Abstract
MRI is invaluable in assessing the neonatal brain following suspected perinatal injury. Good quality imaging requires adaptations to both the hardware and the sequences used for adults or older children. The perinatal and postnatal details often predict the pattern of lesions sustained and should be available to aid interpretation of the imaging findings. Perinatal lesions, the pattern of which can predict neurodevelopmental outcome, are at their most obvious on conventional imaging between 1 and 2 weeks from birth. Very early imaging during the first week may be useful to make management decisions in ventilated neonates but brain abnormalities may still be subtle using conventional sequences. Diffusion-weighted imaging (DWI) is very useful for the early identification of ischaemic tissue in the neonatal brain but may underestimate the final extent of injury, particularly basal ganglia and thalamic lesions. MR imaging is an excellent predictor of outcome following perinatal brain injury and can therefore be used as a biomarker in interventional trials designed to reduce injury and improve neurodevelopmental outcome.
Similar content being viewed by others
Introduction
MR imaging of the brain has an increasingly important role in neonatology with the advent of treatments aimed to decrease perinatally acquired brain injury [1]. The neonatal brain may be injured by many processes including hypoxia-ischaemia, viral and bacterial infection, trauma, haemorrhage and metabolic disorders, acquired conditions such as kernicterus and hypoglycaemia or congenital as in one of the many the inherited disorders of metabolism. It is beyond the scope of this review to cover all aspects of acquired injury and it will focus therefore on the most common cause of perinatal brain injury, hypoxia-ischaemia. In the context of acquired perinatal brain injury MR imaging in the neonatal period will provide information that can:
-
Explain neurological symptoms
-
Confirm the presence or absence of acquired brain lesions
-
Ascertain the timing and aetiology of any brain abnormalities
-
Inform clinical management and guide diagnostic investigations
-
Monitor the effects of interventions
-
Predict outcome and help counsel parents
-
Inform critical incident proceedings
-
Prevent expensive medico legal proceedings
To achieve these ends it is important to obtain both good quality imaging and a correct interpretation of the imaging findings. Image quality is often impaired by motion artefact, poor signal to noise or inappropriate sequence choice. Good preparation is essential to achieve a successful examination and a good account of the necessary practical issues is provided in a recent review [2]. The protocols in our unit are shown below (Table 1).
Image acquisition
Image quality is governed by the signal to noise ratio (SNR) and the absence of artefact. The SNR is maximised by using a closely fitting coil. In the absence of a dedicated neonatal head coil an adult knee coil may be used. Phased-array coils may provide improved benefit in terms of SNR although lack of homogeneity may be an issue. It is important to place the neonatal head in the centre of the coil to avoid image signal inhomogeneity. Sick infants may be ventilator-dependent in the first few days after delivery and it may therefore be necessary to perform the imaging examination using MR-compatible ventilator equipment. In the absence of this, a neonate can be safely hand bagged during a short MR examination if monitoring is in place. In ventilated neonates a larger adult head coil may be necessary to accommodate the endotracheal tube. Imaging of the neonatal brain may be performed on a 1-, 1.5- or 3-T magnet system but MR sequences need to be adapted for the neonatal brain with its higher water content. An ideal neonatal protocol is listed in Table 2 and sequence parameters are suggested in Table 3. The single most common reason for a failed or poor quality image examination is neonatal motion. If the infant is unsettled it may be necessary to repeat the sequence acquisition or to image with a different approach, whilst this may sacrifice some SNR it will allow interpretable images to be acquired. T1- and T2-weighted imaging sequences from fast brain protocols may be used or motion “correction” sequences e.g., PROPELLER, BLADE. Alternatively a single-shot technique such as the snapshot volume reconstruction (SVR) that has been developed for imaging the mobile fetal brain can provide images of sufficient quality to exclude major pathologies [3].
