Antemortem cranial MRI compared with postmortemhistopathologic examination of the brain in terminfants with neonatal encephalopathy followingperinatal asphyxiaThomas Alderliesten,1 Peter G J Nikkels,2 Manon J N L Benders,1 Linda S de Vries,1

Floris Groenendaal1

▸ Additional supplementaryfiles are published online only.To view these files please visitthe journal online (http://dx.doi.org/10.1136/archdischild-2012-301768).1Department of Neonatology,Wilhelmina Children’s Hospital/University Medical CentreUtrecht, Utrecht, TheNetherlands2Department of Pathology,University Medical CentreUtrecht, Utrecht, TheNetherlands

Correspondence toDr Floris Groenendaal,Department of Neonatology,Room KE 04.123.1,Wilhelmina Children’s Hospital,University Medical CentreUtrecht, Lundlaan 6,Utrecht 3584 EA,The Netherlands;F.Groenendaal@umcutrecht.nl

Accepted 15 October 2012

ABSTRACTAim To compare antemortem cranial MRI withpostmortem histopathological examination of the brain infull-term infants with neonatal encephalopathy followingperinatal asphyxia.Patients and methods In this retrospectiveobservational cohort study, 23 infants with neonatalencephalopathy who subsequently died, were analysed.Infants underwent antemortem cranial MRI andpostmortem histopathological examination of the brain. MRIincluded T1, T2 and diffusion-weighted sequences.Histopathology included staining with H&E, and monoclonalantibodies to CD68 and HLA-DR. Histological abnormalitieswere compared with MRI in 10 different brain regions.Results All neonates underwent cranial MRI within 7 daysafter birth (median day 3, IQR 2–4 days). Infants died onmedian day 4 (IQR 2–5 days). Histopathology demonstratedsignificantly (p=0.0016) more abnormal regions (median 10,IQR 7–10) per patient than did MRI (median 8, IQR 5–9). Thenumber of cases with abnormalities in the thalamus, basalganglia, posterior limb of the internal capsule (PLIC), cerebralcortex and cerebellum were not significantly differentbetween MRI and histopathology. By contrast, thehippocampus (70% vs 96%, p=0.047), cerebral whitematter (anterior 65% vs 96%, p=0.022, posterior 61% vs91%, p=0.035) and brainstem (57% vs 96%, p=0.004)were confirmed to be affected more often onhistopathological examination than with MRI.Conclusions Whereas early postnatal MR imaging isexcellent in detecting injury to the basal ganglia andthalamus, PLIC, cortex and cerebellum, it may underestimateinjury to the hippocampus, cerebral white matter, and thebrainstem in term infants with neonatal encephalopathyfollowing perinatal asphyxia.

INTRODUCTIONNeonatal encephalopathy following perinatalasphyxia is one of the leading causes of neonataldeath worldwide.1 MR imaging is considered thestandard technique to visualise various types ofbrain injury in neonates.2–5 It is used on a daily basisfor diagnostic purposes, to aid prediction of neurode-velopmental outcome and to assist in clinical deci-sion making.6 7 Because clinical practice is relyingmore and more on results of MR imaging, it isimportant to know how well brain injury seen onMRI reflects actual brain damage seen on post-mortem examination. There is, however, little avail-able evidence that describes how well brain injury

seen on MRI reflects brain damage seen on post-mortem histopathological examination in full-terminfants with perinatal asphyxia.8 9 The aim of thisstudy is to describe how MR images obtained duringlife correspond with postmortem histopathologicalexamination in term infants with neonatal encephal-opathy following perinatal asphyxia.

METHODSThe ethical committee of the University MedicalCentre, Utrecht, The Netherlands, stated that theneed for informed consent was waived for thisstudy.

In this retrospective observational cohort study,infants with perinatal asphyxia, and a gestational

What is already known on this topic

▸ Neonatal encephalopathy following perinatalasphyxia remains one of the leading causes ofneonatal death worldwide.

