Pj Lally

Diagnostic accuracy and limitations of post-mortem MRI for neurological abnormalities in fetuses and children.

AIM To compare the diagnostic accuracy of non-invasive cerebral post-mortem magnetic resonance imaging (PMMRI) specifically for cerebral and neurological abnormalities in a series of fetuses and children, compared to conventional autopsy. MATERIALS AND METHODS Institutional ethics approval and parental consent was obtained. Pre-autopsy cerebral PMMRI was performed in a sequential prospective cohort (n = 400) of fetuses (n = 277; 185 ≤ 24 weeks and 92 > 24 weeks gestation) and children <16 years (n = 123) of age. PMMRI and conventional autopsy findings were reported blinded and independently of each other. RESULTS Cerebral PMMRI had sensitivities and specificities (95% confidence interval) of 88.4% (75.5 to 94.9), and 95.2% (92.1 to 97.1), respectively, for cerebral malformations; 100% (83.9 to 100), and 99.1% (97.2 to 99.7) for major intracranial bleeds; and 87.5% (80.1 to 92.4) and 74.1% (68 to 79.4) for overall brain pathology. Formal neuropathological examination was non-diagnostic due to maceration/autolysis in 43/277 (16%) fetuses; of these, cerebral PMMRI imaging provided clinically important information in 23 (53%). The sensitivity of PMMRI for detecting significant ante-mortem ischaemic injury was only 68% (48.4 to 82.8) overall. CONCLUSIONS PMMRI is an accurate investigational technique for identifying significant neuropathology in fetuses and children, and may provide important information even in cases where autolysis prevents formal neuropathological examination; however, PMMRI is less sensitive at detecting hypoxic-ischaemic brain injury, and may not detect rarer disorders not encountered in this study.

Does Magnetic Resonance Brain Scanning at 3.0 Tesla Pose a Hyperthermic Challenge to Term Neonates

Next-generation 3-Tesla magnetic resonance (MR) scanners offer improved neonatal neuroimaging, but the greater associated radiofrequency radiation may increase the risk of hyperthermia. Safety data for neonatal 3-T MR scanning are lacking. We measured rectal temperatures continuously in 25 neonates undergoing 3-T brain MR imaging and observed no significant hyperthermic threat.

8.9 Microstructural Changes in Neonatal Encephalopathy Revealed with the Neurite Orientation Dispersion and Density Imaging (NODDI) Model

Background Although diffusion tensor imaging (DTI) fractional anisotropy (FA) is commonly used to quantify neural injury, it is non-specific and affected by a number of microstructural changes. Objective To examine alterations in white matter (WM) associated with neonatal encephalopathy (NE), and relate these to tangible biophysical changes using the neurite orientation dispersion and density imaging (NODDI) model. Design/Methods We recruited with parental consent consecutive encephalopathic neonates (Thompson score ≥6) admitted to Calicut Medical College, India over a 6 month period. At age <3 wk diffusion tensor magnetic resonance imaging (DTI, TR/TE = 2800 ms/94 ms, 20 directions, b = 0&1000 s/mm2, 1.8 × 1.8 × 5 mm3) was performed at 1.5T (Siemens Avanto). Sarnat encephalopathy stage (none, mild, moderate or severe) was allocated at day 3. DTI data were fitted to the NODDI model, generating maps of orientation dispersion index (ODI) and neurite density index (NDI). These were compared between infants grouped by encephalopathy severity using tract-based spatial statistics (TBSS). Results Fifty-four infants were recruited; 31 had usable data. The mean FA skeleton is shown in green (Figure 1a). Compared to normal/mild (n = 22) the moderate/severe encephalopathy group (n = 9) had significantly reduced WM FA (Figure 1b: red p < 0.05; yellow p < 0.01) and increased radial diffusivity (RD, Figure 1c). This corresponded to a decrease in NDI (Figure 1d), but not ODI (Figure 1e). Conclusions In this cohort, NODDI fitting indicates that microstructural changes in NE may be due to a reduced neurite density. Further work will establish whether these findings are consistent with those obtained from gold-standard multi-shell diffusion data. Abstract 8.9 Figure 1

‘Feed and wrap’ or sedate and immobilise for neonatal brain MRI?

