Initial imaging should not only help to make the diagnosis of acute ischemic stroke (AIS) but also indicate the potential of an ischemic stroke to grow and swell. Magnetic resonance imaging (MRI) has been proposed for prognostication of severe stroke, including malignant hemispheric stroke. Studies yielded robust results on the predictive potential of certain sizes of the diffusion-weighted (DWI, >145 cm³) and the apparent diffusion coefficient (ADC, <80%, >82 cm³) lesions [5]. However, MRI is not part of the initial stroke work-up at several institutions and has not yet gained widespread acceptance for assessment and prognostication in these patients. This is due to wide availability of computed tomography (CT). Hence, CT is still the most popular imaging modality to judge the course of severe AIS. Horizontal displacement of the pineal gland of >4 mm within 48 h [6], loss of gray and white matter differentiation within the first hours [7], and hypoattenuation in more than 50% of the territory of the middle cerebral artery (MCA) [8] have all been found predictive of an unfavorable clinical course. However, patients show marked interindividual differences in their edema dynamics. Hence, quite often serial CT imaging during the first days post stroke is necessary to assess and anticipate the course of the patients. In large MCA infarction, total or subtotal hypoattenuation of the MCA territory, hypoattenuation in adjacent vessel territories, (partial) involvement of the basal ganglia, and space-occupying effect (e.g. compressed lateral ventricle or midline shift) are radiological criteria to seriously consider decompressive surgery at our institution.

Extensive AIS leads to the generation of at first cytotoxic and later vasogenic edema, with considerable differences in extent and dynamics between individual patients [10]. On average, clinically relevant edema develops at day 2 or 3 from stroke onset. Basic measures to prevent brain edema comprise restriction of free water, avoidance of hypo-osmolar fluids, avoidance of excess glucose administration, avoidance of hypertension after reperfusion therapies, minimization of hypoxemia and hypercarbia, caution in application of drugs causing cerebral vasodilation, and aggressive treatment of hyperthermia. If clinically relevant edema occurs and is detected by clinical signs (decline in level of consciousness, worsening of neurological deficit, nausea and vomiting, anisocoria), ICP monitoring and cerebral imaging, as well as pharmacological treatment, should be applied. Additionally, patients with severe stroke whose exams are difficult to assess due to sedation should have their ICP measured by parenchymal probe or external ventricular drain (if hydrocephalus is feared or present). The most common potential cause of neurologic worsening in large AIS is edema with mass effect and midline shift. Secondary hemorrhage, seizures, and reduced venous return are additional potential causes. There do not exist specific data on optimal ICP-lowering therapy in severe stroke, so it is recommended to follow a stepwise approach common in other brain injuries (Table 8.3).

The order of steps may vary individually. Hyperventilation and barbiturate administration can be deleterious and should be only considered cautiously.

Occlusion of the distal ICA or proximal MCA leads to infarction of the total or subtotal MCA territory, possibly combined with infarction of the adjacent anterior cerebral artery (ACA) and/or posterior cerebral artery (PCA) territory. This type of brain infarct is called large hemispheric stroke. As mentioned above, MMI refers to a large hemispheric stroke complicated by significant brain tissue swelling, severe enough to compress critical neural structures.

Patients initially present with a severe hemisyndrome, combined with aphasia or neglect depending on the affected hemisphere, and display quite stereotypical deterioration over the first few days, such as decline in level of consciousness and anisocoria, reflecting swelling of the affected hemisphere. Despite maximal conservative critical care efforts, the mortality of MMI is between 70% and 80% due to massive swelling of the infarcted area leading to horizontal displacement of the brain stem, and intracranial pressure increases within the rigid skull (Fig. 8.1). Medical options to prevent or reduce brain edema have been disappointing so far. Classical osmotherapy is not sufficient to improve clinical outcome in MMI [1, 2, 13–17].

The previously bleak perspective of these patients has changed over the last 15 years, since recent evidence has been obtained on the benefits of decompressive craniectomy (DC). This surgical measure, by which a large (>12 cm diameter) bone flap is removed over the affected hemisphere combined with duraplasty, allows for outside swelling of the affected tissue leaving the contralateral unaffected parts of the brain uncompromised (Fig. 8.2). The bone flap is stored frozen or subcutaneously and reinserted after rehabilitation and shrinking of the affected hemisphere, usually between 3 weeks and 3 months from the insult.

After first encouraging results from observational studies [16], the pooled data on 93 patients from 3 European randomized trials showed that DC within 48 h from onset for patients <60 years suffering from MMI reduces mortality by 50%, (number needed to treat (NNT) 2) and severe disability by 40% (NNT 2); on the downside, the number of those surviving with moderate to severe disability increases by 30% [17]. This evidence of survival and outcome benefits from the procedure led to international recognition of DC and to implementation into guidelines for MMI patients under 60 years of age. Recently, the DESTINYII trial on DC for MMI in 112 patients >60 years confirmed the highly significant survival benefit (33% vs. 70%, NNT 4) that was mainly responsible for an “outcome benefit” (mRS 0–4 38% vs. 18%), but showed survival after 6 months with moderate to severe disability (mRS 4) in 32% and with severe disability (mRS 5) in 28% of the surgical patients [18]. These and more randomized trials on DC for MMI have been nicely reviewed by Zha and colleagues [19]. These data provide fairly solid grounds for patient and family counseling on the option of DC, which is a very important element of care in this type of stroke (see below). Surgical treatment of MMI involves optimal and maximal neurocritical care before and after the operation.