▸ MRI is considered the standard technique tovisualise various types of brain injury inneonates.

▸ MRI is increasingly used on a daily basis fordiagnostic purposes, to aid prediction ofneuromotor outcome, and to assist in clinicaldecision making.

What this study adds

▸ Results of MRI and histopathologicexamination are comparable regarding damageto the thalamus, basal ganglia, PLIC, cerebralcortex and cerebellum in infants with perinatalasphyxia.

▸ MRI underestimates damage to the cerebralwhite matter, hippocampus and brainstem ininfants with neonatal encephalopathy followingperinatal asphyxia.

▸ The inferior olive in the medulla oblongata anddentate nucleus in the cerebellum are relativelyspared in infants with neonatal encephalopathyfollowing perinatal asphyxia.

Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768 F1

Original article ADC-FNN Online First, published on November 21, 2012 as 10.1136/archdischild-2012-301768

Copyright Article author (or their employer) 2012. Produced by BMJ Publishing Group Ltd (& RCPCH) under licence.

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

age ≥36 weeks were analysed. All infants were admitted within24 h after birth to the level III neonatal intensive care unit of theWilhelmina Children’s Hospital/University Medical CentreUtrecht, The Netherlands. Between December 1999 andDecember 2010, a total of 303 infants with perinatal asphyxiawere admitted. Perinatal asphyxia was diagnosed when neonateshad at least three of the following criteria, as described previ-ously: (1) fetal heart rate abnormalities, (2) meconium stainingof amniotic fluid, (3) delayed onset of respiration, (4) arterialcord or early postnatal blood pH <7.1, (5 Apgar score at 5 min<7) or (6) multiorgan failure.7 Furthermore, all had signs of neo-natal encephalopathy.10

Neonates were eligible for the study when they had under-gone a cranial MRI during the first days of life and underwentpostmortem histopathological examination. Neonates, whodied and had major congenital abnormalities, a metabolic dis-order or chromosomal abnormality, diagnosed during admissionwere excluded from the analysis (n=15), leaving 52 neonateswho died following perinatal asphyxia. Of these patients, sevenwere excluded because either no MRI was performed, as theirclinical condition did not allow transport to the MRI depart-ment (n=4), a postmortem MRI was performed (n=1), MRIwas performed >7 days after birth (n=1), or the patient died>40 days after birth (n=1). In 22 patients, a postmortemexamination was not performed because parental consent wasnot obtained, in all patients, informed consent for postmortemexamination was asked. There were no differences regardingethnicity and/or religious background between parents whogave and parents who did not give consent.

A total of 23 neonates were eligible for this study, 11 of whomhave been described in previous studies.7 9 11 Parental consentwas obtained in all for both MR imaging and postmortem histo-pathologic examination. In seven neonates admitted afterJanuary 2008, therapeutic hypothermia was used for 72 h. Allpatients died because of irreversible brain injury. Decisions towithdraw treatment/redirect care, were always based on thecombination of abnormal clinical examination, abnormal cere-bral MRI, including diffusion-weighted imaging (DWI), presenceof lactate in the basal ganglia, as determined by proton magneticresonance spectroscopy, abnormal cranial ultrasound and abnor-mal electrophysiology, including continuousamplitude-integrated EEG and EEG, as described previously.12

