Method of predicting cardiovascular behavior, method of use, program, medium and system

Abstract

In particular, the present disclosure relates to a computer-implemented method for predicting cardiovascular behavior of a neonate at birth from fetal physiological parameters. The method comprises obtaining physiological parameters of the fetus, obtaining a non-calibrated alternative cardiovascular model that models the cardiovascular system of the fetus and at least one physiological change caused by birth, applying a data assimilation algorithm to the non-calibrated alternative cardiovascular model by using the obtained physiological parameters of the fetus, calibrating the non-calibrated alternative cardiovascular model to obtain a calibrated alternative cardiovascular model of the fetus, and predicting cardiovascular behavior of the neonate by inducing at least one physiological change in the calibrated alternative cardiovascular model of the fetus to obtain a calibrated alternative cardiovascular model of the neonate.

Description

Technical Field

[0001] This disclosure relates to the field of computer programs and systems, and more specifically to methods, systems, and programs for predicting the cardiovascular behavior of a newborn at birth based on fetal physiological parameters. Therefore, this disclosure relates to the field of pediatric cardiology, with a particular focus on congenital heart disease. Background Technology

[0002] The transition from fetal or prenatal circulation to neonatal circulation is a major event in human life and presents significant challenges to the body. The key physiological changes (morphological and functional) during this transition include umbilical cord clamping, respiration, and intense pressure changes between systemic and pulmonary circulation, which trigger valve and arterial closure to separate systemic and pulmonary circulation. In most cases, the transition from prenatal to neonatal circulation is not problematic for the newborn.

[0003] However, in the presence of congenital heart disease (CHD), this transformation can be fatal if left unattended. In fact, CHD is a broad term for fetal heart malformations. Known malformations include, but are not limited to, hypoplastic left ventricle syndrome, complete transposition of the great arteries (also known as right rotation of the great arteries), tetralogy of Fallot, and truncus arteriosus.

[0004] These malformations alter normal heart physiology as well as lung and systemic circulation, so once the transition from fetus to newborn is complete, the child's life may be threatened even if flow, oxygen, and nutrients are now well distributed.

[0005] An example of a critical scenario is hypoplastic left heart syndrome (HLHS), in which the left ventricle and surrounding cardiac structures do not develop properly to sustain life. In HLHS, the subject requires emergency surgery within days.

[0006] Early detection of CHD can be achieved through prenatal imaging, primarily using ultrasound (US). This detection is confirmed by postnatal US imaging, which can be supplemented by magnetic resonance imaging (MRI). These images provide morphological and functional information about the subject. In the case of HLHS subjects, morphological biomarkers such as aortic size, left ventricle (LV), right ventricle (RV), and valve size are considered. Functional biomarkers include blood flow velocity in pulsed or continuous wave Doppler US and qualitative assessments in color Doppler US.

[0007] These tests allow for confirmation of the malformation and determination of the severity of the condition in the patient. For example, in patients with HLHS, the surgeon must perform the Norwood intervention (or equivalent) within a few days (4 days) after birth because the newborn's heart will not be able to pump the necessary blood into the systemic circulation.

[0008] Decisions regarding the treatment of these subjects are difficult and ambiguous due to several factors, including the complexity of the process; the high risk of death; the irreversibility of the process (many structures are intentionally damaged and rearranged); and the disruption to the patient's entire life caused by subsequent surgeries, regular hospital visits, and medical examinations. Furthermore, the uncertainty of treatment is very high because, in some cases, corrective or less destructive surgeries can be performed on the patient.

[0009] One problem is that both prenatal and postnatal ultrasound imaging require the subject to remain largely still, which is difficult (almost impossible) for the fetus or newborn. Therefore, the image quality may be insufficient (accurate) to obtain a clear view and identify malformations.

[0010] To support healthcare teams in the decision-making process, digital twins have become a support tool for physicians, as seen in works such as "A. Quarteroni, L. Dede', F. Regazzoni, and C. Vergara, 'A mathematical model of the human heart suitable to address clinical problems,'" Jpn. J. Ind. Appl. Math. This is discussed in "Apr. 2023, doi: 10.1007 / s13160-023-00579-6".

[0011] Digital twins can be constructed from the fetal, newborn, or adult stages, as in "PM Trusty" et al. , “Fontan Surgical Planning: Previous Accomplishments, Current Challenges, andFuture Directions,” J Cardiovasc. Transl. Res. This is discussed in "[Article Title], vol. 11, no. 2, pp. 133–144, Apr. 2018, doi: 10.1007 / s12265-018-9786-0". Typically, digital twins are generated from anatomical and functional information obtained from patients, primarily collected from 3D images. Figure 1The diagram illustrates a general pipeline for generating such a 3D model. The inputs are cardiac magnetic resonance (CMR) images used to create detailed illustrations of the subject's heart and arteries, and phase-contrast magnetic resonance (PC-MRI) images used to determine flow velocity. Based on these two inputs, a 3D mesh representing the subject's heart and arteries is computed. Computational fluid dynamics (CFD) calculations are then used to obtain a 3D model with a velocity representation of blood flow. However, this virtual construction method has several drawbacks. First, the full-scale 3D digital twin used in clinical practice for planning congenital heart disease requires time-consuming tasks, including manual preprocessing of several hours to several days, as discussed by PM Trusty et al. These processes include image segmentation, registration, and tracking using specialized tools. Second, growing computational resources must be reserved for long-term computations (from days to weeks). Furthermore, model calibration also takes time to adapt to the subject. In other words, obtaining a model using this method takes too long, making it impossible to perform interventions within the appropriate time window.

