Within the first few hours of an ischemic stroke, imaging features on CT include loss of gray–white matter differentiation and loss of the ability to differentiate the basal ganglia or thalamus from the internal capsule. Mass effect is typically absent. Areas of cortex with poor collateral blood supply, such as the insular ribbon (Fig. 1) are more vulnerable to ischemia [18] and should be carefully inspected to detect early loss of gray–white matter differentiation. The basal ganglia should also be scrutinized for findings associated with hyperacute stroke; relative hypodensity in the lentiform nucleus (with loss of contrast difference to the adjacent white matter in the internal capsule) can be seen in 75 % of patients at 3 h [18] (Fig. 2). The dense vessel sign reflective of acute embolic occlusion can sometimes be seen on unenhanced CT and is helpful in the setting of normal-appearing acutely infarcted parenchyma. Evaluation of CT angiogram source images may make acute infarction more apparent because of absent or decreased enhancement of hypoperfused gray matter (Fig. 3).

Over the first week, hypoattenuation and edema associated with ischemia progressively increase. In the case of a large infarct, substantial mass effect is typical, often causing secondary damage to other areas of the brain, such as in cases of herniation (Fig. 4). Contrast-enhanced scans may demonstrate parenchymal enhancement, which is typically gyriform in cases of cortical infarction. Approximately one week after the onset of cerebral infarction, the swelling starts to subside and there are new cortical petechial hemorrhages and/or hyperemia, which result in elevation of cortical attenuation (pseudonormalization). In such cases, imaging of a stroke at this time can be misleading as the affected cortex will appear normal, known as the “CT fogging phenomenon.” Administration of contrast will demonstrate gyriform enhancement of the “fogged” infarct area and can thus be performed to prevent misinterpretation [19] (Fig. 5). Over the weeks and months that follow, infarction gliosis sets in, which appears as a region of low density with focal volume loss (atrophy), such as in the case of ex vacuo dilatation of the ventricles and associated encephalomalacia (Fig. 6). Cortical mineralization or laminar necrosis may also occur, appearing as curvilinear hyperdensities following gyral anatomy.

Perfusion CT is superior in detecting the presence and extent of possible acute strokes that are inconspicuous on unenhanced CT. It has emerged as a critical tool for the selection of patients optimal for reperfusion therapy, as it can identify the penumbra surrounding the core of an infarct, which is the region that is ischemic but has not yet undergone permanent infarction and may thus be possibly salvaged. The parameters associated with cerebral hemodynamics include cerebral blood flow (CBF), which represents instantaneous capillary flow in tissue; cerebral blood volume (CBV), which describes the blood volume of the cerebral capillaries and venules per cerebral tissue volume; the mean transit time (MTT), which measures the length of time a certain volume of blood spends in the cerebral capillary circulation; and the time to peak (TTP), which is inversely related to CBF in that reduction of blood flow results in an increase in the time needed for contrast to reach its peak in the perfused volume of brain tissue. These parameters are linked by the central volume principle which state that CBF = CBV/MTT. Specifically, areas which demonstrate matched defects in CBV or CBF and MTT represent unsalvageable infarct core, whereas those with prolonged MTT but preserved CBV are considered to be the ischemic penumbra [20]. Recently the value of perfusion imaging in determining success of stroke intervention has come into question [21], leading some investigators to state that, given the importance of time to treatment, a non-contrast head CT is all that is needed to make a therapeutic decision [22] (Fig. 2).

CT angiography (CTA) is utilized to identify possible thrombus within an intracranial vessel and guide intra-arterial thrombolysis or clot retrieval. CTA can also be applied to the carotid and vertebral arteries in the neck to establish stroke etiology, such as atherosclerosis or carotid dissection, and identify any possible endovascular treatment access limitations, such as arterial tortuosity or stenosis (Figs. 2 and 3).

The acuity or chronicity of a stroke is better elucidated on MRI than CT due to different time-sensitive signal characteristics. The typical MRI findings that may help determine the relative ages of cerebral infarcts are displayed in Table 4. MRI is a much more sensitive modality than CT in the identification of hyperacute infarction (<6 h) [23]. In the first 3 h after infarction, CT does not typically demonstrate any parenchymal abnormalities [23, 24]. T2-weighted fluid attenuation inversion recovery (FLAIR) sequences are also relatively insensitive during this hyperacute period (and are less sensitive than CT in the 3–6 h range) but the overall sensitivity of MR is 82 % in the first 24 h, whereas detection rates for acute infarction on CT are 58 % in the first 24 h [25]. Even more impressive is the 99 % sensitivity rate of diffusion-weighted imaging (DWI), associated with detection of an ischemic stroke within 12 h of its onset (including the first 1 h), in contradistinction to the 39 % sensitivity rate associated with CT [26]. In addition, causes of stroke such as venous thrombosis, vascular malformations, infections, and tumors are better identified and characterized with greater accuracy on MRI than is possible with CT.

