Prediction of cognitive and motor outcome of preterm infants based on automatic quantitative descriptors from neonatal MR brain images

Using heart rate variability to predict neurological outcomes in preterm infants: a scoping review

Infants born preterm are at higher risk of neurological complications, including intraventricular haemorrhage and white matter injury. After discharge, these infants may experience adverse neurodevelopmental outcomes and exhibit lower educational attainment. Early detection of brain injury and accurate prediction of neurodevelopmental impairment would allow early intervention and support. Heart rate variability (HRV) describes the variation of time intervals between each subsequent heartbeat. HRV is controlled by the autonomic nervous system, which may be affected by hypoxia and compromised blood flow. While HRV has primarily been investigated in neonatal sepsis, the association between HRV, brain injury and neurodevelopmental outcomes in preterm infants is less established. The present scoping review examines the utility of HRV monitoring for predicting short-term and long-term neurological outcomes in preterm infants. Following systematic search of Medline, Embase, Web of Science and the Cochrane Library, 15 studies were included. Nine studies examined the relationship between HRV and brain injury, with all but two showed an association. Eight studies examined the relationship between HRV and long-term outcomes and all eight found an association. This scoping review suggests that decreased HRV in the neonatal period is associated with short- and long-term neurodevelopmental outcomes in preterm infants. IMPACT: Changes in heart rate variability correlate with the occurrence of intraventricular haemorrhage in preterm infants. A decrease in heart rate variability may precede the development of intraventricular haemorrhage. Alterations in heart rate variability correlate with long-term neurodevelopmental outcomes. Significant variability exists in metrics used in assessing heart rate variability.

Thalamic volume in very preterm infants: associations with severe brain injury and neurodevelopmental outcome at two years

Introduction

Several studies demonstrate the relationship between preterm birth and a reduced thalamus volume at term-equivalent age. In contrast, this study aims to investigate the link between the thalamic growth trajectory during the early postnatal period and neurodevelopment at two years of age.

Methods

Thalamic volume was extracted from 84 early MRI scans at postmenstrual age of 32.33 (± 2.63) weeks and 93 term-equivalent age MRI scans at postmenstrual age of 42.05 (± 3.33) weeks of 116 very preterm infants (56% male) with gestational age at birth of 29.32 (± 2.28) weeks and a birth weight of 1158.92 (± 348.59) grams. Cognitive, motor, and language outcomes at two years of age were assessed with Bayley Scales of Infant and Toddler Development Third Edition. Bivariate analysis was used to describe the clinical variables according to neurodevelopmental outcomes and multilevel linear regression models were used to examine the impact of these variables on thalamic volume and its relationship with neurodevelopmental outcomes.

Results

The results suggest an association between severe brain injury and thalamic growth trajectory (β coef = -0.611; p < 0.001). Moreover, thalamic growth trajectory during early postnatal life was associated with the three subscale scores of the neurodevelopmental assessment (cognitive: β coef = 6.297; p = 0.004; motor: β coef = 7.283; p = 0.001; language: β coeficient = 9.053; p = 0.002).

Discussion

These findings highlight (i) the impact of severe brain injury on thalamic growth trajectory during early extrauterine life after preterm birth and (ii) the relationship of thalamic growth trajectory with cognitive, motor, and language outcomes.

Developing a practical neurodevelopmental prediction model for targeting high-risk very preterm infants during visit after NICU: a retrospective national longitudinal cohort study

BACKGROUND

Follow-up visits for very preterm infants (VPI) after hospital discharge is crucial for their neurodevelopmental trajectories, but ensuring their attendance before 12 months corrected age (CA) remains a challenge. Current prediction models focus on future outcomes at discharge, but post-discharge data may enhance predictions of neurodevelopmental trajectories due to brain plasticity. Few studies in this field have utilized machine learning models to achieve this potential benefit with transparency, explainability, and transportability.

