- Email: [email protected]
A Novel Scoring System for Term-Equivalent-Age Cranial Ultrasound in Extremely Preterm Infants
Ultrasound in Med. & Biol., Vol. 45, No. 3, pp. 786794, 2019 Copyright © 2018 World Federation for Ultrasound in Medicine & Biology. All rights reserved. Printed in the USA. All rights reserved. 0301-5629/$ - see front matter
https://doi.org/10.1016/j.ultrasmedbio.2018.11.005
Original Contribution A NOVEL SCORING SYSTEM FOR TERM-EQUIVALENT-AGE CRANIAL ULTRASOUND IN EXTREMELY PRETERM INFANTS AGEDPD1XT XBEATRICE
€ ,D2X X*,y D3X XBOUBOU HALLBERG,D4Xy,z ,D8X X*,y SKIOLD X D5X XBRIGITTE VOLLMER,D6X X*,1 D7X XULRIKA ADEN y,z y,2 D9X XMATS BLENNOW,D10X X and D1X XSANDRA HORSCHDX12 X TAGEDEN
* Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden; y Department of Neonatology, Karolinska University Hospital, Stockholm, Sweden; and z Department Clinical Science Intervention and Technology (CLINTEC), Karolinska Institutet, Stockholm, Sweden (Received 24 April 2018; revised 9 October 2018; in final from 8 November 2018)
Abstract—The role of term-equivalent-age (TEA) cranial ultrasound (cUS) in predicting outcome in preterm infants is increasingly being recognized. However, a detailed quantitative scoring system that allows comparison of groups and comparison with TEA magnetic resonance imaging (MRI) scoring systems is lacking. Eighty-four extremely preterm infants underwent cUS and MRI at TEA. Cranial US was evaluated using a novel detailed scoring system. Agreement between cUS and MRI scores was good (Spearman’s r = 0.51, p < 0.001). Outcome at 30 mo corrected was assessed in 66 of 84 preterm and 85 term-born infants. Sensitivity was the same for cUS and MRI in prediction of cerebral palsy (75%) and severe cognitive delay (100%); the specificity was slightly higher for MRI (cerebral palsy: 97% vs. 90%, severe cognitive delay: 95% vs. 90%). The proposed novel cUS scoring system is a helpful tool in quantitative assessment of cUS at TEA and prediction of outcome at 30 mo. (E-mail: [email protected]) © 2018 World Federation for Ultrasound in Medicine & Biology. All rights reserved. Key Words: Brain injury, Cerebral palsy, Magnetic resonance imaging, Neonate, Neurodevelopmental outcome, White matter abnormalities, Prognosis, Ultrasound.
INTRODUCTION
neonatal centres, and optimal timing of scanning has been discussed in detail elsewhere (Wezel-Meijler et al. 2014). But when high-resolution cUS is repeated frequently from the first week of life until term-equivalent age (TEA), it is a very sensitive tool in predicting cerebral palsy (CP), with a specificity reaching 95% and a sensitivity of 76% (de Vries et al. 2004). Recently, the value of term/near-term cUS for prediction of outcome has been highlighted in two large studies (Edwards et al. 2018; Hintz et al. 2015). In both of these studies, the definition of an abnormal late cUS was limited to the presence of cystic WMI, porencephalic cysts or ventriculomegaly, whereas more subtle abnormalities such as thinning of corpus callosum and delay in cortical folding, that are evaluated in the most commonly used magnetic resonance imaging (MRI) scoring systems were not taken into account. Making fair comparisons between imaging modalities is difficult, because a corresponding validated quantitative scoring system for TEA cUS is lacking. In the present study, we elaborated on our previous work in which we reported good agreement between a
Extremely preterm infants are at high risk of brain injury and subsequent neurodevelopmental impairments (Pierat et al. 2017; Serenius et al. 2016). Early prediction of long-term outcome remains a challenging task, despite the increased use of neonatal neuroimaging (Janvier and Barrington 2012; van’t Hooft et al. 2015). Serial cranial ultrasound (cUS) scans in the first week after birth are the gold standard for bedside detection of the most important prematurity-related brain lesions, such as germinal matrixintraventricular haemorrhages (GMH-IVHs), periventricular haemorrhagic infarctions (PHIs), post-haemorrhagic ventricular dilation (PHVD) and cystic white matter injury (WMI). Scanning protocols vary considerably between different Address correspondence to: Sandra Horsch, Helios Klinikum Berlin Buch, Berlin, Schwanebecker Chaussee 50, 13125 Berlin, Germany. E-mail: [email protected] 1 Present address: Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southhampton, UK. 2 Present address: Department of Neonatology, Helios Klinkum Berlin Buch, Berlin, Germany.
786
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
tentative cUS scoring system and an established MRI scoring system in extremely preterm infants <27 wk of gestation (Horsch et al. 2010). We here present a novel quantitative scoring system for cUS at TEA that accounts for all prematurity-related forms of brain injury that are relevant for outcome. We study its value for prediction of outcome in a population-based cohort of extremely preterm infants, particularly with respect to CP and severe cognitive, language and motor developmental delay, and compared it with cranial MRI performed on the same day. METHODS Participants The regional ethics review board in Stockholm approved the study, and informed consent was obtained from every participating family. All infants born in Stockholm between August 2004 and April 2007 with a gestational age (GA) <27 wk + 0 d were eligible for inclusion. Perinatal data were collected prospectively. Cranial ultrasound at TEA Cranial US was performed at TEA (3841 wk GA), on the same day as the MRI. Infants were scanned by one of two examiners (B.S. or S.H.) using an Acuson Sequoia ultrasound system (Siemens Medical Solutions, Erlangen, Germany) with a multifrequency sector transducer (58 MHz). The imaging protocol consisted of the standard projections in the coronal and sagittal/parasagittal planes (Ecury-Goossen et al. 2015; Levene et al. 1985). Images were acquired through the anterior fontanel only. Additional acoustic windows, such as mastoid and posterior fontanel, were not used. Cranial US images were stored digitally and analysed offline by three independent observers (B.S., S.H., M.B.; blinded to the clinical course, MRI findings and outcome), The cUS scoring system consists of in total 10 items (Table 1). Five items are subjectively scored (cysts, cortical and deep grey matter (GM) abnormalities, maturation of gyral folding and cerebellar abnormalities). If the scores of observers 1 and 2 differed (this was the case in 2% of all scores), observer 3 served as tie breaker. Five items were measured (size of frontal horns, ventricular mid-body, subarachnoidal space, interhemispheric fissure and corpus callosum) (Fig. 1). For the five measured items, the means of the measurements from the three observers were used for further analysis. The items were chosen to account for all important prematurity-related forms of brain injury (GMHIVH, PHI, cystic and non-cystic WMI and indirect signs of impaired brain growth/development) and then weighted according to their clinical relevance in relation to outcome based on results from previous ultrasound
787
and MRI studies (Brouwers et al. 2008; de Vries et al. 2004; Edwards et al. 2018; Hintz et al. 2015; Horsch et al. 2005; Inder et al. 2003; Woodward et al. 2006). Thus, the possible scores differ between items and reflect the clinical relevance for outcome. The composite cUS scores (sum of all 10 items) were categorized into four severity groups: (i) normal = composite score (CS) 10; (ii) mild abnormalities = CS of 1114; (iii) moderate abnormalities = CS of 1520; (iv) severe abnormalities = CS >20. Magnetic resonance imaging at TEA All scans were performed at TEA on a Philips Intera 1.5-T scanner (Philips International, Amsterdam, Netherlands) at Karolinska University Hospital in Stockholm, as previously described (Ski€old et al. 2010, 2012). Conventional MR images were evaluated by four independent observers (S.H./B.S., B.H./U.A., M.B. and the local neuroradiologist, all blinded to the clinical course, cUS findings and outcome) using a previously described scoring system for structural brain abnormalities (Inder et al. 2003; Woodward et al. 2006). We previously reported an inter-observer agreement of >98% using this scoring system with the same observers. Remaining discrepancies were resolved by discussions, and consensus was thus reached in all cases. Five white matter (WM) items are graded from 1 (normal) to 3 (severe) —abnormal WM signal, reduction in WM volume, cystic changes, myelination/thinning of the corpus callosum and ventricular dilation— resulting in a total score between 5 and 15. According to these scores, data sets were categorised into four groups: no (56), mild (79), moderate (1012) and severe (1315) WM abnormalities. For the GM, three items were graded from 1 (normal) to 3 (severe): abnormal signal in the cortical or deep GM, enlargement of the subarachnoid spaces and delay in gyral maturation. Data sets were then categorised as normal GM (35) and abnormal GM (69). Follow-up at age 30 mo Follow-up was performed at 30 mo corrected age within the framework of the EXPRESS project (Serenius et al. 2013). A control group of 85 term-born infants with similar age and sex distributions underwent the same follow-up assessments. For preterm children who did not attend the research follow up, medical records from routine clinical follow-up (performed by either a neonatologist or a paediatric neurologist) were reviewed for information regarding neurologic status. Neurologic status (movements, posture, reflexes and muscular tone) was assessed by a paediatric neurologist (B.V.) and categorized as “normal” or “abnormal” if neurologic signs of CP were present. A third category
788
Table 1. Cranial ultrasound scoring at term-equivalent age* Possible score Scoring Item
Scored items
Cysts or cavity
Score 1 None
II
Cortical GM abnormalities
Score 1 None
III
Deep GM abnormalities
IV
Maturation of gyral folding
Score 1 None Score 1 Normal
Score 2 (Isolated) Frontal reduction of complex gyral folding
V
Cerebellar abnormalities
Score 1 None
Score 2 Small focal haemorrhage
Score 2 Focal abnormality (one region only)
Score 5 Severe dilatationy or shunt without dilatation
VI
Score 10 Cyst or cavity likely involving CST or bilateral cystic PVL3
Score 10 Bilateral atrophy/ cysts
Score 10 Shunt with persistent dilatation
Score 3 Severely enlarged, >15 mm Score 3 Severely enlarged, at least one >6 mm Score 3 severely enlarged >6 mm Score 3 marked thinning <1.5 mm
CST = cortico-spinal tract; GA = gestational age; GM = grey matter; OR = optic radiation; PVL= periventricular leukomalacia. * Score 1 = no or only one moderately enlarged parameter, no shunt; score 2 = two or more moderately enlarged parameters, no shunt; score 5 = two or more severely enlarged parameters, no shunt or VP shunt in situ, no or only one moderately enlarged parameter; score 10 = VP shunt in situ, two or more enlarged parameters. y Short and long axes of the frontal horns at the level of the foramina Monroe were measured (see Fig. 1). Cutoff values for short axis (also known as frontal horn width): normal = <3 mm, moderate = 36 mm, severe = >6 mm. Cutoff values for long axis: normal = <13 mm, moderate = 1316 mm, severe = >16 mm.
