CT angiography and CT perfusion in acute ischemic stroke

CT angiography and CT perfusion are used in patients with acute ischemic stroke for diagnostic purposes and to select patients for treatment. In this thesis, the reproducibility of CT angiography and CT perfusion is examined, the additional value of CT angiography and CT perfusion for stroke outcome prediction is investigated, and the importance of collaterals is studied. Chapter 2 describes the rationale, aims, and study procedures of the Dutch acute stroke study (DUST). In chapter 3, the reproducibility of the assessment of ischemic changes was investigated for CT angiography and CT perfusion. Good reproducibility is important, especially between observers with varying experience. Agreement for CT perfusion was excellent between observers with a different level of experience and was much better than for non-contrast CT and CT angiography. The agreement for CT angiography was similar to the agreement for non-contrast CT. We concluded therefore that CT perfusion could be a very reliable addition to non-contrast CT. The main results of the DUST are described in chapter 4 and chapter 5. For the prediction of clinical outcome, poor leptomeningeal collaterals and the presence of a proximal intracranial occlusion were the strongest predictors of clinical outcome on CT angiography. On CT perfusion, the extent of ischemic changes on cerebral blood volume (CBV) maps and mean transit time (MTT) maps and the size of the infarct core were the strongest predictors. Multivariable prediction models showed that the combination of clinical information and non-contrast CT already had a high prognostic value. When CT angiography and CT perfusion were added to the multivariable prediction models, there was a very small improvement of the prognostic value, but this difference was too small to be clinically relevant. For the prediction of the infarct volume, paradoxically, CT angiography and CT perfusion improved the prognostic value of the basic prediction model including clinical information and non-contrast CT. We explained this paradox by factors other than infarct volume that also contribute to the clinical outcome, including infarct location, new ischemic events, post-stroke infection, or non-stroke related events. Leptomeningeal collaterals are important in acute ischemic stroke as they can protect the brain against ischemic damage. In chapter 6, possible determinants of leptomeningeal collaterals were explored. Higher admission glucose level, presence of a proximal middle cerebral artery occlusion, and an incomplete posterior circle of Willis were related to poor leptomeningeal collaterals. We confirmed previous findings that patients with poor leptomeningeal collaterals have worse clinical outcome than patients with good leptomeningeal collaterals, and showed that this was also the case after adjusting for the determinants of leptomeningeal collaterals. The circle of Willis was investigated further in chapter 7. We showed that the combination of an incomplete anterior circle of Willis and an incomplete posterior circle of Willis was related to the occurrence of anterior circulation stroke in high-risk patients without prior cerebrovascular disease. In conclusion, this thesis improves our knowledge of the prognostic value of CT angiography and CT perfusion in acute ischemic stroke and provides more insight in the importance of collaterals