Pediatric neuroradiology Common clinical problems
Epilepsy
Epilepsy in childhood is most commonly due to structural changes in the brain that cannot be detected by any neuroradiological method. Most children with epilepsy will make a full recovery before adulthood. Malformations and destructive lesions of the brain can sometimes be associated with epilepsy and diagnostic efforts are often concentrated on identifying the se patients with the help of neuroradiology. Tumors of the brain are uncommon as the cause for epilepsy in childhood. It is neither practically possible, nor desirable to study all children with epilepsy using neuroradiological imaging methods. It is the task of the pediatric neurologist or pediatrician with specific expertise in neurology to select the patient who should be investigated using neuroradiology. It is important to recognize that a request for a neuroradiological investigation shall never replace a referral of the patient with epilepsy to a specialist in pediatric neurology. The most commonly used criteria in selecting children suitable for CT examination of the brain are the following:
- epilepsy with psycho-motor defects in a child below one year of age, particularly infantile spasm
- epilepsy with a changed seizure pattern
- epilepsy with focal EEG-changes, particularly with focal slowing - epilepsy with seizures resistant to medical therapy and in which
surgery is contemplated
- status epilepticus
- post-traumatic epilepsy
It is usually state d that the CT scanning can detect pathology in about 40 % of patients with infantile spasm. The diagnoses encountered include such diseases as phakomatoses, cerebral malformations or destructive brain damage that may have occurred in ufero or perinatally. If possible, this investigation should be delayed until the child has reached the age of six months, in order to maximize the diagnostic yield. It is easier to assess the structure of the brain at the age of six months and malformations are then easier to detect or exclude. Rare cases of a tumor causing epilepsy are found particularly in children with a changing pattern of epilepsy and in those with focal EEG-changes. The neuroradiological work-up is extensive, should surgery be contemplated as treatment for epilepsy. It will include MR-imaging, positron emission tomography and specialized neurophysiological methods (Fig. l). Positive findings are uncommon in patients with status epilepticus but the finding of a normal CT scan is quite helpful and important in the immediate care for the child that is unconscious following status epilepticus. Children developing seizures hours or days following a head trauma should be urgently investigated with a CT -scan, even if such a procedure was performed and found to be negative at the time of admission following the trauma. The possibility for delayed intracranial hemorrhage is significant in the short range, while development of leptomeningeal cysts is a distinct possibility in the long range.
 a
|
Figure 1.
This five-year-old boy has epilepsy with focal EEG-changes generated from the right cerebral hemisphere.
a) This CT-image through the parietal lobes shows a focal area in the right parietal region where the cortex has an abnormal appearance with thicker than usual cortical mantel (arrows). There is a widened sulcus superficial to this abnormality.
b) The corresponding MR proton weighted image shows the white matter in the right-sided centrum semiovale to lack the extensions normally found into each gyrus.
c) This heavily T2-weighted image in the coronal plane shows no abnormal signal characteristics but the rather simplistic pattern of white and gray matter in the right parietal lobe as compared to the left side. These images show a focal abnormality of neuronal migration following which the cortical structures have been malformed.
|
 b
|
 c
|
Abnormal head circumference hydrocephalus or atrophy
The head circumference and growth rate is dependent on the intracranial content as long as the sutures are open. Increased intracranial volume may be due to increased amount of fluid or increased amount of normal or abnormal brain tissue and will result in increased head circumference in small children. By the same mechanism, a decreased size of the brain will give rise to a small head.