Diffusion-weighted imaging (DWI) should include as many directions as possible and the generation of at least an apparent diffusion coefficient (ADC) map, which should be possible on most MR consoles. Whilst there have been one or two reports of the useful role of proton-density sequences in early imaging following perinatal brain injury we have acquired it as part of a dual-echo sequence for years and not found it to add information that was not evident on the T1- and T2-W sequences [4]. Similarly a fluid-attenuated inversion recovery (FLAIR) sequence may accentuate white matter (WM) injury in the neonatal brain but seldom adds new information [5]. In suspected infection the administration of contrast may be considered, the concern being that conventional imaging without contrast enhancement may be negative if performed early in the disease [6]. The majority of neonates with suspected bacterial or viral infection are treated empirically and imaging may not be requested until after the acute illness. At that stage any injury should be established and a negative imaging examination should be truly negative and the administration of contrast would be unnecessary. With the recent discovery of the association between gadolinium-based contrast agents and nephrogenic systemic fibrosis (NSF), the possible advantages of giving contrast at any stage need to be weighed up against the small risk of precipitating NSF [7]. The youngest cases reported to date have been eight years old and all the cases reported have had chronic renal failure. Neonates have an immature renal system and following a hypoxic-ischaemic event often demonstrate both transient renal and hepatic dysfunction placing them at theoretically increased risk of NSF. Where contrast administration is considered the responsible clinician will need to check renal function and for the presence of metabolic acidosis prior to administration. It would seem prudent to keep to the minimum dose, to restrict the use to those neonates with a high suspicion of intracranial infection or those not responding to treatment and to consider contrast agents with a cyclic structure not associated, as yet, with NSF.
The role of susceptibility-weighted imaging (SWI) of the brain has been highlighted recently in the investigation of potential thrombosis [8]. We would recommend a standard MR sinus venogram in every neonate with suspected perinatal brain injury. Many of these neonates have subdural haemorrhage and it can be difficult to differentiate from sinus thrombosis. In term born neonates with haemorrhagic lesions the underlying diagnosis may well be thrombosis of smaller draining veins [9, 10]. In later stage sinus thrombosis the diagnosis may be confirmed by the presence of characteristic cortical and subcortical haemorrhagic lesions. Of interest is the potential role of SWI in detecting calcium [8, 11]. Whilst this is relatively easy with cranial US, MR underdiagnoses the presence of calcium when compared to CT. The presence of calcium may be sought in suspected congenital infection e.g., cytomegalovirus, but many acquired hypoxic-ischaemic lesions may be calcified at some stage during their evolution. As for many sequences in neonatal imaging, optimisation is required because of the small size and altered tissue properties of the immature brain.
MR angiography can be used without contrast medium and allows depiction of the proximal cerebral arteries in the neonatal brain. Imaging at 3T improves SNR and therefore the quality of the images (Fig. 1) [12]. Angiography allows the variations of the circle of Willis to be documented, although it remains unclear how these might relate to lesion acquisition and lesion evolution. In neonates with middle cerebral artery infarction there may be transient occlusion of the artery, but in later imaging the vessels on the affected side are often involved within the infarct itself and so an asymmetry in the appearance of the vessels is a common finding.
MR spectroscopy has progressed little in the last ten years compared to imaging and its role in the clinical examination is limited. Following a hypoxic-ischaemic brain insult proton spectroscopy may demonstrate an increased lactate peak; this is useful to confirm the presence of a significant tissue injury, particularly in early examinations before conventional imaging becomes overtly abnormal. Proton spectroscopy has also been used to predict outcome in the research environment and has a good sensitivity and specificity for poor outcomes in neonates with hypoxic-ischaemic encephalopathy (HIE) [13]. However the technique requires the spectra to be processed and the metabolite ratios quantified, which may not be practical in clinical practice. In addition, as will be noted later in this review, important information for determining the nature of the neurodevelopmental impairment is obtained from the site and pattern of the lesions sustained. Occasionally proton spectroscopy may identify the presence of an abnormal peak in a neonate with a metabolic disorder, masquerading as HIE encephalopathy (HIE). Most commonly this is glycine, demonstrated at short echo times, in an infant with non-ketotic hyperglycinaemia [14].