MR imagingStandard MRI was performed using 1.5 or 3.0 Tesla units(Gyroscan ACS-NT or Achieva, Philips Medical Systems, Best,The Netherlands). A vacuum pillow (MedVac-Kohlbrat & BunzGmbH, Radstadt, Austria) was used to prevent patient move-ment during MR examinations. Ear protection was providedusing neonatal ear muffs (MiniMuffs; Natus Medical, SanCarlos, California, USA), heart rate and transcutaneous oxygensaturation were monitored using a pulse oxymeter (NoninMedical Inc, Plymouth, Minnesota, USA), and respiratory rateby using an abdominal transducer (Phillips Medical Systems,Best, The Netherlands). All patients received mechanical venti-lation and were sedated with morphine 10 μg/kg intravenously.A neonatologist was present throughout the procedure. MRIincluded T1-weighted sagittal sequences, and axial T1-,T2-weighted sequences, axial inversion recovery (IR) and axialdiffusion-weighted sequences (DWI). Axial IR was performedin 1.5 Tesla units, and axial T1 in 3.0 Tesla MR units. Slicethickness was 2 mm in cases examined in a 1.5 T environmentand 1.2 mm in cases examined in a 3.0 T environment. Aftercompletion of imaging, an apparent diffusion coefficient

(ADC) map was generated based on the DWI sequence, asreported previously.7

Injury, as seen on MRI and DWI, was scored without priorknowledge of histopathological findings (authors FG and LSdV,each with >20 years of experience in neonatal MRI reading),using a previously described scoring system.13 This scoringsystem includes the basal ganglia and thalamus (BGT), water-shed areas (WS) and the posterior limb of the internal capsule(PLIC).6 14 In addition, the brainstem, hippocampus, the cerebel-lum, anterior and posterior white matter and cortex were scoredas either normal or abnormal, as scores per region could not bedistilled from the scoring system. A total MRI brain injury scorewas constructed from the BGT, WS and PLIC scores combinedwith brainstem, hippocampus and the cerebellum abnormalities.A maximum of 14 points could be obtained, with higher scoresindicating more severe brain injury. The scoring system is pro-vided as an online supplementary table E1. Representativeimages on which the hippocampus, cerebellum and brainstemwere scored as either normal or abnormal are provided as onlinefigures (see online supplementary figures E1 and E2).

HistopathologyAt autopsy, brains were removed and fixed in a 4% buffered for-malin solution for 4–6 weeks before representative samples ofvarious brain regions were obtained. These samples includedthe thalamus, basal ganglia, the posterior limb of internalcapsule, anterior and posterior white matter, anterior and pos-terior cortex, brain stem, hippocampus and the cerebellum.Histological sections, 5–6 mm thick, were cut and mounted oncoated slides. Tissues were stained with H&E. Monoclonalmouse antibodies against CD68 (Novo Castra, Newcastle, UK)and HLA-DR (Dako, Glostrup, Denmark) were used to deter-mine the presence of activated microglia and macrophages.CD68 staining is seen 1–3 days after the hypoxic-ischaemicevent, while HLA-DR staining is seen from the fourth dayonwards.11 15 16 The slides were scored by two authors (PGJNand TA), without prior knowledge of MRI findings (PGJN is aneonatal/paediatric pathologist with 20 years of experience inneonatal neuropathology). Final scores were based on CD68and HLA-DR staining, and made by consensus agreement.A score of 0–3 points was given per-brain region for both theCD68 and the HLA-DR slides: 0 points when there was nostaining, 1 point when high magnification (total magnification200×) was necessary to visualise staining, 2 points when mod-erate magnification (100×) was necessary, and three pointswhen staining was visible on low magnification (25×).11 H&Estaining was not used for scoring of injury, since it mightdetect injury sustained very recently before death, which wasnot yet present at the time of the MRI.

Figure 1 shows an example of HLA-DR staining (score 3) ofthe hippocampus and basal ganglia. The brainstem and cerebel-lum were assessed in subregions as part of routine histopatho-logical workup; four subregions for the brainstem (area of thecranial nerves in the medulla and pons, the inferior olives andthe ventral pons) and two for the cerebellum 2 (cerebellarcortex/white matter and dentate nucleus). For each region, thehighest score, either CD68 or HLA-DR, was used to representthe amount of brain damage in that region. In brain regionswith more than one slide or subregion available (cerebralcortex/white matter, brainstem and cerebellum), the slide orsubregion with the highest score was used to represent thewhole region. This enabled comparison of 10 brain regionsbetween MRI and histopathology per patient: thalamus, basal

F2 Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768

Original article

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

ganglia, PLIC, anterior and posterior white matter, anterior andposterior cortex, brain stem, hippocampus and the cerebellum.