METHODS

We developed four prediction models for cognitive or motor function at 24 months CA separately at each follow-up visits, two for the 6-month and two for the 12-month CA visits, using hospitalized and follow-up data of VPI from the Taiwan Premature Infant Follow-up Network from 2010 to 2017. Regression models were employed at 6 months CA, defined as a decline in The Bayley Scales of Infant Development 3rd edition (BSIDIII) composite score > 1 SD between 6- and 24-month CA. The delay models were developed at 12 months CA, defined as a BSIDIII composite score < 85 at 24 months CA. We used an evolutionary-derived machine learning method (EL-NDI) to develop models and compared them to those built by lasso regression, random forest, and support vector machine.

RESULTS

One thousand two hundred forty-four VPI were in the developmental set and the two validation cohorts had 763 and 1347 VPI, respectively. EL-NDI used only 4-10 variables, while the others required 29 or more variables to achieve similar performance. For models at 6 months CA, the area under the receiver operating curve (AUC) of EL-NDI were 0.76-0.81(95% CI, 0.73-0.83) for cognitive regress with 4 variables and 0.79-0.83 (95% CI, 0.76-0.86) for motor regress with 4 variables. For models at 12 months CA, the AUC of EL-NDI were 0.75-0.78 (95% CI, 0.72-0.82) for cognitive delay with 10 variables and 0.73-0.82 (95% CI, 0.72-0.85) for motor delay with 4 variables.

CONCLUSIONS

Our EL-NDI demonstrated good performance using simpler, transparent, explainable models for clinical purpose. Implementing these models for VPI during follow-up visits may facilitate more informed discussions between parents and physicians and identify high-risk infants more effectively for early intervention.

The past, current, and future of neonatal intensive care units with artificial intelligence: a systematic review

Machine learning and deep learning are two subsets of artificial intelligence that involve teaching computers to learn and make decisions from any sort of data. Most recent developments in artificial intelligence are coming from deep learning, which has proven revolutionary in almost all fields, from computer vision to health sciences. The effects of deep learning in medicine have changed the conventional ways of clinical application significantly. Although some sub-fields of medicine, such as pediatrics, have been relatively slow in receiving the critical benefits of deep learning, related research in pediatrics has started to accumulate to a significant level, too. Hence, in this paper, we review recently developed machine learning and deep learning-based solutions for neonatology applications. We systematically evaluate the roles of both classical machine learning and deep learning in neonatology applications, define the methodologies, including algorithmic developments, and describe the remaining challenges in the assessment of neonatal diseases by using PRISMA 2020 guidelines. To date, the primary areas of focus in neonatology regarding AI applications have included survival analysis, neuroimaging, analysis of vital parameters and biosignals, and retinopathy of prematurity diagnosis. We have categorically summarized 106 research articles from 1996 to 2022 and discussed their pros and cons, respectively. In this systematic review, we aimed to further enhance the comprehensiveness of the study. We also discuss possible directions for new AI models and the future of neonatology with the rising power of AI, suggesting roadmaps for the integration of AI into neonatal intensive care units.

Early brain morphometrics from neonatal MRI predict motor and cognitive outcomes at 2-years corrected age in very preterm infants