There are few, if any, clinical indications for using perfusion-weighted imaging in suspected perinatal brain injury. There are several reports in the literature using contrast-enhanced methods and one or two exploiting 3T systems to perform arterial spin labelling (ASL) [15, 16]. Both techniques allow the relative differences in perfusion between white and gray matter in both apparently normal and abnormal tissue to be assessed. Absolute quantification is always a problem, where factors such as levels of haemoglobin, blood viscosity and blood carbon dioxide levels need to be addressed prior to quantification. These techniques remain very much in the research arena and may have a role in interventional studies although current concerns over the use of contrast agents favours developing the ASL approach at higher field strength.
When time is limited, because of heavy demand on the scanner or concerns about the neonate waking up, it is important to prioritise sequences for each neonate depending on the clinical history and the assumed underlying diagnosis.
Image interpretation should be performed by an experienced person, with knowledge of the normal appearances of the term neonatal brain (Fig. 1), although published data with normative values is not prolific, largely because of ethical restrictions on imaging normal term neonates in some centres [17, 18]. The accurate interpretation of image findings also requires information on the antenatal, perinatal and neonatal details, a sound knowledge of the range of lesions seen in perinatal brain injury and the effect of timing from injury on their appearance.
Clinical history
The pattern of injury sustained by a neonate is influenced by the nature of the insult and to some extent the gestational age. The pattern of injury may be predicted by the clinical history and the clinical presentation [21-31]. Term born neonates who sustain perinatal brain injury may present with an encephalopathy or with isolated seizures. HIE describes those neonates who become encephalopathic following an apparent hypoxic-ischaemic event that may be evidenced by signs of fetal distress on cardiotocograph (CTG), a low cord pH and a necessity for resuscitation. In a minority of neonates with HIE there will be a well documented sentinel event, such as a uterine rupture or cord prolapse, but in the majority fetal distress develops without obvious cause. Those neonates with a history of a sentinel event are likely to sustain basal ganglia and thalamic (BGT) lesions (Fig. 2) [28]. These are usually accompanied by abnormalities in the appearance of the intervening posterior limb of the internal capsule (PLIC). In severe cases there may also be brainstem involvement (Figs. 2 and 3). This pattern is seen irrespective of gestational age [29]. Term born neonates with this imaging signature of an acute hypoxic-ischaemic insult may also develop abnormal signal intensity (SI) in specific cortical regions, usually around the central sulcus, the interhemispheric fissure and the insula (Figs. 4 and 5). These cortical changes are most obvious on T1-W images and are usually accompanied by abnormal intensity in the adjacent subcortical WM. Both the cortical and subcortical changes may not be obvious until the end of the first week following birth (Fig. 5). It is unusual to identify isolated cortical lesions in the absence of lesions in the BGT or WM. The presence of isolated cortical lesions should raise suspicion of a sinus thrombosis.
In those neonates who fulfil the clinical criteria for HIE, but without a history of a sentinel event, BGT lesions are still the most prevalent, but these are accompanied by more widespread WM injury in approximately a half of cases (Fig. 6).
Isolated WM and cortical injury with normal BGT is a rare pattern of injury in neonates with HIE and is more likely to be detected in neonates with an atypical clinical course and without the need for major resuscitation following delivery. Neonates with HIE following a history of decreased fetal movements tend to sustain WM injury either in isolation or in combination with BGT lesions (Fig. 7) [30]. Many infants with cortical and WM injury present with isolated seizures rather than HIE. Injury may take the form of a focal WM infarct or “stroke” or bilateral parasagittal infarction (Figs. 7 and 8) [21]. Focal infarction is more likely to occur in infants born to primigravida women; delivery may be difficult and some resuscitation may be required but the infant usually recovers sufficiently and does not require admission to neonatal intensive care. Within this group with mainly WM and cortical injury there are also more likely to be neonates with alternative or additional pathologies e.g., cerebral malformations, infection, metabolic disorders, hypoglycaemia (Fig. 9) [23, 31].
The clinical presentation therefore serves as a guide to the lesions that will have been sustained and it is important to obtain all relevant details prior to image interpretation. This should include gestational age, antenatal history, type of delivery, Apgar scores, resuscitation, neonatal course, family history.
