Magnetic resonance imaging in neonatal hypoxic-ischemic brain injury
Liauw, L.
Citation
Liauw, L. (2009, March 19). Magnetic resonance imaging in neonatal hypoxic-ischemic brain injury. Retrieved from
https://hdl.handle.net/1887/13690
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/13690
Note: To cite this publication please use the final published version (if
applicable).
Chapter 1
Introduction and aims
Chapter 1
10
Hypoxic-ischemic brain injury
In young infants different types of hypoxic-ischemic brain injury can occur (3,5, 7,9,10,39,101,114,136,155,164,165,167,177,198). Brain injury can be localized or diffuse (3,5,7,9,10,39,101,136,155,164,165,167,177,198). In localized brain injury arterial or venous periventricular infarctions are encountered. Causes of arterial infarction in the pediatric population are coagulopathies, vasculopathies, polycythemia, and cyanotic congenital heart disease (16). In the pediatric population, including young infants, in more than half of cases the cause of arterial infarction is unknown (16,70,136). The infant may present with seizures, hypotonia, lethargy, or motor asymmetry or the infarction may be asymptomatic, which may remain undetected until neuroimaging is performed for other reasons (16,136). Compared to adults, the prognosis in young infants with arterial infarction is generally better (16,132,133,135,136). When no underlying cause is found, the child is unlikely to have another stroke in the future and neurological prognosis, apart from the local injury, is good (16). Venous periventricular infarction is typically seen after germinal matrix hemorrhage in preterm infants and remains beyond the scope of this thesis (16,113,193).
Diffuse ischemic brain injury is the result of decreased brain perfusion or hypoxemia.
This may result in ischemic injury in full-term infants and in germinal matrix hemorrhage and/or white matter injury in preterm infants (3,4,5,7,9,10,39,41, 46,68,101,114,136,155,164,165,167,177,191,193,198,199). In full-term infants the clinical entity is referred to as hypoxic-ischemic encephalopathy, and has been classified into three grades according to Sarnat (171). In general, infants with signs of hypoxic-ischemic encephalopathy show fetal distress prior to delivery, have abnormal Apgar scores, require resuscitation at birth, and have neurological abnormalities within the first days of life, such as feeding difficulties, irritability, abnormality of tone, seizures, and decreased level of consciousness (167). Mild encephalopathy (grade 1 according to Sarnat) is characterized by alternating level of consciousness, including periods of lethargy, irritability, and hyperalertness, exaggerated Moro and tendon reflexes, and sympathetic overreactivity. The infants are jittery and may feed poorly, muscle tone is normal or high and seizures are absent (192). Recovery is usually complete within two days (45,192). Children may show mild abnormalities in tone and may show tremor and/or abnormal motor development in early childhood, but usually there are no long-term sequelae (16,45,167,192). In moderate encephalopathy (grade 2 according to Sarnat), the infant may be lethargic, hypotonic with increased tendon reflexes and abnormal complex reflexes. Seizures occur frequently within the first 24 hours (192). The infant feeds poorly and the infant may have bradycardia. Infants with moderate encephalopathy have a 10% risk of death, and survivors have a 30% risk of disabilities (175). Surviving children may have spastic and/or dyskinetic motor impairment, generally with intellectual preservation (16,167). Infants with severe encephalopathy (grade 3 according to Sarnat), with coma, flaccid muscle tone, brainstem and autonomic dysfunction, seizures, and possible increased intracranial pressure, generally die. Sixty percent of infants with severe encephalopathy die, and many survivors, if not all, are handicapped (175). Those who do survive
suffer from severe neurological abnormalities, including microcephaly, mental retardation, spastic quadriparesis, choreoathetosis, and visual impairment (192).
The clinical picture, neuroimaging findings, and electroencephalography (EEG) results help to prognosticate neurological outcome in (near term) infants with hypoxic-ischemic encephalopathy. Severe hypoxic-ischemic encephalopathy, basal ganglia and/or widespread cortical abnormalities on neuroimaging, and severe EEG abnormalities portend a poor outcome (24,116). Fifteen to twenty percent of all infants with hypoxic-ischemic encephalopathy die in the neonatal period (64), and another 25% develop significant neurologic sequelae (64,74).
Imaging modalities
Different imaging modalities are available to detect neonatal brain injury, including cranial ultrasonography, computer tomography (CT), and magnetic resonance imaging (MR imaging) (2,190).
Cranial ultrasonography can be performed at the bedside, which is an advantage in unstable and/or very preterm infants. Furthermore, this method is non-invasive and relatively low-cost. Cranial ultrasonography is particularly suitable for screening and follow-up examinations (144,206). It is generally performed using the anterior fontanelle as an acoustic window. The posterior fontanelle and mastoid fontanelles can be used as acoustic windows to study the posterior fossa and brainstem. The superficial part and deep regions of the brain as well as the intracranial blood vessels can be studied with the array of available transducers. The sensitivity and reliability of cranial ultrasonography for detection of germinal matrix hemorrhage, intraventricular hemorrhage, and periventricular leukomalacia in preterm infants is well-known (16,195,198). Diffuse white matter injury in preterm infants is less well detected by cranial ultrasonography (38,51,91,123,139,141,149,154).