Volume 45, Number 3, 2019
Measured items (Fig. 1 illustrates how measurements were performed) Score 1 Score 2 moderate dilatationy Size of frontal hornsy Measurement of long and short Normaly (anterior horn width) axesy VII Size of ventricular midbody Score 1 Score 2 Normal <10 mm Mildly to moderately enlarged, 1015 mm VIII Size of subarachnoidal spaces (sinu- Score 1 Score 2 cortical width) Normal <4 mm Mildly enlarged, 46 mm in 2 of 3 measurements IX Size of inter-hemispheric fissure Score 1 Score 2 Normal <3 mm Mildly enlarged 36 mm X Thickness of corpus callosum Score 1 Normal 1.5 mm Composite score = sum of scores for items IX 010 = no abnormalities; 1114 = mild abnormalities; 1520 = moderate abnormalities; >20 = severe abnormalities
Score 3 Score 5 Focal cyst or cavity but not involv- Unilateral cyst or cavity ing CST or involving more than one region but not involving CST1 or OR2 Score 5 Extensive abnormalities (more than one region) Score 5 Unilateral atrophy/ cysts Score 3 Global reduction of complex gyral folding/ delayed gyration for GA Score 3 Score 5 Unilateral extensive lobar Bilateral extensive lobar haemorrhage haemorrhage
Ultrasound in Medicine & Biology
I
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
789
Fig. 1. Cranial ultrasound measurements: (1) long axis of lateral ventricles; (2) short axis of lateral ventricles; (3) size of ventricular midbody; (4) size of subarachnoidal spaces (sinu-cortical width); (5) size of the interhemispheric fissure; (6) thickness of corpus callosum.
comprised children with “unspecific signs,” for example, mild muscular hypotonia and brisk reflexes, but not fulfilling the Surveillance of Cerebral Palsy criteria for diagnosis of CP (Surveillance of Cerebral Palsy in Europe Working Group 2000). The Bayley Scales of Infant and Toddler Development-III (BSITD-III) (Bayley 2006) were performed on the same day as the neurologic examination, assessing cognitive, language and motor development. The cutoff level for severe delay (<2 SD) was applied to the mean values of the regional term-born control group, which had a mean cognitive CS of 104 (SD § 9.6, i.e., severe delay <85), a mean language CS of 108 (SD § 11.7, i.e., severe delay <85) and a mean motor CS of 117 (SD §14.7, i.e., severe delay <88) (Ski€ old et al. 2012). Statistics Statistical analysis was performed with PASW Statistics Software 21.0 (IBM, Armonk, NY, USA). Infants were divided based on the MRI scoring categories for WM and GM. Similarly, infants were categorized based on cUS scoring groups. Furthermore, infants were grouped based on neurologic status (normal, abnormal or unspecific signs) and developmental outcomes (no or severe delay: i.e., a CS >2SD below the mean of the control group).
Student’s t-test was used for continuous variables, and Pearson’s x2-test and Fisher’s exact test were used for dichotomous data. Non-parametric tests were used for ranked data and unequal group sizes. Spearman’s rank coefficient was used to assess correlations between the CSs of the two scoring systems. A p value < 0.05 was chosen as the cutoff level for significance. RESULTS Eighty-four preterm infants were included in the study. Patient characteristics and perinatal factors are summarized in Table 2. Neuroimaging at TEA Brain abnormalities on MRI and cUS in the sample in the present study did not differ from data previously reported for the whole Stockholm cohort (Horsch et al. 2010; Ski€old et al. 2010, 2012). Incidence of brain abnormalities on cUS and WM injury on MRI is illustrated in Figure 2. Grey matter abnormalities on MRI were seen in 4 patients. There was overall good agreement between the CSs on cUS and MRI, both between total cUS score and WM abnormality score (Spearman’s r = 0.51, p < 0.001) and between total cUS score and GM abnormality score
790
Ultrasound in Medicine & Biology
Volume 45, Number 3, 2019
Table 2. Demographic data and neonatal clinical characteristics*
Gestational age, wk + d Birth weight, g Gender, males Time on CPAP, d Time on mechanical ventilation, d Bronchopulmonary dysplasia, oxygen required at 36 wk GA Retinopathy of prematurity, treated Patent ductus arteriosus, surgical ligation Necrotizing enterocholitis, Bells grade 23 No intraventricular haemorrhagey Intraventricular haemorrhage, grade 12y Intraventricular haemorrhage, grade 3y Parenchymal haemorrhagic infarctiony Periventricular leukomalaciay
All infants (n = 84)
Attending FU at age 30 mo (n = 66)
FU info from clinical records (n = 18)
P Value
(25 + 4) § 1 ([23 + 1][26 + 6]) 808 § 153 (4941161) 45/84 (54%) 39 § 11 (1964) 13 § 13 (055) 33/84 (39%)
(25 + 4) § 1 ([23+1][26+6]) 826 § 147 (4941161) 34/66 (52%) 38 § 11 (1964) 13 § 13 (055) 25/66 (38%)
(25 + 3) § 1 ([23 + 3][26+6]) 744 § 162 (4991001) 11/18 (61%) 40 § 10 (2060) 12 § 12 (036) 8/18 (40%)
ns ns ns ns ns ns
12/84 (14%) 23/84 (27%)
8/66 (12%) 21/66 (32%)
4/18 (22%) 2/18 (11%)
ns ns
5/84 (6%)
5/66 (8%)
0
ns
46/84 (55%) 28/84 (33%)
33/66 (50%) 23/66 (35%)
13/18 (72%) 5/18 (28%)
ns ns
4/84 (5%)
4/66 (6%)
0
ns
7/84 (8%)
7/66 (11%)
0
ns
3/84 (4%)
2/66 (3%)
1/18 (6%)
ns
FU = follow-up; CPAP = continuous positive airway pressure; GA = gestational age; ns = nonsignificant. * Values are expressed as number/total (%) or mean § standard deviation (range). y Diagnosis from routine cranial ultrasound examinations performed repeatedly during the hospital stay.
(Spearman’s r = 0.58, p < 0.001). Of the 34 infants with an entirely normal cUS, 68% (23/34) had normal and 32% (11/34) had mild WM abnormalities on MRI, none had moderate or severe WM abnormalities and none had abnormal GM. The two infants with severe WM abnormalities on MRI were identified as having moderate or severe abnormalities on cUS. Neurodevelopmental outcome At age 30 mo, 79% (66/84) of the preterm-born children attended follow-up. Of these, 7 had incomplete
BSITD-III assessments. For preterm children who did not attend the research follow-up (18/84), medical records from routine clinical follow-up (mean age at follow-up = 24 mo, SD: §8, range: 1136) were reviewed for information regarding neurologic status. This group did not differ from the group of children that did attend follow-up with respect to perinatal characteristics, neuroimaging status or neurologic status (Table 2). On the BSITD-III, the mean cognitive CS of the preterm study sample was 95 (SD: §8.3, range: 65120), the mean language CS was 97 (SD: §14,
Fig. 2. Abnormalities on magnetic resonance imaging and cranial ultrasound. cUS = cranial ultrasound; MRI = magnetic resonance imaging; WM = white matter.