The strict definition of hydrocephalus is increased amount of fluid in the cranial cavity. The meaning of the word has gradually changed and the contemporary definition is that of abnormal hydrodynamics influencing the cerebro-spinal fluid circulation and causing increased intracranial pressure. Dilatation of the ventricles and sometimes the extracerebral subarachnoid spaces is secondary to the increased pressure and the increased amounts of fluid will cause an increase in the head circumference. On the other hand, destructive damage to the brain will not only cause loss of brain tissue but also a decrease in further growth of the brain. The consequence will be a small head, microcephaly. Loss of brain tisssue, atrophy, may also cause an increased size of the intracranial CSF-spaces but this increase in size is secondary to a decreased size of the brain. The degree of atrophy in a normo-cephalic adult can be directly assessed by comparing the size of the brain to the size of the skull, Decreased skull growth in a child, that is a consequence of a brain damage with tissue loss, will tend to hide even extensive loss of brain tissue as the size of the skull is adapted to the size of the brain and the degree of atrophy can then no longer be assessed simply by comparing the size of the brain to size of the skull, For this reason, the degree of atrophy in a child is often underestimated.
Wide CSF-spaces are very common before the age of 18 months. However, this finding does not necessarily indicate neither hydrocephalus nor atrophy. The finding of wide CSF-spaces cannot and must not be interpreted without knowledge about the head circumference, head growth and clinical symptoms. This situation must be recognized by the radiologist who must never make the diagnosis of atrophy without being absolutely convinced of the presence of loss of brain tissue, as in a child with wide CSF-spaces and reduced head growth or microcephaly. Similarly one should never make the diagnosis of hydrocephalus unless head growth is accelerated or head size significantly increased. Ventriculomegaly is common and may be present in many other clinical situations than hydrocephalus. A combination of atrophy and hydrocephalus may therefore be very difficult or even impossible to evaluate using neuroradiology. Hence, every request for a neuroradiological investigation in children must include information about history, size and growth of the head, as well as a clinical evaluation of the intracranial pressure.
Clinical suspicion of increased intracranial pressure is a clear indication for neuroradiological investigation. The task of the radiologist is first to confirm presence of hydrocephalus and then to assess the cause of this condition. Increased intracranial pressure can usually be radiologically confirmed but it may sometimes be impossible to diagnose increased intracranial pressure by neuroradiological methods alone. The investigation must be carried as far as possible and may include both CT and MRI before and after contrast injection, if the cause for hydrocephalus or ventriculomegaly is unknown. The diagnosis of aqueductal stenosis, previously a diagnosis by exclusion, has been common in the past. With the capability of MRI to much more precisely detect the cause for hydrocephalus, aqueductal stenosis is now a rare diagnosis. If the reason for the hydrocephalus in an infant is known, i.e. intracranial hemorrhage or meningitis, a neurosonographic examination may be quite sufficient to establish the size of the ventricular system. No further neuroradiological investigation is then indicated.
Assessment of shunt patency and function becomes necessary when a patient with hydrocephalus has been treated with a ventriculo-peritoneal shunt. Direct verification that the intracranial pressure has been normalized is often difficult. However, neuroradiology can be of assistance by assessing indirect signs such as the size of the ventricles. Although assessment of the ventricular size is the most simple method, the ideal size of the ventricles varies from patient to patient and the assessment is therefore very difficult without access to a base-line examination carried out when the patient is clinically well and without any clinical signs of either under- or over-function of the shunt. Routine examinations of the brain are not indicated.
Delayed psycho-motor development
Delayed psycho-motor development is in most cases thought to be due to brain damage that occurred in utero or during the perinatal period. Although few of the conditions that may cause psycho-motor delay are treatable, the indication for neuroradiological investigation is obvious. Establishing a correct diagnosis has great impact for the parents of the child and might relieve feelings of guilt and assist in genetic counselling. The finding of a congenital malformation of the brain has important consequences in medico-legal situations when quality of obstetrical care may be in question. The importance of neuroradiological investigation is
|
Figure 2.
This boy had a larger than normal head at two years of age. He had a marked delay in his psycho-motor development.
