Transforming neonatal care with artificial intelligence: challenges, ethical consideration, and opportunities.

Artificial intelligence (AI) offers tremendous potential to transform neonatology through improved diagnostics, personalized treatments, and earlier prevention of complications. However, there are many challenges to address before AI is ready for clinical practice. This review defines key AI concepts and discusses ethical considerations and implicit biases associated with AI. Next we will review literature examples of AI already being explored in neonatology research and we will suggest future potentials for AI work. Examples discussed in this article include predicting outcomes such as sepsis, optimizing oxygen therapy, and image analysis to detect brain injury and retinopathy of prematurity. Realizing AI's potential necessitates collaboration between diverse stakeholders across the entire process of incorporating AI tools in the NICU to address testability, usability, bias, and transparency. With multi-center and multi-disciplinary collaboration, AI holds tremendous potential to transform the future of neonatology.

A Nitric Oxide-Releasing Self-Assembled Peptide Amphiphile Nanomatrix for Improving the Biocompatibility of Microporous Hollow Fibers.

Oxygenators are critical components of extracorporeal circuits used frequently in cardiopulmonary bypass and intensive care, but platelet activation and induction of a complex inflammatory response are usually observed with their use. To improve the biocompatibility of oxygenators, we developed a nitric oxide (NO)-releasing, self-assembled peptide amphiphile nanomatrix. The nanomatrix formed a homogenous coating over the microporous hollow fibers as demonstrated by scanning electron microscopy. We quantitated platelet adhesion to the artificial fibers by measuring absorbance/area of platelets (Abs/A; nm/m2) using acid phosphatase assay. There was a 17-fold decrease in platelet adhesion to the nanomatrix (Abs/A = 0.125) compared with collagen controls (Abs/A = 2.07; p < 0.05) and a 22-fold decrease compared with uncoated fibers (Abs/A = 2.75; p < 0.05). Importantly, the nanomatrix coating did not impede oxygen transfer in water through coated fiber modules (p > 0.05) in a benchtop test circuit at different flow rates as estimated by change in partial pressure of oxygen in relation to water velocity through fibers. These findings demonstrate the feasibility of coating microporous hollow fibers with a NO-releasing self-assembled amphiphile nanomatrix that may improve the biocompatibility of the hollow fibers without affecting their gas exchange capacity.

The value of Pa(CO2) in relation to outcome in congenital diaphragmatic hernia.

BACKGROUND: Postnatal assessment of disease severity is critical for analysis of mortality rates and development of future interventions in congenital diaphragmatic hernia (CDH). OBJECTIVE: The objective of this study was to stratify the risk of mortality based on arterial Paco 2. METHODS: Retrospective analysis of infants (n = 133) with CDH admitted to a regional extracorporeal membrane oxygenation (ECMO) center in two different periods: period I (1987-1996; n = 46) and period II (2002-2010; n = 87). RESULTS: The mortality rate (37%) was similar in both periods (p = 0.98). Paco 2 < 60 mm Hg in the first arterial blood gas (ABG) was an independent predictor of survival in both periods (p = 0.03). The predicted survival rate was 84% if initial Paco 2 was < 55 mm Hg. For infants with initial Paco 2 > 55 mm Hg treated with ECMO (n = 83), the predicted survival rate was 11% if the Paco 2 was > 88 mm Hg before the initiation of ECMO. CONCLUSION: Paco 2, a surrogate of lung hypoplasia, may be useful for risk stratification in CDH. Paco 2 < 60 mm Hg in the first ABG may indicate milder pulmonary hypoplasia. A Paco 2 > 80 mm Hg in the first ABG and/or before ECMO may indicate severe pulmonary hypoplasia.

ABO blood group is associated with response to inhaled nitric oxide in neonates with respiratory failure.

BACKGROUND: Inhaled nitric oxide (iNO) reduces death or need for extracorporeal membrane oxygenation (ECMO) in infants with persistent pulmonary hypertension of the newborn (PPHN). However, the response to iNO is variable and only 50-60% of infants demonstrate a response to iNO. It is not known why only some infants respond to iNO. Adults and children with blood groups B or AB do not respond as well to iNO as those with blood groups O/A. METHODS/PRINCIPAL FINDINGS: To determine if blood group was associated with iNO response in newborn infants, a retrospective medical record review was done of infants admitted to a regional NICU from 2002-9 with a diagnosis of PPHN. Data were collected during the first twelve hours post-initiation of treatment. Of 86 infants diagnosed with PPHN, 23 infants had blood group A [18 received iNO], 21 had group B [18 with iNO], 40 had group O [36 with iNO], and 2 had group AB [both received iNO]. Change in PaO(2)/FiO(2) was less in infants with blood group A, of whom less than half were responders (ΔPaO(2)/FiO(2)>20%) at 12 h versus 90% of infants with either O or B. Race, sex, birth weight, gestational age, Apgar scores at 1 and 5 minutes, and baseline PaO(2)/FiO(2) were similar among groups. Outcomes including need for ECMO, death, length of ventilatory support, length of iNO use, and hospital stay were statistically not different by blood groups. CONCLUSIONS/SIGNIFICANCE: Our results indicate that blood group influences iNO response in neonates. We hypothesize that either there is genetic linkage of the ABO gene locus with vasoregulatory genes, or that blood group antigens directly affect vascular reactivity.