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
791
Table 3. Neurological status
Normal Unspecific signs Abnormal, cerebral palsy
All infants (n = 84)
Attending FU at age 30 mo (n = 66)
FU info from clinical records (n = 18)
56/84 (66%) 24/84 (29%) 4/84 (5%)
42/66 (64%) 20/66 (30%) 4/66 (6%)
14/18 (78%) 4/18 (22%) 0
P Value ns ns ns
FU = follow-up; ns = nonsignificant.
range: 53124) and the mean motor CS was 104 (SD: §17, range: 45130). Two children (3%) had severe cognitive delay, 8 children (13%) had severe language delay and 8 children (13%) had severe motor delay (Table 3). Four infants (6%) fulfilled the criteria for CP (Tables 3 and 4). Associations between cUS scoring groups and outcome For infants with either moderate or severe abnormalities on cUS, associations were observed with CP (Fisher’s exact test, p = 0.007) and with severe cognitive delay (Fisher’s exact test, p = 0.015); other investigated associations (unspecific signs and BSITD-III language and motor CSs) were not significant (Table 5). Of the 4 infants with severe abnormalities, 2 developed CP, and 1, severe cognitive delay. With respect to individual cUS items, cysts were associated with unspecific neurologic signs, CP and severe cognitive delay (Pearson’s x2, p < 0.001); and parieto-occipital WM loss was associated with severe cognitive delay (Pearson’s x2, p < 0.001). The negative predictive values of a normal or mildly abnormal TEA cUS regarding normal
neurologic status and cognitive outcome were 98% and 100%, respectively (Table 6). One infant that developed CP had a normal CUS and MRI. Associations between MRI scoring groups and outcome For infants with either moderate or severe WM abnormalities on MRI, associations were seen both with CP (Fisher’s exact test, p = 0.001) and with severe cognitive delay (Fisher’s exact test, p = 0.005); other investigated associations (unspecific neurologic signs and BSITD-III language and motor CSs) were not significant (Table 5) (previously published [Ski€old et al. 2012]). Both infants with severe WM abnormalities developed CP and severe cognitive delay. When infants with mild and moderate WM abnormalities were pooled together, no associations with neurologic status or developmental outcome were seen. Abnormal GM (irrespective of WM abnormalities) was not associated with CP or severe developmental delay. Adding “abnormal GM” to the pooled analyses of infants with moderate and severe WM abnormalities on MRI did not alter the associations with outcome. When
Table 4. Imaging findings and outcome in infants with cerebral palsy BSITD-III Gender F M F M
GA
Magnetic resonance imaging
Cranial ultrasound
Cognitive composite
Motor composite
GMFCS
26 + 5 25 + 4 26 + 4 25 + 1
No WMAs Normal GM Severe WMAs Normal GM Severe WMAs Abnormal GM Moderate WMAs Normal GM
No abnormalities Moderate abnormalities Severe abnormalities Severe abnormalities
105 80 65 95
85 103 45 97
I I IV I
BSITD-III = Bayley Scales of Infant and Toddler Development-III; GA = gestational age; GM = grey matter; GMFCS = Gross Motor Function Classification System; WMAs = white matter abnormalities.
Table 5. Associations between scoring of imaging findings and outcome Scoring results at term-equivalent age MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities
Outcome at age 30 mo
p value (Fisher’s exact test)
Cerebral palsy Cerebral palsy Severe cognitive delay, regional cutoff* Severe cognitive delay, regional cutoff*
0.001 0.007 0.005 0.015
CUS = cranial ultrasound; MRI = magnetic resonance imaging; WMAs = white matter abnormalities. * Bayley Scales of Infant and Toddler Development-III cognitive composite score 2SD below the mean of matched regional control infants (Stockholm).
792
Ultrasound in Medicine & Biology
Volume 45, Number 3, 2019
Table 6. MRI and CUS scores in predicting cerebral palsy and/or severe cognitive delay Outcome*
Sensitivity
Specificity
PPV
NPV
p value (Fisher’s exact test)
Cerebral palsy Cerebral palsy Severe cognitive delay regional cutoffy Severe cognitive delay regional cutoffy
75% 75% 100% 100%
97% 90% 95% 90%
60% 33% 40% 25%
98% 98% 100% 100%
0.001 0.007 0.005 0.015
Assessment MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities
CUS = cranial ultrasound; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; WMAs = white matter abnormalities. * Predictive values for the combination of cerebral palsy and/or severe cognitive delay (irrespective of cutoff) were the same as predictive values for cerebral palsy on its own (data not shown). * Bayley Scales of Infant and Toddler Development-III composite score 2SD below the mean of matched regional control infants (Stockholm).
investigating individual scoring items, WM reduction, delayed myelination and cysts were associated with unspecific neurologic signs, CP and severe cognitive delay (Pearson’s x2, p < 0.001). Abnormal signal in the cortical or deep GM was associated with severe cognitive delay (Pearson x2, p < 0.001). The two scoring systems are compared with respect to diagnostic accuracy in Table 6. Cerebellar haemorrhage Six infants had small cerebellar haemorrhages. All 6 were only detected on MRI. None was detected by cUS via the anterior fontanel. DISCUSSION We describe a novel comprehensive scoring system for cUS at TEA that allows quantification of overall brain injury comparable to that of established MRI scoring systems. This cUS scoring system covers the most relevant prematurity-related brain injuries such as residuals and sequelae of haemorrhages, periventricular infarctions, post-haemorrhagic hydrocephalus, WMI and brain atrophy. The TEA cUS scores exhibited good agreement with TEA MRI scores and reached similar predictive values for severe impairments. Magnetic resonance imaging and cUS are the two most important neuroimaging modalities in newborn infants. Neonatal MRI has been increasingly used in the last two decades and has tremendously enhanced our understanding of neonatal brain development and injury (Inder et al. 2003; Keunen et al. 2016; Woodward et al. 2006). Nevertheless, MRI remains an expensive method that is time and staff consuming. Access to MRI is limited in many countries and centres; therefore, patients who undergo MRI need to be carefully selected. The increasing use of MRI at term age for screening of extremely preterm infants remains controversial. In 2015, the American Academy of Pediatrics’ (AAP) section on Neonatal Perinatal Medicine undertook an expert consensus building process to identify “use of unnecessary
tests and treatments that contribute to health care waste” as part of the AAP Choosing Wisely campaign. The authors concluded that routine screening with TEA MRIs in preterm infants is related to, but not proven to improve outcomes (Ho et al. 2015). This was in brief supported by a recent randomized trial by Edwards et al. (2018) investigating predictive power, maternal anxiety and costs of TEA neuroimaging (cUS and MRI), concluding that MRI “increased costs and provided only modest benefits.” On the other hand, experts have opposed the AAP Choosing Wisely statement (Cheong et al. 2018). This controversy underlines the need for thorough research in this field and careful assessment of cost-effective alternatives to routine TEA MRI in preterm infants. Early studies that compared MRI and cUS in the prediction of outcome of extremely preterm infants found TEA MRI to be superior to cUS (Maalouf et al. 2001; Mirmiran et al. 2004; Woodward et al. 2006). In many of these studies, cUS was performed during the first week after birth and not at term age, using a variety of cUS diagnoses (IVH, PHVD, periventricular leukomalacia) to be compared with data from complex MRI scoring systems. Later studies using serial cUS and (near-) term data reported improved accuracy of cUS in the prediction of outcome (de Vries et al. 2004; Edwards et al. 2018). Hintz et al. (2015) reported that adverse cUS findings on the late near-term cUS scan were associated with outcomes at 1822 mo corrected age and reached predictive values in the range of TEA MRI (Hintz et al. 2015). Interestingly, although a detailed MRI scoring system that was developed for white and GM abnormalities has been widely used, refined and updated over the years (Inder et al. 2003; Kidokoro et al. 2013; Woodward et al. 2006), no detailed or systematic scoring system for cUS images has, to our knowledge, yet been published. The aim of the proposed cUS scoring was to create a quantitative assessment tool that reliably detects clinically relevant sequelae of preterm brain injury and impaired brain development at TEA. The examination does not differ significantly from a clinical routine scan
€ et al. New scoring system for cranial US in preterm infants B. SKIOLD
(usually takes approximately 10 min for an experienced user) as only standard projections according to Levene et al. (1985) are used. Cranial US images are stored, and scoring may be performed on a later occasion, that is, not interfering with clinical work on the ward, which also facilitates training of less experienced cUS users. The inter-observer agreement in the present study was high. Similar to the updated MRI scoring system of Kidokoro et al. (2013), the present cUS scoring includes a combination of subjectively scored items and objective metrics, resulting in a composite (“global”) brain abnormality score. With this novel scoring system, not only adverse findings but also normal TEA cUSs and mild abnormalities are clearly defined and easily reproducible for users. This is important as the negative predictive values regarding all investigated outcomes were high when TEA cUS scoring was normal or revealed only mild abnormalities. This underlines the value of TEA cUS for screening infants that highly unlikely will profit from further investigation with MRI. We believe that this proposed cUS scoring system can be integrated into clinical routine, but also enables uniform structured assessment of brain injuries in the research setting, such as comparison across cohorts. The low rates of severe brain abnormalities and severe impairments in our cohort are important limitations to this study. Although it is of course gratifying that the cohort is doing well, it restricts the prognostic value of the tool, a problem experienced by others in similar settings (Brouwer et al. 2017). Therefore, validation of the presented cUS scoring system in a cohort with higher incidences of both brain injury and impaired outcomes is warranted. Another limitation is that cUS images were acquired only via the anterior fontanel, which most likely contributed to the low rate of detection of cerebellar injury. In future studies, additional acoustic windows such as the mastoid fontanel, posterior fontanel, temporal window and foramen magnum should be used to improve visualization of areas less accessible via the anterior fontanel. Moreover, a comparison with the revised version of the InderKidokoro MRI scoring system (Kidokoro et al. 2013) would also be of great interest. Another major limitation is the relatively short follow-up, which might naturally underestimate more subtle, later-emerging cognitive impairments. We believe that the major strength of the present scoring system is that it assesses the full spectrum of prematurity-related brain injury and uses weighted subscores that reflect the relevance for outcome of each item. Moreover, the cUS scoring system is clearly structured and easy to use also for non-experts. The population-based design is another strength of our study, along with the large sample size of infants with a GA at birth
793
<27 wk. In addition, to increase accuracy of the outcome comparisons, we used a regional control group of term-born infants, as the third edition of the Bayley Scales has been reported to underestimate developmental delay (Anderson et al. 2010; Vohr et al. 2012). CONCLUSIONS This article has described a novel comprehensive scoring system for cUS at TEA, allowing quantification of preterm brain injury comparable to that of established MRI scoring systems. It may thus enable uniform structured assessment of brain injuries in the research setting, such as comparison across cohorts. Moreover, we believe this cUS scoring system may be integrated into study protocols and clinical routine. The negative predictive values regarding all investigated outcomes were high; that is, cUS scoring may potentially be used to screen for infants that are unlikely to profit from further investigation with MRI. Acknowledgments—The present study was supported by the following grants: European Society for Pediatric Research Young Investigator Exchange Program, Jerring Foundation, S€allskapet Barnavard, M€arta och Gunnar V Philipsson Foundation, Swedish Medical Research Council, Swedish Brain Foundation (Hj€arnfonden), Foundation Samar iten, Free Masonry Foundation Barnhuset in Stockholm, Ake Wiberg Foundation, Jeansson Foundation and Karolinska University Hospital. These sources financed the research activity of S.H. and B.S. We thank Mikael Mosskin for his participation in the magnetic resonance imaging scoring sessions, helpful advice and support of this study.