The CT image shows a generalized malformation of the brain, called holoprosencephaly. In this malformation there is an abnormality of the cleavage into two cerebral hemispheres which may be only partial. Note the absent interhemispheric fissure and the bridging of white and gray matter between the two cerebral hemispheres anteriorly (arrows). The lateral ventricles are fused into a monoventricle communicating posteriorly with a midline parietal cyst. Absence of falx cerebri is secondary to absence of the interhemispheric fissure anteriorly.
|
much less in the normo-cephalic mentally retarded child without motor handicap.
The neuroradiological investigation of children with non-progressive psycho-motor handicap, cerebralpalsy, should include CT or MRI without contrast and is best carried out after the child has reached the age of at least eighteen months. The radiologist should be particularly observant towards loss of brain tissue that may cause asymmetries between the cerebral hemispheres and abnormal structure of the brain that may indicate abnormal neuronal migration or other congenital malformations, when investigating a child on the indication of cerebral palsy (Fig. 2). Phakomatoses, such as tuberous sclerosis or destructive damages after intra-uterine infection, may also be seen as well as brain tissue loss caused by ischemic brain damage in the pre-, peri- or postnatal period.
Metabolic diseases should be suspected in children with progressive symptoms of brain dysfunction. This damage may be widespread and involve both white and grey matter. It can be seen by CT but it is much better assessed using MRI. These diseases can produce complicated patterns of decreased attenuation on CT or changes in signal on MRI. Although typical patterns can be recognized in specific diseases, assessment in the individual case of these findings is often difficult and the findings are quite frequently non-specific.
Headaches
Headache is a common symptom in children and adolescents but it is an uncommon symptom of intracranial pathology and as such occurs almost exclusively in situations with increased intracranial pressure. CT or MRI without contrast injection can, with certainty, exclude the presence of a mass lesion as the reason for the headaches. However, CT or MRI both with or without contrast can never exclude the presence of the rare aneurysm or a small arterio-venous malformation that has not bled. However, headaches occur in this situation when the aneurysm or A VM has ruptured and the fresh subarachnoid hemorrhage can then be diagnosed using CT without contrast. A large AVM with a significant arterio-venous shunt may cause headaches, even without rupture, but such a large arterio-venous malformation is also obvious on a CT scan without contrast. As with epilepsy, a request for a neuroradiological investigation, CT or MRI, must never replace the referral of a child with headaches to a specialist in pediatric neurology. A normal result is the expected finding on CT or MRI in a child with headaches. As a normal result may be very important as support in the further care of the child with headaches, the request for neuroradiology in a child with headaches should come from a physician trained in pediatric neurology.
Severe, often a unilateral headache in a child who has experienced a recent head trauma should lead to an urgent investigation on the suspicion of an epidural hematoma.
Suspected or ongoing meningitis is a common reason to request a neuroradiological examination. However, it is reasonable to suggest that indication for imaging is present only when the detection of pathology will lead to change of therapy. Hence, investigations done with the sole purpose of confirming the strong clinical suspicion that a complication such as an infarct has happened, is of rather limited value. Apart from infarction, subdural effusions and contrast enhancement in the leptomeninges are common findings. In addition, every child suffering from bacterial meningitis has to a certain degree a disturbed CSF-circulation, recognized as slight increase in the size of the ventricular system and subarachonid spaces (Fig. 3). The medical treatment given to the child should not be influenced by these findings on CT or MRI as the causal relationship between the clinical findings and these radiological findings are
 a | Figure 3.This 4-months-old girl had been diagnosed with a H. influenzae meningitis a week prior to this CT scan. Early seizures provided the indication for this study.
a) Note the extensive area of decreased brain tissue attenuation corresponding to an infarct in the areas supplied by the left middle cerebral artery (arrows) in this pre-contrast image. Two small subdural effusions are seen adjacent to the anterior portion of the interhemispheric fissure (arrowheads).
b) No break down of the blood-brain barrier is noted in the infarct following contrast enhancement but dense enhancement is seen in the leptomeninges over both frontal lobes (arrows). This enhancement is due to meningeal inflammation secondary to infection and does not indicate an empyema.
