Blood-brain barrier permeability by CT perfusion predicts parenchymal hematoma after recanalization with thrombectomy

Xinyi Chen,Jie Xu,Shunyuan Guo,Sheng Zhang,Huiyuan Wang,Panpan Shen,Yafei Shang,Mingming Tan,Yu Geng
DOI: https://doi.org/10.1111/jon.13172
Abstract:Background and purpose: Parenchymal hematoma is a dreaded complication of mechanical thrombectomy after acute ischemic stroke. This study evaluated whether blood-brain barrier permeability measurements based on CT perfusion could be used as predictors of parenchymal hematoma after successful recanalization and compared the predictive value of various permeability parameters in patients with acute ischemic stroke. Methods: We enrolled 53 patients with acute ischemic stroke who underwent mechanical thrombectomy and achieved successful recanalization. Each patient underwent CT, CT angiography, and CT perfusion imaging before treatment. We used relative volume transfer constant (rKtrans ) values, relative permeability-surface area product (rP·S), and relative extraction fraction (rE) to evaluate preoperative blood-brain barrier permeability in the delayed perfusion area. Results: Overall, 22 patients (37.7%) developed hemorrhagic transformation after surgery, including 10 patients (16.9%) with hemorrhagic infarction and 11 patients (20.8%) with parenchymal hematoma. The rP·S, rKtrans , and rE of the hypoperfusion area in the parenchymal hematoma group were significantly higher than those in the hemorrhagic infarction and no-hemorrhage transformation groups (p < .01). We found that rE and rP·S were superior to rKtrans in predicting parenchymal hematoma transformation after thrombectomy (P·S area under the curve [AUC] .844 vs. rKtrans AUC .753, z = 2.064, p = .039; rE AUC .907 vs. rKtrans AUC .753, z = 2.399, p = .017). Conclusions: Patients with parenchymal hematoma after mechanical thrombectomy had higher blood-brain barrier permeability in hypoperfusion areas. Among blood-brain barrier permeability measurement parameters, rP·S and rE showed better accuracy for parenchymal hematoma prediction.
What problem does this paper attempt to address?