REFERENCES Anderson PJ, De Luca CR, Hutchinson E, Roberts G, Doyle LW. Victorian Infant Collaborative Group. Underestimation of developmental delay by the new Bayley-III Scale. Arch Pediatr Adolesc Med 2010;164:352–356. Bayley N. Bayley scales of infant and toddler development. 3rd ed. Harcourt: Assessment Inc; 2006. Brouwer A, Greoendaal F, van Haastert IL, Rademaker K, Hanlo P, De Vries L. Neurodevelopmental outcome of preterm infants with severe intraventricular hemorrhage and therapy for post-haemorrhagic ventricular dilation. J Pediatr 2008;152:648–654. Brouwer MJ, Kersbergen KJ, van Kooij BJM, Benders MJNL, van Haastert IC, Koopman-Esseboom C, Neil JJ, de Vries LS, Kidokoro H, Inder TE, Groenendaal F. Preterm brain injury on term-equivalent age MRI in relation to perinatal factors and neurodevelopmental outcome at two years. PLoS One 2017;12: e0177128. Cheong JLY, Miller SP. Imaging the neonatal brain in the 21st century: Why, when and how? Arch Dis Child Fetal Neonatal Ed 2018;103: F4–F5. De Vries LS, Van Haastert IL, Rademaker KJ, Koopman C, Groenendaal F. Ultrasound abnormalities preceding cerebral palsy in highrisk preterm infants. J Pediatr 2004;144:815–820. Ecury-Goossen GM, Camfferman FA, Leijser LM, Govaert P, Dudink J. State of the art cranial ultrasound imaging in neonates. J Vis Exp 2015;96:e52238. Edwards AD, Redshaw ME, Kennea N, Rivero-Arias O, GonzalesCinca N, Nongena P, Ederies M, Falconer S, Chew A, Omar O, Hardy P, Harvey ME, Eddama O, Hayward N, Wurie J, Azzopardi D, Rutherford MA, Counsell S. ePrime Investigators. Effect of MRI on preterm infants and their families: A randomised trial with
794
Ultrasound in Medicine & Biology
nested diagnostic and economic evaluation. Arch Dis Child Fetal Neonatal Ed 2018;103:F15–F21. Hintz SR, Barnes PD, Bulas D, Slovis TL, Finer NN, Wrage LA, Das A, Tyson JE, Stevenson DK, Carlo WA, Walsh MC, Laptook AR, Yoder BA, Van Meurs KP, Faix RG, Rich W, Newman NS, Cheng H, Heyne RJ, Vohr BR, Acarregui MJ, Vaucher YE, Pappas A, Peralta-Carcelen M, Wilson-Costello DE, Evans PW, Goldstein RF, Myers GJ, Poindexter BB, McGowan EC, Adams-Chapman I, Fuller J, Higgins RD. SUPPORT Study Group of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Neuroimaging and neurodevelopmental outcome in extremely preterm infants. Pediatrics 2015;135:e32–e42. Ho T, Dukhovny D, Zupancic JA, Goldmann DA, Horbar JD, Pursley DM. Choosing wisely in newborn medicine: Five opportunities to increase value. Pediatrics 2015;136:e482–e489. Horsch S, Muentjes C, Franz A, Roll C. Ultrasound diagnosis of brain atrophy is related to neurodevelopmental outcome in preterm infants. Acta Paediatr 2005;94:1815–1822. Horsch S, Skiold B, Hallberg B, Nordell B, Nordell A, Mosskin M, Lagercrantz H, Aden U, Blennow M. Cranial ultrasound and MRI at term age in extremely preterm infants. Arch Dis Child Fetal Neonatal Ed 2010;95:F310–F314. Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature of the cerebral abnormalities in the premature infant: A qualitative magnetic resonance imaging study. J Pediatr 2003;143:171–179. Janvier A, Barrington K. Trying to predict the future of ex-preterm infants: Who benefits from a brain MRI at term? Acta Paediatr 2012;101:1016–1017. Keunen K, Isgum I, van Kooij BJ, Anbeek P, van Haastert IC, Koopman-Esseboom C, Fieret-van Stam PC, Nievelstein RA, Viergever MA, de Vries LS, Groenendaal F, Benders MJ. Brain volumes at term-equivalent age in preterm infants: Imaging biomarkers for neurodevelopmental outcome through early school age. J Pediatr 2016;172:88–95. Kidokoro H, Neil JJ, Inder TE. New MR imaging assessment tool to define brain abnormalities in very preterm infants at term. AJNR Am J Neuroradiol 2013;34:2208–2214. Levene MI, Williams JL, Fawer CL. Ultrasound of the infant brain. Clin Dev Med 1985;92:9–33. Maalouf EF, Duggan PJ, Counsell SJ, Rutherford MA, Cowan F, Azzopardi D, Edwards AD. Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 2001;107:719–727. Mirmiran M, Barnes PD, Keller K, Constantinou JC, Fleisher BE, Hintz SR, Ariagno RL. Neonatal brain magnetic resonance imaging before discharge is better than serial cranial ultrasound in predicting
Volume 45, Number 3, 2019 cerebral palsy in very low birth weight preterm infants. Pediatrics 2004;114:992–998. Pierrat V, Marchand-Martin L, Arnaud C, Kaminski M, Resche-Rigon M, Lebeaux C, Bodeau-Livinec F, Morgan AS, Goffinet F, Marret S, Ancel PY. EPIPAGE-2 Writing Group. Neurodevelopmental outcome at 2 years for preterm children born at 22 to 34 weeks’ gestation in France in 2011: EPIPAGE-2 cohort study. BMJ 2017;358:j3448. Serenius F, K€allen K, Blennow M, Ewald U, Fellman V, Holmstr€om G, Lindberg E, Lundqvist P, Marsal K, Norman M, Olhager E, Stigson L, Stjernqvist K, Vollmer B, Str€omberg B. EXPRESS Group. Neurodevelopmental outcome in extremely preterm infants at 2.5 years after active perinatal care in Sweden. JAMA 2013;309:1810–1820. Serenius F, Ewald U, Farooqi A, Fellman V, Hafstr€om M, Hellgren K, Marsal K, Ohlin A, Olhager E, Stjernqvist K, Str€omberg B, Aden U, K€allen K. Extremely Preterm Infants in Sweden Study Group. Neurodevelopmental outcomes among extremely preterm infants 6.5 years after active perinatal care in Sweden. JAMA Pediatr 2016;170:954–963. Skiold B, Horsch S, Hallberg B, Engstr€om M, Nagy Z, Mosskin M, Blennow M, Aden U. White matter changes in extremely preterm infants, a population-based diffusion tensor imaging study. Acta Paediatr 2010;99:842–849. Skiold B, Vollmer B, B€ohm B, Hallberg B, Horsch S, Mosskin M, Lagercrantz H, Aden U, Blennow M. Neonatal magnetic resonance imaging and outcome at age 30 months in extremely preterm infants. J Pediatr 2012;160:559–566. Surveillance of Cerebral Palsy in Europe (SCPE) Working Group. Surveillance of cerebral palsy in Europe: A collaboration of cerebral palsy surveys and registers. Dev Med Child Neurol 2000;42:816–824. Van’t Hooft J, van der Lee JH, Opmeer BC, Aarnoudse-Moens CS, Leenders AG, Mol BW, de Haan TR. Predicting developmental outcomes in premature infants by term equivalent MRI: Systematic review and meta-analysis. Syst Rev 2015;4:71. Vohr BR, Stephens BE, Higgins RD, Bann CM, Hintz SR, Das A, Newman JE, Peralta-Carcelen M, Yolton K, Dusick AM, Evans PW, Goldstein RF, Ehrenkranz RA, Pappas A, Adams-Chapman I, Wilson-Costello DE, Bauer CR, Bodnar A, Heyne RJ, Vaucher YE, Dillard RG, Acarregui MJ, McGowan EC, Myers GJ, Fuller J. Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Are outcomes of extremely preterm infants improving? Impact of Bayley assessment on outcomes. J Pediatr 2012;161:222–228 e3. Wezel-Meijler G, de Vries LS. Cranial ultrasound—Optimizing utility in the NICU. Curr Pediatr Rev 2014;10:16–27. Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med 2006;355:685–694.