c) An extensive area with loss of brain tissue, atrophy, corresponding to the infarct in a) and b) is seen in this image obtained four months later. |
 b |
 c |
at best weak or even non-existing. Hence, prolonged fever and early convulsions represent rather doubtful indications for imaging. The possibility that an abscess can present as a complication to a hematogenous spread meningitis is unlikely. However, neonatal meningitis caused by various gram-negative bacteria may frequently be complicated with the formation of brain abscesses. Severe neurological depression and unconsciousness as presenting symptoms during meningitis does indicate early and extensive brain damage and the
CT scan done at the time of, or soon after, admission may provide important prognostic information, as evidence of widespread ischemic brain damage early in the disease process carries a poor prognosis. Late onset of convulsions or persistent seizures later than two weeks after presentation of the disease, represent good indications for neuroradiological investigation with
CT, as subdural effusions may have become secondarily infected and developed into an infected subdural empyema. It is therefore quite clear from the discussion above, that routine imaging is not indicated in children with meningitis, but that each case must be evaluated and the indications for imaging assessed individually.
The clinical symptoms are often dramatic in smaller children who suffer encephalitis. Nevertheless, neuroradiological findings are often rather unimpressive. Neither CT nor MRI seem to have a great potential in the early diagnosis of the disease. The diagnosis is usually confirmed clinically long before neuroradiological findings may be present. The clinical course is similarly quite dramatic in patients with postinfectious encephalitis, acute disseminated encephalo-myelitis (AD EM), and there are usually good reasons to suspect an intracranial or intraspinal mass lesion. CT scanning of the brain is capable of excluding an intracranial mass lesion, while MRI may show, even during the early stages of the disease, the demyelinating plaque either in the brain or in the spinal cord. The clinical picture is often typical with alarming neurological symptoms approximately ten days to two weeks following an uncomplicated upper respiratory infection or after an immunization. Acute onset of transverse myelitis is the presenting symptom should the demyelinating plaque be located in the spinal cord. These patients must be investigated on an urgent basis, 24 hours a day, in order to exclude an intraspinal abscess or hemorrhage amenable to surgical intervention. MRI is clearly the procedure of choice in transverse myelitis as it can not only exclude a mass lesion but also clearly demonstrate the demyelinating plaque, if present.
Stroke in childhood
Acute onset of hemiplegia or other neurological deficits, represent a serious disease in a child. The first aim of the investigations is to differentiate between a hemorrhagic or an ischemic les ion as the cause for the patient's symptoms. A hemorrhage may be caused by an arterio-venous malformate ion or by a tumor. Further investigation with cerebral angiography is indicated, should an AVM be suspected. An immediate angiogram is indicated if urgent surgery is needed. Otherwise, such angiograms should be delayed and performed two or three months after the acute episode. The hematoma may compress the arterio-venous malformation which then escape may detection altogether. Hence, an angiogram performed in the acute stage of the disease can fail to show all components of the AVM or fail to show any trace of the AVM. Whether or not the angiogram is normal, it must be repeated when the hematoma has been reabsorbed and prior to elective treatment.
The indication for angiographic evaluation in the ischemic injury is controversial. An urgent angiogram is only indicated if the findings can lead to an altered therapy. Alterations in the child's coagulation system may cause an infarction but can also increase the risk for complications during an angiogram. Although treatment of childhood stroke usually is conservative, an angiogram can demonstrate an injury in the cervical vessels that may be amenable to surgery if the stroke has a causal relationship with neck trauma.
Neonatal cerebral disease
Neuroradiological investigation aimed at identifying intracranial hemorrhage or pathology amenable to surgery during the neonatal period is of assistance in the immediate care for the neonate. In addition, neuroradiological investigation can assist in the early assessment of long term prognosis in terms of possible future psycho-motor handicap.