On the otherhand, in full-term infants, infarction and white matter injury are well detected (50). Cranial ultrasonography is less sensitive to abnormalities of convexity structures and the posterior fossa (16,206), it does not give detailed information about myelination and it shows difficulty in detecting lesions of the posterior limb of the internal capsule (PLIC) (16). There are limited data in the literature regarding comparison of the accuracy of cranial ultrasonography and MR imaging for the detection of hypoxic-ischemic lesions in full-term infants (26,50,163). Duplex Doppler examination provides additional information on cerebral perfusion.
CT is not sensitive in depicting edema and infarction in hypoxic-ischemic brain injury because of the high water content in the newborn brain, resulting in poor contrast resolution (37). CT should therefore not be used to visualize hypoxic- ischemic brain injury in infants. On the otherhand, CT affords excellent depiction of hemorrhage, such as intraventricular hemorrhage (14). However, especially in these young infants, the use of ionizing radiation is a disadvantage (14). Since cranial ultrasonography also depicts most hemorrhages, the role of CT in young infants is very limited.
Chapter 1
12
MR imaging has the advantage of superbly displaying soft tissue contrast differentiation and moreover displaying the exact extent and site of brain injury better than cranial ultrasonography (14,124). In addition, the myelination process can be assessed by MR imaging while cranial ultrasonography does not visualize myelination (104). However, performing MR imaging is logistically more challenging in infants than cranial ultrasonography. Consequently, MR imaging, especially serial, is not performed routinely. In imaging neonatal hypoxic-ischemic brain injury, conventional T1- and T2-weighted pulse sequences are used most often (14,96,124). Due to the myelination changes associated with the maturation process of the brain, T1- and T2- relaxation times change over time in the maturing brain, leading to changing contrast between grey and white matter on T1- and T2-weighted images (14,46,104,124). During the first six months of life the myelination process is best visualized on T1-weighted images, whereas after six months T2-weighted images better reflect this process (16,104). Because of the high water content of the immature brain, fluid-attenuated inversion recovery (FLAIR) imaging is of less use in the first year after birth than in older children and adults (14,119,167,169). Only limited data is available on the value of contrast enhanced imaging in imaging hypoxic-ischemic brain injury (203).
Currently, the following more advanced MRI techniques are available for assessing hypoxic-ischemic brain injury and its sequelae:
1) Diffusion-weighted imaging (DWI) makes use of the hydrogen molecule’s physical property of diffusion. Unless restricted (for instance by cell membranes and white matter fibres), diffusion occurs in all directions. DWI is sensitive in showing cytotoxic edema (13,44,90,120,151,168). DWI may not detect brain injury if performed in the first hours after the incident, this as delayed injury and late glial swelling due to delayed energy failure (also called bi-phasic energy failure) has not yet taken place. Therefore, if performed in the first 24 hours, DWI may underestimate the extent of brain injury (16,167), also when compared to MR spectroscopy (13). DWI is probably most sensitive two to five days after hypoxia-ischemia (16,167). In cases without visual abnormalities on DWI, diffuse hypoxic-ischemic brain injury can not be excluded with certainty (207).
In contrast to that in adults, DWI may be falsely negative in neonates due to the very long T2 values of the infant brain (16). It has therefore been advocated to perform apparent diffusion coefficient (ADC) measurements routinely in young infants (121,143,168,183,207). In addition to random thermal motion (Brownian motion), the motion of water molecules in biological tissue is influenced by several factors, such as concentration gradients, pressure gradients, or ionic interactions (145,172). With DWI no correction takes place for the available volume fraction or the increases in distance travelled due to tortuous pathways, thus the water diffusion measured in tissue with DWI is called the ADC (145,172). In adults, normal brain ADC values have been published (80,142). In young infants less information is available about normal brain ADC values (168,174,183,184,207). ADC values measured after acute ischemia are mostly below those of normal tissue and often remain markedly reduced for 3-5 days. ADC values generally return to baseline at 1-4 weeks. This normalisation is often referred to as ‘pseudonormalization’
(108,125). It probably reflects persistence of cytotoxic edema, associated with
decreased diffusion, and at the same time development of vasogenic edema and cell membrane disruption, leading to increased extracellular water, associated with increased diffusion (172). The affected tissue is damaged and probably nonviable with persistence of high signal intensity on T2-weighted images, but with normal ADC values. This increased diffusion corresponds to elevated ADC values. Elevated ADC values return to baseline after weeks-months. After pseudonormalization, because of increased diffusion due to increased extracellular water the ADC values continue to increase and remain elevated during a longer period of months, thereafter returning to baseline.
2) A more sophisticated method of diffusion imaging is diffusion tensor imaging (DTI). DTI is a mathematical description of water motion in a voxel. As water motion is primarily influenced by brain structure, DTI reflects brain structure. In conjunction with connectivity algorithms, DTI allows axonal tracts to be mapped.