Statistical analysisStatistical analysis was performed using PASW Statistics forWindows 17.0 (SPSS Inc, Chicago, Illinois, USA). Differencesbetween MRI and histopathology were assessed using theχ2 test and Mann–Whitney U test, where appropriate. For ana-lysis, histopathology scores (0–3 points) were dichotomised:scores 0 and 1 were graded as ‘normal-mildly abnormal’ andscores 2 and 3 were graded as ‘moderate-severely abnormal’. Forthe PLIC score in MRI (0–2 points), both impaired (1) andabsent (2) myelination were scored as abnormal.

RESULTSClinical data is presented in table 1. One patient suffered fromsevere blood loss (Hb 5.6 mmol/l) and hypoglycaemia(0.7 mmol/l), in addition to perinatal asphyxia, and one patientsuffered from severe pneumonia. Infants died a median of 1 day(IQR 0–1) after the MRI.

MR imagingNeonates were examined at a median postnatal age of 3 days(IQR 2–4 days). The median imaging score was 10 (IQR 6–12).Postnatal age at scan and imaging score were no different

between hypothermic and normothermic patients (p=0.192and p=0.498, respectively).

HistopathologyAll but one patient had representative blocks available from allbrain regions; one patient had no hippocampal specimens avail-able. The scores per brain region (maximum value, either CD68or HLA-DR) were not significantly different between hypother-mic and normothermic patients, except for the anterior whitematter (hypothermia median score 3, range 2–3 vs normother-mia median score 2, range 1–3, p=0.038). Results per brainregion, presented in figure 2. Figure 3, demonstrate the percent-age of histopathologic abnormalities in brainstem and cerebel-lum subregions. The patient who also had severe blood lossand hypoglycaemia, in addition to perinatal asphyxia, sustainedsevere damage to both the dentate nucleus and the inferiorolives. The patient with severe pneumonia died 2 days (50 h)after birth, and demonstrated abnormalities in only three outof 10 brain regions, according to CD68/HLA-DR staining.

Figure 1 Histopathology images of tissue stained with of HLA-DR antibodies. (A) hippocampus with HLA-DR-staining score 3 (total magnification25×), (B) basal ganglia with HLA-DR staining score 3 (total magnification 25×) and C) basal ganglia in more detail (total magnification 100×).

Figure 2 Percentage abnormalities per-brain region for MRI (left sideof image, blue) and histopathology (right side of image, pink).Percentages placed in designated brain regions: basal ganglia,thalamus, posterior limb of the internal capsule, cerebral cortex,cerebral white matter, hippocampus, brainstem and cerebellum. Forhistopathology CD68/HLA-DR staining scores 2 and 3(moderate-severely abnormal) were graded as abnormal.

Table 1 Patient characteristics

Gender (male/female) 14/9

Gestational age (weeks)* 40.3 (0.9)Mode of deliveryEmergency caesarean section 10 (43%)Normal vaginal delivery 11 (48%)Vacuum extraction 2 (9%)

Apgar 1 min† 2 (1– 4)Apgar 5 min† 3 (2–6)Cord pH* 6.97 (0.17)Birth weight (grams)* 3469 (424)SarnatModerate 13 (57%)Severe 10 (43%)

Hypothermia (yes/no) 7/16MRI (1.5 Tesla/3 Tesla) 19/4Age at MRI (days)† 3 (2–4)Age at death (days)† 4 (2–5)Days after MR until death (days)† 1 (0–1)

*Mean (SD). †Median (IQR).

Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768 F3

Original article

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

There were, however, severe abnormalities seen with H&Estaining in this patient. Neurones in all regions but the dentatenucleus showed severe cytotoxic oedema, apoptosis and necro-sis. Layer V of the cerebral cortex showed marked neuronal lossand white matter (cerebral, PLIC and cerebellum) showedoedema. The other seven patients who died within 2 days afterbirth had a minimum of seven out of 10 abnormal brainregions. In two cases, CD68/HLA-DR staining indicated a pre-natal onset of the hypoxic-ischaemic event (approximately 2–3 days before birth).

MRI versus histopathologyThe number of abnormal brain regions per patient was signifi-cantly lower with MRI, as compared with histopathology(MRI median score 8, IQR 5–9 vs histopathology median score10, IQR 7–10, p=0.016). Per-brain region, the number of caseswith abnormalities in the thalamus, basal ganglia, PLIC, cere-bral cortex and cerebellum was not significantly different

between MRI and histopathology. The hippocampus(p=0.047), anterior and posterior cerebral white matter(p=0.022 and p=0.035) and brainstem (p=0.004), were con-firmed to be more often affected on histopathology (figure 2).Figure 4 shows MRI images of an example of a discrepancybetween MR images and histopathology: there is no apparentdamage to the hippocampus on MRI, while histopathologydemonstrated severe staining with HLA-DR antibodies in allhippocampal regions (figure 1, panel A). In three cases, MRIshowed a higher total number of abnormalities per patientthan histopathology, in two there was one additional abnormalbrain region (PLIC) seen on MRI. In the infant with severepneumonia, MRI demonstrated an additional five abnormalbrain regions. In this case, MRI showed abnormalities in thethalamus, basal ganglia, PLIC, cerebral white matter and cortexand the hippocampus. CD68/HLA-DR staining only demon-strated moderate-severe staining in the PLIC, cerebellum andbrainstem.

DISCUSSIONIn the majority of cases, histopathological examination showedmore abnormal brain regions per patient (median 10) than MRimaging (median 8). Results, per-brain region, demonstrate thathistopathological examination shows more abnormalities in thehippocampus, the cerebral white matter and the brain stem, ascompared with MRI. To our knowledge, this is the largestpatient group, to date, in which MRI is compared with histo-pathological examination in term infants with neonatalencephalopathy following perinatal asphyxia.8 9

In a series presented by Jouvet et al,8 there were virtually nonormal brain regions seen on histopathological examination. Asa consequence, MRI had a low negative predictive value. Thepositive predictive value, however, was 100%, indicating thatwhen MRI marks a brain region as abnormal, one can be surethat histology is abnormal. DWI was not included in their evalu-ation, a technique that nowadays is regarded as part of standardMRI, and is invaluable in early assessment of brain damage instroke and perinatal asphyxia.7 The additional value of DWI inneonates with a variety of brain lesions, including five with neo-natal encephalopathy following perinatal asphyxia, was assessedby our group a decade ago. Abnormalities seen on DWI werealways correlated with abnormalities on histopathology; DWI,

Figure 4 MR images acquired on a 1.5 Tesla unit. (A) Axial T2-weighted image, and (B) axial diffusion-weighted image (calculated ADC map). MRimages show abnormal signal intensities in the brainstem (T2: marked anterior/posterior contrast; ADC: decreased signal intensity throughout thebrainstem) and cerebral hemispheric white matter (ADC). The hippocampus shows no apparent damage, while histopathological examination of thehippocampus in this patient demonstrated severe damage (figure 1).

Figure 3 Percentage of abnormalities seen on histopathologydisplayed per subregion for the brainstem and cerebellum: area of thecranial nerves in the medulla and pons, the inferior olives, the ventralpons, cerebellar cortex/white matter and dentate nucleus. CD68/HLA-DR staining scores 2 and 3 (moderate-severely abnormal) weregraded as abnormal.

F4 Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768

Original article

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

however, underestimated the extent of the damage.9 In thepresent study, DWI was not analysed separately. Results showthat even when including DWI, MRI still tends to underesti-mate the amount of brain damage demonstrated with histo-pathological examination in the hippocampus, cerebral whitematter and the brain stem.