https://doi.org/10.1016/j.ultrasmedbio.2018.11.005
Original Contribution A NOVEL SCORING SYSTEM FOR TERM-EQUIVALENT-AGE CRANIAL ULTRASOUND IN EXTREMELY PRETERM INFANTS AGEDPD1XT XBEATRICE
€ ,D2X X*,y D3X XBOUBOU HALLBERG,D4Xy,z ,D8X X*,y SKIOLD X D5X XBRIGITTE VOLLMER,D6X X*,1 D7X XULRIKA ADEN y,z y,2 D9X XMATS BLENNOW,D10X X and D1X XSANDRA HORSCHDX12 X TAGEDEN
* Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden; y Department of Neonatology, Karolinska University Hospital, Stockholm, Sweden; and z Department Clinical Science Intervention and Technology (CLINTEC), Karolinska Institutet, Stockholm, Sweden (Received 24 April 2018; revised 9 October 2018; in final from 8 November 2018)
Abstract—The role of term-equivalent-age (TEA) cranial ultrasound (cUS) in predicting outcome in preterm infants is increasingly being recognized. However, a detailed quantitative scoring system that allows comparison of groups and comparison with TEA magnetic resonance imaging (MRI) scoring systems is lacking. Eighty-four extremely preterm infants underwent cUS and MRI at TEA. Cranial US was evaluated using a novel detailed scoring system. Agreement between cUS and MRI scores was good (Spearman’s r = 0.51, p < 0.001). Outcome at 30 mo corrected was assessed in 66 of 84 preterm and 85 term-born infants. Sensitivity was the same for cUS and MRI in prediction of cerebral palsy (75%) and severe cognitive delay (100%); the specificity was slightly higher for MRI (cerebral palsy: 97% vs. 90%, severe cognitive delay: 95% vs. 90%). The proposed novel cUS scoring system is a helpful tool in quantitative assessment of cUS at TEA and prediction of outcome at 30 mo. (E-mail: [email protected]) © 2018 World Federation for Ultrasound in Medicine & Biology. All rights reserved. Key Words: Brain injury, Cerebral palsy, Magnetic resonance imaging, Neonate, Neurodevelopmental outcome, White matter abnormalities, Prognosis, Ultrasound.
INTRODUCTION
neonatal centres, and optimal timing of scanning has been discussed in detail elsewhere (Wezel-Meijler et al. 2014). But when high-resolution cUS is repeated frequently from the first week of life until term-equivalent age (TEA), it is a very sensitive tool in predicting cerebral palsy (CP), with a specificity reaching 95% and a sensitivity of 76% (de Vries et al. 2004). Recently, the value of term/near-term cUS for prediction of outcome has been highlighted in two large studies (Edwards et al. 2018; Hintz et al. 2015). In both of these studies, the definition of an abnormal late cUS was limited to the presence of cystic WMI, porencephalic cysts or ventriculomegaly, whereas more subtle abnormalities such as thinning of corpus callosum and delay in cortical folding, that are evaluated in the most commonly used magnetic resonance imaging (MRI) scoring systems were not taken into account. Making fair comparisons between imaging modalities is difficult, because a corresponding validated quantitative scoring system for TEA cUS is lacking. In the present study, we elaborated on our previous work in which we reported good agreement between a
Extremely preterm infants are at high risk of brain injury and subsequent neurodevelopmental impairments (Pierat et al. 2017; Serenius et al. 2016). Early prediction of long-term outcome remains a challenging task, despite the increased use of neonatal neuroimaging (Janvier and Barrington 2012; van’t Hooft et al. 2015). Serial cranial ultrasound (cUS) scans in the first week after birth are the gold standard for bedside detection of the most important prematurity-related brain lesions, such as germinal matrixintraventricular haemorrhages (GMH-IVHs), periventricular haemorrhagic infarctions (PHIs), post-haemorrhagic ventricular dilation (PHVD) and cystic white matter injury (WMI). Scanning protocols vary considerably between different Address correspondence to: Sandra Horsch, Helios Klinikum Berlin Buch, Berlin, Schwanebecker Chaussee 50, 13125 Berlin, Germany. E-mail: [email protected] 1 Present address: Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southhampton, UK. 2 Present address: Department of Neonatology, Helios Klinkum Berlin Buch, Berlin, Germany.
786
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
tentative cUS scoring system and an established MRI scoring system in extremely preterm infants <27 wk of gestation (Horsch et al. 2010). We here present a novel quantitative scoring system for cUS at TEA that accounts for all prematurity-related forms of brain injury that are relevant for outcome. We study its value for prediction of outcome in a population-based cohort of extremely preterm infants, particularly with respect to CP and severe cognitive, language and motor developmental delay, and compared it with cranial MRI performed on the same day. METHODS Participants The regional ethics review board in Stockholm approved the study, and informed consent was obtained from every participating family. All infants born in Stockholm between August 2004 and April 2007 with a gestational age (GA) <27 wk + 0 d were eligible for inclusion. Perinatal data were collected prospectively. Cranial ultrasound at TEA Cranial US was performed at TEA (3841 wk GA), on the same day as the MRI. Infants were scanned by one of two examiners (B.S. or S.H.) using an Acuson Sequoia ultrasound system (Siemens Medical Solutions, Erlangen, Germany) with a multifrequency sector transducer (58 MHz). The imaging protocol consisted of the standard projections in the coronal and sagittal/parasagittal planes (Ecury-Goossen et al. 2015; Levene et al. 1985). Images were acquired through the anterior fontanel only. Additional acoustic windows, such as mastoid and posterior fontanel, were not used. Cranial US images were stored digitally and analysed offline by three independent observers (B.S., S.H., M.B.; blinded to the clinical course, MRI findings and outcome), The cUS scoring system consists of in total 10 items (Table 1). Five items are subjectively scored (cysts, cortical and deep grey matter (GM) abnormalities, maturation of gyral folding and cerebellar abnormalities). If the scores of observers 1 and 2 differed (this was the case in 2% of all scores), observer 3 served as tie breaker. Five items were measured (size of frontal horns, ventricular mid-body, subarachnoidal space, interhemispheric fissure and corpus callosum) (Fig. 1). For the five measured items, the means of the measurements from the three observers were used for further analysis. The items were chosen to account for all important prematurity-related forms of brain injury (GMHIVH, PHI, cystic and non-cystic WMI and indirect signs of impaired brain growth/development) and then weighted according to their clinical relevance in relation to outcome based on results from previous ultrasound
787
and MRI studies (Brouwers et al. 2008; de Vries et al. 2004; Edwards et al. 2018; Hintz et al. 2015; Horsch et al. 2005; Inder et al. 2003; Woodward et al. 2006). Thus, the possible scores differ between items and reflect the clinical relevance for outcome. The composite cUS scores (sum of all 10 items) were categorized into four severity groups: (i) normal = composite score (CS) 10; (ii) mild abnormalities = CS of 1114; (iii) moderate abnormalities = CS of 1520; (iv) severe abnormalities = CS >20. Magnetic resonance imaging at TEA All scans were performed at TEA on a Philips Intera 1.5-T scanner (Philips International, Amsterdam, Netherlands) at Karolinska University Hospital in Stockholm, as previously described (Ski€old et al. 2010, 2012). Conventional MR images were evaluated by four independent observers (S.H./B.S., B.H./U.A., M.B. and the local neuroradiologist, all blinded to the clinical course, cUS findings and outcome) using a previously described scoring system for structural brain abnormalities (Inder et al. 2003; Woodward et al. 2006). We previously reported an inter-observer agreement of >98% using this scoring system with the same observers. Remaining discrepancies were resolved by discussions, and consensus was thus reached in all cases. Five white matter (WM) items are graded from 1 (normal) to 3 (severe) —abnormal WM signal, reduction in WM volume, cystic changes, myelination/thinning of the corpus callosum and ventricular dilation— resulting in a total score between 5 and 15. According to these scores, data sets were categorised into four groups: no (56), mild (79), moderate (1012) and severe (1315) WM abnormalities. For the GM, three items were graded from 1 (normal) to 3 (severe): abnormal signal in the cortical or deep GM, enlargement of the subarachnoid spaces and delay in gyral maturation. Data sets were then categorised as normal GM (35) and abnormal GM (69). Follow-up at age 30 mo Follow-up was performed at 30 mo corrected age within the framework of the EXPRESS project (Serenius et al. 2013). A control group of 85 term-born infants with similar age and sex distributions underwent the same follow-up assessments. For preterm children who did not attend the research follow up, medical records from routine clinical follow-up (performed by either a neonatologist or a paediatric neurologist) were reviewed for information regarding neurologic status. Neurologic status (movements, posture, reflexes and muscular tone) was assessed by a paediatric neurologist (B.V.) and categorized as “normal” or “abnormal” if neurologic signs of CP were present. A third category
788
Table 1. Cranial ultrasound scoring at term-equivalent age* Possible score Scoring Item
Scored items
Cysts or cavity
Score 1 None
II
Cortical GM abnormalities
Score 1 None
III
Deep GM abnormalities
IV
Maturation of gyral folding
Score 1 None Score 1 Normal
Score 2 (Isolated) Frontal reduction of complex gyral folding
V
Cerebellar abnormalities
Score 1 None
Score 2 Small focal haemorrhage
Score 2 Focal abnormality (one region only)
Score 5 Severe dilatationy or shunt without dilatation
VI
Score 10 Cyst or cavity likely involving CST or bilateral cystic PVL3
Score 10 Bilateral atrophy/ cysts
Score 10 Shunt with persistent dilatation
Score 3 Severely enlarged, >15 mm Score 3 Severely enlarged, at least one >6 mm Score 3 severely enlarged >6 mm Score 3 marked thinning <1.5 mm
CST = cortico-spinal tract; GA = gestational age; GM = grey matter; OR = optic radiation; PVL= periventricular leukomalacia. * Score 1 = no or only one moderately enlarged parameter, no shunt; score 2 = two or more moderately enlarged parameters, no shunt; score 5 = two or more severely enlarged parameters, no shunt or VP shunt in situ, no or only one moderately enlarged parameter; score 10 = VP shunt in situ, two or more enlarged parameters. y Short and long axes of the frontal horns at the level of the foramina Monroe were measured (see Fig. 1). Cutoff values for short axis (also known as frontal horn width): normal = <3 mm, moderate = 36 mm, severe = >6 mm. Cutoff values for long axis: normal = <13 mm, moderate = 1316 mm, severe = >16 mm.