Evaluation of a pumpless lung assist device in hypoxia-induced pulmonary hypertension in juvenile piglets.

Persistent pulmonary hypertension is an important cause of mortality and morbidity in term infants. The lung assist device (LAD) is a novel, pumpless, low-resistance extracorporeal oxygenator to supplement mechanical ventilation. The LAD may be associated with fewer complications compared with conventional extracorporeal membrane oxygenation. The objective was to test the feasibility and efficacy of the LAD in juvenile piglets with hypoxia-induced pulmonary hypertension. Pulmonary hypertension was acutely induced by hypoxia in six 3- to 4-wk-old acutely instrumented and intubated piglets. The LAD was attached between a carotid artery and jugular vein. Gas exchange and hemodynamic variables, including pulmonary arterial pressure (PAP) and cardiac output (CO), were measured. Successful LAD cannulation was achieved without complications in all animals. Extracorporeal shunt flow through the device averaged 18% of CO. The LAD achieved oxygen delivery of 20% of total oxygen consumption. PAP was reduced by 35% from 28 +/- 5 to 18 +/- 4 mm Hg (p < 0.05) and systemic Pao2 increased by 33% from 27 +/- 2 to 36 +/- 4 mm Hg (p < 0.05). Other hemodynamic variables remained stable. The novel LAD shows feasibility and efficacy in improving gas exchange and reducing PAPs in a juvenile animal model of hypoxia-induced pulmonary hypertension.

A pumpless lung assist device reduces mechanical ventilation-induced lung injury in juvenile piglets.

Respiratory failure is a major contributor to mortality and morbidity in newborn infants. The lung assist device (LAD) is a novel gas exchange device that supplements mechanical ventilation. The objective is to test the effect of the LAD on pulmonary histopathology in juvenile piglets with acute lung injury caused by saline lung lavage (SLL) followed by intermittent mandatory ventilation (IMV). Three- to 4-wk-old piglets were randomized to no intervention (control group), SLL alone (SLL group), SLL + IMV (IMV group), or SLL + IMV + LAD (LAD group) (n = 6 per group). The carotid artery and jugular vein were cannulated and an arteriovenous circuit completed, and the LAD was inserted into this circuit. Gas exchange via the LAD was initiated by passage of 100% oxygen over the blood-carrying hollow fibers of the LAD. Hemodynamic variables were recorded. Mechanical ventilation was systematically weaned. Lung histology was scored by two observers masked to treatment group. There were no differences in hemodynamic variables between the study groups. There was a significant increase in the total lung injury score in the IMV group compared with the LAD group. The novel pumpless low-resistance LAD has shown feasibility and potential to decrease ventilator-induced lung injury in a juvenile animal model.

Nitric oxide administration using an oxygen hood: a pilot trial.

BACKGROUND: We have shown earlier that inhaled nitric oxide (iNO) administered by oxygen hood reduces pulmonary hypertension in an animal model (J Perinatol 2002; 22:50-6). Our objective in this study was to determine feasibility of iNO by oxygen hood in neonates with elevated alveolar-arterial oxygen gradients (A-aDO(2)). METHODS/PRINCIPAL FINDINGS: Masked randomized controlled pilot trial. Inclusion criteria were: gestation>or=34 weeks, age<7 days, with post-ductal arterial line, and A-aDO(2) 400-600. Infants were randomized to study gas (iNO 20 ppm or equivalent O(2) flow) for 1 hr which was then weaned over the next 4 hours. Primary outcome was PaO(2) one hour post-randomization. Four infants each were randomized to iNO or O(2) (controls). Two of the four infants given iNO had an increase in PaO(2) of >100 torr, while oxygenation was unchanged in the controls. Methemoglobinemia and other adverse effects were not noted in any infant. Environmental levels of NO and NO(2) were minimal (<1 ppm) at >0.3 m from the hood. CONCLUSIONS: Administration of iNO by oxygen hood is feasible. Larger randomized controlled trials are required to measure the efficacy and determine an appropriate target population for this technique. TRIAL REGISTRATION: ClinicalTrials.gov NCT00041548.