[0012] Mitigating these technical problems relies on surrogate models, such as 0D or lumped parameter modeling (LPM), which are effective alternatives to 3D. A 0D model of the cardiovascular system is a simplified representation of the system's components, constructed from equivalent sets of electrical elements such as resistors, capacitors, and inductors. Figure 2 and Figure 2 (Continued 1) Figure 2 (Continued 2) Depicts the combination of 0D and 3D models of the human circulatory system, as in S. Pant. et al. , “Multiscale modeling of Potts shunt as a potential palliative treatment for suprasystemicidiopathic pulmonary artery hypertension: a paediatric case study,” Biomech. Model. Mechanobiol. As discussed in , vol. 21, no. 2, pp. 471–511, Apr. 2022, doi: 10.1007 / s10237-021-01545-2. The main advantage is the reduced computational cost compared to 3D models. However, as Figure 2 and Figure 2 (Continued 1) Figure 2As shown in (continued 2), there is still a 3D representation of the cardiovascular system, which means that more computational resources are still needed. Furthermore, while OD models allow for flexible modeling and rapid simulation of patient physiology under different clinical conditions (e.g., prenatal and postnatal, preoperative and postoperative), as seen in S. Pant et al., “Multiscale modeling of Potts shunt as a potential palliative treatment for suprasystemic idiopathic pulmonary artery hypertension: a paediatric casestudy,” Biomech. Model. Mechanobiol., vol. 21, no. 2, pp. 471–511, Apr. 2022, doi: 10.1007 / s10237-021-01545-2, S. Pant, C. Corsini, C. Baker, T.-Y. Hsia, G. Pennati, and IE Vignon-Clementel, “Inverse problems in reduced order models of cardiovascular haemodynamics: aspects of data assimilation and heart rate variability,” JR Soc. Interface, vol. 14, no. 126, p. 20160513, Jan. 2017, doi: 10.1098 / rsif.2016.0513 , S. Pant, C. Corsini, C. Baker, T.-Y.Hsia, G. Pennati, and IE Vignon-Clementel, “A Lumped Parameter Model to Study Atrioventricular Valve Regurgitation in Stage 1 and Changes AcrossStage 2 Surgery in Single Ventricle Patients,” IEEE Trans. Biomed. Eng., vol.65, no. 11, pp. 2450–2458, Nov. 2018, doi: 10.1109 / TBME.2018.The literature discussed in 2797999, CD Sá-Couto, P. Andriessen, WL Van Meurs, D. Ayres-De-Campos, and PM Sá-Couto, “A Model for Educational Simulation of Hemodynamic Transitions at Birth,” Pediatr. Res., vol. 67, no. 2, pp. 158–165, Feb. 2010, doi: 10.1203 / PDR.0b013e3181c2def3, focuses on general modeling of common patients or modeling of pathology-specific aspects. However, the routine clinical use of digital twins requires a comprehensive pipeline for their construction and calibration, and evaluation of general biomarkers by clinicians.

[0013] In this context, there is still a need for an improved method for predicting neonatal cardiovascular behavior from fetal physiological parameters. Summary of the Invention

[0014] Therefore, a computer-implemented method is provided for predicting neonatal cardiovascular behavior at birth based on fetal physiological parameters. The method includes: - Obtain the physiological parameters of the fetus described in (A), particularly those inferred from medical images; - Obtain (C1) a non-calibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus and for modeling at least one physiological change caused by the birth; - By applying a data assimilation algorithm to the uncalibrated alternative cardiovascular model using the obtained physiological parameters of the fetus, the uncalibrated alternative cardiovascular model is calibrated (C2) to obtain (D1) the calibrated alternative cardiovascular model of the fetus; -A calibrated alternative cardiovascular model of the newborn is obtained by predicting the cardiovascular behavior of the newborn (E1) by inducing at least one physiological change in the calibrated alternative cardiovascular model of the fetus.

[0015] In the example, the method may include one or more of the following: - The at least one physiological change resulting from the birth is selected from: placental clamping of the newborn; respiration of the newborn; closure of the foramen ovale of the newborn; and / or closure of the ductus arteriosus of the newborn; - Prior to obtaining (C1), obtaining (B) a general uncalibrated alternative cardiovascular model for simulating the cardiovascular system of a general fetal population; and further comprising: constructing an uncalibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus by combining the general uncalibrated alternative cardiovascular model obtained (B) with the obtained fetal physiological parameters (A), such that the obtained uncalibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus includes at least all of the obtained physiological parameters of the fetus; - The general fetal group of the universal uncalibrated alternative cardiovascular model includes at least one cardiovascular behavior that represents a cardiovascular malformation, preferably one of the at least one cardiovascular behavior that represents a cardiovascular malformation in the fetus; - The obtained uncalibrated alternative cardiovascular model used to model the cardiovascular system of the fetus is a lumped parameter model; - The data assimilation algorithm used for calibration (C1) is an unscented Kalman filter; - Predictive biomarkers of the fetus are extracted from the calibrated alternative cardiovascular model of the fetus (D2), and each predicted biomarker provides a quantitative assessment of the biomarker; - Predictive biomarkers of the newborns are extracted from the calibrated alternative cardiovascular model of the newborns (E2), and each of the predicted biomarkers provides a quantitative assessment of the biomarker; - Calculate the transition of biomarkers for each extraction from the biomarkers of the fetus (D2) and the neonatal extraction (E2); - The biomarkers are: morphological biomarkers, such as fetal and / or neonatal size, heart rate, aortic diameter, left ventricular size, right ventricular size, or valve size; and / or functional biomarkers, such as blood flow velocity; - The uncalibrated alternative cardiovascular model is an alternative model, preferably an OD alternative model.

[0016] A computer-implemented method is also provided, which uses the method described above to simulate surgical interventions on the cardiovascular system of a fetus and / or a newborn. The method includes: inputting physical changes in the cardiovascular system of the fetus by modifying the calibrated fetal alternative cardiovascular model and / or inputting physical changes in the cardiovascular system of the newborn by modifying the calibrated alternative cardiovascular model of the newborn; and re-performing the prediction of the newborn's cardiovascular behavior.

[0017] A computer program is also provided, including instructions for performing the method and / or the method of use.

[0018] A computer-readable storage medium having the computer program recorded thereon is also provided.

[0019] A system is also provided, including a processor coupled to a memory and a graphical user interface, wherein the computer program is recorded on the memory. Attached Figure Description

[0020] A non-limiting example will now be described with reference to the accompanying drawings, in which: - Figure 1 A known example of a pipeline for generating 3D models from image data is shown; - Figure 2 and Figure 2 (Continued 1) Figure 2 (Continued 2) shows known examples of 0D models; - Figure 3 An example of a flowchart of the present invention is shown; - Figure 4 An example of an OD model of the fetal cardiovascular system is shown; - Figure 5 It shows Figure 4 An example of a 0D model, which has a representation of the physical changes triggered by birth; - Figure 6 An example of the use of an unscented Kalman filter is shown; - Figure 7 Another example of the flowchart of the present invention is shown; - Figure 8 and Figure 8 (Continued) Examples of different paths to the abnormal origin of the coronary artery and their corresponding equivalent circuit representations are shown; - Figure 9a , Figure 9a (Continued) and Figure 9b Examples of predicted postnatal biomarkers are shown; - Figures 10a to 10c Examples of the transition of circulation from fetal to neonatal in key vascular structures are shown; and - Figure 11 An example of a computerized system is shown. Detailed Implementation

[0021] refer to Figure 3 The flowchart presents a computer-implemented method for predicting neonatal cardiovascular behavior at birth based on fetal physiological parameters.