Determination of the time of onset of brain injury remains dif-ficult, and it is known that the hypoxic-ischemic event does notalways coincide with birth.17 The presence of different stages ofactivated microglial cells/macrophages, as assessed with CD68/HLA-DR, can be used to roughly determine time since injury(mostly CD68: <3 days; mostly HLA-DR: >3 days).11 15 In ourseries, the hypoxic-ischaemic event commenced 2–3 days beforebirth in two cases, which correlated with patient history. In theinfant with a severe pneumonia, only three out of 10 brainregions were abnormal using dichotomised CD68/HLA-DRscores, the extent of damage seen with H&E staining, however,was substantial. This suggests that most of the brain damage wasdue to a hypoxic-ischaemic event shortly before death, probablycaused by the severe pneumonia. This led to more brain injurydemonstrated by MRI, compared with CD68/HLA-DR in severalregions in this neonate (also explaining the higher basal gangliapercentage for MRI in figure 2).

Even though MRI, rather than ultrasound as a neuroimagingtechnique, and a 3-Tesla magnet, rather than a 1.5-Tesla magnetis showing more detailed information, histopathology is, atpresent, still considered to be the gold standard. In experimen-tal research, histopathology is almost exclusively used for asses-sing brain injury following hypoxia-ischaemia.18 The highersensitivity of MRI to demonstrate injury to the PLIC is due todichotomisation of the scores. In all cases, both MRI and histo-pathology scores were 1, but these scores were subsequentlydichotomised as abnormal for MRI, but as normal-mild forhistopathology. In all other regions, MRI showed fewer abnor-malities than histopathology, we speculate that the use ofthinner slices and higher field strengths may address this issueappropriately, especially in the hippocampus and brain stem.

One neonate suffered from severe blood loss and hypogly-caemia in addition to perinatal asphyxia, and demonstratedsevere damage to the dentate nucleus and the inferior olives, acombination that was infrequently seen in the other patients.Although this is only a single observation, it suggests that bloodloss and hypoglycaemia, in addition to perinatal asphyxia, causeeven more brain damage than asphyxia alone.19 20

During the study period, parental consent for autopsy wasnot obtained in 22 infants with perinatal asphyxia, whichreflects declining autopsy rates worldwide.21 22 MRI wasrecently put forward by others as a minimally invasive substi-tute for conventional autopsy. Minimally invasive autopsiesuse whole-body imaging techniques (MRI or CT), or a combin-ation of imaging techniques (mostly MRI), withendoscopic-assisted internal examination for tissue samplingand organ dissection.23–25 Evidence to support the use of min-imally invasive autopsy instead of conventional autopsy in theneonatal population is currently lacking.25 Postmortem MRexamination was not the focus of our study. A study evaluatingthe accuracy of postmortem MRI in fetuses, infants and chil-dren is in progress.26

There are some limitations that need to be addressed. First ofall, parental consent for autopsy was not obtained in 22patients. As this was supposedly conducted at random, it has,arguably, not influenced our results. The inability to obtain anMRI in four severely ill cases could, however, have influencedthe results, as these four may represent the most severe end of

the spectrum. Furthermore, infants underwent MR examin-ation at different postnatal ages (median day 3, IQR 2–4) and,thus, at different times after the hypoxic-ischaemic event(birth). This could have influenced the results, since injury seenwith conventional MR becomes more apparent after the firstweek. In our study, DWI was included to optimise sensitivityof MRI during the first week after the hypoxic-ischaemicevent.7 The third limitation is that terminal events could haveinduced additional brain injury. This does indeed directly affectdamage seen with H&E staining, however, activation/influx ofmicroglial cells and macrophages takes time, hence,CD68 staining only comes up approximately 1 day after theevent.11 Moreover, time between actual withdrawal of intensivecare and death was always limited (hours at the most).Therefore, damage sustained the day before death should nothave influenced our histopathology results which are based onCD68 or HLA-DR staining. Finally, as our study populationonly consists of full-term neonates with neonatal encephalop-athy following asphyxia, results are difficult to generalise toneonates with other types of brain injury.

