Pediatr Radiol DOI 10.1007/s00247-014-2945-9

ORIGINAL ARTICLE

Prognostic value of diffusion-weighted imaging summation scores or apparent diffusion coefficient maps in newborns with hypoxic-ischemic encephalopathy Francesca Cavalleri & Licia Lugli & Marisa Pugliese & Roberto D’Amico & Alessandra Todeschini & Elisa Della Casa & Claudio Gallo & Rossella Frassoldati & Fabrizio Ferrari

Received: 13 November 2013 / Revised: 23 January 2014 / Accepted: 19 February 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Background The diagnostic and prognostic assessment of newborn infants with hypoxic-ischemic encephalopathy (HIE) comprises, among other tools, diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) maps. Objective To compare the ability of DWI and ADC maps in newborns with HIE to predict the neurodevelopmental outcome at 2 years of age. Materials and methods Thirty-four term newborns with HIE admitted to the Neonatal Intensive Care Unit of Modena University Hospital from 2004 to 2008 were consecutively enrolled in the study. All newborns received EEG, conventional MRI and DWI within the first week of life. DWI was analyzed by means of summation (S) score and regional ADC measurements. Neurodevelopmental outcome was assessed with a standard 1–4 scale and the Griffiths Mental Developmental Scales - Revised (GMDS-R). Results When the outcome was evaluated with a standard 1–4 scale, the DWI S scores showed very high area under the curve (AUC) (0.89) whereas regional ADC measurements in specific

subregions had relatively modest predictive value. The lentiform nucleus was the region with the highest AUC (0.78). When GMDS-R were considered, DWI S scores were good to excellent predictors for some GMDS-R subscales. The predictive value of ADC measurements was both region- and subscalespecific. In particular, ADC measurements in some regions (basal ganglia, white matter or rolandic cortex) were excellent predictors for specific GMDS-R with AUCs up to 0.93. Conclusions DWI S scores showed the highest prognostic value for the neurological outcome at 2 years of age. Regional ADC measurements in specific subregions proved to be highly prognostic for specific neurodevelopmental outcomes. Keywords Hypoxic-ischemic encephalopathy . Diffusion-weighted imaging . Apparent diffusion coefficient values . Receiver operating characteristic analysis . Magnetic resonance imaging . Newborns

Introduction F. Cavalleri (*) : A. Todeschini Neuroradiology Unit, Department of Neuroscience, Nuovo Ospedale Civile S. Agostino Estense di Modena, Azienda Unità Sanitaria Locale di Modena, via Giardini 1355, Modena 41100, Italy e-mail: [email protected] L. Lugli : M. Pugliese : E. Della Casa : C. Gallo : R. Frassoldati : F. Ferrari Institute of Pediatrics and Neonatal Medicine and NICU, Modena University Hospital, Modena, Italy R. D’Amico Department of Clinical and Diagnostic Medicine and Public Health, University of Modena and Reggio Emilia, Modena, Italy

Hypoxic-ischemic encephalopathy (HIE) is a leading cause of mortality and morbidity in the neonatal period and occurs in 1 to 2 per 1,000 live births [1]: 10–15% of the newborn infants with HIE will die, 10–15% will develop cerebral palsy and up to 40% will have other neurodevelopmental disabilities. Different patterns of hypoxic-ischemic (HI) brain injury may be observed depending on three primary factors: brain maturity at the time of insult, severity and duration of HI insult. In the full-term neonate, a protracted, mild to moderate HI insult results in injury to the parasagittal corticosubcortical frontoparietal and occipital regions (watershed areas) whereas in the event of severe HI insult, often preceded by a sentinel event (e.g., uterine rupture, cord prolapse or placental

