TY - JOUR
T1 - Mathematical modelling of haemorrhagic transformation in the human brain
AU - Chen, Xi
AU - Wang, Jiayu
AU - van Kranendonk, Katinka R.
AU - Józsa, Tamas.I.
AU - el-Bouri, Wahbi K.
AU - Kappelhof, Manon
AU - van der Sluijs, Matthijs
AU - Dippel, Diederik
AU - Roos, Yvo B. W. M.
AU - Marquering, Henk A.
AU - Majoie, Charles B. L. M.
AU - Payne, Stephen J.
N1 - Publisher Copyright: © 2023
PY - 2023/9/1
Y1 - 2023/9/1
N2 - Objective: Haemorrhagic transformation (HT) is one of the most common complications after ischaemic stroke. HT can be the result of stroke progression or a complication of reperfusion treatment for stroke. The aim of this study is to apply a previously proposed HT mathematical model within a computational whole brain model to determine the factors that affect the severity of HT. In addition, these simulations are directly compared with neuroimaging data. Approach: The MR CLEAN–NO IV trial assessed the effect of endovascular therapy (EVT) alone compared with intravenous alteplase treatment (IVT) followed by EVT for patients with acute ischaemic stroke due to anterior circulation large vessel occlusion. We included imaging data of 15 HT patients from the MR CLEAN–NO IV trial, 5 patients suffered from haemorrhagic infarction type 1, 5 from haemorrhagic infarction type 2 and 5 had parenchymal haematoma type 1. The comparison of simulations with patient image data is carried out by comparing the haematoma locations and haematoma volume. The parameters of the model are then optimised to improve agreement with clinical data. Finally, the model is used to investigate the factors that affect the severity of HT. Main results: Based on the computational whole brain model, we found that perfusion reduced by 5–16% after HT onset. The results are in good agreement with the clinical data. We then showed that 1% increase of blood viscosity reduces perfusion by 0.04% and increases haematoma volume by 10.35% from baseline, and 1% increase of blood pressure reduces perfusion by 0.80% and increases haematoma volume by 4.73% from baseline. These results indicate that increased blood glucose and hypertension (among other factors) both appear to lead to a higher severity of HT. Significance: This model, by enabling us to bridge the gap between the mathematical HT model and clinical imaging data, provides the first whole brain prediction model for HT severity assessment.
AB - Objective: Haemorrhagic transformation (HT) is one of the most common complications after ischaemic stroke. HT can be the result of stroke progression or a complication of reperfusion treatment for stroke. The aim of this study is to apply a previously proposed HT mathematical model within a computational whole brain model to determine the factors that affect the severity of HT. In addition, these simulations are directly compared with neuroimaging data. Approach: The MR CLEAN–NO IV trial assessed the effect of endovascular therapy (EVT) alone compared with intravenous alteplase treatment (IVT) followed by EVT for patients with acute ischaemic stroke due to anterior circulation large vessel occlusion. We included imaging data of 15 HT patients from the MR CLEAN–NO IV trial, 5 patients suffered from haemorrhagic infarction type 1, 5 from haemorrhagic infarction type 2 and 5 had parenchymal haematoma type 1. The comparison of simulations with patient image data is carried out by comparing the haematoma locations and haematoma volume. The parameters of the model are then optimised to improve agreement with clinical data. Finally, the model is used to investigate the factors that affect the severity of HT. Main results: Based on the computational whole brain model, we found that perfusion reduced by 5–16% after HT onset. The results are in good agreement with the clinical data. We then showed that 1% increase of blood viscosity reduces perfusion by 0.04% and increases haematoma volume by 10.35% from baseline, and 1% increase of blood pressure reduces perfusion by 0.80% and increases haematoma volume by 4.73% from baseline. These results indicate that increased blood glucose and hypertension (among other factors) both appear to lead to a higher severity of HT. Significance: This model, by enabling us to bridge the gap between the mathematical HT model and clinical imaging data, provides the first whole brain prediction model for HT severity assessment.
KW - Cerebral blood flow
KW - Finite element method
KW - Haemorrhagic transformation
KW - Ischaemic stroke
UR - http://www.scopus.com/inward/record.url?scp=85153868806&partnerID=8YFLogxK
U2 - https://doi.org/10.1016/j.apm.2023.04.031
DO - https://doi.org/10.1016/j.apm.2023.04.031
M3 - Article
SN - 0307-904X
VL - 121
SP - 96
EP - 110
JO - APPLIED MATHEMATICAL MODELLING
JF - APPLIED MATHEMATICAL MODELLING
ER -