Extreme preterm birth
Annually, 15 million children worldwide are born prematurely. In addition to short-term complications, such as respiratory distress syndrome (RDS), bronchopulmonary dysplasia (BPD), necrotic enterocolitis (NEC) and intraventricular haemorrhage (IVH), there are also long-term complications, such as neurocognitive developmental disorders .
To improve the life expectancy and quality of these children, research is underway to work on a solution that mimics the womb environment better than the current incubator. In this way, the organs of extremely premature babies can continue to grow just as they would in the womb. This fundamentally new vision of the incubator offers a fluid-filled environment combined with an artificial placenta [3,7] (Figure 1). The combination of the artificial placenta and amniotic fluid sac facilitates further organ maturation while avoiding the deleterious effects of mechanical ventilation of the lungs. The goal of this new technology is to increase survival rates, reduce the risk of morbidity and help the extremely premature children through the critical period of 24 to 28 weeks. In Europe, the multidisciplinary Perinatal Life Support (PLS) consortium was formed to develop such an incubator, or PLS system. The consortium has a wide variety of knowledge and expertise: TU Eindhoven, RWTH Aachen, Politecnico di Milano, LifeTec Group, Nemo Healthcare, Medsim and Máxima Medisch Centrum.
Figure 1: Schematic representation of the perinatal life support system.
To design and develop the PLS system – and ecosystem of protocols and other tools – it is necessary to test the situation as realistically as possible. Pre-clinical animal models have been used for a proof-of-principle to test this new generation of incubators in foetal lambs [2,3,7]. Nevertheless, the step from animal testing with a fully developed PLS system to human studies is still immense. Both animal experiments and human clinical studies during PLS system development are difficult to achieve because of ethical concerns. Much of the attention in the ethical debate surrounding PLS system development has focused on the long-term implications, but short-term ethical concerns, such as the use of animal studies, should not be overlooked either .
Given that we want to avoid conducting animal experiments for technology development as much as possible, validation of technical prototypes is a major challenge. A promising alternative for successfully developing medical devices, improving the quality of care and facilitating training without animal testing is medical simulation. Therefore, within this project we are developing foetal manikins (Faculty of Industrial Design, TU Eindhoven) and digital twins (Faculty of Biomedical Engineering, TU Eindhoven) to mimic the foetal patient both anatomically and physiologically. In this way, testing on laboratory animals and validation in humans can be postponed to a stage where research results are significantly less unpredictable.
Methods & Results
To develop a realistic physical 3D model of a foetus, we used MRI images of a foetus with a gestational age of 24 weeks as the basis for the design of the manikin. 3D printing techniques and moulding methods allowed it to be fabricated in silicone. Literature research identified the most relevant physiological and physical components involved in using this PLS system. Based on this, a selection of sensors and actuators (motors to simulate symptoms) were integrated into the manikin. These include temperature monitoring, as well as the simulation of cyanosis, heart rate, breathing and movement. Incorporating foetal cardiorespiratory physiology into the manikin with features such as blood flow and heart rate can inform us whether sustained circulation has been established between the PLS system and the foetus.
Figure 2: Physical prototype of the manikin, based on MRI data from a 24-week-old fetus.
The digital twin in this project is a virtual representation of the baby in the PLS system that interacts with the real baby in the PLS system through measurements of the baby. The digital twin is a mathematical model consisting of a detailed vessel structure, where blood flow and pressure in each vessel can be calculated. Further, the mathematical model can simulate oxygen distribution, metabolism, foetal growth, baro- and chemoreceptor reflexes, and interaction between baby and the PLS components. Information for, and verification of this mathematical model is based on literature and retrospective clinical studies. During the development phase of the PLS system, the function of the “digital twin” is to gather more information about the interaction between uterus and baby to develop the most optimal environment. Once the PLS system is used in the clinic then the digital twin can be used to support clinical decision-making by predicting the effect of a device setting or intervention.
Figure 3: Schematic overview of the mathematical model that is part of the digital twin.