Volume 45, Number 3, 2019
Measured items (Fig. 1 illustrates how measurements were performed) Score 1 Score 2 moderate dilatationy Size of frontal hornsy Measurement of long and short Normaly (anterior horn width) axesy VII Size of ventricular midbody Score 1 Score 2 Normal <10 mm Mildly to moderately enlarged, 1015 mm VIII Size of subarachnoidal spaces (sinu- Score 1 Score 2 cortical width) Normal <4 mm Mildly enlarged, 46 mm in 2 of 3 measurements IX Size of inter-hemispheric fissure Score 1 Score 2 Normal <3 mm Mildly enlarged 36 mm X Thickness of corpus callosum Score 1 Normal 1.5 mm Composite score = sum of scores for items IX 010 = no abnormalities; 1114 = mild abnormalities; 1520 = moderate abnormalities; >20 = severe abnormalities
Score 3 Score 5 Focal cyst or cavity but not involv- Unilateral cyst or cavity ing CST or involving more than one region but not involving CST1 or OR2 Score 5 Extensive abnormalities (more than one region) Score 5 Unilateral atrophy/ cysts Score 3 Global reduction of complex gyral folding/ delayed gyration for GA Score 3 Score 5 Unilateral extensive lobar Bilateral extensive lobar haemorrhage haemorrhage
Ultrasound in Medicine & Biology
I
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
789
Fig. 1. Cranial ultrasound measurements: (1) long axis of lateral ventricles; (2) short axis of lateral ventricles; (3) size of ventricular midbody; (4) size of subarachnoidal spaces (sinu-cortical width); (5) size of the interhemispheric fissure; (6) thickness of corpus callosum.
comprised children with “unspecific signs,” for example, mild muscular hypotonia and brisk reflexes, but not fulfilling the Surveillance of Cerebral Palsy criteria for diagnosis of CP (Surveillance of Cerebral Palsy in Europe Working Group 2000). The Bayley Scales of Infant and Toddler Development-III (BSITD-III) (Bayley 2006) were performed on the same day as the neurologic examination, assessing cognitive, language and motor development. The cutoff level for severe delay (<2 SD) was applied to the mean values of the regional term-born control group, which had a mean cognitive CS of 104 (SD § 9.6, i.e., severe delay <85), a mean language CS of 108 (SD § 11.7, i.e., severe delay <85) and a mean motor CS of 117 (SD §14.7, i.e., severe delay <88) (Ski€ old et al. 2012). Statistics Statistical analysis was performed with PASW Statistics Software 21.0 (IBM, Armonk, NY, USA). Infants were divided based on the MRI scoring categories for WM and GM. Similarly, infants were categorized based on cUS scoring groups. Furthermore, infants were grouped based on neurologic status (normal, abnormal or unspecific signs) and developmental outcomes (no or severe delay: i.e., a CS >2SD below the mean of the control group).
Student’s t-test was used for continuous variables, and Pearson’s x2-test and Fisher’s exact test were used for dichotomous data. Non-parametric tests were used for ranked data and unequal group sizes. Spearman’s rank coefficient was used to assess correlations between the CSs of the two scoring systems. A p value < 0.05 was chosen as the cutoff level for significance. RESULTS Eighty-four preterm infants were included in the study. Patient characteristics and perinatal factors are summarized in Table 2. Neuroimaging at TEA Brain abnormalities on MRI and cUS in the sample in the present study did not differ from data previously reported for the whole Stockholm cohort (Horsch et al. 2010; Ski€old et al. 2010, 2012). Incidence of brain abnormalities on cUS and WM injury on MRI is illustrated in Figure 2. Grey matter abnormalities on MRI were seen in 4 patients. There was overall good agreement between the CSs on cUS and MRI, both between total cUS score and WM abnormality score (Spearman’s r = 0.51, p < 0.001) and between total cUS score and GM abnormality score
790
Ultrasound in Medicine & Biology
Volume 45, Number 3, 2019
Table 2. Demographic data and neonatal clinical characteristics*
Gestational age, wk + d Birth weight, g Gender, males Time on CPAP, d Time on mechanical ventilation, d Bronchopulmonary dysplasia, oxygen required at 36 wk GA Retinopathy of prematurity, treated Patent ductus arteriosus, surgical ligation Necrotizing enterocholitis, Bells grade 23 No intraventricular haemorrhagey Intraventricular haemorrhage, grade 12y Intraventricular haemorrhage, grade 3y Parenchymal haemorrhagic infarctiony Periventricular leukomalaciay
All infants (n = 84)
Attending FU at age 30 mo (n = 66)
FU info from clinical records (n = 18)
P Value
(25 + 4) § 1 ([23 + 1][26 + 6]) 808 § 153 (4941161) 45/84 (54%) 39 § 11 (1964) 13 § 13 (055) 33/84 (39%)
(25 + 4) § 1 ([23+1][26+6]) 826 § 147 (4941161) 34/66 (52%) 38 § 11 (1964) 13 § 13 (055) 25/66 (38%)
(25 + 3) § 1 ([23 + 3][26+6]) 744 § 162 (4991001) 11/18 (61%) 40 § 10 (2060) 12 § 12 (036) 8/18 (40%)
ns ns ns ns ns ns
12/84 (14%) 23/84 (27%)
8/66 (12%) 21/66 (32%)
4/18 (22%) 2/18 (11%)
ns ns
5/84 (6%)
5/66 (8%)
0
ns
46/84 (55%) 28/84 (33%)
33/66 (50%) 23/66 (35%)
13/18 (72%) 5/18 (28%)
ns ns
4/84 (5%)
4/66 (6%)
0
ns
7/84 (8%)
7/66 (11%)
0
ns
3/84 (4%)
2/66 (3%)
1/18 (6%)
ns
FU = follow-up; CPAP = continuous positive airway pressure; GA = gestational age; ns = nonsignificant. * Values are expressed as number/total (%) or mean § standard deviation (range). y Diagnosis from routine cranial ultrasound examinations performed repeatedly during the hospital stay.
(Spearman’s r = 0.58, p < 0.001). Of the 34 infants with an entirely normal cUS, 68% (23/34) had normal and 32% (11/34) had mild WM abnormalities on MRI, none had moderate or severe WM abnormalities and none had abnormal GM. The two infants with severe WM abnormalities on MRI were identified as having moderate or severe abnormalities on cUS. Neurodevelopmental outcome At age 30 mo, 79% (66/84) of the preterm-born children attended follow-up. Of these, 7 had incomplete
BSITD-III assessments. For preterm children who did not attend the research follow-up (18/84), medical records from routine clinical follow-up (mean age at follow-up = 24 mo, SD: §8, range: 1136) were reviewed for information regarding neurologic status. This group did not differ from the group of children that did attend follow-up with respect to perinatal characteristics, neuroimaging status or neurologic status (Table 2). On the BSITD-III, the mean cognitive CS of the preterm study sample was 95 (SD: §8.3, range: 65120), the mean language CS was 97 (SD: §14,
Fig. 2. Abnormalities on magnetic resonance imaging and cranial ultrasound. cUS = cranial ultrasound; MRI = magnetic resonance imaging; WM = white matter.
€ New scoring system for cranial US in preterm infants B. SKIOLD et al.