Pediatr Radiol

abruption), the vulnerable zones are the basal ganglia, thalamus, hippocampus, dorsal brainstem, corticospinal tracts and perirolandic cortex (basal ganglia and thalami or diffuse patterns of injury) [2–4]. In recent years, neonatal brain conventional MRI and diffusion-weighted imaging (DWI) in the first days of life have been widely used to study the effect of perinatal HI insult and the patterns of brain damage have been accurately described. It has been repeatedly shown that DWI detects HI injury at an earlier stage than conventional MRI. All authors agree on the sensitivity of DWI for early diagnosis of HIE and on the importance of performing DWI during the first week [5–9], though DWI may underestimate the extent of final basal ganglia and thalamic injury [10]. On the other hand, the time lag between the HI insult and the MRI study should not exceed 7–10 days since the DWI findings are affected by the pseudonormalisation phenomenon [7, 11]. The MRI results are powerful predictors of the neurodevelopmental outcome in HIE. Therefore, in association with the results of the early neurological assessment, MRI is a major element for selecting those infants who need a strict neurological follow-up and early intervention strategies. The prognostic relevance of apparent diffusion coefficient (ADC) maps is more discussed, though in general decreased ADC values in specific regions (such as the lentiform nucleus and the posterior limb of the internal capsule) [12, 13] seem to be well correlated with increased severity of the long-term outcome. Overall, a recent meta-analysis indicated that MRI is a biomarker prognosticator of neurodevelopmental outcome in HIE [14]. Yet, the prognostic value of specific diagnostic tools and imaging examinations performed in the neonatal period is still under discussion and awaits further investigation. The primary aim of the study was to compare the ability of several current diagnostic tools performed in the early neonatal period in HIE newborns to predict the neurodevelopmental outcome at 2 years of age. As diagnostic approaches we considered routine DWI as well as measurements of regional ADC values. To assess neurodevelopment outcome we used a simple four-level score for the neurological outcome and the Griffiths Mental Developmental Scales - Revised (GMDS-R) [15] at 2 years of age. A second aim was to test in our patients the hypothesis put forward by Liauw et al. [16] that ADC values in basal ganglia and brainstem that have normal appearance on DWI are good predictors of neurological outcome, while ADC values in visibly abnormal regions on DWI did not have any predictive value.

Materials and methods The study was approved by the Ethical Committee of Modena Province (protocol n. 240.2009).

Subjects We retrospectively collected data about all term newborns (gestational age ≥37 weeks of gestations, n=34) with neonatal HIE consecutively admitted to the Neonatal Intensive Care Unit of Modena University Hospital from 2004 to 2008. Inclusion criteria required presentation with neonatal encephalopathy (abnormal tone patterns, feeding difficulties, altered alertness with or without seizures) and at least two of the following: Apgar score ≤5 at 1 min or ≤7 at 5 min; umbilical artery pH ≤7.0 or base deficit ≥16 mmol/l, abnormal cardiotocography or meconium-stained amniotic fluid. Exclusion criteria were congenital malformations, chromosomal abnormalities, congenital infections and defined metabolic syndromes. Magnetic resonance imaging Infants underwent MRI examinations as part of a detailed neurological evaluation conducted in our unit. All infants underwent conventional MRI and DWI examinations within the first week of life and a second conventional MRI within the first 2 years of life. Infants were sedated with oral chloral hydrate (50 mg/Kg) and monitored with pulse oximetry during the scan. MRI was performed using a 1.5-T scanner (Intera; Philips Medical Systems, Best, The Netherlands). The mean age at neonatal MRI examination was 3.9±1.7 days (range: 1– 7 days). Conventional images were obtained with T1weighted (SE 600/20), T2-weighted (FSE 6,000/150) and inversion recovery (IR 3,500/15/950). DWI was investigated in n = 3 non-collinear directions with repetition times 3,500 ms, echo time 90 ms and b factor 700 s/mm2. For all sequences, FOV=200 mm, matrix=256×256, and spatial resolution = 0.781 × 0.781 mm/pixel. The ADC map was derived with monoexponential fit. MRI images were independently assessed for abnormal signal intensities by two experienced observers (F.C. and A.T., 13 and 10 years of experience, respectively), blinded to perinatal data and outcome. In case of disagreement, the images were re-examined by the two observers together to find a consensus. Consensus was reached in all cases. DWI abnormalities were scored according to the summation (S) scoring system of Barkovich et al. [2] that ranges from 0 to 9 (Table 1). DWI and ADC trace maps were visually assessed for abnormal signal, and ADC values within regions of interest (ROIs) were evaluated for each scan. ROIs were identified on the original T2-weighted images and visually matched and positioned on the b=0 image of the DWI acquisition. All the analysis was performed by the same operator