Interaction digital twin and manikin
By connecting the digital twin to the manikin, the effect of the PLS system and the actions of the medical staff on the baby can be predicted. In this way, the system can be adjusted to the most optimal environment for the baby. In addition, real-life situations can be simulated to train healthcare professionals. Figure 4 describes an example of a possible interaction between a manikin and digital twin. The manikin contains actuators that can generate lifelike signals of oxygen saturation in the brain, heart rate and blood flow in the umbilical cord. These signals are read by sensors and used by the digital twin. Peripheral oxygen saturation is then estimated. The estimated and measured signals are used to make a prediction of the baby’s well-being. If the condition of the (simulated) baby deteriorates, relevant symptoms can be simulated by the manikin. In this way, health care providees can react appropriately during training or by optimizing settings of the PLS system to create an ideal environment for the baby.
Figure 4: Potential scenario describing the interaction between manikin and digital twin.
The purpose of the simulation is threefold; first, it can serve as a method to optimize and validate early prototypes and medical protocols of the PLS system. Second, the functional prototype can be tested with (new) clinical protocols so that it can be used to train medical personnel with this new technology in a safe setting. Using the PLS system involves complex tasks that require precision, teamwork, and correct time management. By using a simulation, clinical scenarios can be completely controlled, putting neither the (simulated) patient nor the healthcare team in an unsafe situation (psychologically or physically). In addition, the simulated environment has many degrees of freedom. Different scenarios can be simulated with low incidence but high risk. Within medical training, this freedom comes in handy, as one can become very familiar with complex procedures and their associated tools in a short period of time. Abnormalities, patients with specific complications or even different medical team compositions can be simulated effortlessly and frequently in this way. Third, once the PLS system is implemented in the clinic, the simulations can support clinical decision-making. Signals measured in the infant can be used for a patient-specific simulation. This simulation can then estimate signals that cannot be measured. Based on the estimated and measured signals, advice can be given to the physician about PLS system settings and, if necessary, possible interventions.
Although final validation of the PLS system will need to include animal testing, we propose a simulation-based validation by combining the physical manikin with a digital twin. The use of simulation technology to validate a “life-support” system is unprecedented in neonatology. A proof-of-concept prototype is expected in about 2 years, when the Horizon 2020 PLS project comes to an end.
Translated from the original article on MTintegraal.nl
This translated version is published with permission from MTintegraal. Redistribution of this English version is not permitted.
 Glass, HC, Costarino, AT, Stayer, SA, Brett, CM, Cladis, F., & Davis, PJ Outcomes for extremely premature infants. Anesthesia and analgesia (2015). 120(6), 1337–1351.
 Hornick, MA, Davey, MG, Partridge, EA, Mejaddam, AY, McGovern, PE, Olive, AM, … Flake, AW Umbilical cannulation optimizes circuit flows in preterm lambs supported by the EXTra-uterine Environment for Neonatal Development ( EXTEND). The Journal of Physiology (2018). 596(9), 1575–1585. doi:10.1113/jp275367
 Partridge, EA, Davey, MG, Hornick, MA, McGovern, PE, Mejaddam, AY, Vrecenak J. D… Flake, AW An extrauterine system to physiologically support the extreme premature lamb. Nature Communications. (2017). 8, 1–15;
 Patel RM Short- and Long-Term Outcomes for Extremely Preterm Infants. American Journal of Perinatology. (2016). 33 (3), 318–328.
 Romanis EC. Artificial womb technology and clinical translation: Innovative treatment or medical research? Bioethics. (2019). 00:1–11.
 Torchin, H., & Ancel, P.-Y. Epidemiologie et facteurs de risque de la prématurité. Journal de Gynécologie Obstétrique et Biologie de la Reproduction. (2016). 45(10), 1213–1230.
 Usuda, H., Watanabe, S., Saito, M., Sato, S., Musk, GC, Fee, ME, Kemp, MW Successful use of an artificial placenta to support extremely preterm ovine fetuses at the border of viability. American Journal of Obstetrics and Gynecology. (2019). 221 (1).