Stroke: imaging and differential diagnosis.

Structural and vascular imaging helps to differentiate haemorrhagic from acute ischemic stroke (AIS) and rule out non-stroke causes, as well as identify specific subtypes of stroke such as carotid dissection and venous thrombosis. However, it is negative in most AIS patients within 3-6 hrs of onset and thus does not allow efficient patient classification for management purposes. Physiologic neuroimaging with PET, SPECT and combined diffusion- and perfusion-weighted MR gives access to tissue perfusion and cell function/homeostasis. It has near 100% sensitivity in AIS, even in small cortical or brainstem strokes. In middle-cerebral artery (MCA) stroke, physiologic imaging also allows pathophysiological differentiation into four tissue subtypes: i) already irreversibly damaged ("core"); ii) severely hypoperfused ("penumbra"), which represents the main target for therapy; iii) mildly hypoperfused ("oligaemia"), not at risk of infarction unless secondary complications arise; and iv) reperfused/hyperperfused. PET studies have evidenced the penumbra in man, shown its largely cortical topography, documented its anticipated impact on both acute-stage neurological deficit and recovery therefrom, and shown its persistence up to 16 hrs after stroke onset in some patients. However, some patients acutely exhibit extensive irreversible damage, which places them at considerable risk of malignant MCA infarction, and others early spontaneous reperfusion, which is almost invariably associated with rapid and complete recovery. Thrombolytics and/or neuroprotective agents would therefore be expected to benefit, and hence should ideally be reserved to, only those patients in whom a substantial penumbra is documented by physiologic neuroimaging, even perhaps beyond the 3 to 6 hrs rule. In addition, excluding from thrombolytic therapy those patients with substantial necrotic core should avoid many instances of symptomatic haemorrhagic transformations. Finally, patients with extensive core might benefit from early decompressive surgery, and those with early extensive reperfusion from anti-inflammatory agents. Overall, therefore, the pathophysiologic heterogeneity underlying AIS may account for both the complications from thrombolysis and the limited success of clinical trials of neuroprotective agents, despite apparent benefit in the laboratory. Pathophysiological diagnosis as afforded by neuroimaging should now be incorporated in the design of clinical trials as well as in the routine management of stroke.