Data on the long-term outcome of children who are exposed to maternal cancer with or without treatment during pregnancy are lacking. In this multicenter case–control study, we compared children whose mothers received a diagnosis of cancer during the pregnancy with matched children of women without a cancer diagnosis. We used a health questionnaire and medical files to collect data regarding neonatal and general health. All children were prospectively assessed (by means of a neurologic examination and the Bayley Scales of Infant Development) at 18 months, 36 months, or both. A cardiac assessment was performed at 36 months. A total of 129 children (median age, 22 months; range, 12 to 42) were included in the group whose mother had cancer (prenatal-exposure group) with a matching number in the control group. During pregnancy, 96 children (74.4%) were exposed to chemotherapy (alone or in combination with other treatments), 11 (8.5%) to radiotherapy (alone or in combination), 13 (10.1%) to surgery alone, 2 (1.6%) to other drug treatments, and 14 (10.9%) to no treatment. Birth weight was below the 10th percentile in 28 of 127 children (22.0%) in the prenatal-exposure group and in 19 of 125 children (15.2%) in the control group (P=0.16). There was no significant between-group difference in cognitive development on the basis of the Bayley score (P=0.08) or in subgroup analyses. The gestational age at birth was correlated with the cognitive outcome in the two study groups. Cardiologic evaluation among 47 children at 36 months of age showed normal cardiac findings. Prenatal exposure to maternal cancer with or without treatment did not impair the cognitive, cardiac, or general development of children in early childhood. Prematurity was correlated with a worse cognitive outcome, but this effect was independent of cancer treatment. (Funded by Research Foundation–Flanders and others; ClinicalTrials.gov number, NCT00330447.)

This review focuses on the clinical use of near infrared spectroscopy (NIRS) to assess brain oxygenation by the tissue oxygenation index (TOI), and monitoring regional cerebral oxygen saturation (rScO2), cerebral fractional tissue oxygen extraction (cFTOE), which is derived from rScO2, and systemic oxygen saturation. Its precision and pitfalls are discussed. At this stage, it is clear that NIRS-monitored oxygenation of the brain by rScO2 or TOI lacks the precision required to be used as a robust quantitative variable to monitor cerebral oxygenation. Intra- and especially interpatient variability are too large for this aim. On the other hand, when used merely as a trend monitor in the individual patient, substantial changes in rScO2 or TOI and consequently of cFTOE, larger than the limits of agreement, can yield important clinical information that suggest an intervention. Since neonatal intensive care is for a substantial part ‘brain orientated’ this approach seems conceivable. This gives rise to the conclusion that NIRS-monitored TOI, rScO2 and cFTOE increasingly will have a role in clinical practice as semiquantitative indicators of changes in cerebral oxygenation and oxygen extraction. Combination with other (cerebral) parameters such as amplitude-integrated EEG and blood pressure seems promising for further optimization of monitoring the immature brain.

Identifying a core set of features is one of the most important steps in the development of an automated seizure detector. In most of the published studies describing features and seizure classifiers, the features were hand-engineered, which may not be optimal. The main goal of the present paper is using deep convolutional neural networks (CNNs) and random forest to automatically optimize feature selection and classification. The input of the proposed classifier is raw multi-channel EEG and the output is the class label: seizure/nonseizure. By training this network, the required features are optimized, while fitting a nonlinear classifier on the features. After training the network with EEG recordings of 26 neonates, five end layers performing the classification were replaced with a random forest classifier in order to improve the performance. This resulted in a false alarm rate of 0.9 per hour and seizure detection rate of 77% using a test set of EEG recordings of 22 neonates that also included dubious seizures. The newly proposed CNN classifier outperformed three data-driven feature-based approaches and performed similar to a previously developed heuristic method.

OBJECTIVES: The goals were to investigate how many subclinical seizures in full-term neonates with hypoxic-ischemic encephalopathy (HIE) would be missed without continuous amplitude-integrated electroencephalography (aEEG) and whether immediate treatment of both clinical and subclinical seizures would result in a reduction in the total duration of seizures and a decrease in brain injury, as seen on MRI scans. METHODS: In this multicenter, randomized, controlled trial, term infants with moderate to severe HIE and subclinical seizures were assigned randomly to either treatment of both clinical seizures and subclinical seizure patterns (group A) or blinding of the aEEG registration and treatment of clinical seizures only (group B). All recordings were reviewed with respect to the duration of seizure patterns and the use of antiepileptic drugs (AEDs). MRI scans were scored for the severity of brain injury. RESULTS: Nineteen infants in group A and 14 infants in group B were available for comparison. The median duration of seizure patterns in group A was 196 minutes, compared with 503 minutes in group B (not statistically significant). No significant differences in the number of AEDs were seen. Five infants in group B received AEDs when no seizure discharges were seen on aEEG traces. Six of 19 infants in group A and 7 of 14 infants in group B died during the neonatal period. A significant correlation between the duration of seizure patterns and the severity of brain injury in the blinded group, as well as in the whole group, was found. CONCLUSIONS: In this small group of infants with neonatal HIE and seizures, there was a trend for a reduction in seizure duration when clinical and subclinical seizures were treated. The severity of brain injury seen on MRI scans was associated with a longer duration of seizure patterns.

BACKGROUND AND OBJECTIVES: Available data on survival rates and outcomes of extremely low gestational age (GA) infants (22–25 weeks’ gestation) display wide variation by country. Whether similar variation is found in statements by national professional bodies is unknown. The objectives were to perform a systematic review of management from scientific and professional organizations for delivery room care of extremely low GA infants. METHODS: We searched Embase, PubMed, and Google Scholar for management guidelines on perinatal care. Countries were included if rated by the United Nations Development Programme’s Human Development Index as “very highly developed.” The primary outcome was rating of recommendations from “comfort care” to “active care.” Secondary outcomes were specifying country-specific survival and considering potential for 3 biases: limitations of GA assessment; bias from different definitions of stillbirths and live births; and bias from the use of different denominators to calculate survival. RESULTS: Of 47 highly developed countries, 34 guidelines from 23 countries and 4 international groups were identified. Of these, 3 did not state management recommendations. Of the remaining 31 guidelines, 21 (68%) supported comfort care at 22 weeks’ gestation, and 20 (65%) supported active care at 25 weeks’ gestation. Between 23 and 24 weeks’ gestation, much greater variation was seen. Seventeen guidelines cited national survival rates. Few guidelines discussed potential biases: limitations in GA (n = 17); definition bias (n = 3); and denominator bias (n = 7). CONCLUSIONS: Although there is a wide variation in recommendations (especially between 23 and 24 weeks’ GA), there is general agreement for comfort care at 22 weeks’ GA and active care at 25 weeks’ GA.