In conclusion, damage to the thalamus, basal ganglia, PLIC,cerebral cortex and the cerebellum can be reliably determinedby antemortem MRI, even when MRI was performed as earlyas at a median postnatal age of 3 days. MRI underestimateddamage to the hippocampus, brain stem and cerebral whitematter. Our results underscore the need for histological examin-ation to assess the full extent of hypoxic-ischaemic braininjury.

Contributors Guarantors of integrity of entire study, TA and FG; study concepts/studydesign, TA, PGJN and FG; data acquisition, all authors; data analysis/interpretation, allauthors; Manuscript drafting or manuscript revision for important intellectual content,all authors; Manuscript final version approval, all authors; statistical analysis, TA, FG;and manuscript editing, all authors.

Competing interests None.

Funding None.

Ethics approval The Ethical Committee of the University Medical Centre Utrecht,The Netherlands.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES1. Lawn JE, Kerber K, Enweronu-Laryea C, et al. 3.6 million neonatal deaths—what is

progressing and what is not? Semin Perinatol 2010;34:371–86.2. de Vries LS, Dubowitz LM, Dubowitz V, et al. Predictive value of cranial ultrasound

in the newborn baby: a reappraisal. Lancet 1985;2:137–40.3. Vermeulen RJ, Fetter WP, Hendrikx L, et al. Diffusion-weighted MRI in severe

neonatal hypoxic ischaemia: the white cerebrum. Neuropediatrics 2003;34:72–6.4. Groenendaal F, van der Grond J, Witkamp TD, et al. Proton magnetic resonance

spectroscopic imaging in neonatal stroke. Neuropediatrics 1995;26:243–8.5. de Vries LS, Groenendaal F. Patterns of neonatal hypoxic-ischaemic brain injury.

Neuroradiology 2010;52:555–66.6. Rutherford MA, Pennock JM, Counsell SJ, et al. Abnormal magnetic resonance

signal in the internal capsule predicts poor neurodevelopmental outcome in infantswith hypoxic-ischemic encephalopathy. Pediatrics 1998;102:323–8.

7. Alderliesten T, de Vries LS, Benders MJ, et al. MR imaging and outcome of termneonates with perinatal asphyxia: value of diffusion-weighted MR imaging and 1HMR spectroscopy. Radiology 2011;261:235–42.

8. Jouvet P, Cowan FM, Cox P, et al. Reproducibility and accuracy of MR imaging ofthe brain after severe birth asphyxia. AJNR Am J Neuroradiol 1999;20:1343–8.

9. Roelants-van Rijn AM, Nikkels PG, Groenendaal F, et al. Neonataldiffusion-weighted MR imaging: relation with histopathology or follow-up MRexamination. Neuropediatrics 2001;32:286–94.

10. Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal distress. A clinicaland electroencephalographic study. Arch Neurol 1976;33:696–705.

11. Groenendaal F, Lammers H, Smit D, et al. Nitrotyrosine in brain tissue of neonatesafter perinatal asphyxia. Arch Dis Child Fetal Neonatal Ed 2006;91:F429–33.

12. de Vries LS, Toet MC. Amplitude integrated electroencephalography in the full-termnewborn. Clin Perinatol 2006;33:619–32.

Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768 F5

Original article

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

13. van Rooij LG, Toet MC, van Huffelen AC, et al. Effect of treatment of subclinicalneonatal seizures detected with aEEG: randomized, controlled trial. Pediatrics2010;125:e358–66.

14. Barkovich AJ, Hajnal BL, Vigneron D, et al. Prediction of neuromotor outcome in perinatalasphyxia: evaluation of MR scoring systems. AJNR Am J Neuroradiol 1998;19:143–9.