[0022] "He / She" refers to a fetus that can be male or female, but is an intersex fetus.

[0023] "Predicting cardiovascular behavior" means providing a quantitative statement that predicts what will be observed under specific conditions. "Cardiovascular behavior" refers to the actions or responses of the heart and blood vessels (in this case, the birth of a newborn).

[0024] "Newborn birth" involves the transition from the (fetal) prenatal environment to the external world at birth. This transition involves significant changes in several physiological systems, including the cardiovascular system. Regarding fetal circulation, before birth, the fetus receives oxygen and nutrients from the mother through the placenta. Fetal circulation is characterized by certain unique physiological parameters, such as the presence of fetal shunts, like the ductus arteriosus and foramen ovale, which allow blood to bypass the lungs, as they are non-functional in the uterus. At birth, several changes occur to facilitate the newborn's adaptation to breathing air and independent circulation: 1) Closure of fetal shunts: As the newborn begins to breathe and the lungs expand, allowing blood to flow into them for oxygenation, the ductus arteriosus and foramen ovale typically close within hours to days after birth. 2) Increased pulmonary blood flow: With the onset of breathing, blood flow to the lungs increases rapidly to facilitate oxygenation. 3) Increased systemic vascular resistance: As the newborn's lungs expand and oxygen levels rise, systemic vascular resistance increases, contributing to the establishment of independent pulmonary and systemic circulation. 4) Blood pressure fluctuations: As the cardiovascular system adjusts to the extrauterine environment, blood pressure may fluctuate. 5) Neonatal circulation: After the transition period, the neonatal cardiovascular system is configured similarly to that of an adult, with blood flowing in parallel through the lungs and systemic circulation.

[0025] The cardiovascular system continues to mature over the next few weeks and month.

[0026] Physiological parameters are measurable characteristics or factors related to the function of an organism's bodily systems. Generally, these parameters provide valuable information about an organism's health, function, and homeostasis. Fetal physiological parameters can vary throughout pregnancy and can be monitored to assess fetal health during pregnancy. Physiological parameters may include vital signs such as fetal heart rate (FHR), fetal oxygenation, and blood pressure.

[0027] Still referencing Figure 3 The method includes obtaining (A) physiological parameters of the fetus, particularly physiological parameters inferred from medical images. "Obtaining" means that the system performing the method (e.g., a computerized system) has access to the physiological parameters (their data representations), which have been measured by known means, particularly but not limited to medical imaging techniques. The physiological parameters (their data representations) may be stored in the system's memory for use in subsequent steps of the method.

[0028] Examples of physiological parameters inferred from medical images will now be discussed, and it should be understood that they are not limited to these. Magnetic resonance imaging (MRI) can provide detailed images of different types of tissues based on their density and components; it can distinguish muscles, fat, and various organs based on the MRI signal characteristics of muscles, fat, and various organs. Dynamic contrast-enhanced MRI (DCE-MRI) can be used to assess blood flow and tissue perfusion in organs such as the brain, heart, and liver. Diffusion-weighted MRI (DW-MRI) measures the diffusion of water molecules within tissues. As another example of medical imaging techniques, ultrasound (US) imaging can assess blood flow and velocity within blood vessels. For example, B-mode ultrasound (also known as brightness-mode ultrasound) is a medical imaging technique that uses sound waves to produce real-time two-dimensional images of internal structures of the body. For example, color Doppler ultrasound is a medical imaging technique that combines conventional B-mode ultrasound imaging with Doppler ultrasound to visualize blood flow within blood vessels in the body. It provides real-time color-coded images that depict the direction and velocity of blood flow, superimposed on grayscale anatomical images obtained from B-mode ultrasound. For example, PW-Doppler, or pulsed-wave Doppler, is a specific mode of Doppler ultrasound used in medical imaging to measure the velocity of blood flow within blood vessels. Unlike color Doppler, which provides information about the direction and velocity of blood flow over a wide range, PW-Doppler allows for precise measurement of blood flow at specific locations within a blood vessel.

[0029] Fetal physiological parameters may include, but are not limited to, gestational age, estimated weight, and size. Some of these parameters can also be obtained through imaging techniques.

[0030] Next, (C1) an uncalibrated alternative cardiovascular model is obtained for modeling the fetal cardiovascular system and for modeling at least one physiological change induced by birth. Obtaining means that the system performing the method (e.g., a computerized system) has access to the uncalibrated alternative cardiovascular model (its data representation) modeling the fetal cardiovascular system. The uncalibrated alternative cardiovascular model (its data representation) modeling the fetal cardiovascular system can be stored in the system's memory for use in subsequent steps of the method. The uncalibrated alternative cardiovascular model is a simplified representation of the fetal cardiovascular system that does not depend on a specific calibration procedure or patient-specific data. Instead of precisely modeling individual fetal physiological parameters such as blood pressure, heart rate, or vascular resistance, the uncalibrated alternative cardiovascular model of the fetus uses general fetal parameters or group-based parameters to simulate cardiovascular dynamics.

[0031] Uncalibrated alternative cardiovascular models for modeling the fetal cardiovascular system are a type of alternative model, i.e., a simplified and computationally faster model representing the fetal cardiovascular system. In one example, an uncalibrated alternative cardiovascular model could be an 0D model. An 0D model is an alternative to a high-fidelity model that lacks spatial dimensions and has parameters describing the overall system dynamics, such as blood flow rate, pressure, and volume. A high-fidelity model is a complex and complete 3D model. In another example, an uncalibrated alternative cardiovascular model could be an AI-based model that replaces the 0D model. Examples of AI-based models include trained algorithms learned from past experiences (solutions from one or more high-fidelity models), as described in P. Vurtur Badarinath, M. Chierichetti, and F. Davoudi Kakhki, “AMachine Learning Approach as a Surrogate for a Finite Element Analysis: Status of Research and Application to One Dimensional Systems,” Sensors This is discussed in vol.21, no. 5, p. 1654, Feb. 2021.

[0032] In the example, the 0D model is a lumped-element model, a simplified representation of the fetal cardiovascular system that assumes all components are concentrated at a single point and their behavior can be described by an idealized mathematical model. Lumped-element models are known in the art.