Infants born very preterm face a range of neurodevelopmental challenges in cognitive, language, behavioural and/or motor domains. Early accurate identification of those at risk of adverse neurodevelopmental outcomes, through clinical assessment and Magnetic Resonance Imaging (MRI), enables prognostication of outcomes and the initiation of targeted early interventions. This study utilises a prospective cohort of 181 infants born <31 weeks gestation, who had 3T MRIs acquired at 29-35 weeks postmenstrual age and a comprehensive neurodevelopmental evaluation at 2 years corrected age (CA). Cognitive, language and motor outcomes were assessed using the Bayley Scales of Infant and Toddler Development - Third Edition and functional motor outcomes using the Neuro-sensory Motor Developmental Assessment. By leveraging advanced structural MRI pre-processing steps to standardise the data, and the state-of-the-art developing Human Connectome Pipeline, early MRI biomarkers of neurodevelopmental outcomes were identified. Using Least Absolute Shrinkage and Selection Operator (LASSO) regression, significant associations between brain structure on early MRIs with 2-year outcomes were obtained (r = 0.51 and 0.48 for motor and cognitive outcomes respectively) on an independent 25% of the data. Additionally, important brain biomarkers from early MRIs were identified, including cortical grey matter volumes, as well as cortical thickness and sulcal depth across the entire cortex. Adverse outcome on the Bayley-III motor and cognitive composite scores were accurately predicted, with an Area Under the Curve of 0.86 for both scores. These associations between 2-year outcomes and patient prognosis and early neonatal MRI measures demonstrate the utility of imaging prior to term equivalent age for providing earlier commencement of targeted interventions for infants born preterm.

Machine Learning Prediction Models for Neurodevelopmental Outcome After Preterm Birth: A Scoping Review and New Machine Learning Evaluation Framework

BACKGROUND AND OBJECTIVES

Outcome prediction of preterm birth is important for neonatal care, yet prediction performance using conventional statistical models remains insufficient. Machine learning has a high potential for complex outcome prediction. In this scoping review, we provide an overview of the current applications of machine learning models in the prediction of neurodevelopmental outcomes in preterm infants, assess the quality of the developed models, and provide guidance for future application of machine learning models to predict neurodevelopmental outcomes of preterm infants.

METHODS

A systematic search was performed using PubMed. Studies were included if they reported on neurodevelopmental outcome prediction in preterm infants using predictors from the neonatal period and applying machine learning techniques. Data extraction and quality assessment were independently performed by 2 reviewers.

RESULTS

Fourteen studies were included, focusing mainly on very or extreme preterm infants, predicting neurodevelopmental outcome before age 3 years, and mostly assessing outcomes using the Bayley Scales of Infant Development. Predictors were most often based on MRI. The most prevalent machine learning techniques included linear regression and neural networks. None of the studies met all newly developed quality assessment criteria. Studies least prone to inflated performance showed promising results, with areas under the curve up to 0.86 for classification and R2 values up to 91% in continuous prediction. A limitation was that only 1 data source was used for the literature search.

CONCLUSIONS

Studies least prone to inflated prediction results are the most promising. The provided evaluation framework may contribute to improved quality of future machine learning models.

Cognitive Outcome Prediction in Infants With Neonatal Hypoxic-Ischemic Encephalopathy Based on Functional Connectivity and Complexity of the Electroencephalography Signal

Impaired neurodevelopmental outcome, in particular cognitive impairment, after neonatal hypoxic-ischemic encephalopathy is a major concern for parents, clinicians, and society. This study aims to investigate the potential benefits of using advanced quantitative electroencephalography analysis (qEEG) for early prediction of cognitive outcomes, assessed here at 2 years of age. EEG data were recorded within the first week after birth from a cohort of twenty infants with neonatal hypoxic-ischemic encephalopathy (HIE). A proposed regression framework was based on two different sets of features, namely graph-theoretical features derived from the weighted phase-lag index (WPLI) and entropies metrics represented by sample entropy (SampEn), permutation entropy (PEn), and spectral entropy (SpEn). Both sets of features were calculated within the noise-assisted multivariate empirical mode decomposition (NA-MEMD) domain. Correlation analysis showed a significant association in the delta band between the proposed features, graph attributes (radius, transitivity, global efficiency, and characteristic path length) and entropy features (Pen and SpEn) from the neonatal EEG data and the cognitive development at age two years. These features were used to train and test the tree ensemble (boosted and bagged) regression models. The highest prediction performance was reached to 14.27 root mean square error (RMSE), 12.07 mean absolute error (MAE), and 0.45 R-squared using the entropy features with a boosted tree regression model. Thus, the results demonstrate that the proposed qEEG features show the state of brain function at an early stage; hence, they could serve as predictive biomarkers of later cognitive impairment, which could facilitate identifying those who might benefit from early targeted intervention.

