Abstract Body: Introduction
Hemorrhagic transformation is a common complication of mechanical thrombectomy (MT), often as a result of reperfusion injury. We investigated the correlation between automated CT perfusion (CTP) parameters and Hemorrhagic transformation after MT in acute ischemic stroke.
Method
We conducted a retrospective cohort analysis of patients consecutively admitted to the Providence Health System between January 2018 and February 2022 with acute ischemic stroke who underwent MT and had a CTP on admission. We also addressed each CTP parameter’s predictive power through bivariate logistic regression and ROC/AUC analysis. We compared the median and IQR ranges of volumes of automated CTP parameters between patients with and without ICH. Any degree of hemorrhagic transformation in the ECASS II classification system was considered ICH. Symptomatic intracranial hemorrhage (sICH) was defined by The Heidelberg Bleeding Classification. (4 point increase in NIHSS, or 2 in one NIHSS subcategory, major medical/surgical intervention involved)
Results
We included 513 patients with a ischemic stroke who had an MT. Median age was 75(63, 85) years old. Higher blood sugar (p=0.003) and chronic kidney disease (p=0.001) were significantly associated with sICH Table1. Cerebral blood flow( CBF)<30%, cerebral blood volume (CBV)<38%, CBV<42%, mismatch ratio, and CBV index were associated to sICH as shown in Table 2. ROC analysis for CBV index showed an AUC of 0.64 in the bivariate analysis. Other CTP parameter logistic regression and an ROC/AUC analysis are shown in Table 3.
Conclusion
Higher CBF<30%, CBV<38%, and CBV<42% volumes were associated with an increased frequency of ICH and sICH. Higher mismatch ratios correlated with a lower frequency of ICH and sICH. Lower hypoperfusion index; and, elevated Tmax volumes >10s and >8s, were associated to a higher frequency of ICH. However, these parameters were insufficiently powered to show a significant association with sICH. CBV index outperformed the other CTP parameters in the ROC analysis. Larger studies are needed to clarify these associations.