[0033] In the context of the fetal cardiovascular system, lumped parameter models are used to represent the complex network of blood vessels, valves, and the heart as a simplified system with discrete compartments or “lumped” elements. Each element represents a set of similar components with similar behaviors. The entire systemic circulation can be represented by several compartments. For example, an arterial compartment can represent a group of large arteries, such as the descending aorta, iliac, and femoral arteries. As another example, a capillary compartment represents a network of tiny blood vessels where nutrient and gas exchange occurs. A further example is a venous compartment, which represents the large veins that return blood to the heart. In these examples, each compartment is characterized by parameters such as compliance (stretchability), resistance (flow hindrance), and volume (capacity to store blood). It should be understood that lumped parameter models of the fetal cardiovascular system can include many other types of compartments.

[0034] Figure 4 An example of a lumped parameter model of the fetal cardiovascular system is shown. This universal prenatal digital twin architecture has been used with Dymola.TM Modeling was performed, Dymola TM This is a commercial software tool for modeling and simulating dynamic systems, released by Dassault Systèmes™. This example of a lumped parameter model has the ability to simulate birth transitions. In the center of the figure, sphere 400 represents the right ventricle, and sphere 402 represents the left ventricle. Arrow 410 points to the ductus arteriosus (DA), and arrow 412 points to the foramen ovale (FO). The model also includes the coronary artery tree 420 supplying the myocardium. On the left side of the figure, the respiratory circulation is modeled using blocks of the pulmonary artery (PA) 430 and pulmonary vein (PV) 432. Figure 4 On the right side, the systemic circulation is modeled using the internal thoracic vein 440 and external thoracic vein 442 (coronary artery tree 420 is at the top of the figure). The systemic circulation also includes the placenta (PL) 444.

[0035] Referring again to C1, at least one physiological change induced by birth is modeled. At C1, the modeled changes have not yet been applied to an uncalibrated alternative cardiovascular model for modeling the fetal cardiovascular system, but it is stored, for example, on a storage medium so that it can be applied to subsequent steps of the method, as will be discussed below.

[0036] In a normal birth, the physiological changes triggered by birth begin at the start of labor and are crucial for the transition from intrauterine to extrauterine life.

[0037] Now let's discuss an example of the physiological changes triggered by birth. In this example, the clamping of the placenta in the newborn can be modeled. The clamping of the placenta in the newborn is a common process during birth and is usually performed by clamping the umbilical cord that connects the newborn to the placenta. Now refer to... Figure 5 An example, which represents Figure 4 The model has been applied to model examples of physiological changes triggered by birth. Placental clamping triggers... Figure 4 Complete inhibition of flow on the right branch 444 (enclosed by an ellipse). Therefore, in this example, the modeling of neonatal placental clamping includes the closure of the right branch 444.

[0038] In the example, the physiological changes triggered by birth may include the newborn's breathing. Gas exchange in the lungs is activated. Figure 5 In the example, oxygen uptake changes from placenta 444 (oxygen uptake occurs in the placenta before the umbilical cord clamps) to lungs 432 (oxygen uptake occurs in the lungs after the umbilical cord clamps), for example, increasing linearly over one minute. Therefore, modeling neonatal breathing involves, for example, activating element lungs 432 as oxygen is introduced into the neonatal cardiovascular system with a linear increase (e.g., from 0% to 100%) over approximately one minute.

[0039] In the example, physiological changes triggered by birth may include the closure of the foramen ovale in newborns. Figure 5 In the example, the closure of segment 412 in the circuit represents this physiological change, which is controlled by the pressure gradient between the left and right atria. Therefore, modeling the foramen ovale closure in newborns involves closing segment 412 in an uncalibrated surrogate cardiovascular model simulating the fetal cardiovascular system.

[0040] In the example, physiological changes triggered by birth may include the closure of the ductus arteriosus in newborns. This change can be represented by signals that decrease the diameter of the arteries, thereby increasing their resistance. Figure 5 In the example, element 414, representing a circuit that decreases the diameter of the artery, activates the closure of the ductus arteriosus in the newborn.

[0041] The examples of physiological changes presented individually should be understood to mean that any combination of two or more of these examples can be performed when at least one physiological change induced by birth is modeled (C1).

[0042] Back Figure 3 At point C2, the uncalibrated alternative cardiovascular model is calibrated using data assimilation algorithms, employing physiological parameters obtained from the fetus on the uncalibrated alternative cardiovascular model. The uncalibrated alternative cardiovascular model is suitable for capturing the desired components of the fetal cardiovascular system. Functional data, i.e., the obtained fetal physiological parameters, are used for this purpose.

[0043] The obtained fetal physiological parameters include all numerical model parameters, such as resistance or elasticity defined in the uncalibrated alternative cardiovascular model. Calibration thus enables the digital twin to reproduce the fetal (prenatal) physiological state. To achieve this, a data assimilation algorithm is used. As is known in the art, the data assimilation algorithm combines the observations (the obtained fetal physiological parameters) with the uncalibrated alternative cardiovascular model to produce an estimate of the state of the uncalibrated alternative cardiovascular model; as a result of calibration, a calibrated alternative cardiovascular model of the fetus is obtained.

[0044] Numerous data assimilation methods exist in the literature, including variational, sequence, and deep learning methods, as discussed in RW Mayet al., “From fetus to neonate: A review of cardiovascular modeling in early life,” WIREs Mech. Dis., p. e1608, Mar. 2023, doi: 10.1002 / wsbm.1608. Any of these methods can be used in this invention.

[0045] In the example, the data assimilation algorithm used to calibrate the uncalibrated alternative cardiovascular model is the unscented Kalman filter (UKF), as discussed, for example, in E.A. Wan and R. Van Der Merwe, “The unscented Kalman filter for nonlinear estimation,” in Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat.No.00EX373), Lake Louise, Alta., Canada: IEEE, 2000, pp. 153–158. doi:10.1109 / ASSPCC.2000.882463. The unscented Kalman filter is a sequential and recursive algorithm that assimilates dynamical systems to noisy observational time-series data. This is an adaptation of the classical Kalman filter to nonlinear models. This is particularly well-suited for cardiovascular systems because cardiovascular lumped-parameter models are nonlinear.

[0046] Figure 6 The results of applying the UKF to an equivalent but simplified inverse problem (with only one variable) are shown, where the length of the pendulum is found (on the right-hand plot). Figure 6 In this diagram, angular velocity, x, and y coordinates are measurements represented by points on the left-hand plot. These measurements are generated using a mathematical model represented by a continuous line (on the left-hand curve plot) after adding some noise. This is a very simple example, and the process is the same when the UKF is applied to calibrate a cardiovascular system. The main difference is that the number of parameters to be calibrated is ~200.