Successful neonatal brain MRI relies on having a settled infant within the scanner to permit acquisition of the necessary sequences and good-quality images. Unsettledness may lead to incomplete or unsuccessful scans, and costly rescheduled scans with concomitant parental anxiety and inconvenience. Significant motion artefact may confound or preclude interpretation, leading to diagnostic errors.1 Use of premedication to assist neonatal MRI is controversial: routine sedation can be used safely and effectively in neonates2 though some report successful whole-body MRI without sedation.3 It is unclear how widely routine sedation premedication is practised. In October 2013, our centre (Norwich) electively introduced routine use of chloral hydrate sedation along with a body splint-immobilising device for neonatal MRI. We aimed to: (a) review our experience with MRI quality and success in epochs before and after introducing sedation and vacuum immobilisation, (b) determine current UK practices regarding use of sedation for neonatal brain MRI. We …

PFM.33 Estimating Maceration Severity Using Whole Body Magnetic Resonance T2 Relaxometry

Background Magnetic resonance (MR) imaging is an ideal modality to observe gross global changes in tissue structure, as is present with maceration. As tissue degrades, its MR transverse relaxation time (T2) should increase, with relaxometry methods enabling quantitative measurement of this. Objective To use T2 relaxometry to non-invasively classify maceration and compare this with the assessment of a pathologist. Design/Methods We performed post-mortem imaging in foetuses and neonates at 1.5T (Siemens Avanto). T2 relaxometry was performed using an eight-echo turbo spin echo sequence (TR = 2400 ms, TE = 44/88/132/176/220/264/308/352 ms), with even-echo images fitted to a mono-exponential decay function. Voxel-wise T2 values were generated from the model fit in each voxel to generate quantitative T2 maps, on which the lungs and liver were delineated (as shown in the figure). Mean T2 values in each region were evaluated against global pathology maceration scores (1 to 4; none to severe) based on findings of paediatric pathologists at autopsy. Abstract PFM.33 Figure Results T2 was increased in both the lungs and liver for maceration scores of 3 or 4, when compared to those with grades 1 and 2. However, T2 also correlated inversely to the gestational age at death. Conclusions Increasing maceration is associated with the prolongation of T2 values in the liver and lungs, but limitations arise due to conflation with gestational age effects, as younger foetuses were more macerated. T2 dependencies on gestational age are provided in the liver and lungs, allowing optimisation of post-mortem imaging for future studies.

PC.26 Feasibility of Magnetic Resonance Spectroscopy in Examining Thalamic Metabolite Concentrations in a Multi-Centre Study of Neonatal Encephalopathy

Background Proton magnetic resonance spectroscopy (MRS) has high prognostic value in hypoxic ischaemic encephalopathy (HIE), however its multi-centre application is limited by inconsistencies between scanners and protocols. N-acetylaspartate (NAA) is predominantly neuronal: cerebral NAA concentration may be a more reliable HIE-severity biomarker than lactate/NAA. Objective To quantify the inter-site and inter-subject variability of NAA concentration measurements. Design/Methods We recruited 5 healthy adult volunteers (aged 24–38, 2 male, 3 female) whom we scanned at 3 UK sites participating in a multi-centre neonatal-brain study (University Hospital, Coventry (UHC); Alder Hey Children’s Hospital, Liverpool (AH); University College Hospital, London (UCLH)). Thalamic NAA concentration was measured using the neonatal brain-water concentration reference protocol (15 × 15 × 15mm3 voxel; PRESS; water-suppressed TR/TE = 2s/288&60ms; TR/TE = 5s/60ms; non-water-suppressed TR = 10s, TE = 60/124/205/316/495/1000 ms). Spectra were post-processed in jMRUI and metabolites fitted with LCModel. Results One volunteer was unavailable for scanning at AH. The mean (SD) NAA concentrations across the remaining four subjects were 15.5(3.4)mmol/kg wet weight, 14.4(0.1) mmol/kg wet weight, and 15.2(0.1) mmol/kg wet weight at UHC, AH and UCLH respectively. This corresponds to a maximum inter-site group mean variation (range/mean) of 8%. The inter-subject variability of mean NAA concentration (SD/mean) was 10%. Conclusions The variation of thalamic NAA concentration between sites was 8% and the standard deviation across subjects was 10%, hence [NAA] may be suitable for multi-centre studies. Abstract PC.26 Figure

