Perfusion computed tomography in the evaluation of blood-brain barrier disruption in acute ischaemic stroke

Abstract

Breakdown of the blood-brain barrier, one of the pathological changes that occurs following ischaemic stroke, may predispose to various complications, such as haemorrhagic transformation, massive cerebral oedema, and infarct expansion, and may ultimately affect patient outcome. This research project was carried out to investigate if blood-brain barrier permeability could be estimated using first-pass perfusion computed tomography (perfusion CT) data and to examine the ability of blood-brain barrier permeability estimates to predict stroke complications and patient outcome. Ninety-three patients with acute stroke were recruited at the Royal Brisbane & Women’s Hospital. Patients underwent a perfusion CT at admission and a follow-up CT or MRI subsequently. Blood-brain barrier permeability was measured using the Gjedde-Patlak method with first-pass and delayed phase perfusion CT data (43 datasets with first-pass and delayed phase while 50 with the first-pass only). We assessed the ability of the permeability value, which was measured using the first-pass data (K1FP), to predict delayed phase permeability estimates (K1DP). On follow-up scans, final infarct volume was measured and the presence of stroke complications, such as haemorrhagic transformation and severity of brain oedema were also diagnosed. Patient outcome was assessed at hospital discharge and three months after stroke onset. This study subsequently evaluated the use of blood-brain barrier permeability to predict stroke complications, infarct expansion, and patient outcome. Analyses showed a strong linear correlation between K1FP and K1DP across stroke compartments. Using a linear model, K1DP predicted from K1FP was comparable with the measured K1DP. A similar result was obtained from analysis of an independent validation group. Predicted K1DPwas found to be a significant predictor of stroke haemorrhagic transformation. Predictive ability was comparable with the reference, measured K1DP. Predicted K1DP also correlated with the development of severe brain oedema and infarct expansion, particularly in patients treated conservatively. Tissue viability in the penumbral region could also be predicted at different permeability thresholds although there was large between-subject variance. In this study, clinical outcome assessed at hospital discharge and at three months after stroke onset, were found to be unrelated to increased blood-brain barrier permeability. However other factors, such as age, initial clinical course, initial ischaemic volume, and stroke complications were associated with patient outcome. In summary, the current study has developed a model to generate a reliable measure of blood-brain barrier permeability using first-pass perfusion CT data. The generated values were validated with an independent comparison group and the values were used to predict stroke complications and patient radiographic outcome. Haemorrhagic transformation, severe brain oedema, tissue viability, and infarct expansion could be predicted at the early phase of stroke using blood-brain barrier permeability values generated by the model developed in this study.