Lövblad et al. first demonstrated the usefulness of diffusion-weighted images (DWI) in the diagnosis of brain death. They found a significant drop in the apparent diffusion coefficient (ADC) values compared with those of normal controls, and these observations have been later confirmed by other authors [7, 10, 11]. With DWI and ADC mapping, it is possible to identify areas corresponding to cytotoxic edema and ischemic damage – see Fig. 3b, d.

Selcuk et al. [11] observed greater decrease in ADC values in white matter than in gray matter for both the cerebral and cerebellar hemispheres. The sensitivity and specificity as well as positive and negative predictive cutoff values of ADC to distinguish patients with brain death from controls were 100 % in their study, but they recommend to determine ADC cutoff values for different MRI scanners and DWI protocols. They believe that a 100 % sensitivity and specificity in their results indicates a potential role of ADC maps as confirmatory test in diagnosis of brain death.

However, Luchtmann et al. evaluating the reliability of DWI in the diagnosis of brain death concluded that effects of pseudonormalization of ADC values (after 5–28 days after brain ischemic damage) and possible susceptibility artifacts due to microhemorrhages preclude a leading role of DWI in the diagnosis of brain death [10].

Other sophisticated technique offered by MRI is MR spectroscopy (MRS), allowing to study metabolic changes in brain tissue. In previously used method of 31P MRS, the most striking feature was a complete absence of adenosine triphosphate (ATP) and domination of intense inorganic phosphate signal [2, 3]. Kato et al. [3] observed also an absence of phosphocreatine and stated that MRS may be helpful in diagnostics of brain death particularly in children, when other examinations are not sufficient to establish defined diagnosis.

Falini et al. [12] performing sequential MR imaging and proton (1H) MRS in a patient with severe hypoxic-ischemic brain injury found that observed structural and biochemical changes reflected the known evolution of brain degeneration after asphyxial injury, particularly the massive decrease of cortical N-acetylaspartate in the acute phase, suggesting the severity of the neuronal damage, and the subsequent progressive increase of choline related to the delayed degeneration of white matter.

To our best knowledge, no other study on application of MRS in human brain death is available.

At raising ICP the cerebral arteries and veins fill with contrast with increasing delay in comparison to extracranial vessels of the head. At the early stage of cerebral circulatory arrest, nonfilling of the cortical arterial branches and the deep cerebral veins (the internal cerebral vein, ICV; the great cerebral vein, GCV; and the straight sinus) occurs. At this stage an oscillation of contrast column in the proximal intracranial arteries may be detected. The pendulum movement of the contrast is accompanied by its slow antegrade propagation and intracranial stasis. The contrast is gradually eliminated from the intracranial vessels through the arteriovenous shunts and the interrupted blood-brain barrier. Noteworthy the superior sagittal sinus and the transverse sinuses may fill from the extracranial vessels through the emissary veins. With further increasing ICP, the contrast column is stopped in more proximal arterial segments. A total intracranial nonfilling occurs when the ICP reaches systolic blood pressure.

Catheter cerebral angiography for the determination of brain death is performed with an injection of iodinated contrast material to reach both anterior and posterior cerebral circulation. The propagation of contrast is assessed in a filming sequence acquired with the usage of DSA technique. The duration of the acquisition varies among national guidelines, but it should not be shorter than 15 s to confirm cerebral circulatory arrest (personal experience).

The national guidelines permit two different techniques of contrast injection:

While performing the selective angiography, one must take into account that the contrast injected under high pressure into the ICA or VA may be pushed forward to intracranial arteries resulting in an artifactual intracranial filling. This can be avoided if the injection from the aortic arch is applied. Moreover, the selective technique requires higher manual skills and experience than aortic arch angiography. Finally, the risk of procedure-related complications is higher for selective technique. Because both techniques are characterized by equally high accuracy, the aortic arch angiography is a preferred option.