Shape variability of the central sulcus in the developing brain: a longitudinal descriptive and predictive study in preterm infants

Despite growing evidence of links between sulcation and function in the adult brain, the folding dynamics, occurring mostly before normal-term-birth, is vastly unknown. Looking into the development of cortical sulci in infants can give us keys to address fundamental questions: what is the sulcal shape variability in the developing brain? When are the shape features encoded? How are these morphological parameters related to further functional development? In this study, we aimed to investigate the shape variability of the developing central sulcus, which is the frontier between the primary somatosensory and motor cortices. We studied a cohort of 71 extremely preterm infants scanned twice using MRI - once around 30 weeks post-menstrual age (w PMA) and once at term-equivalent age, around 40w PMA -, in order to quantify the sulcus's shape variability using manifold learning, regardless of age-group or hemisphere. We then used these shape descriptors to evaluate the sulcus's variability at both ages and to assess hemispheric and age-group specificities. This led us to propose a description of ten shape features capturing the variability in the central sulcus of preterm infants. Our results suggested that most of these features (8/10) are encoded as early as 30w PMA. We unprecedentedly observed hemispheric asymmetries at both ages, and the one captured at term-equivalent age seems to correspond with the asymmetry pattern previously reported in adults. We further trained classifiers in order to explore the predictive value of these shape features on manual performance at 5 years of age (handedness and fine motor outcome). The central sulcus's shape alone showed a limited but relevant predictive capacity in both cases. The study of sulcal shape features during early neurodevelopment may participate to a better comprehension of the complex links between morphological and functional organization of the developing brain.

Deep Multimodal Learning From MRI and Clinical Data for Early Prediction of Neurodevelopmental Deficits in Very Preterm Infants

The prevalence of disabled survivors of prematurity has increased dramatically in the past 3 decades. These survivors, especially, very preterm infants (VPIs), born ≤ 32 weeks gestational age, are at high risk for neurodevelopmental impairments. Early and clinically effective personalized prediction of outcomes, which forms the basis for early treatment decisions, is urgently needed during the peak neuroplasticity window—the first couple of years after birth—for at-risk infants, when intervention is likely to be most effective. Advances in MRI enable the noninvasive visualization of infants' brains through acquired multimodal images, which are more informative than unimodal MRI data by providing complementary/supplementary depicting of brain tissue characteristics and pathology. Thus, analyzing quantitative multimodal MRI features affords unique opportunities to study early postnatal brain development and neurodevelopmental outcome prediction in VPIs. In this study, we investigated the predictive power of multimodal MRI data, including T2-weighted anatomical MRI, diffusion tensor imaging, resting-state functional MRI, and clinical data for the prediction of neurodevelopmental deficits. We hypothesize that integrating multimodal MRI and clinical data improves the prediction over using each individual data modality. Employing the aforementioned multimodal data, we proposed novel end-to-end deep multimodal models to predict neurodevelopmental (i.e., cognitive, language, and motor) deficits independently at 2 years corrected age. We found that the proposed models can predict cognitive, language, and motor deficits at 2 years corrected age with an accuracy of 88.4, 87.2, and 86.7%, respectively, significantly better than using individual data modalities. This current study can be considered as proof-of-concept. A larger study with external validation is important to validate our approach to further assess its clinical utility and overall generalizability.

