Abstract TP44: Low Cerebral Blood Flow (CBF) at Baseline Best Predicts Parenchymal Hematoma (PH) Post Revascularization Therapy

    loading  Checking for direct PDF access through Ovid


Introduction: In patients with acute ischemic stroke (AIS), parenchymal hematoma (PH) after revascularization therapy can lead to clinical deterioration and is often unpredictable. We sought to examine the association between admission CT perfusion (CTP) parameters (cerebral blood volume (CBV), cerebral blood flow (CBF), and Tmax) and PH in a case-controlled sample of patients receiving intravenous tPA ± endovascular treatment when using an industry standard CTP algorithm.Hypothesis: We hypothesize that CTP derived parameters (CBV, CBF, and Tmax) can predict PH in AIS.Methods: PH was classified according to ECASS II criteria. Two-phase (150s) and single-phase (66s) CTP acquisitions were performed within 12hrs of ictus. CTP 4D(GE Healthcare) delay-insensitive software was used to calculate CBF, CBV, and Tmax maps. Ipsilateral hemisphere gray and white matter (GM, WM) were flooded to determine volumes (mm3) for 3 lesion types: 1) patient-specific very low CBV (vlCBV) thresholds derived from the lower 10th, 5th and 2.5th percentiles of the contralateral hemisphere, 2) very low CBF threshold of ≤7ml/(min·100g), and 3) very high Tmax threshold of ≥16s. To correct for varying scan coverage, ratio of threshold output volume by ipsilateral hemisphere volume within slices that contained lesions was obtained. Receiver operating characteristic (ROC) analysis was used to compare models to determine which CTP parameter best predicted PH.Results: 34 AIS patients (18 PH, 16 no hemorrhage) were included. Very low CBF threshold of ≤7ml/(min·100g) best predicted the occurrence of PH post revascularization therapy (p < 0.001; comparison of c-statistic). CBV and Tmax parameters were not discriminative of PH. (See Figure 1 for comparison of c- statistics).Conclusion: When using a delay insensitive industry standard CTP paradigm, very low CBF≤7ml/(min·100g) has the ability in predicting PH post revascularization therapy.

    loading  Loading Related Articles