791
Table 3. Neurological status
Normal Unspecific signs Abnormal, cerebral palsy
All infants (n = 84)
Attending FU at age 30 mo (n = 66)
FU info from clinical records (n = 18)
56/84 (66%) 24/84 (29%) 4/84 (5%)
42/66 (64%) 20/66 (30%) 4/66 (6%)
14/18 (78%) 4/18 (22%) 0
P Value ns ns ns
FU = follow-up; ns = nonsignificant.
range: 53124) and the mean motor CS was 104 (SD: §17, range: 45130). Two children (3%) had severe cognitive delay, 8 children (13%) had severe language delay and 8 children (13%) had severe motor delay (Table 3). Four infants (6%) fulfilled the criteria for CP (Tables 3 and 4). Associations between cUS scoring groups and outcome For infants with either moderate or severe abnormalities on cUS, associations were observed with CP (Fisher’s exact test, p = 0.007) and with severe cognitive delay (Fisher’s exact test, p = 0.015); other investigated associations (unspecific signs and BSITD-III language and motor CSs) were not significant (Table 5). Of the 4 infants with severe abnormalities, 2 developed CP, and 1, severe cognitive delay. With respect to individual cUS items, cysts were associated with unspecific neurologic signs, CP and severe cognitive delay (Pearson’s x2, p < 0.001); and parieto-occipital WM loss was associated with severe cognitive delay (Pearson’s x2, p < 0.001). The negative predictive values of a normal or mildly abnormal TEA cUS regarding normal
neurologic status and cognitive outcome were 98% and 100%, respectively (Table 6). One infant that developed CP had a normal CUS and MRI. Associations between MRI scoring groups and outcome For infants with either moderate or severe WM abnormalities on MRI, associations were seen both with CP (Fisher’s exact test, p = 0.001) and with severe cognitive delay (Fisher’s exact test, p = 0.005); other investigated associations (unspecific neurologic signs and BSITD-III language and motor CSs) were not significant (Table 5) (previously published [Ski€old et al. 2012]). Both infants with severe WM abnormalities developed CP and severe cognitive delay. When infants with mild and moderate WM abnormalities were pooled together, no associations with neurologic status or developmental outcome were seen. Abnormal GM (irrespective of WM abnormalities) was not associated with CP or severe developmental delay. Adding “abnormal GM” to the pooled analyses of infants with moderate and severe WM abnormalities on MRI did not alter the associations with outcome. When
Table 4. Imaging findings and outcome in infants with cerebral palsy BSITD-III Gender F M F M
GA
Magnetic resonance imaging
Cranial ultrasound
Cognitive composite
Motor composite
GMFCS
26 + 5 25 + 4 26 + 4 25 + 1
No WMAs Normal GM Severe WMAs Normal GM Severe WMAs Abnormal GM Moderate WMAs Normal GM
No abnormalities Moderate abnormalities Severe abnormalities Severe abnormalities
105 80 65 95
85 103 45 97
I I IV I
BSITD-III = Bayley Scales of Infant and Toddler Development-III; GA = gestational age; GM = grey matter; GMFCS = Gross Motor Function Classification System; WMAs = white matter abnormalities.
Table 5. Associations between scoring of imaging findings and outcome Scoring results at term-equivalent age MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities
Outcome at age 30 mo
p value (Fisher’s exact test)
Cerebral palsy Cerebral palsy Severe cognitive delay, regional cutoff* Severe cognitive delay, regional cutoff*
0.001 0.007 0.005 0.015
CUS = cranial ultrasound; MRI = magnetic resonance imaging; WMAs = white matter abnormalities. * Bayley Scales of Infant and Toddler Development-III cognitive composite score 2SD below the mean of matched regional control infants (Stockholm).
792
Ultrasound in Medicine & Biology
Volume 45, Number 3, 2019
Table 6. MRI and CUS scores in predicting cerebral palsy and/or severe cognitive delay Outcome*
Sensitivity
Specificity
PPV
NPV
p value (Fisher’s exact test)
Cerebral palsy Cerebral palsy Severe cognitive delay regional cutoffy Severe cognitive delay regional cutoffy
75% 75% 100% 100%
97% 90% 95% 90%
60% 33% 40% 25%
98% 98% 100% 100%
0.001 0.007 0.005 0.015
Assessment MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities MRI: Moderate to severe WMAs CUS: Moderate to severe abnormalities
CUS = cranial ultrasound; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; WMAs = white matter abnormalities. * Predictive values for the combination of cerebral palsy and/or severe cognitive delay (irrespective of cutoff) were the same as predictive values for cerebral palsy on its own (data not shown). * Bayley Scales of Infant and Toddler Development-III composite score 2SD below the mean of matched regional control infants (Stockholm).
investigating individual scoring items, WM reduction, delayed myelination and cysts were associated with unspecific neurologic signs, CP and severe cognitive delay (Pearson’s x2, p < 0.001). Abnormal signal in the cortical or deep GM was associated with severe cognitive delay (Pearson x2, p < 0.001). The two scoring systems are compared with respect to diagnostic accuracy in Table 6. Cerebellar haemorrhage Six infants had small cerebellar haemorrhages. All 6 were only detected on MRI. None was detected by cUS via the anterior fontanel. DISCUSSION We describe a novel comprehensive scoring system for cUS at TEA that allows quantification of overall brain injury comparable to that of established MRI scoring systems. This cUS scoring system covers the most relevant prematurity-related brain injuries such as residuals and sequelae of haemorrhages, periventricular infarctions, post-haemorrhagic hydrocephalus, WMI and brain atrophy. The TEA cUS scores exhibited good agreement with TEA MRI scores and reached similar predictive values for severe impairments. Magnetic resonance imaging and cUS are the two most important neuroimaging modalities in newborn infants. Neonatal MRI has been increasingly used in the last two decades and has tremendously enhanced our understanding of neonatal brain development and injury (Inder et al. 2003; Keunen et al. 2016; Woodward et al. 2006). Nevertheless, MRI remains an expensive method that is time and staff consuming. Access to MRI is limited in many countries and centres; therefore, patients who undergo MRI need to be carefully selected. The increasing use of MRI at term age for screening of extremely preterm infants remains controversial. In 2015, the American Academy of Pediatrics’ (AAP) section on Neonatal Perinatal Medicine undertook an expert consensus building process to identify “use of unnecessary
tests and treatments that contribute to health care waste” as part of the AAP Choosing Wisely campaign. The authors concluded that routine screening with TEA MRIs in preterm infants is related to, but not proven to improve outcomes (Ho et al. 2015). This was in brief supported by a recent randomized trial by Edwards et al. (2018) investigating predictive power, maternal anxiety and costs of TEA neuroimaging (cUS and MRI), concluding that MRI “increased costs and provided only modest benefits.” On the other hand, experts have opposed the AAP Choosing Wisely statement (Cheong et al. 2018). This controversy underlines the need for thorough research in this field and careful assessment of cost-effective alternatives to routine TEA MRI in preterm infants. Early studies that compared MRI and cUS in the prediction of outcome of extremely preterm infants found TEA MRI to be superior to cUS (Maalouf et al. 2001; Mirmiran et al. 2004; Woodward et al. 2006). In many of these studies, cUS was performed during the first week after birth and not at term age, using a variety of cUS diagnoses (IVH, PHVD, periventricular leukomalacia) to be compared with data from complex MRI scoring systems. Later studies using serial cUS and (near-) term data reported improved accuracy of cUS in the prediction of outcome (de Vries et al. 2004; Edwards et al. 2018). Hintz et al. (2015) reported that adverse cUS findings on the late near-term cUS scan were associated with outcomes at 1822 mo corrected age and reached predictive values in the range of TEA MRI (Hintz et al. 2015). Interestingly, although a detailed MRI scoring system that was developed for white and GM abnormalities has been widely used, refined and updated over the years (Inder et al. 2003; Kidokoro et al. 2013; Woodward et al. 2006), no detailed or systematic scoring system for cUS images has, to our knowledge, yet been published. The aim of the proposed cUS scoring was to create a quantitative assessment tool that reliably detects clinically relevant sequelae of preterm brain injury and impaired brain development at TEA. The examination does not differ significantly from a clinical routine scan
€ et al. New scoring system for cranial US in preterm infants B. SKIOLD
(usually takes approximately 10 min for an experienced user) as only standard projections according to Levene et al. (1985) are used. Cranial US images are stored, and scoring may be performed on a later occasion, that is, not interfering with clinical work on the ward, which also facilitates training of less experienced cUS users. The inter-observer agreement in the present study was high. Similar to the updated MRI scoring system of Kidokoro et al. (2013), the present cUS scoring includes a combination of subjectively scored items and objective metrics, resulting in a composite (“global”) brain abnormality score. With this novel scoring system, not only adverse findings but also normal TEA cUSs and mild abnormalities are clearly defined and easily reproducible for users. This is important as the negative predictive values regarding all investigated outcomes were high when TEA cUS scoring was normal or revealed only mild abnormalities. This underlines the value of TEA cUS for screening infants that highly unlikely will profit from further investigation with MRI. We believe that this proposed cUS scoring system can be integrated into clinical routine, but also enables uniform structured assessment of brain injuries in the research setting, such as comparison across cohorts. The low rates of severe brain abnormalities and severe impairments in our cohort are important limitations to this study. Although it is of course gratifying that the cohort is doing well, it restricts the prognostic value of the tool, a problem experienced by others in similar settings (Brouwer et al. 2017). Therefore, validation of the presented cUS scoring system in a cohort with higher incidences of both brain injury and impaired outcomes is warranted. Another limitation is that cUS images were acquired only via the anterior fontanel, which most likely contributed to the low rate of detection of cerebellar injury. In future studies, additional acoustic windows such as the mastoid fontanel, posterior fontanel, temporal window and foramen magnum should be used to improve visualization of areas less accessible via the anterior fontanel. Moreover, a comparison with the revised version of the InderKidokoro MRI scoring system (Kidokoro et al. 2013) would also be of great interest. Another major limitation is the relatively short follow-up, which might naturally underestimate more subtle, later-emerging cognitive impairments. We believe that the major strength of the present scoring system is that it assesses the full spectrum of prematurity-related brain injury and uses weighted subscores that reflect the relevance for outcome of each item. Moreover, the cUS scoring system is clearly structured and easy to use also for non-experts. The population-based design is another strength of our study, along with the large sample size of infants with a GA at birth
793
<27 wk. In addition, to increase accuracy of the outcome comparisons, we used a regional control group of term-born infants, as the third edition of the Bayley Scales has been reported to underestimate developmental delay (Anderson et al. 2010; Vohr et al. 2012). CONCLUSIONS This article has described a novel comprehensive scoring system for cUS at TEA, allowing quantification of preterm brain injury comparable to that of established MRI scoring systems. It may thus enable uniform structured assessment of brain injuries in the research setting, such as comparison across cohorts. Moreover, we believe this cUS scoring system may be integrated into study protocols and clinical routine. The negative predictive values regarding all investigated outcomes were high; that is, cUS scoring may potentially be used to screen for infants that are unlikely to profit from further investigation with MRI. Acknowledgments—The present study was supported by the following grants: European Society for Pediatric Research Young Investigator Exchange Program, Jerring Foundation, S€allskapet Barnavard, M€arta och Gunnar V Philipsson Foundation, Swedish Medical Research Council, Swedish Brain Foundation (Hj€arnfonden), Foundation Samar iten, Free Masonry Foundation Barnhuset in Stockholm, Ake Wiberg Foundation, Jeansson Foundation and Karolinska University Hospital. These sources financed the research activity of S.H. and B.S. We thank Mikael Mosskin for his participation in the magnetic resonance imaging scoring sessions, helpful advice and support of this study.