Intravenous angiography can be applied in the determination of cerebral circulatory arrest as well [15]. However, when performing the exam with i.v. injection, one must consider the jugular and dural venous sinus reflux of contrast. Such finding is not indicative of preserved cerebral circulation [15]. Another disadvantage of this technique is a low concentration of contrast achieved in the head and neck vessels due to the effect of dilution. This could make the detection of possible intracranial opacification equivocal. Therefore arterial injection is preferred.

Cerebral circulatory arrest can be documented by catheter cerebral angiography if all of the following conditions are met:

However, delayed opacification of proximal segments of the cerebral arteries with contrast stasis and without deep venous filling does not preclude the diagnosis of brain death – see Fig. 5.

Catheter angiography is widely accepted as a reference test in the diagnosis of brain death. Despite this several disadvantages limit its application. The catheterization of aortic arch, especially ICAs and VAs, requires high expertise. The exam must be performed in the angiographic lab; thus, transportation of a critically ill patient is necessary. The procedure is invasive, accompanied by the risk of serious complications, such as an injury of a catheterized vessel or vasospasm. Besides catheter angiography is the most costly procedure among the blood flow tests. Potential risk of a damage of transplantable organs caused by iodinated contrast, although postulated, has not been confirmed yet. Moreover, it was revealed that contrast medium administration to donors does not affect kidney graft function after transplant [17]. No false-positive results of catheter angiography in the diagnosis of brain death have been reported (i.e., angiography revealed cerebral circulatory arrest, yet the patient survived), which means the 100 % specificity. The sensitivity is close to 100 % as well. There are casuistic reports of false-negative results in cases in which residual preservation or restoration of intracranial blood flow was observed after ICP was relieved through some decompressive mechanism, e.g., craniectomy or skull fractures (see “Limitations” section) [15]. There are several reports of preserved blood flow in the posterior fossa with cerebral circulatory arrest in the anterior circulation [15, 18]. This points to the protective function of the cerebellar tentorium, which causes an uneven increase in ICP. In all false-negative exams, patients met all clinical criteria of brain death, and in all cases intracranial circulation was markedly slowed.

Unfortunately CTA findings were not studied in non-brain-dead patients with intracranial hypertension. The earliest sign of cerebral circulatory arrest in CTA is a lack of opacification of deep veins – the ICVs and the GCV. The sensitivity of this finding in the diagnosis of brain death in CTA is 98–100 % [19–23]. A lack of opacification of cortical branches of the middle cerebral arteries (MCA-M4) is slightly less sensitive indicator of brain death with the sensitivity of 86–100 % [19, 21, 22]. The basilar artery (BA) and cortical branches of the posterior cerebral arteries (PCA-P2) are more frequently opacified in brain-dead patients than MCA-M4 – see Fig. 6. Their sensitivity is 85–94 % and 79 %, respectively [19, 21]. The least sensitive finding of brain death is a lack of opacification of cortical branches of the anterior cerebral arteries (ACA-A3, the pericallosal arteries) with the sensitivity of 64 % [21] – see Fig. 7. Such a sequence is explained by the highest susceptibility of cortical branches of the MCA to intracranial hypertension and the lowest CPP in these arteries. In the course of cerebral circulatory arrest, proximal segments of the cerebral arteries frequently show delayed opacification for some time. These vessels appear as thin and faint. The intracranial stasis of contrast is observed with its extremely slow elimination through the arteriovenous shunts and extravasation due to the interrupted blood-brain barrier. The latest sign of cerebral circulatory arrest is intracranial nonfilling – see Fig. 8.

The technique of CTA involves a rapid, intravenous administration of iodinated contrast in the bolus at a flow rate of 3–5 ml/s. Volume of contrast medium used for CTA for the diagnosis of cerebral circulatory arrest is 60–120 ml. The contrast can be followed by the infusion of 30–40 ml of normal saline at the same rate (i.e., saline flush) to push the contrast medium forward and to optimize contrast enhancement.

For the diagnosis of brain death, at least three acquisitions should be performed:

Intermediate post-contrast scans may be added between the early and the late phase for a better follow-up of the contrast dynamics.

There are no widely accepted criteria of CTA in the diagnosis of brain death. Four methods of evaluation have been proposed so far; they are summarized in Table 4 and Fig. 9.