[0047] In the example, filtering in the frequency domain (LO Muller, A. Caiazzo, and PJ Blanco, “Reduced-Order Unscented Kalman Filter With Observations in the Frequency Domain: Application to Computational Hemodynamics,” IEEE Trans. Biomed. Eng., vol. 66, no. 5, pp. 1269–1276, May 2019, doi: 10.1109 / TBME.2018.2872323), reduction formulas, and sensitivity analysis procedures (A. Caiazzo, F. Caforio, G. Montecinos, LO Muller, PJ Blanco, and EF Toro, “Assessment of reduced-order unscented Kalman filter for parameter identification in 1-dimensional blood flow models using experimental data: Kalman filter for 1D blood flow models and in vitro data,” Int. J. Numer. Methods Biomed. Eng., vol. 33, no. 8, p. e2843, Aug. 2017, doi: 10.1002 / cnm.2843), a specialized version of the UKF cardiovascular application can be used that takes into account different heart rates from multiple medical examinations (S. Pant, C. Corsini, C. Baker, T.-Y. Hsia, G. Pennati, and IE Vignon-Clementel, “Inverse problems in reduced order models of cardiovascular haemodynamics: aspects of data assimilation and heart rate variability,” JR Soc. Interface, vol. 14, no. 126, p. 20160513, Jan. 2017, doi: 10.1098 / rsif).(May 13, 2016) to optimize filtering performance.

[0048] Calibration has been performed at D1, yielding a substitute cardiovascular model of the fetus (and thus usable for further steps in the method). Therefore, D1 corresponds to a calibrated prenatal digital twin, whose architecture is... Figure 4 As shown in the figure. Based on this calibrated alternative cardiovascular model of the fetus, prenatal biomarkers can be delivered to the medical team for assessment, such as the state of the fetal cardiovascular system. These prenatal biomarkers are already valuable to physicians because they represent quantitative assessments of a patient's health status (and are not the primary qualitative assessments known in the art), and provide information on variables not directly measured from the fetus, such as oxygen saturation and flow rate.

[0049] Still referencing Figure 3 At E1, neonatal cardiovascular behavior is predicted by inducing at least one physiological change in a calibrated alternative cardiovascular model of the fetus. Inducing at least one physiological change means modifying the obtained calibrated alternative cardiovascular model of the fetus to reflect a model (A) that reflects at least one physiological change. Examples of birth-induced physiological changes have been discussed. Inducing at least one physiological change means modifying the calibrated alternative model by applying a model of the physiological change (e.g., parameters of the calibrated alternative cardiovascular model) with new values ​​provided by the model of at least one birth-induced physiological change.

[0050] As a result of the changes, a calibrated alternative cardiovascular model of the newborn is obtained. Therefore, based on the prenatal digital twin, the birth of the fetus can be simulated, and a postnatal digital twin of him / her is generated.

[0051] Therefore, this method proposes generating a personalized digital twin of the subject at the prenatal age (i.e., fetal age) capable of predicting the neonatal (postnatal) physiological condition, such as warning of infeasible situations after birth. To this end, subject data (medical imaging and non-image-derived data) will be collected from the subject during the fetal stage. The digital twin of the fetal cardiovascular system will then be calibrated to the patient data using a data assimilation algorithm. Once the digital twin is personalized, the neonatal physiological condition is predicted by simulating the patient's birth using the digital twin. This postnatal simulation can provide postnatal physiological conditions, such as oxygen saturation of major / critical blood vessels (e.g., aorta, coronary arteries, and veins). Thus, if the subject's postnatal condition jeopardizes his / her survival, all necessary measures can be planned prenatally to ensure the subject's postnatal viability. Furthermore, this digital twin reduces subjective interpretation of patient data analysis and provides a platform for physicians to simulate (and test through simulation) different treatment strategies by combining prenatal and postnatal stages, as described below.

[0052] This method improves the prediction of neonatal cardiovascular behavior at birth from fetal physiological parameters. Notably, the method provides a quantitative assessment of the subject's condition, rather than a qualitative interpretation by a medical expert. This quantification reduces the time required to understand the subject's cardiovascular status, increases confidence in any decisions made by the medical team, and, importantly, provides additional information on the optimal treatment for a given subject. The method can be used as a tool for comparison to generalize and understand CHD, thereby facilitating comparisons between subject conditions. For example, this method according to the invention enables successful corrective interventions instead of palpation / destructive procedures (such as the Norwood intervention), thereby reducing additional interventions and medical visits to the subject and lowering costs. Furthermore, the choice of OD model keeps simulation time sufficiently short to evaluate multiple scenarios under time pressure. It should be understood that this method only provides instructions and options; the choice of using them remains the physician's discretion.

[0053] Now for reference Figure 7 Other examples of the method will be discussed. Figure 7 It is a combination of these other examples, and it should be understood that these examples can be used individually with... Figure 3 The method combination, or any combination of these examples, can be used with... Figure 3 A combination of methods.

[0054] In the example, the uncalibrated alternative cardiovascular model used to model the cardiovascular system of the fetus was obtained from (B) a general uncalibrated alternative cardiovascular model that models the cardiovascular system of a general fetal population and the (A) physiological parameters of the fetus.

[0055] A universal uncalibrated alternative cardiovascular model models the cardiovascular system of a general fetal population. This means that a universal uncalibrated alternative cardiovascular model includes all the discrete compartments or “lumps” elements required to model a given general fetal population using an uncalibrated alternative cardiovascular model.

[0056] Therefore, the genericity of a general uncalibrated alternative cardiovascular model depends on the definition of a general fetal population. In the examples, a general fetal cohort can be validated on cohort data to reflect the average subjects, such as in, for example, CD Sá-Couto, P. Andriessen, WL Van Meurs, D. Ayres-De-Campos, and PM Sá-Couto, “A Model for Educational Simulation of Hemodynamic Transitions at Birth,” Pediatr. Res., vol. 67, no. 2, pp. 158–165, Feb. 2010, doi: 10.1203 / PDR.0b013e3181c2def3, or AG Munneke, J. Lumens, and T. Delhaas, “Cardiovascular fetal-to-neonatal transition: an in-silico model,” Pediatr. Res., vol. 91, no. 1, pp. 116–128, Jan. 2022, doi: In 10.1038 / s41390-021-01401-0. In another example, the universal model may be limited to one or more CHDs; the general fetal group is limited to a group of fetuses suffering from said one or more CHDs. Restricting the fetal group allows for obtaining a universal uncalibrated model that more closely approximates the physiological reality of the subject. In particular, obtaining a universal uncalibrated model of the cardiovascular system of a restricted fetal group (e.g., fetuses suffering from at least one (or more) CHDs) improves the calibration process.