Timing of imaging examination
Perinatal brain lesions are at their most visually obvious on conventional MR sequences between 1 and 2 weeks from birth. This is therefore the most useful time to image to assess the extent of injury. In addition, neonates are more likely to be clinically stable and off assisted ventilation by this time and in those infants receiving treatment with hypothermia (that is continued for 72 h) this will have been completed. Earlier imaging, within the first few days, may be required to make a diagnosis or to assist clinical management. Imaging within the first couple of days may demonstrate only subtle abnormalities on conventional sequences in the presence of significant brain injury (Figs. 7 and 10). Early image examinations should always therefore include a diffusion-weighted sequence. DWI, which is available on most modern scanners, should identify ischaemic WM (Figs. 7 and 8) but is not so reliable at detecting significant injury to the BGT, where it may underestimate the extent of the lesions (Fig. 11).
The visual appearances of infarcted tissue on DWI are obvious very early and last for 7–14 days (Fig. 12) [17]. By this time conventional imaging should be overtly abnormal. Occasionally visual analysis of the DWI is unremarkable or difficult to interpret particularly in the presence of widespread injury where there is no normal tissue for comparison. Most scanners have the software necessary to obtain ADC values from the trace diffusion images. Several studies have reported their own centres values for ADC within the normal term brain (Table 4) [17, 18]. It is recommended that ADC values are measured in all infants even when the DWI appears unremarkable (Fig. 13). During the first week from injury ADC values are usually decreased in the presence of ischaemic WM and then show pseudonormalisation. In clinically significant BGT lesions however, ADC values may be reduced, normal or elevated (Fig. 11) [17]. In addition, abnormalities consistent with restricted diffusion appear to involve different regions within the BGT over time (Fig. 12) [19]. Any single diffusion examination is therefore likely to underestimate the total lesion load in the BGT. We take a value of <1 × 10-3/mm2/s in term WM to be consistent with impending infarction. The range of abnormal values for ADC within the BGT is much narrower and therefore it is less easy to predict irreversible infarction but levels <0.7 × 10-3/mm2/s are associated with significant lesions (Table 4). In addition, because values within the central grey matter can be both abnormally low and abnormally high in adjacent tissue, an ADC value may be measured as within the normal range because of partial volume effects. Measures of fractional anisotropy (FA) may be useful in detecting persistent abnormalities within tissue after the first week from injury. FA values have been shown to continue to decrease in abnormal tissue that shows pseudonormaliation of the ADC value [20].
Serial imaging provides information about the secondary effects of an acute hypoxic-ischaemic insult. In neonates with BGT lesions the WM may show initially normal ADC values that increase over time as part of a secondary WM injury. The WM then shows signs of atrophy on subsequent scans and the infant shows poor head growth (Fig. 11).
There has been speculation that treatment with hypothermia will delay the appearance of abnormalities on MRI. As the evolution and pattern of lesions may be very variable between individual infants, any potential effect on the evolution of abnormalities would need to be studied in large numbers of infants. It is possible that with the routine use of hypothermia as a treatment strategy in clinical practice we will be in a better position to assess the effect of hypothermia on imaging appearances. With or without hypothermia the second week from delivery is the ideal time to detect the extent of acquired lesions on conventional imaging. In the recently published TOBY trial the presence of hypothermia did not influence the ability of a neonatal MRI (median 8 days) to predict neurodevelopmental outcome [1].
There is a predictable evolution of perinatal brain lesions and so serial imaging may allow the timing of injury to be assessed. Repeat imaging may also be useful to identify atypical evolution of imaging abnormalities, particularly when an additional or different diagnosis such as a metabolic disorder is suspected. In some metabolic disorders there are additional congenital malformations of the brain such as agenesis of the corpus callosum in non-ketotic hyperglycinaemia. A normal scan or an isolated delay in myelination in a neonate with persisting seizures should raise the possibility of a metabolic disorder (Fig. 9).
Patterns of injury and prediction of outcome
MR imaging provides detailed information about the pattern of lesions following perinatal brain injury and conventional sequences may provide excellent predictions of outcome.