REFERENCES Anderson PJ, De Luca CR, Hutchinson E, Roberts G, Doyle LW. Victorian Infant Collaborative Group. Underestimation of developmental delay by the new Bayley-III Scale. Arch Pediatr Adolesc Med 2010;164:352–356. Bayley N. Bayley scales of infant and toddler development. 3rd ed. Harcourt: Assessment Inc; 2006. Brouwer A, Greoendaal F, van Haastert IL, Rademaker K, Hanlo P, De Vries L. Neurodevelopmental outcome of preterm infants with severe intraventricular hemorrhage and therapy for post-haemorrhagic ventricular dilation. J Pediatr 2008;152:648–654. Brouwer MJ, Kersbergen KJ, van Kooij BJM, Benders MJNL, van Haastert IC, Koopman-Esseboom C, Neil JJ, de Vries LS, Kidokoro H, Inder TE, Groenendaal F. Preterm brain injury on term-equivalent age MRI in relation to perinatal factors and neurodevelopmental outcome at two years. PLoS One 2017;12: e0177128. Cheong JLY, Miller SP. Imaging the neonatal brain in the 21st century: Why, when and how? Arch Dis Child Fetal Neonatal Ed 2018;103: F4–F5. De Vries LS, Van Haastert IL, Rademaker KJ, Koopman C, Groenendaal F. Ultrasound abnormalities preceding cerebral palsy in highrisk preterm infants. J Pediatr 2004;144:815–820. Ecury-Goossen GM, Camfferman FA, Leijser LM, Govaert P, Dudink J. State of the art cranial ultrasound imaging in neonates. J Vis Exp 2015;96:e52238. Edwards AD, Redshaw ME, Kennea N, Rivero-Arias O, GonzalesCinca N, Nongena P, Ederies M, Falconer S, Chew A, Omar O, Hardy P, Harvey ME, Eddama O, Hayward N, Wurie J, Azzopardi D, Rutherford MA, Counsell S. ePrime Investigators. Effect of MRI on preterm infants and their families: A randomised trial with
794
Ultrasound in Medicine & Biology
nested diagnostic and economic evaluation. Arch Dis Child Fetal Neonatal Ed 2018;103:F15–F21. Hintz SR, Barnes PD, Bulas D, Slovis TL, Finer NN, Wrage LA, Das A, Tyson JE, Stevenson DK, Carlo WA, Walsh MC, Laptook AR, Yoder BA, Van Meurs KP, Faix RG, Rich W, Newman NS, Cheng H, Heyne RJ, Vohr BR, Acarregui MJ, Vaucher YE, Pappas A, Peralta-Carcelen M, Wilson-Costello DE, Evans PW, Goldstein RF, Myers GJ, Poindexter BB, McGowan EC, Adams-Chapman I, Fuller J, Higgins RD. SUPPORT Study Group of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Neuroimaging and neurodevelopmental outcome in extremely preterm infants. Pediatrics 2015;135:e32–e42. Ho T, Dukhovny D, Zupancic JA, Goldmann DA, Horbar JD, Pursley DM. Choosing wisely in newborn medicine: Five opportunities to increase value. Pediatrics 2015;136:e482–e489. Horsch S, Muentjes C, Franz A, Roll C. Ultrasound diagnosis of brain atrophy is related to neurodevelopmental outcome in preterm infants. Acta Paediatr 2005;94:1815–1822. Horsch S, Skiold B, Hallberg B, Nordell B, Nordell A, Mosskin M, Lagercrantz H, Aden U, Blennow M. Cranial ultrasound and MRI at term age in extremely preterm infants. Arch Dis Child Fetal Neonatal Ed 2010;95:F310–F314. Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature of the cerebral abnormalities in the premature infant: A qualitative magnetic resonance imaging study. J Pediatr 2003;143:171–179. Janvier A, Barrington K. Trying to predict the future of ex-preterm infants: Who benefits from a brain MRI at term? Acta Paediatr 2012;101:1016–1017. Keunen K, Isgum I, van Kooij BJ, Anbeek P, van Haastert IC, Koopman-Esseboom C, Fieret-van Stam PC, Nievelstein RA, Viergever MA, de Vries LS, Groenendaal F, Benders MJ. Brain volumes at term-equivalent age in preterm infants: Imaging biomarkers for neurodevelopmental outcome through early school age. J Pediatr 2016;172:88–95. Kidokoro H, Neil JJ, Inder TE. New MR imaging assessment tool to define brain abnormalities in very preterm infants at term. AJNR Am J Neuroradiol 2013;34:2208–2214. Levene MI, Williams JL, Fawer CL. Ultrasound of the infant brain. Clin Dev Med 1985;92:9–33. Maalouf EF, Duggan PJ, Counsell SJ, Rutherford MA, Cowan F, Azzopardi D, Edwards AD. Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 2001;107:719–727. Mirmiran M, Barnes PD, Keller K, Constantinou JC, Fleisher BE, Hintz SR, Ariagno RL. Neonatal brain magnetic resonance imaging before discharge is better than serial cranial ultrasound in predicting
Volume 45, Number 3, 2019 cerebral palsy in very low birth weight preterm infants. Pediatrics 2004;114:992–998. Pierrat V, Marchand-Martin L, Arnaud C, Kaminski M, Resche-Rigon M, Lebeaux C, Bodeau-Livinec F, Morgan AS, Goffinet F, Marret S, Ancel PY. EPIPAGE-2 Writing Group. Neurodevelopmental outcome at 2 years for preterm children born at 22 to 34 weeks’ gestation in France in 2011: EPIPAGE-2 cohort study. BMJ 2017;358:j3448. Serenius F, K€allen K, Blennow M, Ewald U, Fellman V, Holmstr€om G, Lindberg E, Lundqvist P, Marsal K, Norman M, Olhager E, Stigson L, Stjernqvist K, Vollmer B, Str€omberg B. EXPRESS Group. Neurodevelopmental outcome in extremely preterm infants at 2.5 years after active perinatal care in Sweden. JAMA 2013;309:1810–1820. Serenius F, Ewald U, Farooqi A, Fellman V, Hafstr€om M, Hellgren K, Marsal K, Ohlin A, Olhager E, Stjernqvist K, Str€omberg B, Aden U, K€allen K. Extremely Preterm Infants in Sweden Study Group. Neurodevelopmental outcomes among extremely preterm infants 6.5 years after active perinatal care in Sweden. JAMA Pediatr 2016;170:954–963. Skiold B, Horsch S, Hallberg B, Engstr€om M, Nagy Z, Mosskin M, Blennow M, Aden U. White matter changes in extremely preterm infants, a population-based diffusion tensor imaging study. Acta Paediatr 2010;99:842–849. Skiold B, Vollmer B, B€ohm B, Hallberg B, Horsch S, Mosskin M, Lagercrantz H, Aden U, Blennow M. Neonatal magnetic resonance imaging and outcome at age 30 months in extremely preterm infants. J Pediatr 2012;160:559–566. Surveillance of Cerebral Palsy in Europe (SCPE) Working Group. Surveillance of cerebral palsy in Europe: A collaboration of cerebral palsy surveys and registers. Dev Med Child Neurol 2000;42:816–824. Van’t Hooft J, van der Lee JH, Opmeer BC, Aarnoudse-Moens CS, Leenders AG, Mol BW, de Haan TR. Predicting developmental outcomes in premature infants by term equivalent MRI: Systematic review and meta-analysis. Syst Rev 2015;4:71. Vohr BR, Stephens BE, Higgins RD, Bann CM, Hintz SR, Das A, Newman JE, Peralta-Carcelen M, Yolton K, Dusick AM, Evans PW, Goldstein RF, Ehrenkranz RA, Pappas A, Adams-Chapman I, Wilson-Costello DE, Bauer CR, Bodnar A, Heyne RJ, Vaucher YE, Dillard RG, Acarregui MJ, McGowan EC, Myers GJ, Fuller J. Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Are outcomes of extremely preterm infants improving? Impact of Bayley assessment on outcomes. J Pediatr 2012;161:222–228 e3. Wezel-Meijler G, de Vries LS. Cranial ultrasound—Optimizing utility in the NICU. Curr Pediatr Rev 2014;10:16–27. Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med 2006;355:685–694.