[0057] A general uncalibrated alternative cardiovascular model is obtained by constructing a fetal uncalibrated alternative cardiovascular model by combining the obtained (B) general uncalibrated alternative cardiovascular model with the obtained (A) fetal physiological parameters. This construction allows for obtaining a simplified mathematical representation of the fetal cardiovascular system, which may include specific morphological abnormalities of the fetus, depending on the selected general fetal population as previously discussed.

[0058] In the example, the construction may include performing manual, semi-automatic modifications to the obtained (B) generic uncalibrated alternative cardiovascular model based on the acquired images and / or available patient-specific data. For example, valves and cardiac chambers can be automatically labeled and segmented from ultrasound images, as shown in the following documents: T. Kim, M. Hedayat, VV Vaitkus, M. Belohlavek, V. Krishnamurthy, and I. Borazjani, “Automatic segmentation of the left ventricle in echocardiographic images using convolutional neural networks,” Quant. Imaging Med. Surg., vol. 11, no. 5, pp. 1763–1781, May 2021, doi: 10.21037 / qims-20-745; P. Carnahan, J. Moore, D. Bainbridge, M. Eskandari, ECS Chen, and TM Peters, “DeepMitral: Fully Automatic 3DEchocardiography Segmentation for Patient Specific Mitral Valve Modelling,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2021; B. Yuan et al., “Automatic valve segmentation in cardiac ultrasound timeseries data,” in Medical Imaging 2018: Image Processing, ED Angelini and B. A. Landman, Eds., Houston, United States: SPIE, Mar. 2018, p. 69. doi:10.1117 / 12.2293255; A. Arafati et al., "Generalizable fully automated multi-label segmentation of four-chamber view echocardiograms based on deepconvolutional adversarial networks," JR Soc. Interface, vol. 17, no. 169, p. 20200267, Aug. 2020, doi: 10.1098 / rsif.2020.0267. .

[0059] As another example, if specific physiological parameters of a subject are not present in the general model, such as, for example, the origin of the coronary arteries, which can vary widely, predefined artery configuration groups can be used in the user's treatment to modify the general uncalibrated alternative model. From this artery configuration group, the user can select the closest representation of the subject's artery tree and insert it into the main architecture. For example, in the case of coronary heart disease, this is done by replacing the corresponding general coronary block with the selected coronary artery tree in the artery configuration group. Another way to construct specific arterial segments with unique branches and / or duplicated arteries is by drawing connections, which can be done using Dymola™.

[0060] refer to Figure 8 and Figure 8 (Continued) An example of a predefined group of arterial configurations for handling users is shown. Each configuration shows a different path of the coronary artery, where the RCA is the right coronary artery, the LAD is the left anterior descending artery, and the LCX is the left circumflex artery. The physiological configuration labeled A represents the normal anatomy; its equivalent circuit representation is shown in the upper right of the figure. The physiological configuration labeled B represents the RCA traveling interarterially from the tip of the left coronary artery; its equivalent circuit representation is shown in the middle of the right side of the figure. The physiological configuration labeled C represents the left main coronary artery (LM) traveling anteriorly from the tip of the right coronary artery. The configuration labeled D represents the LM traveling interarterially from the tip of the right coronary artery, and the configuration labeled E represents the LM traveling posteriorly from the tip of the right coronary artery. These three configurations C, D, and E have equivalent circuit representations shown in the lower right of the figure; the arrows indicate the major resistance relative to the other two arterial segment capture configurations C, D, and E.

[0061] Return to reference Figure 4 The model includes a coronary artery tree 420 supplying the myocardium. Users can utilize this model in cases where fetal subjects have CHD not represented in general uncalibrated alternative cardiovascular models. Figure 8 and Figure 8(Continued) The model can be modified using an artery configuration from a predefined group of artery configurations to adapt to a general model. For example, if the fetus has a malformation represented by a physiological configuration labeled C, the coronary tree 420 of the general model can be replaced by an equivalent circuit representation of the physiological configuration labeled C, such as... Figure 8 and Figure 8 (Continued)

[0062] Therefore, the resulting uncalibrated alternative cardiovascular model for modeling the fetal cardiovascular system includes at least all physiological parameters of the fetus observed from fetal physiological parameters. If the fetus includes malformations, such as CHD, the uncalibrated alternative cardiovascular model models the physiological parameters of the fetal malformation.

[0063] As mentioned above, the genericity of the universal uncalibrated alternative cardiovascular model depends on the definition of the general fetal cohort. In the example, the general fetal cohort is selected based on specific physiological parameters of the fetus. For example, the general cohort may be limited to fetuses exhibiting at least one cardiovascular behavior that represents a cardiovascular malformation in the fetal subject. In this case, the universal uncalibrated alternative model is suitable for one (or at least one) physiological fault, such as CHD.

[0064] Return to reference Figure 7 Examples of the previous examples will now be discussed, involving blocks labeled D2, E2, and G.

[0065] In this example, predicted biomarkers for the D2 fetus are extracted from a calibrated surrogate cardiovascular model of the fetus. Biomarkers are measurable biological characteristics of a biological system. Examples of human biomarkers include weight, size, the amount of specific molecules in the blood, and blood pressure.

[0066] In the example, the extracted biomarkers can be morphological biomarkers. Morphological biomarkers refer to physical properties or characteristics of biological entities that can be objectively measured, quantified, and analyzed. These biomarkers are typically associated with the structure, shape, size, or appearance of cells, tissues, organs, or organisms. Morphological biomarkers can be, but are not limited to, fetal and / or neonatal size, heart rate, aortic diameter, left ventricular size, right ventricular size, and valve size.

[0067] In this example, the extracted biomarkers can be functional biomarkers. Functional biomarkers provide insights into the physiological or functional state of a biological system. They reflect dynamic processes or activities occurring within cells, tissues, organs, or organisms. Functional biomarkers can be, but are not limited to, blood flow velocity.