PFM.25 Assessment of visceral maceration using post-mortem magnetic resonance imaging in fetuses

Background Post-mortem magnetic resonance imaging (PM MRI) is increasingly used as an alternative for perinatal autopsy, however the artefacts related to maceration has not been described. Objective To compare the severity of maceration assessed by PM MRI with that estimated by histopathology in fetuses. Design/Methods We performed PM MRI in 75 fetuses using a 1.5 Tesla Siemens Avanto MR scanner (Erlangen, Germany) before conventional autopsy. PM MRI images were reported blinded to the clinical history and autopsy data using a maceration scale of 0–3 for 6 different visceral organs, and then added to make a composite score (0 to 18). The extent of maceration on histopathology was categorised from 1–4 (none to severe). Results PM MRI maceration score showed a good correlation with the histopathology maceration score (Figure A), however no relation was seen with the delivery (death) to PM MRI interval. Figure B and C shows PM MRI of fetuses with minimal and extensive maceration, respectively. Abstract PFM.25 Figure Conclusions PM MRI can be used to quantify the extent of maceration in fetuses. The time from delivery to PM MRI did not affect the MRI assessment of maceration if bodies are stored in the mortuary at low temperature.

PC.45 Quantification of N-Acetylaspartate Concentration in the Neonatal Brain: Initial Results from the Multi-Centre Marble Study

Background Early cerebral proton magnetic resonance spectroscopy (MRS) predicts medium-term outcomes in neonatal encephalopathy (NE). Metabolite peak-area ratios are most commonly used for prognosis, but conflate pathological information from different metabolites. N-acetylaspartate (NAA) is predominantly neuronal and neuronal loss should result in reduced NAA absolute-concentration ([NAA]). Thus, thalamic [NAA] should offer significant prognostic value but is difficult to measure in a clinical setting. We have established a protocol for multi-centre [NAA] measurement with the aim to use it as a surrogate biomarker in phase II clinical trials. Objective To investigate the feasibility and utility of [NAA] quantitation across multiple centres. Design/Methods We recruited cooled, term neonates with NE (by Sarnat grade) with parental consent across participating sites. Using various 3T scanners, thalamic MRS was performed aged 7 ± 4d (PRESS; water-suppressed TR = 2s/TE = 288/60 ms;TR/TE = 5s/60 ms; non-water-suppressed TR = 10s, TE = 60/124/205/316/495/1000 ms, ~30min acquisition). Spectra were post-processed in jMRUI and metabolite contributions determined with LCModel. [NAA] was calculated, correcting for T2 effects and cerebrospinal fluid partial volume. Results Ten cases had sufficient data for [NAA] quantification. Sarnat grading <6h identified infants with highest [NAA] (median (IQR) 10.0(9.7–10.3) mmol/kg wet weight) as mild NE, with a lower [NAA] range for moderate NE (7.0(5.0–7.8) mmol/kg wet weight). Conclusions [NAA] quantification is achievable in a multi-centre setting, and agrees with clinical NE grading during the therapeutic window. Follow-up examinations will allow comparison of neonatal [NAA] with later neurodevelopmental outcomes. Abstract PC.45 Figure