In neonates with a global hypoxic-ischaemic insult, lesions are usually detected within the BGT with abnormal SI in the intervening posterior limb of the internal capsule (PLIC) (Fig. 3). Abnormal SI within the PLIC is an excellent predictor of abnormal motor outcome in term infants with HIE [32]. BGT lesions give rise to motor impairment in the form of cerebral palsy. The severity of the BGT lesions dictates the severity and nature of the cerebral palsy (Fig. 2) [28]. Lesions within the brainstem are usually found in those neonates with the most severe form of BGT injury and are often associated with early death (Fig. 14) [33]. In those neonates that survive, persisting feeding difficulties and a necessity for gastrostomy are common. In the presence of severe BGT injury there are usually secondary WM changes with subsequent atrophy, poor head growth with the development of a secondary microcephaly and cognitive impairment [34]. In approximately 50% of neonates with BGT lesions there will be more extensive WM abnormalities seen on the initial scan consistent with a primary injury (Figs. 6 and 13) [23, 24]. The motor outcome for these children is still dictated by the BGT lesions but the more severe WM involvement may exacerbate any cognitive deficit and results in very poor head growth [34].
In some neonates who present with what is thought to be HIE there is no BGT involvement but only WM lesions (Fig. 7). These may be haemorrhagic (Fig. 15). These lesions give rise to tissue atrophy, poor head growth and the development of a secondary microcephaly and are associated with later cognitive impairment. The more severe the WM lesions the worse the cognitive outcome and the greater the likelihood of some motor impairment [35, 36]. In neonates who have a term perinatally acquired WM injury, follow-up imaging may be indistinguishable from what would normally be called periventricular leucomalacia (PVL), an injury thought specific to the more preterm brain (Fig. 15). Without knowledge of the clinical neonatal history, the injury may therefore be wrongly attributed to antenatal damage at an earlier gestation. However in an infant with signs of PVL, on imaging and/or clinically, who was born at term it is probably only reasonable to implicate perinatal events if there were neonatal symptoms such as encephalopathy or seizures.
Focal infarction usually involves the territory of the middle cerebral artery, the left side being more commonly affected than the right (Fig. 16). The outcome following an MCA infarct depends on the extent and sites involved. If there are abnormalities within the parenchyma, BGT and PLIC then the child is likely to develop a later hemiplegia (Fig. 16) [37].
The relationship between ADC values and outcome is complicated. Detection of WM infarction may not relate closely to later outcome as it depends how this is defined and when it is measured [38]. Abnormally low ADC values within the PLIC or the BGT are more likely to predict the development of a significant motor impairment that will be detectable within the first two years [13, 38, 39]. Neonates with reduced ADC values within the WM consistent with acute infarction may have a good short term outcome as in the presence of normal BGT and a normal PLIC, motor development may be normal. Poor head growth should raise suspicion and these children may present in early childhood with neurocognitive and behavioural impairments.
Post mortem imaging
Infants with severe encephalopathy should always have brain imaging. This may be difficult and it may not be possible to perform before an infant dies. In such circumstances and particularly if no autopsy is performed post mortem MR imaging should be considered. The usual neonatal protocol can be followed; increasing the signal averages will improve the SNR. T2-W images are usually of better quality than T1-W images. Post mortem imaging may allow confirmation of abnormalities consistent with a hypoxic-ischaemic insult or suggest an alternative diagnosis (Fig. 17) [40]. This would clearly have important genetic counselling and medicolegal considerations.