[0068] Therefore, the predicted biomarkers provide a quantitative assessment of the biomarkers, and this quantitative assessment can be performed at a given time point. It should be understood that the extraction of predicted biomarkers is performed on the obtained calibrated alternative cardiovascular model (D1) of the fetus, such as... Figure 7 As shown. The extraction of predictive biomarkers is performed as is known in the art. For example, if the obtained calibrated alternative cardiovascular model of the fetus is an 0D model, such as a lumped parameter model, the input can be a scalar value (a single-dimensional input, typically represented by a single value) passed through the model to obtain the output (the extracted biomarker).

[0069] In this example, predictive biomarkers for E2 newborns were extracted from a calibrated alternative cardiovascular model of the newborn. This was done in a similar manner to the extraction of fetal biomarkers, and comments apply.

[0070] In the example, the predicted biomarkers for the fetus (E1) and the predicted biomarkers for the newborn (E2) are combined to calculate the transition for each extracted biomarker. Calculating the transition for each extracted biomarker involves applying the same biomarker to both the fetus and the newborn. Figure 9a , Figure 9a (Continued) and Figure 9b An example of a combined report is shown, in which flow rate and oxygen saturation are depicted. Figure 9a and Figure 9a (Continued) In this section, flow rate and oxygen saturation have been predicted for different levels of ventricular hypoplasia in different circulatory components (aorta, pulmonary artery, systemic veins, and left coronary artery). On the right, the horizontal dashed line indicates the critical level of oxygen for life-threatening feasibility. Figure 9b The [NOUS ALLONS REFAIRE FAIRE LES FIGURES] indicates the evolution of the pressure volume loop (PV loop) in healthy subjects from prenatal to 2 minutes, 1 hour, 12 hours after delivery, and in the neonatal cycle.

[0071] Figure 9a , Figure 9a (Continued) and Figure 9b Different scenarios of left ventricular hypoplasia are depicted. For certain levels of congenital malformation, the patient's life is at risk because of systemic venous oxygen saturation (as determined by...). Figure 9a and Figure 9a(The horizontal dashed line in the continued text depicts a value below the physiologically feasible value.) This assessment demonstrates an explicit physiological response to a combination of cardiac malformations, meaning that any set of measurements can be collected via digital twins and visualized using this type of physiological response. Currently, this physiological response is obtained indirectly through a set of measurements (morphology and function), but it cannot compare different CHD categories, especially if different morphological and functional measurements are recorded. In contrast, this functional assessment allows for direct comparison of any CHD as well as healthy patients.

[0072] Figure 10a The diagram shows the changes in extracted biomarker flow rates during the prenatal to neonatal transition, with the left side showing the first hour after umbilical cord clamping and the right side showing the subsequent hours after birth. Figure 10a The closure of the foramen ovale and shunt of the ductus arteriosus (DA) are shown in normal subjects after birth. Figure 10b The shift in biomarker oxygen saturation extracted during the prenatal to neonatal transition in normal subjects is shown. Figure 10c The evolution of parameters pre-ductal (DA) 100, pulmonary vascular resistance (PVR) 110, and whole-body resolution (TSR) 120 applied to the calibrated alternative cardiovascular model for the fetus and newborn during the prenatal to neonatal transition is shown.

[0073] Return to reference Figure 7 Now we discuss virtual intervention G. A virtual intervention can be simulated if, based on an examination of the predicted postnatal condition (E2), changes in extracted biomarkers indicate that the subject's life is infeasible or at risk after birth. Virtual modification G allows modification of the prenatal digital twin by sequentially simulating interventions at the prenatal stage (D1) or postnatal surgery (E1), or both. Methods for predicting neonatal cardiovascular behavior at birth based on fetal physiological parameters are used to simulate surgical interventions on the cardiovascular system of the fetus and / or newborn. Physical changes in the fetal cardiovascular system are input by modifying a calibrated alternative cardiovascular model of the fetus, to simulate the outcome of the surgical intervention. Physical changes in the neonatal cardiovascular system are input by modifying a calibrated alternative cardiovascular model of the newborn, to simulate the outcome of the surgical intervention. The input calibrated alternative cardiovascular models of the fetus and / or newborn are modified, and prediction E1 is performed again. In the example, extractions D2 and E2 can be performed to calculate new changes in each extracted biomarker.

[0074] Simulated surgical interventions do not affect real surgical interventions and must be considered a decision support tool; the execution of surgical interventions remains at the discretion of the physician.

[0075] Examples of prenatal interventions include intrauterine angioplasty. Examples of postnatal interventions include the Norwood procedure or specific drug dosing regimens to prevent ductus arteriosus closure in patients with HLHS. The combined use of prenatal and postnatal digital twins allows for quantitative assessment to correct for borderline conditions detected during the fetal stage.

[0076] In a particular example, this disclosure may include a computer-implemented method for predicting fetal cardiovascular behavior from physiological parameters, the method being: - Obtain the physiological parameters of the fetus described in (A), particularly those inferred from medical images; - Obtain (C1) a non-calibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus and for modeling at least one physiological change caused by the birth; - By applying a data assimilation algorithm to the uncalibrated alternative cardiovascular model using the obtained physiological parameters of the fetus, the uncalibrated alternative cardiovascular model is calibrated (C2) to obtain (D1) the calibrated alternative cardiovascular model of the fetus; - Predictive biomarkers for the fetus (D2) are extracted from the calibrated alternative cardiovascular model of the fetus, and each predicted biomarker provides a quantitative assessment of the biomarker at a given time point.

[0077] It should be understood that this can be used as a reference. Figure 3 and Figure 7 Examples of each of the blocks A, B, C1, C2, and D1 discussed also apply.

[0078] It should be understood that this invention allows for the reuse of physiological parameters obtained after model calibration. Some physiological parameters can be extracted during the prenatal stage and then directly used in the neonatal stage. Some of these parameters are more readily available during the fetal stage, such as cerebral arterial flow.

[0079] The method is implemented by a computer. This means that the steps (or substantially all steps) of the method are performed by at least one computer or any similar system. Therefore, the execution of the steps by a computer may be fully automatic or semi-automatic. In the example, at least some steps of the method may be triggered via user-computer interaction. The required level of user-computer interaction may depend on the anticipated level of automation and be balanced with the need to fulfill the user's wishes. In the example, this level may be user-defined and / or predefined.

[0080] A typical example of a computer implementation of a method is to utilize a system suitable for this purpose to perform the method. The system may include a processor coupled to memory and a graphical user interface (GUI), on which a computer program containing instructions for performing the method is stored. The memory may also store a database. The memory is any hardware suitable for such storage and may include several physically distinct parts (e.g., one for the program, and possibly one for the database).

[0081] Figure 11 An example of the system is shown, wherein the system is a client computer system, such as a user's workstation.

