to predict parenchymal hematoma (PH) following endovascular thrombectomy (EVT) in patients with large core infarcts.

We retrospectively enrolled 133 consecutive patients with large core anterior circulation infarctions receiving EVT from March 2018 to December 2024. Following univariable analysis (p < 0.1), LASSO and multivariable logistic regression were utilized to investigate predictive factors and establish a PH prediction nomogram. Model performance was assessed via ROC, calibration, decision, and clinical impact curves, with 1000-bootstrap internal validation. Subgroup analyses were stratified by age, baseline ASPECTS, and prior intravenous thrombolysis (IVT).

PH occurred in 26 patients (19.5%). Baseline NIHSS score (OR 1.109, 95% CI 1.017-1.221; p = 0.023), monocytes (OR 0.002, 95% CI 9.74 × 10^-6-0.174; p = 0.012), platelets (OR 0.989, 95% CI 0.978-0.998; p = 0.028), post-thrombectomy fasting blood glucose (FBG) (OR 1.326, 95% CI 1.121-1.616; p = 0.002), and number of passes (OR 2.312, 95% CI 1.443-3.990; p = 0.001) were significantly associated with PH and incorporated to construct the nomogram. The model showed good discrimination (AUC = 0.876; adjusted AUC = 0.844) and acceptable calibration (Hosmer-Lemeshow p = 0.928; adjusted Brier score = 0.128). Subgroup analyses indicated a significant interaction between number of passes and IVT (P for interaction = 0.020): increased passes elevated PH risk only in patients without bridging IVT (OR 2.055, 95% CI 1.318-3.204; p = 0.001).

This nomogram provides a robust tool for predicting post-thrombectomy PH in large core infarctions. The interaction between IVT and the number of passes potentially indicates minimizing unnecessary passes in direct thrombectomy. Multicenter prospective validation is warranted.