(C-82) Stochastic Modeling of Pulmonary Blood Flow in Neonatal Congenital Diaphragmatic Hernia Patients: Insights into the Impact of Modified Branching Architectures on Arteriole Flow and Pressure

Introduction:: Congenital diaphragmatic hernia (CDH) is a severe congenital birth defect characterized by the protrusion of abdominal organs through the diaphragm, resulting in restricted lung growth and the development of pulmonary hypoplasia and hypertension (PH)¹. CDH-PH is associated with impaired lung function caused by changes in the mechanical environment and blood flow within the pulmonary circulation². However, the small size of the affected vessels makes it challenging to measure these mechanical alterations directly. Consequently, our focus has been on modeling pulmonary blood flows in the affected vessels by analyzing architectural differences in CDH pathology to identify correlations between the severity of CDH-PH and current clinical stratifications based on hernia size and hypoplasia measurements. Current approaches for modeling blood flow use MRI data to simulate pressures and blood velocities in generated vascular networks³. However, these methods are inadequate for assessing neonatal vessel properties due to their small size. Moreover, they do not capture trends across entire patient populations, as they are typically used for individual patient treatment plans. Therefore, we have integrated clinical echocardiogram data and patient histology samples into a stochastic framework to construct virtual patient libraries of pulmonary vascular architectures and to calculate vascular flows subsequently. With these networks, we can model the mechanical alterations that occur due to CDH-PH and correlate them to current clinical stratifications that do not take PH into account in hopes of identifying improved metrics for predicting CDH outcomes.


Materials and Methods:: Pulmonary vascular velocities, pressures, and large vessel diameter measurements were compiled from clinical echocardiogram scans of CDH patients on the day of birth. Parent-daughter branch radius ratios were randomly selected at each bifurcation for each tree based on average values found in literature⁴. They were used along with echocardiogram diameter measurements of the main pulmonary artery (MPA) to construct bifurcating vessel networks stochastically. Each network consisted of large compliant vessels connected to fractal resistance vessels. The population of virtual patient networks was then sorted into stratifications based on the physiological relevance of tree structures. Input velocities, pressures, and large vessel diameters were used as boundary conditions for calculating the flow across every tree in each group of generated vascular geometry distributions. Average arteriole flow and pressure for each tree were calculated for comparison against other virtual patients.


Results, Conclusions, and Discussions:: Branching architectures consisting of large compliant arteries connected to smaller resistance vessels were assembled stochastically using bifurcating trees (Figure 1A). MPA diameters were randomly selected for each tree using a normal distribution with average and standard deviations from measured echocardiogram data. Parent-daughter branch radius ratios were chosen randomly using a beta distribution at each bifurcation. Alpha and beta parameters for the beta distributions were the same across each tree and were randomly selected by generating a mean utilizing a uniform distribution. Large virtual patient populations were sorted into sub-populations of physiologically relevant branching architectures grouped based on the arteriole density in each tree (Figure 1B). These numbers were experimentally measured density distributions found in CDH histology samples that were clinically stratified based on herniation and hypoplastic severity⁴. Arteriole blood flows and shear stresses were calculated for each tree within a population, demonstrating that changes in arteriole density or MPA diameter result in shifts in the distribution of flows (Figure 1C).
Our team has devised a computational framework that utilizes clinical echocardiogram data and stochastic techniques to estimate flow in neonatal pulmonary arterioles. This innovative approach enables us to model the impact of modified branching architectures in different CDH severities on arteriole flow and pressure patterns. By investigating the correlations between the severity of CDH-PH and clinical stratifications based solely on factors unrelated to PH, we aim to gain valuable insights that can inform treatment strategies and establish predictive measures for further clinical stratification.




Acknowledgements (Optional): :

References (Optional): : [1] Dillon, W. P et al. J. Pediatr. Surg, 2004;39:307-304. [2] Taira, Y. et al. Pediatr. Surg. Int, 1998. [3] Olufsen, M. S. et al. Ann. Biomed. Eng, 2000;28:1281-1299. [4] Levin, D. L. J. Pediatr, 1978;92:805-809.