Cerebral damage in the form of ischemic or hemorrhagic lesions caused by cerebral vascular disease can occur during late pregnancy, in the perinatal or neonatal period. The lesions may be located in white matter (periventricular leukomalacia) or grey matter. Hemorrhage or ischemic damage to the white matter is thought to occur almost exclusively in the immature brain while damage to the grey matter is more common after 34 weeks of gestation and in the child born at term. The pathophysiolsogy of these injuries is intimately related to the changes in vascular physiology in the developing brain. The vascular architecture of the brain in a 30 weeks fetus is profoundly different from that of a full term neonate. Consequently, the pathology of brain damage before and after 34 weeks gestation is quite different, even though the mechanism of injury may be the same. Hence, different types of injury can be related to the stage of brain development at the time of injury and should not be related to the time of the delivery. The pathology in the neonatal brain injury is quite different in the mature and immature brain, respectively. This fact has an important impact on the indications for investigation and choice of imaging modality; the modality used for neuroradiological investigation should depend mainly on the maturity of the neonate.
Neuroradiological investigation in the prematurely born neonate is best carried out using neurosonography. Significant hemorrhage is easily identified while evidence of parenchymal damage may be rather subtle and often impossible to show with either neurosonography, CT or MRI. It is important to identify intraventricular hemorrhage, as large hemorrhages are almost always associated with damage to parenchyma and hence future handicap. lntracerebral hemorrhages are uncommon in neonates more mature than 34 weeks gestational age. lschemic damage to parenchyma is in these children often located in cortical or subcortical brain tissue. Extensive ischemic brain damage can give rise to generalized cerebral edema, a condition that almost always can be demonstrated by CT, while the interpretation of findings of edema on neurosonography may be more difficult. Hence, CT scanning should be the method of choice in neuroradiological investigation of mature newborns with clinical evidence of brain injury. The cerebral edema developing as a consequence to ischemic damage is known to peak at about 72 hours following the injury. Therefore, the CT should if possible be carried out during the third day of life, assuming an injury at the time of delivery. ACT scan performed at this time can provide important information useful in assessing the long term prognosis, albeit only in terms of severe handicap. It is clear from the discussion above that a request for neuroradiological investigation of a neonate must include information about the maturity of the neonate to allow the radiologist to choose the appropriate mode of investigation.
Congenital malformations of the spine include both mesodermal (muscular or skeletal) and neuro-ectodermal (neurogenic) structures. The malformations typically occur in combination with varying degrees of involvement. However, malformations can occur in each of these structures independent of the other. Abnormal segmentation of the spine can be associated with neurogenic malformations but can also be isolated.
The neuroradiological investigation in a patient with suspected spinal malformation should aim first at confirming or excluding the possibility of neurogenic malformation. If a malformation is found, further investigation should concentrate on demonstrating all components of the malformation. This very important knowledge allows the neurosurgeon to plan reconstructive surgery. Plain films of the spine, sometimes including conventional tomography, may be very important for the orthopedic surgeon but are of little value in assessing a possible association with neurogenic malformations. This evaluation must focus on the content of the spinal canal by direct visualization of the subarachnoid space and it's content. Spinal ultrasound can provide useful information during the neonatal period but this mode of investigation does not allow confident exclusion of a suspected malformation. Myelography followed by CT is capable of providing all necessary information. MR imaging of the spine is also capable of showing all components of a malformation. However, a complete MR examination of the spine is a rather time-consuming procedure and, although non-invasive, quite often requires heavy sedation or general anesthesia. Modem MR scanners, and with the use of modem imaging sequences, are capable of showing most components of these malformations but some types of malformations, e.g. diastematomyelia and dermoid tumors, can be very difficult to detect, despite the use of excellent MR technique. It is often so that both MRI and myelography followed by CT, are required not only to describe all aspects of a malformation but also, in other patients, to completely exclude the possibility of a congenital malformation of neural tissue.
Olof Flodmark and Derek Harwood-Nash