Precision Medicine in Neonates: A Tailored Approach to Neonatal Brain Injury

Despite advances in neonatal care to prevent neonatal brain injury and neurodevelopmental impairment, predicting long-term outcome in neonates at risk for brain injury remains difficult. Early prognosis is currently based on cranial ultrasound (CUS), MRI, EEG, NIRS, and/or general movements assessed at specific ages, and predicting outcome in an individual (precision medicine) is not yet possible. New algorithms based on large databases and machine learning applied to clinical, neuromonitoring, and neuroimaging data and genetic analysis and assays measuring multiple biomarkers (omics) can fulfill the needs of modern neonatology. A synergy of all these techniques and the use of automatic quantitative analysis might give clinicians the possibility to provide patient-targeted decision-making for individualized diagnosis, therapy, and outcome prediction. This review will first focus on common neonatal neurological diseases, associated risk factors, and most common treatments. After that, we will discuss how precision medicine and machine learning (ML) approaches could change the future of prediction and prognosis in this field.

White matter tracts related to memory and emotion in very preterm children

BACKGROUND

Very preterm (VP) children are at risk of memory and emotional impairments; however, the neural correlates remain incompletely defined. This study investigated the effect of VP birth on white matter tracts traditionally related to episodic memory and emotion.

METHODS

The cingulum, fornix, uncinate fasciculus, medial forebrain bundle and anterior thalamic radiation were reconstructed using tractography in 144 VP children and 33 full-term controls at age 7 years.

RESULTS

Compared with controls, VP children had higher axial, radial, and mean diffusivities and neurite orientation dispersion, and lower volume and neurite density in the fornix, along with higher neurite orientation dispersion in the medial forebrain bundle. Support vector classification models based on tract measures significantly classified VP children and controls. Higher fractional anisotropy and lower diffusivities in the cingulum, uncinate fasciculus, medial forebrain bundle and anterior thalamic radiation were associated with better episodic memory, independent of key perinatal risk factors. Support vector regression models using tract measures did not predict episodic memory and emotional outcomes.

CONCLUSIONS

Altered tract structure is related to adverse episodic memory outcomes in VP children, but further research is required to determine the ability of tract structure to predict outcomes of individual children.

IMPACT

We studied white matter fibre tracts thought to be involved in episodic memory and emotion in VP and full-term children using diffusion magnetic resonance imaging and machine learning. VP children have altered fornix and medial forebrain bundle structure compared with full-term children. Altered tract structure can be detected using machine learning, which accurately classified VP and full-term children using tract data. Altered cingulum, uncinate fasciculus, medial forebrain bundle and anterior thalamic radiation structure was associated with poorer episodic memory skills using linear regression. The ability of tract structure to predict episodic memory and emotional outcomes of individual children based on support vector regression was limited.

MRI of the Neonatal Brain: A Review of Methodological Challenges and Neuroscientific Advances

In recent years, exploration of the developing brain has become a major focus for researchers and clinicians in an attempt to understand what allows children to acquire amazing and unique abilities, as well as the impact of early disruptions (eg, prematurity, neonatal insults) that can lead to a wide range of neurodevelopmental disorders. Noninvasive neuroimaging methods such as MRI are essential to establish links between the brain and behavioral changes in newborns and infants. In this review article, we aim to highlight recent and representative studies using the various techniques available: anatomical MRI, quantitative MRI (relaxometry, diffusion MRI), multiparametric approaches, and functional MRI. Today, protocols use 1.5 or 3T MRI scanners, and specialized methodologies have been put in place for data acquisition and processing to address the methodological challenges specific to this population, such as sensitivity to motion. MR sequences must be adapted to the brains of newborns and infants to obtain relevant good soft-tissue contrast, given the small size of the cerebral structures and the incomplete maturation of tissues. The use of age-specific image postprocessing tools is also essential, as signal and contrast differ from the adult brain. Appropriate methodologies then make it possible to explore multiple neurodevelopmental mechanisms in a precise way, and assess changes with age or differences between groups of subjects, particularly through large-scale projects. Although MRI measurements only indirectly reflect the complex series of dynamic processes observed throughout development at the molecular and cellular levels, this technique can provide information on brain morphology, structural connectivity, microstructural properties of gray and white matter, and on the functional architecture. Finally, MRI measures related to clinical, behavioral, and electrophysiological markers have a key role to play from a diagnostic and prognostic perspective in the implementation of early interventions to avoid long-term disabilities in children. EVIDENCE LEVEL: 2 TECHNICAL EFFICACY STAGE: 1.