[0082] The client computer in this example includes a central processing unit (CPU) 1010 connected to an internal communication bus 1000 and random access memory (RAM) 1070 also connected to the bus. The client computer is also provided with a graphics processing unit (GPU) 1110 associated with a video random access memory 1100 connected to the bus. The video RAM 1100 is also referred to in the art as a frame buffer. A mass storage device controller 1020 manages access to mass storage devices such as a hard disk drive 1030. Mass storage devices suitable for tangibly representing computer program instructions and data include all forms of non-volatile memory, such as semiconductor memory devices like EPROM, EEPROM, and flash memory devices; disks such as internal hard disks and removable disks; and magneto-optical disks. Any of the above can be supplemented or incorporated by a specially designed ASIC (Application-Specific Integrated Circuit). A network adapter 1050 manages access to a network 1060. The client computer may also include a haptic device 1090 such as a cursor control device, a keyboard, etc. A cursor control device is used on the client computer to allow the user to selectively position the cursor at any desired location on the monitor 1080. Furthermore, the cursor control device allows the user to select various commands and input control signals. The cursor control device includes multiple signal generating devices for inputting control signals to the system. Typically, the cursor control device can be a mouse, with mouse buttons used to generate signals. Alternatively or additionally, the client computer system may include a sensitive pad and / or a sensitive screen.

[0083] The computer program may include computer-executable instructions, which include means for causing the system to perform the method. The program may be recorded on any data storage medium, including the system's memory. The program may be implemented, for example, in digital electronic circuitry, or in computer hardware, firmware, software, or a combination thereof. The program may be implemented as a device, such as a product tangibly contained in a machine-readable storage device for execution by a programmable processor. The method steps may be executed by a programmable processor that executes the program of instructions to perform the function of the method by manipulating input data and generating output. The processor may therefore be programmable and coupled to receive data and instructions from the data storage system, at least one input device, and at least one output device, and to send data and instructions to the data storage system and at least one output device. The application program may be implemented in a high-level procedural or object-oriented programming language, or, if desired, in assembly or machine language. In any case, the language may be a compiled language or an interpreted language. The program may be a complete installation program or update program. The application of the program on the system will in any case generate instructions for performing the method. Alternatively, the computer program can be stored and executed on a server in a cloud computing environment that communicates with one or more clients over a network. In this case, the processing unit executes the instructions included in the program, thereby enabling the method to be performed in the cloud computing environment.

Claims

1. A computer-implemented method for predicting the cardiovascular behavior of a newborn at birth based on fetal physiological parameters, the method comprising: - Obtain the physiological parameters of the fetus described in (A), particularly those inferred from medical images; - Obtain (C1) a non-calibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus and for modeling at least one physiological change caused by the birth; - By applying a data assimilation algorithm to the uncalibrated alternative cardiovascular model using the obtained physiological parameters of the fetus, the uncalibrated alternative cardiovascular model is calibrated (C2) to obtain (D1) the calibrated alternative cardiovascular model of the fetus; -To obtain a calibrated alternative cardiovascular model of the newborn by predicting (E1) the cardiovascular behavior of the newborn by inducing at least one physiological change in the calibrated alternative cardiovascular model of the fetus.

2. The computer-implemented method of claim 1, wherein, The at least one physiological change triggered by the birth is selected from: -The placenta of the newborn is clamped; -The newborn's breathing; - The closure of the foramen ovale in the newborn; and / or - The closure of the ductus arteriosus in the newborn.

3. The computer-implemented method of claim 1 or 2, further comprising: Before obtaining (C1), - Obtain (B) a general uncalibrated alternative cardiovascular model for simulating the cardiovascular system of a general fetal population; It also includes: constructing the uncalibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus by combining the general uncalibrated alternative cardiovascular model obtained (B) with the fetal physiological parameters obtained (A), such that the obtained uncalibrated alternative cardiovascular model for modeling the cardiovascular system of the fetus includes at least all the physiological parameters of the fetus obtained.

4. The computer-implemented method of claim 3, wherein, The general fetal group of the universal uncalibrated alternative cardiovascular model includes at least one cardiovascular behavior representing a cardiovascular malformation, preferably, one of the at least one cardiovascular behavior representing a cardiovascular malformation in the fetus.

5. The computer-implemented method of any one of claims 1 to 4, wherein, The obtained uncalibrated alternative cardiovascular model used to model the cardiovascular system of the fetus is a lumped parameter model.

6. The computer-implemented method of any one of claims 1 to 5, wherein, The data assimilation algorithm used for calibration (C1) is an unscented Kalman filter.

7. The computer-implemented method according to any one of claims 1 to 6, further comprising: - Predictive biomarkers of the fetus are extracted from the calibrated alternative cardiovascular model of the fetus (D2), and each predicted biomarker provides a quantitative assessment of the biomarker.

8. The computer-implemented method according to any one of claims 1 to 7, further comprising: - Predictive biomarkers of the newborns are extracted from the calibrated alternative cardiovascular model of the newborns (E2), and each predicted biomarker provides a quantitative assessment of the biomarker.

9. The computer-implemented method according to claim 7 or 8, further comprising: - Calculate the transition of each extracted biomarker from the extracted (D2) fetal biomarkers and the extracted (E2) neonatal biomarkers.

10. The computer-implemented method according to any one of claims 7 to 9, wherein, The biomarker is: - Morphological biomarkers, including fetal and / or neonatal size, heart rate, aortic diameter, left ventricular size, right ventricular size, or valve size; and / or - Functional biomarkers, including blood flow velocity.

11. The computer-implemented method according to any one of claims 1 to 10, wherein, The uncalibrated alternative cardiovascular model is an alternative model, and preferably, the uncalibrated alternative cardiovascular model is an OD alternative model.

12. A computer-implemented method for simulating surgical intervention on the cardiovascular system of a fetus and / or newborn, the method using the method according to any one of claims 1 to 11, comprising: - Input the physical changes of the fetal cardiovascular system by modifying the calibrated fetal alternative cardiovascular model and / or input the physical changes of the neonatal cardiovascular system by modifying the calibrated alternative cardiovascular model of the neonate; - Repeat the prediction of neonatal cardiovascular behavior.

13. A computer program comprising instructions for performing the method of any one of claims 1 to 11 and / or claim 12.

14. A computer-readable storage medium having the computer program of claim 13 recorded thereon.

15. A system comprising a processor coupled to a memory and a graphical user interface, wherein the memory stores the computer program of claim 13.

Images