Acute stroke imaging: predicting response to therapy

© 2012 Dr. Bruce C. V. CampbellAcute ischemic stroke is caused by a blocked blood vessel in the cerebral circulation. It is the most common form of stroke worldwide and a major cause of disability and death. Treatments to re-open the blocked blood vessel and reperfuse the brain are available but their effectiveness, when applied to all patients, rapidly decreases over the first few hours after stroke onset. However, there is significant pathophysiological heterogeneity within acute stroke patients which can be revealed using advanced MRI and CT techniques. The principle of “ischemic penumbra” – hypoperfused and often non-functioning brain that will, nonetheless, potentially recover if reperfused – underlies all therapies aiming to restore blood flow in acute ischemic stroke. Perfusion-diffusion mismatch using MRI is a surrogate marker of ischemic penumbra that has been refined over the last decade. This thesis examines the validity of the mismatch paradigm and confirms the use of diffusion imaging as a reliable indicator of irreversibly damaged brain. Diffusion imaging at 24 hours (a commonly used timepoint to assess reperfusion and hemorrhage) is also established as an accurate measure of final infarct volume. This allows calculation of infarct growth as a surrogate outcome whilst minimising loss to follow-up and is a strong predictor of clinical recovery. A less predictable outcome is the proportion of hypoperfused brain that will proceed to infarction in the absence of reperfusion. Collateral blood flow is shown to be a dynamic phenomenon with alterations correlating with infarct growth. The relationship between collateral flow and perfusion-diffusion mismatch is explored. The mismatch paradigm is then translated to CT perfusion which is more widely accessible in most centres but has, until recently, lacked thorough validation. Perfusion thresholds such as Tmax>6sec translate directly to CT. The best correlate of diffusion imaging for infarct core is shown to be relative cerebral blood flow (relCBF), with the exact threshold highly dependent on the software used in the analysis. This is a shift from previous work which had suggested cerebral blood volume (CBV) was the optimal parameter. Applying mismatch-based treatment decisions in clinical practice is also examined with a comparison of simple visual assessment of mismatch with fully automated volumetric software and manual volumetric calculation. The risk of bleeding after reperfusion (hemorrhagic transformation) is the chief concern when considering reperfusion therapies. This thesis examines predictors of hemorrhage and how they may be applied in clinical practice. The ultimate aim is to move beyond simple time-based windows for treatment to an individualized treatment decision based on the particular pathophysiology revealed by imaging. The amount of potentially salvageable brain tissue can be weighed against the risk of hemorrhage to make an informed treatment decision