Robust Cortical Thickness Morphometry of Neonatal Brain and Systematic Evaluation Using Multi-Site MRI Datasets

The human brain grows the most dramatically during the perinatal and early post-natal periods, during which pre-term birth or perinatal injury that may alter brain structure and lead to developmental anomalies. Thus, characterizing cortical thickness of developing brains remains an important goal. However, this task is often complicated by inaccurate cortical surface extraction due to small-size brains. Here, we propose a novel complex framework for the reconstruction of neonatal WM and pial surfaces, accounting for large partial volumes due to small-size brains. The proposed approach relies only on T1-weighted images unlike previous T2-weighted image-based approaches while only T1-weighted images are sometimes available under the different clinical/research setting. Deep neural networks are first introduced to the neonatal magnetic resonance imaging (MRI) pipeline to address the mis-segmentation of brain tissues. Furthermore, this pipeline enhances cortical boundary delineation using combined models of the cerebrospinal fluid (CSF)/GM boundary detection with edge gradient information and a new skeletonization of sulcal folding where no CSF voxels are seen due to the limited resolution. We also proposed a systematic evaluation using three independent datasets comprising 736 pre-term and 97 term neonates. Qualitative assessment for reconstructed cortical surfaces shows that 86.9% are rated as accurate across the three site datasets. In addition, our landmark-based evaluation shows that the mean displacement of the cortical surfaces from the true boundaries was less than a voxel size (0.532 ± 0.035 mm). Evaluating the proposed pipeline (namely NEOCIVET 2.0) shows the robustness and reproducibility across different sites and different age-groups. The mean cortical thickness measured positively correlated with post-menstrual age (PMA) at scan (p < 0.0001); Cingulate cortical areas grew the most rapidly whereas the inferior temporal cortex grew the least rapidly. The range of the cortical thickness measured was biologically congruent (1.3 mm at 28 weeks of PMA to 1.8 mm at term equivalent). Cortical thickness measured on T1 MRI using NEOCIVET 2.0 was compared with that on T2 using the established dHCP pipeline. It was difficult to conclude that either T1 or T2 imaging is more ideal to construct cortical surfaces. NEOCIVET 2.0 has been open to the public through CBRAIN (https://mcin-cnim.ca/technology/cbrain/), a web-based platform for processing brain imaging data.

Automated brain morphometric biomarkers from MRI at term predict motor development in very preterm infants

•

Nearly 1/3 of very preterm (VPT) infants develop motor impairments later in life.

•

Better early biomarkers are needed for risk-stratification and early intervention.

•

We used MRI morphometrics at term to predict 2-year motor ability in VPT infants.

•

Inner cortical curvature at term is a novel biomarker of early motor aptitude.

•

In regression models, morphometrics explained nearly 50% of motor score variance.

Very preterm infants are at high risk for motor impairments. Early interventions can improve outcomes in this cohort, but they would be most effective if clinicians could accurately identify the highest-risk infants early. A number of biomarkers for motor development exist, but currently none are sufficiently accurate for early risk-stratification. We prospectively enrolled very preterm (gestational age ≤31 weeks) infants from four level-III NICUs. Structural brain MRI was performed at term-equivalent age. We used a established pipeline to automatically derive brain volumetrics and cortical morphometrics – cortical surface area, sulcal depth, gyrification index, and inner cortical curvature – from structural MRI. We related these objective measures to Bayley-III motor scores (overall, gross, and fine) at two-years corrected age. Lasso regression identified the three best predictive biomarkers for each motor scale from our initial feature set. In multivariable regression, we assessed the independent value of these brain biomarkers, over-and-above known predictors of motor development, to predict motor scores. 75 very preterm infants had high-quality T2-weighted MRI and completed Bayley-III motor testing. All three motor scores were positively associated with regional cortical surface area and subcortical volumes and negatively associated with cortical curvature throughout the majority of brain regions. In multivariable regression modeling, thalamic volume, curvature of the temporal lobe, and curvature of the insula were significant predictors of overall motor development on the Bayley-III, independent of known predictors. Objective brain morphometric biomarkers at term show promise in predicting motor development in very preterm infants.

