Prise en charge des attaques cérébrales à l'aide du CT de perfusion

Given the high variability of stroke mechanisms, size, localisation and degree of the penumbra, clinicians treating acute stroke patients need tools in addition to clinical information to match the individual patient with different available treatment strategies. Among the two types of perfusion CTs the dynamic perfusion CT is preferable over the whole-brain technique as it can generate truly quantitative regional cerebral blood volume and cerebral blood-flow values and threshold maps that may differentiate reversible from non-reversible ischaemia. Some drawbacks of perfusion CTs, such as the impossibility of serial examinations (amount of contrast, radiation) or failure to show lacunar or posterior fossa lesions, are counterbalanced by the availability of CTs in most emergency rooms, its easy accessibility, simple monitoring of patients and the possibility to quantify perfusion deficits. Perfusion CT has a sensitivity and positive predictive value above 90% for territorial infarcts in the supratentorial regions, even in the earliest phase of stroke. Dynamic perfusion CT reliably identifies penumbra and core tissue and closely predicts final stroke volume. The final stroke size is usually close to the initial core volume if early arterial recanalisation occurs, and close or equal to the initial core plus penumbra size if early recanalisation does not occur (with or without thrombolysis). TIAs or migraine, but not focal seizures or transient global amnesia, may rarely show minor focal hypoperfusion. In regard to information about supratentorial brain perfusion, it appears at least equivalent to MRI perfusion methods. Regarding treatment decisions, the degree of penumbra on perfusion CTs as well as an arterial occlusion on the angio CT could help to select between intraarterial and intravenous thrombolysis. Further studies might show that the current time windows thrombolyses are too long for patients with little penumbra and too short for patients with a persistent penumbra. Therefore, perfusion imaging may allow to replace a rigid time window for acute interventions and the saying "time is brain" might be replaced by "penumbra is brain". Thus, even patients with unknown onset of stroke, waking up with a stroke or having an epileptic seizure at stroke onset may become candidates for treatment based on the demonstration of a significant penumbra on perfusion imaging. In case neuroprotective treatment becomes available, only patients with a significant penumbra involving the brain matter for which the substance is active (grey versus white brain matter) may be exposed to benefits, costs and side effects of these treatments. When testing new acute stroke therapies, a potential treatment effect may be better shown thanks to selecting patients with perfusion CTs who are more likely to respond to the treatment strategy. In such a study it could be the amount of salvage of the initial radiological penumbra on perfusion CTs that may be used as a surrogate marker to test the efficacy of a new intervention, rather than final infarct size.