In this way brain structure information is obtained after hypoxemia-ischemia and in a variety of other neurological diseases, such as metabolic diseases and congenital developmental abnormalities. DTI holds great promise in depicting brain structure and visualizing the sequelae of hypoxic-ischemic brain injury (14).
3) Physicochemical interactions between free water protons and macromolecular protons, such as diffusion of free water spins, chemical exchange, and dipolar magnetization transfer are the basis of magnetization transfer imaging (MTI), which can be applied to augment tissue contrast and for tissue characterization (31).
In infants, MTI provides opportunities for evaluating the process of myelination (14). The amount of magnetization transfer increases during myelination and an increase in magnetization transfer parallels an increase of myelination. It has been demonstrated that brain maturation can be captured and quantified using magnetization transfer ratios (33,63,209). A quantitative measure of the amount of magnetization transfer is the magnetization transfer ratio (MTR) (32). However, because of differences in software and hardware, standardization of quantitative MTI has to be accomplished before comparison of quantitative MTI data generated on different MR imaging machines is possible (32).
4) Perfusion studies assess cerebral blood flow. These studies can be performed using contrast agents (‘dynamic susceptibility weighted contrast enhanced perfusion MR imaging’) or without contrast agents (‘arterial spin labelling perfusion MR imaging’, and on an anatomical level, ‘blood oxygenation level dependent’ (BOLD) imaging) (36,201). In adults, contrast enhanced perfusion imaging is mainly used in the evaluation of infarction and neoplasms. In young infants, limited injection velocities and small contrast agent doses limit the changes in T2 and T2* due to the passage of contrast agent, which therefore results in decreased imaging resolution (14,16,201). Consequently, contrast enhanced perfusion imaging is technically more challenging to perform in young infants (201). Compared to contrast enhanced perfusion imaging, arterial spin labelling techniques have limited resolution due to the intrinsic small labelling effect (201). However, unlike to contrast enhanced perfusion imaging, repeating arterial spin labelling perfusion imaging is easy. For the reasons mentioned above, the preferred technique of perfusion study in young infants is arterial spin labelling perfusion imaging. It is used for assessment of cerebral blood flow in hypoxia-ischemia and evaluation of brain perfusion in congenital heart disease.
Chapter 1
14
5) MR spectroscopy (MRS) provides information on the concentrations of numerous neurochemicals. MRS is used in the evaluation of neoplasms, metabolic disease, and, especially in young infants, of hypoxia-ischemia. Although MRS is one of the most sensitive methods to detect hypoxic-ischemic brain injury, it is still not performed routinely on a large scale because of technical difficulties and lack of experience (12,13,72,73,78,85,126,150,160). It is not possible to study the whole brain with MRS because spectra are obtained in specified delineated areas of the brain (187). Long echo times provide optimal evaluation and quantification of metabolites such as lactate and N-acetylaspartate (NAA) (14,16). Hypoxia- ischemia triggers anaerobic metabolism with production of lactate. Elevated lactate and reduction of NAA is seen in the first few hours after a hypoxic-ischemic incident. If still present after 24 hours, lactate is a good indicator of permanent brain injury (16). However, lactate is not specific for hypoxic-ischemic brain injury.
Presence of lactate may be normal in the immature brain of full-term neonates and varies with maturity (106,111). As different brain regions mature differently, it is important to know the infant’s gestational age and region where the MRS spectrum was obtained (16).
Nowadays many advanced MR imaging techniques are available for detection of hypoxic-ischemic brain injury in young infants. In tertiary referral centres and large children’s hospitals, these techniques are used on a regular basis in these young patients. However, a considerable proportion of young infants with hypoxic- ischemic brain injury are imaged in primary and secondary care centres where experience with these advanced neuroimaging techniques is limited. Consequently, cranial ultrasonography and the more conventional MR imaging techniques and pulse sequences, such as T1- and T2-weighted imaging remain the mainstay of neonatal neuroimaging today. Detection of hypoxic-ischemic brain injury in young infants on these conventional sequences can be challenging.
This thesis
The main purpose of this thesis was to study diagnostic and predictive MR imaging features of hypoxic-ischemic brain injury in young infants. We decided to focus on T1-, T2-, and diffusion-weighted imaging, since these techniques are applied on a large scale in primary and secondary care centres. It may be anticipated that in the near future, the use of these techniques in young infants will probably increase as experience with these techniques increases and access to MR imaging machines becomes more widespread.
The specific aims of this thesis were:
1) To explore the optimal MR imaging protocol in newborns with hypoxic- ischemic encephalopathy,
2) To develop criteria to distinguish (subtle) hypoxic-ischemic brain injury from normal myelination and late changes due to periventricular leukomalacia from normal ‘terminal zones’ on MR imaging,
3) To assess the predictive value of MR imaging for outcome and to compare the predictive value of MR imaging with other diagnostic tools (electroencephalography (EEG) and cranial ultrasonography) in young infants with hypoxic-ischemic brain injury, and
4) To assess the value of cranial ultrasonography as compared to MR imaging to demonstrate hypoxic-ischemic white matter injury in preterm infants.