References
Rutherford M, Ramenghi LA, Edwards AD et al (2010) Assessment of brain tissue injury after moderate hypothermia in neonates with hypoxic-ischaemic encephalopathy: a nested substudy of a randomised controlled trial. Lancet Neurol 9:39–45
van Wezel-Meijler G, Leijser LM, de Bruïne FT et al (2009) Magnetic resonance imaging of the brain in newborn infants: practical aspects. Early Hum Dev 85:85–92
Jiang S, Xue H, Glover A et al (2007) MRI of moving subjects using multislice snapshot images with volume reconstruction (SVR): application to fetal, neonatal, and adult brain studies. IEEE Trans Med Imaging 26:967–980
Barkovich AJ, Hajnal BL, Vigneron D et al (1998) Prediction of neuromotor outcome in perinatal asphyxia: evaluation of MR scoring systems. AJNR 19:143–149
Liauw L, van der Grond J, van den Berg-Huysmans AA et al (2008) Hypoxic-ischemic encephalopathy: diagnostic value of conventional MR imaging pulse sequences in term-born neonates. Radiology 247:204–212
Coren ME, Buchdahl RM, Cowan FM et al (1999) Imaging and laboratory investigation in herpes simplex encephalitis. J Neurol Neurosurg Psychiatry 67:243–245
Mendichovszky IA, Marks SD, Simcock CM et al (2008) Gadolinium and nephrogenic systemic fibrosis: time to tighten practice. Pediatr Radiol 38:489–496
Tong KA, Ashwal S, Obenaus A et al (2008) Susceptibility-weighted MR imaging: a review of clinical applications in children. AJNR 29:9–17
Lequin MH, Dudink J, Tong KA et al (2009) Magnetic resonance imaging in neonatal stroke. Semin Fetal Neonatal Med 14:299–310
Ramenghi LA, Govaert P, Fumagalli M et al (2009) Neonatal cerebral sinovenous thrombosis. Semin Fetal Neonatal Med 14:278–283
Wu Z, Mittal S, Kish K et al (2009) Identification of calcification with MRI using susceptibility-weighted imaging: a case study. J Magn Reson Imaging 29:177–182
Malamateniou C, Adams ME, Srinivasan L (2009) The anatomic variations of the circle of Willis in preterm-at-term and term-born infants: an MR angiography study at 3T. AJNR 30:1955–1962
Thayyil S, Chandrasekaran M, Taylor A et al (2010) Cerebral magnetic resonance biomarkers in neonatal encephalopathy: a meta-analysis. Pediatrics 125:e382–395
Shah DK, Tingay DG, Fink AM et al (2005) Magnetic resonance imaging in neonatal nonketotic hyperglycinemia. Pediatr Neurol 33:50–52
Laswad T, Wintermark P, Alamo L et al (2009) Method for performing cerebral perfusion-weighted MRI in neonates. Pediatr Radiol 39:260–264
Miranda MJ, Olofsson K, Sidaros K (2006) Noninvasive measurements of regional cerebral perfusion in preterm and term neonates by magnetic resonance arterial spin labeling. Pediatr Res 60:359–363
Rutherford M, Counsell S, Allsop J et al (2004) Diffusion weighted MR imaging in term perinatal brain injury: a comparison with site of lesion and time from birth. Pediatrics 114:1004–1014
Bartha AI, Yap KR, Miller SP et al (2007) The normal neonatal brain: MR imaging, diffusion tensor imaging, and 3D MR spectroscopy in healthy term neonates. AJNR 28:1015–1021
Barkovich AJ, Miller SP, Bartha A et al (2006) MR imaging, MR spectroscopy, and diffusion tensor imaging of sequential studies in neonates with encephalopathy. AJNR 27:533–547
Ward P, Counsell S, Allsop J et al (2006) Reduced fractional anisotropy on diffusion tensor magnetic resonance imaging after hypoxic-ischemic encephalopathy. Pediatrics 117:e619–30
Mercuri E, Cowan F, Rutherford M et al (1995) Ischaemic and haemorrhagic brain lesions in newborns with seizures and normal Apgar scores. Arch Dis Child Fetal Neonatal Ed 73:F67–74
Mercuri E, Rutherford M, Barnett A et al (2002) MRI lesions and infants with neonatal encephalopathy. Is the Apgar score predictive? Neuropediatrics 33:150–156
Cowan F, Rutherford M, Groenendaal F et al (2003) Origin and timing of brain lesions in term infants with neonatal encephalopathy. Lancet 361:713–714