Described as the “light bulb sign,” hyperintensity associated with acute infarction usually occurs within 3 h and is easily detected on DWI [7] (Figs. 7 and 8). This marked hyperintensity is initially due to diffusion restriction in the area of cerebral ischemia. Cytotoxic edema occurs as a result of depletion of adenosine triphosphate, which in turn leads to breakdown of normal cell membrane integrity. As a consequence, there is intracellular swelling and damage to intracellular proteins causing increased intracellular viscosity. These processes are indicators of cell death. After 4–6 h T2 increases within the area of infarction due to vasogenic edema. Restricted diffusion can be assessed independent of T2 by evaluating apparent diffusion coefficient (ADC) maps. Restricted diffusion usually peaks at around 3 days and disappears by 5–10 days (pseudonormalization) [27, 28]. Thereafter diffusion increases in the area of infarction due to removal of dead cells and development of gliosis. Diffusion restriction and increased T2 combine to produce hyperintensity on DWI images that lasts for up to 2 weeks after infarction. In circumstances where diffusion is equal to normal brain but T2 is increased, such as in the case of a subacute infarction, the tissue will appear bright on DWI and isointense on apparent diffusion coefficient (ADC) maps, a phenomenon known as T2 shine through (Fig. 9).

DWI nearly doubles the likelihood of detecting acute ischemic strokes compared to FLAIR [29]. Although T2-weighted imaging and FLAIR are less sensitive in detecting parenchymal change in the first few hours, loss of normal flow voids in large arteries may be visible immediately [30]. The MRI correlate of the early CT dense vessel sign in acute embolic occlusion is intraluminal hypointensity without associated flow-related hyperintensity on gradient echo or susceptibility-weighted images. Normally, large vessels are centrally hypointense with associated hyperintensity adjacent to the vessel due to signal from spatially displaced blood secondary to flow effects. In the presence of acute clot, however, there is marked hypointensity that often extends beyond the lumen of the vessel, referred to as “blooming,” and there is no corresponding flow-related hyperintensity [7] (Fig. 10).

Hyperacute infarction is T1 isointense and T2 isointense to mildly hypertense, which is best appreciated on FLAIR and is usually confined to the gray matter in cases of thromboembolic infarction. It typically takes 4–5 h for hyperintensity to develop on FLAIR, which is actually less sensitive than CT in the 3–6 h period [7]. The presence of restricted diffusion and normal signal on FLAIR (diffusion-FLAIR mismatch) may potentially be used to determine the onset of infarction in patients in whom the time on onset is unknown (referred to as “wake-up” stroke), potentially allowing for therapeutic intervention in these individuals [31] (Fig. 7).

After 6–12 h, infarcted tissue becomes high signal on T2-weighted images, while low intensity on T1-weighted images typically mirrors this high T2/FLAIR signal. In thromboembolic infarcts, T2 hyperintensity is confined to the gray matter. Although the extent and degree of T2 hyperintensity increase during the acute phase of infarction, the extent of the DWI abnormality remains relatively unchanged (unless there has been interval progression of the infarct) (Fig. 10). In 50 % of cases, contrast enhancement can be seen [2]. Enhancement can be intra-arterial due to slow flow, in which case it can be seen at the time of vascular occlusion. Subsequently transient leptomeningeal enhancement may occur and usually resolves by the end of the first week. Parenchymal enhancement due to blood–brain barrier damage and ingrowth of new immature vessels may occur by the end of the first week and last for several weeks [2].

Blood flow to the affected portion of the brain is typically reestablished 24–72 h after infarction, during the early subacute phase. By day 3 or 4, new ingrowth of immature vessels with more permeable blood–brain barriers result in increased vasogenic edema, which is extracellular swelling caused by leakage of fluid out of capillaries. In large infarcts, mass effect typically peaks at around 5 days and can lead to transfalcine or transtentorial herniation. Infarcted brain is mildly T1 hypointense and markedly T2 hyperintense, involving both gray and white matters with ill-defined margins. Intensity on DWI and ADC maps is variable at this stage, reflecting the degree of cytotoxic edema (decreased ADC) and vasogenic edema (increased ADC) (Figs. 10 and 11). If the infarct involves the corticospinal tract, Wallerian degeneration may occur and manifest as mild T2 hyperintensity and restricted diffusion acutely (Fig. 8). In such cases, signal abnormality will be present in the ipsilateral cerebral peduncle and pons and should not be mistaken for an additional area of infarction. In chronic infarcts Wallerian degeneration is manifested by volume loss in the ipsilateral cerebral peduncle without signal abnormality (Fig. 6d).

Within 5–14 days, DWI demonstrates isointensity to mild hyperintensity, while ADC maps display hyperintensity reflective of increased diffusion [7] (Fig. 9). T1 hypointensity and T2 hyperintensity persist into the late subacute and chronic healing phases. Cortical laminar necrosis may be visualized as a ribbon of intrinsic T1 hyperintensity and is usually apparent after 2 weeks [2] (Fig. 12). Cystic encephalomalacia and lacunar infarcts appear as central regions of T1 hypointensity, T2 hyperintensity, and FLAIR hypointensity (similar to spinal fluid intensity) surrounded by T2 hyperintensity and are best seen on FLAIR imaging (Fig. 6). On DWI, chronic infarction is isointense to mildly hypointense, while the infarct appears hyperintense on ADC maps due to increased diffusion in the hypocellular infarcted brain (Fig. 12). If contrast is administered, parenchymal enhancement usually begins near the end of the first week and lasts less than 12 weeks – if it lasts longer, the presence of an underlying lesion should be considered [2].