Brain Damage and Visuospatial Impairments: Exploring Early Structure-Function Associations in Children Born Very Preterm

BACKGROUND

To provide insight into early neurosensory development in children born very preterm, we assessed the association between early structural brain damage and functional visuospatial attention and motion processing from one to two years corrected age.

METHODS

In 112 children born at less than 32 weeks gestational age, we assessed brain damage and growth with a standardized scoring system on magnetic resonance imaging (MRI; 1.5 Tesla) scans performed at 29 to 35 weeks gestational age. Of the children with an MRI scan, 82 participated in an eye tracking-based assessment of visuospatial attention and motion processing (Tobii T60XL) at one year corrected age and 59 at two years corrected age.

RESULTS

MRI scoring showed good intra- and inter-rater reproducibility. At one year, 10% children had delayed attentional reaction times and 23% had delayed motion reaction times. Moderate to severe brain damage significantly correlated with slower visuospatial reaction times. At two years, despite attention and motion reaction times becoming significantly faster, 20% had delayed attentional reaction times and 35% had delayed motion reaction times, but no correlations with MRI scores were found. The presence of structural brain damage was associated with abnormal functional performance over age.

CONCLUSIONS

The present study indicates an association between moderate to severe brain damage and visuospatial attention and motion processing dysfunction at one year corrected age. This provides a new perspective on comprehensive MRI scoring and quantitative functional visuospatial assessments and their applicability in children born very preterm in their first years of life.

Predicting motor outcome in preterm infants from very early brain diffusion MRI using a deep learning convolutional neural network (CNN) model

BACKGROUND AND AIMS

Preterm birth imposes a high risk for developing neuromotor delay. Earlier prediction of adverse outcome in preterm infants is crucial for referral to earlier intervention. This study aimed to predict abnormal motor outcome at 2 years from early brain diffusion magnetic resonance imaging (MRI) acquired between 29 and 35 weeks postmenstrual age (PMA) using a deep learning convolutional neural network (CNN) model.

METHODS

Seventy-seven very preterm infants (born <31 weeks gestational age (GA)) in a prospective longitudinal cohort underwent diffusion MR imaging (3T Siemens Trio; 64 directions, b ​= ​2000 ​s/mm²). Motor outcome at 2 years corrected age (CA) was measured by Neuro-Sensory Motor Developmental Assessment (NSMDA). Scores were dichotomised into normal (functional score: 0, normal; n ​= ​48) and abnormal scores (functional score: 1-5, mild-profound; n ​= ​29). MRIs were pre-processed to reduce artefacts, upsampled to 1.25 ​mm isotropic resolution and maps of fractional anisotropy (FA) were estimated. Patches extracted from each image were used as inputs to train a CNN, wherein each image patch predicted either normal or abnormal outcome. In a postprocessing step, an image was classified as predicting abnormal outcome if at least 27% (determined by a grid search to maximise the model performance) of its patches predicted abnormal outcome. Otherwise, it was considered as normal. Ten-fold cross-validation was used to estimate performance. Finally, heatmaps of model predictions for patches in abnormal scans were generated to explore the locations associated with abnormal outcome.

RESULTS

For the identification of infants with abnormal motor outcome based on the FA data from early MRI, we achieved mean sensitivity 70% (standard deviation SD 19%), mean specificity 74% (SD 39%), mean AUC (area under the receiver operating characteristic curve) 72% (SD 14%), mean F1 score of 68% (SD 13%) and mean accuracy 73% (SD 19%) on an unseen test data set. Patch-based prediction heatmaps showed that the patches around the motor cortex and somatosensory regions were most frequently identified by the model with high precision (74%) as a location associated with abnormal outcome. Part of the cerebellum, and occipital and frontal lobes were also highly associated with abnormal NSMDA/motor outcome.