Miller SP, Ramaswamy V, Michelson D et al (2005) Patterns of brain injury in term neonatal encephalopathy. J Pediatr 146:453–460
de Vries LS, Cowan FM (2009) Evolving understanding of hypoxic-ischemic encephalopathy in the term infant. Semin Pediatr Neurol 16:216–225
Foran A, Cinnante C, Groves AM et al (2009) Patterns of brain injury and outcome in term neonates presenting with postnatal collapse. Arch Dis Child Fetal Neonatal Ed 94:F168–177
Cheong JL, Cowan FM (2009) Neonatal arterial ischaemic stroke: obstetric issues. Semin Fetal Neonatal Med 14:267–271
Okereafor A, Allsop J, Counsell SJ et al (2008) Patterns of brain injury in neonates exposed to perinatal sentinel events. Pediatrics 121:906–914
Logitharajah P, Rutherford M, Cowan F (2009) Hypoxic-ischaemic encephalopathy in the preterm infant: antecedent factors, brain imaging and outcome. Pediatr Res 66:222–229
Sirgiovanni I, Patel A, Allsop J et al (2010) Patterns of brain injury in neonates presenting with encephalopathy following a history of decreased fetal movements. Abstract 753398, PAS Vancouver May 1–4
Burns CM, Rutherford MA, Boardman JP et al (2008) Patterns of cerebral injury and neurodevelopmental outcomes after symptomatic neonatal hypoglycemia. Pediatrics 122:65–74
Rutherford MA, Pennock JM, Counsell SJ et al (1998) Abnormal magnetic resonance signal in the internal capsule predicts poor neurodevelopmental outcome in infants with hypoxic-ischemic encephalopathy. Pediatrics 102:323–328
Martinez Biarge M, Diez-Sebastian J, Kapellou O et al (2010) Predicting motor outcome and death in infants with central gray matter damage following hypoxic-ischaemic encephalopathy (HIE). Abstract 754500 PAS Vancouver May 1–4
Mercuri E, Ricci D, Cowan F et al (2000) Head growth in infants with hypoxic-ischaemic encephalopathy: Correlation with neonatal magnetic resonance imaging. Pediatrics 106:235–243
Cowan F, Dubowitz L, Mercuri E et al (2003) White matter injury can lead to cognitive without major motor deficits following perinatal asphyxia and early encephalopathy. Dev Med Child Neurol Suppl 93(45):14
Steinman KJ, Gorno-Tempini ML, Glidden DV et al (2009) Neonatal watershed brain injury on magnetic resonance imaging correlates with verbal IQ at 4 years. Pediatrics 123:1025–1030, Erratum in: Pediatrics 123:1611
Mercuri E, Rutherford M, Cowan F et al (1999) Early prognostic indicators of outcome in infants with neonatal cerebral infarction: a clinical electroencephalogram and magnetic resonance imaging study. Paediatrics 103:39–46
Liauw L, van Wezel-Meijler G, Veen S et al (2009) Do apparent diffusion coefficient measurements predict outcome in children with neonatal hypoxic-ischemic encephalopathy? AJNR 30:264–270
Vermeulen RJ, van Schie PE, Hendrikx L et al (2008) Diffusion-weighted and conventional MR imaging in neonatal hypoxic ischemia: two-year follow-up study. Radiology 249:631–639
Nicholl RM, Balasubramaniam VP, Urquhart DS et al (2007) Postmortem brain MRI with selective tissue biopsy as an adjunct to autopsy following neonatal encephalopathy. Eur J Paediatr Neurol 11:167–174
Roelants-van Rijn AM, Nikkels PG, Groenendaal F et al (2001) Neonatal diffusion weighted imaging: relation with histopathology and follow-up MR examination. Neuropediatrics 32:286–294
Boardman JP, Counsell SJ, Rueckert D et al (2007) Early growth in brain volume is preserved in majority of preterm infants. Ann Neurol 62:185–182
Acknowledgements
We would like to thank all the staff of the Robert Steiner MR Unit, Hammersmith Hospital and the neonatal units of Hammersmith and Queen Charlottes Hospital. We are also grateful to The Medical Research Council, Philips Medical Systems and the Biomedical Research Centre for their support.