Pediatr Radiol Table 1 Scoring system (from Barkovich and coworkers [2]) Basal ganglia (BG) 0=Normal or isolated focal cortical infarct 1=Abnormal signal in thalamus 2=Abnormal signal in thalamus and lentiform nucleus 3=Abnormal signal in thalamus, lentiform nucleus and perirolandic cortex 4=More extensive involvement Watershed (W) 0=Normal 1=Single focal infarction 2=Abnormal signal in anterior or posterior watershed white matter 3=Abnormal signal in anterior or posterior watershed cortex and white matter 4=Abnormal signal in both anterior and posterior watershed zones 5=More extensive cortical involvement Summation (S) Arithmetic sum of BG and W

(FC) using the JHU Neonate Brain Atlas [17] as reference atlas. Circular, bilateral, 30-mm (20 mm2 for the posterior limb of the internal capsule) wide ROIs were identified on reference images and then transferred to ADC maps. ROIs included (Fig. 1): 1. Frontal cortex, anterior to the rolandic cortex near the vertex 2. Parietal cortex, posterior to the rolandic cortex near the vertex 3. Rolandic cortex near the vertex 4. Anterior white matter of the centrum semiovale 5. Posterior white matter of the centrum semiovale 6. Lentiform nucleus 7. Thalamus 8. Posterior limb of the internal capsule 9. Entorhinal cortex.

Note that, in order to be consistent, we chose to use a small circular ROI to sample the posterior limb of the internal capsule. Given the small size of this region, we cannot exclude that a small amount of lentiform nucleus or thalamus has been included in the measurement. For each region, ROIs were positioned in both brain sides. The mean ADC values of the right and left side were averaged for further analysis.

All infants were included in a strict neurological follow-up program up to 24 months that consisted of a detailed neurological evaluation, based on the protocol of Amiel-Tison and Grenier [18] and an extension of Touwen criteria [19], performed by two neonatologists experienced in follow-up studies (E.D. and C.G.), and the GMDS-R [15], performed by a developmental psychologist (M.P.). All infants with overt neurological signs in the first months were referred for early treatment. The neurodevelopmental outcome was evaluated at 24 months with a neurological examination and with the GMDS-R. The neurological examination was classified as normal (score=1) if there were no abnormal neurological signs, moderately abnormal (score=2) in case of mild neurological signs (tone abnormalities, clumsiness) or severely abnormal (score=3) in case of cerebral palsy. Infants who died in the first months of life received a score=4 [20]. The GMDS-R provides an overall developmental quotient and subquotient scores for five subscales. The subscales assess, respectively: Subscale A, locomotor (Lo, gross motor skills including the ability to balance and to coordinate and control movements); Subscale B, personal-social (PS, proficiency in the activities of daily living, level of independence and interaction with other children); Subscale C, language (La, receptive and expressive language); Subscale D, eye and hand coordination (EH, fine motor skills, manual dexterity and visual monitoring skills); Subscale E, performance (Pe, visuospatial skills including speed of working and precision).

Statistical analysis The prognostic ability of the different diagnostic approaches on the neurodevelopmental outcomes of neonatal HIE was assessed by means of receiver operating characteristic (ROC) analysis. Mean regional ADC values in different conditions were compared by non-parametric Kruskall-Wallis or MannWhitney U test. A probability level

Prognostic value of diffusion-weighted imaging summation scores or apparent diffusion coefficient maps in newborns with hypoxic-ischemic encephalopathy.

The diagnostic and prognostic assessment of newborn infants with hypoxic-ischemic encephalopathy (HIE) comprises, among other tools, diffusion-weighte...
5MB Sizes 0 Downloads 3 Views