DISCUSSION/CONCLUSION

This study established the potential of an early brain MRI-based deep learning CNN model to identify preterm infants at risk of a later motor impairment and to identify brain regions predictive of adverse outcome. Results suggest that predictions can be made from FA maps of diffusion MRIs well before term equivalent age (TEA) without any prior knowledge of which MRI features to extract and associated feature extraction steps. This method, therefore, is suitable for any case of brain condition/abnormality. Future studies should be conducted on a larger cohort to re-validate the robustness and effectiveness of these models.

Automatic extraction of the intracranial volume in fetal and neonatal MR scans using convolutional neural networks

•

Automatic intracranial volume segmentation.

•

Fetal and neonatal MRI.

•

Deep learning.

MR images of infants and fetuses allow non-invasive analysis of the brain. Quantitative analysis of brain development requires automatic brain tissue segmentation that is typically preceded by segmentation of the intracranial volume (ICV). Fast changes in the size and morphology of the developing brain, motion artifacts, and large variation in the field of view make ICV segmentation a challenging task.

We propose an automatic method for segmentation of the ICV in fetal and neonatal MRI scans. The method was developed and tested with a diverse set of scans regarding image acquisition parameters (i.e. field strength, image acquisition plane, image resolution), infant age (23–45 weeks post menstrual age), and pathology (posthaemorrhagic ventricular dilatation, stroke, asphyxia, and Down syndrome). The results demonstrate that the method achieves accurate segmentation with a Dice coefficient (DC) ranging from 0.98 to 0.99 in neonatal and fetal scans regardless of image acquisition parameters or patient characteristics. Hence, the algorithm provides a generic tool for segmentation of the ICV that may be used as a preprocessing step for brain tissue segmentation in fetal and neonatal brain MR scans.

Preterm Neuroimaging and School-Age Cognitive Outcomes

BACKGROUND AND OBJECTIVES

Children born extremely preterm are at risk for cognitive difficulties and disability. The relative prognostic value of neonatal brain MRI and cranial ultrasound (CUS) for school-age outcomes remains unclear. Our objectives were to relate near-term conventional brain MRI and early and late CUS to cognitive impairment and disability at 6 to 7 years among children born extremely preterm and assess prognostic value.

METHODS

A prospective study of adverse early and late CUS and near-term conventional MRI findings to predict outcomes at 6 to 7 years including a full-scale IQ (FSIQ) <70 and disability (FSIQ <70, moderate-to-severe cerebral palsy, or severe vision or hearing impairment) in a subgroup of Surfactant Positive Airway Pressure and Pulse Oximetry Randomized Trial enrollees. Stepwise logistic regression evaluated associations of neuroimaging with outcomes, adjusting for perinatal-neonatal factors.

RESULTS

A total of 386 children had follow-up. In unadjusted analyses, severity of white matter abnormality and cerebellar lesions on MRI and adverse CUS findings were associated with outcomes. In full regression models, both adverse late CUS findings (odds ratio [OR] 27.9; 95% confidence interval [CI] 6.0-129) and significant cerebellar lesions on MRI (OR 2.71; 95% CI 1.1-6.7) remained associated with disability, but only adverse late CUS findings (OR 20.1; 95% CI 3.6-111) were associated with FSIQ <70. Predictive accuracy of stepwise models was not substantially improved with the addition of neuroimaging.

CONCLUSIONS

Severe but rare adverse late CUS findings were most strongly associated with cognitive impairment and disability at school age, and significant cerebellar lesions on MRI were associated with disability. Near-term conventional MRI did not substantively enhance prediction of severe early school-age outcomes.