PC.106 Cerebral Injury and Early Childhood Neurodevelopmental Outcome following Neonatal Encephalopathy in a Middle-income Country

Background Although neonatal encephalopathy (NE), accounts for 1 million neonatal deaths annually in low-and middle-income countries (LMIC), underlying brain injury and long term outcomes are not well characterised in LMIC. Objective To examine cerebral injury (using magnetic resonance (MR) biomarkers), and early childhood outcomes after NE in a government hospital in India. Design/Methods We recruited 54 newborns (>36 wk and >1.8 kg) with NE (Thompson score ≥6) at age <6h, admitted to the neonatal unit at Calicut Medical College, India over 6 months. Conventional MRI (1.5T, Siemens Avanto), diffusion tensor MR imaging and thalamic proton MR spectroscopy (MRS) were performed aged <3 wk. Cerebral injury was graded and group-wise differences in white matter (WM) fractional anisotropy (FA) were examined using tract-based spatial statistics (TBSS). In survivors, adverse neurodevelopmental outcome at mean (SD) 3.4(0.2) years was defined as Bayley-III composite cognitive/motor score ≤85, slow head growth or cerebral palsy. Results MR data available from 44 cases showed evidence of acute perinatal injury. WM changes were seen in 40(91%), basal ganglia/thalamic (BGT) injury in 12(27%). Six infants died neonataly, 16(42%) and had adverse neurodevelopmental outcome. TBSS showed a reduction in FA with adverse neurological outcomes, and in those who had moderate/severe BGT or cortical injury. Conclusions Cerebral injury in this cohort appears to be of perinatal origin and may be amenable to treatment. Although NE stage and injury pattern was mild in the majority of infants, adverse outcomes were seen in 42% at 3½ years. Reduced WM FA was associated with adverse neurodevelopmental outcomes.

8.8 Neonatal Encephalopathy in the Cooling Therapy era – Preliminary Cerebral Magnetic Resonance results from the Marble Consortium

Background Although cerebral metabolic changes during neonatal encephalopathy (NE) have been well characterised using magnetic resonance spectroscopy (MRS) in single-centre studies, the widespread effect of therapeutic hypothermia is less clear. Objective To describe patterns of brain injury in a prospective multi-centre cohort of infants with NE who had rescue hypothermic neuroprotection. Design After MRS harmonisation on 3T MR scanners (Phillips, GE, Siemens), at six participating sites we performed magnetic resonance imaging (MRI) and MRS (single-voxel PRESS, 15 x 15 × x 15 mm3 thalamic voxel, TR = 2s/TE = 288 ms) on infants <2 wks who had whole-body cooling for NE at any of the MARBLE centres. We used jMRUI to process MRS data and calculate metabolite peak-area ratios. Results Forty-three infants were recruited; 11(26%) had mild, 26(60%) moderate and 6(14%) had severe NE (<6h Sarnat staging). Moderate/severe basal ganglia and cortex were detected by MRI in 4(9%) and 1(2%) cases respectively, and 15(34%) had moderate/severe white matter injury. Thirteen (30%) had normal MRI. Elevations in lactate (Lac)/N-acetylaspartate (NAA) were noted in 6(13%) cases, Lac/choline (Cho) in 5(11%) and Lac/ creatine (Cr) in 8(18%). MRS was normal in 28(65%) infants. Abstract 8.8 Figure 1 Figure 1 displays metabolite peak-area ratios across clinical NE stages. Mean Lac ratios were higher and NAA ratios lower (p < 0.05) between severe and both mild and moderate NE, but not between mild and moderate groups Conclusions In this predominantly moderate NE cohort, white matter injury was most common, and the majority had normal MRS metabolite ratios. Ongoing work seeks to establish whether these brain injury patterns are characteristic in the cooling therapy era.