15. Ellis WG, Goetzman BW, Lindenberg JA. Neuropathologic documentation ofprenatal brain damage. Am J Dis Child 1988;142:858–66.

16. Roggendorf W, Strupp S, Paulus W. Distribution and characterization ofmicroglia/macrophages in human brain tumors. Acta Neuropathol 1996;92:288–93.

17. Cowan F, Rutherford M, Groenendaal F, et al. Origin and timing of brain lesions interm infants with neonatal encephalopathy. Lancet 2003;361:736–42.

18. Fan X, Heijnen CJ, van der Kooij MA, et al. Beneficial effect of erythropoietin onsensorimotor function and white matter after hypoxia-ischemia in neonatal mice.Pediatr Res 2011;69:56–61.

19. Salhab WA, Wyckoff MH, Laptook AR, et al. Initial hypoglycemia and neonatalbrain injury in term infants with severe fetal acidemia. Pediatrics2004;114:361–6.

20. Nadeem M, Murray DM, Boylan GB, et al. Early blood glucose profile andneurodevelopmental outcome at two years in neonatal hypoxic-ischaemiccephalopathy. BMC Pediatr 2011;11:10.

21. Shojania KG, Burton EC. The vanishing nonforensic autopsy. N Engl J Med2008;358:873–5.

22. Confidential Enquiry into Maternal and Child Health (CEMACH) 2009. ConfidentialEnquiry into Maternal and Child Health (CEMACH): Perinatal Mortality 2007: UnitedKingdom. Dorchester, Dorset Press, 2009.

23. Brookes JA, Hall-Craggs MA, Sams VR, et al. Non-invasive perinatal necropsy bymagnetic resonance imaging. Lancet 1996;348:1139–41.

24. Griffiths PD, Variend D, Evans M, et al. Postmortem MR imaging of the fetal andstillborn central nervous system. AJNR Am J Neuroradiol 2003;24:22–7.

25. Thayyil S, Chandrasekaran M, Chitty LS, et al. Diagnostic accuracy of post-mortemmagnetic resonance imaging in fetuses, children and adults: a systematic review.Eur J Radiol 2010;75:e142–8.

26. Thayyil S, Robertson NJ, Chitty LS, et al. Protocol 08PRT/5409: Postmortemmagnetic resonance imaging in the fetus, infant, and child: a comparative studywith conventional autopsy (UKCRN: 6794):01-08-2011.

F6 Arch Dis Child Fetal Neonatal Ed 2012;0:F1–F6. doi:10.1136/archdischild-2012-301768

Original article

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

doi: 10.1136/archdischild-2012-301768 published online November 21, 2012Arch Dis Child Fetal Neonatal Ed

 Thomas Alderliesten, Peter G J Nikkels, Manon J N L Benders, et al. asphyxiaencephalopathy following perinatal the brain in term infants with neonatalpostmortem histopathologic examination of Antemortem cranial MRI compared with

http://fn.bmj.com/content/early/2012/11/20/archdischild-2012-301768.full.htmlUpdated information and services can be found at:

These include:

Data Supplement http://fn.bmj.com/content/suppl/2012/11/20/archdischild-2012-301768.DC1.html

"Supplementary Data"

References http://fn.bmj.com/content/early/2012/11/20/archdischild-2012-301768.full.html#ref-list-1

This article cites 24 articles, 8 of which can be accessed free at:

P<P Published online November 21, 2012 in advance of the print journal.

serviceEmail alerting

the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from

CollectionsTopic

(268 articles)Immunology (including allergy)   � (557 articles)Epidemiologic studies   �

(41 articles)Pathology   � (125 articles)Trauma   �

(120 articles)Injury   � Articles on similar topics can be found in the following collections

Notes

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on November 26, 2012 - Published by fn.bmj.comDownloaded from