PS-011 Quantitative Assessment Of Preterm Left Ventricular Anatomical Development And Remodelling Using Neonatal Cardiac Mri And Atlasing Techniques

David J. Cox, Wenjia Bai, An Price, A. David Edwards, Am Groves, Daniel RueckertArchives 2014 · DOI

Abstract

Cardiac development and myocardial maturation continue through the third trimester of gestation. This study aims to quantify neonatal cardiac development in preterm infants compared to healthy term controls using 3 Tesla (3T) cardiac MRI and computational atlas construction. Analysis of 10 preterm neonates (born at 32–35 weeks) scanned at birth and term-corrected age, alongside 4 term controls, revealed that preterm infants at term-corrected age exhibit significantly higher weight-corrected left ventricular mass (LVM) and end-diastolic volume (EDV) than term-born infants. These findings suggest that ex-utero cardiac development in preterm infants involves significant structural remodelling, which can be effectively quantified using MRI and atlasing techniques.

Introduction

The third trimester is a critical period for the maturation of the left ventricle (LV). Preterm birth interrupts this process, forcing the heart to adapt to the high-pressure postnatal environment earlier than expected. Clinical evidence suggests that such early adaptation may lead to long-term cardiovascular risks, including altered LV geometry in adulthood. Understanding the timing and nature of this remodelling is essential for identifying potential pathological changes versus physiological adaptations.

Methods

The study utilized high-resolution 3T cardiac MRI to image the hearts of preterm neonates and healthy term controls. The primary methodology involved:

  1. Manual Segmentation: Using ITK-SNAP software, researchers manually segmented the LV myocardium and blood pool from short-axis stack sequences at end-diastole. This provided precise volumetric measurements of heart structure.
  2. Computational Atlasing: Provisional computational atlases were constructed to move beyond simple volumetrics, allowing for the characterization of 3D shape and geometry.
Manual segmentation of the neonatal left ventricle using ITK-SNAP software.

Results and Discussion

Quantitative Remodelling

The study found stark differences in cardiac metrics when corrected for body weight. By the time preterm infants reached their term-corrected age, their hearts were significantly larger and heavier than those of infants born at full term:

  • Left Ventricular Mass (LVM): Preterm infants at term-corrected age had a mean LVM of 1.89 g/kg, nearly double that of healthy term controls (0.95 g/kg, p=0.001).
  • End-Diastolic Volume (EDV): The LV chamber volume was also significantly larger in the preterm group at term-corrected age (3.42 cm³/kg) compared to term controls (2.16 cm³/kg, p=0.0006).

These results indicate an aggressive growth phase once the infant is outside the womb. The increase in both mass and volume suggests the heart is both thickening and dilating to meet circulatory demands.

Computational Atlasing and Shape Analysis

While volumetric data provides a snapshot of growth, computational atlasing offers a more nuanced view of remodelling. These atlases allow researchers to visualize and mathematically compare the 3D geometry and wall thickness of the heart across different developmental stages.

3D computational atlases visualizing left ventricular wall thickness and geometry.

This approach highlights that the way the heart changes shape may be as important as how much it grows. The provisional atlases currently enable visual comparisons, with future analysis aimed at more sophisticated quantification of these geometric shifts.

Conclusion

Neonatal MRI combined with manual segmentation and computational atlasing provides a powerful tool for quantifying the impact of preterm birth on cardiac development. The observed increases in LVM and EDV at term-corrected age highlight a distinct ex-utero growth trajectory compared to in-utero development. While the sample size is small, these findings lay the groundwork for larger studies to determine if this early remodelling predicts long-term cardiovascular health outcomes, potentially allowing for earlier clinical interventions.

Summary updated

Abstract

<h3>Background and aims</h3> Cardiac development and myocardial maturation continues through the third trimester of gestation. Moderately preterm birth at 30–36 weeks adversely affects long-term cardiovascular health and impacts left ventricular size and geometry at young adulthood. We aim to quantify and characterise normal and pathological neonatal cardiac development using cardiac MRI and computational atlas construction. <h3>Methods</h3> Preterm neonates and healthy term controls underwent neonatal 3 Tesla cardiac MRI. Data from short axis stack sequences was manually segmented at end-diastole (ITK-SNAP software) providing volumetric measurements of the left ventricular myocardium and blood pool. Sub-group analysis compared 10 preterm neonates born at 32–35 weeks, scanned within 7 days of birth and at term-corrected age, with 4 healthy term controls (39–42 weeks). <h3>Results</h3> Weight-corrected left ventricular mass (LVM, g/kg) and end-diastolic volume (EDV, cm<sup>3</sup>/kg) for the preterm cohort at term-corrected age (LVM - mean 1.89, 95% CI 1.89 ± 0.21; EDV – mean 3.42, 95% CI 3.42 ± 0.34) were significantly greater than both the preterm cohort at birth (LVM 1.05, 1.05 ± 0.08, p = 0.0002; EDV 4.89, 4.89 ± 0.59, p = 0.0008) and healthy term controls (LVM 0.95, 0.95 ± 0.18, p = 0.001; EDV 2.16, 2.16 ± 0.38, p = 0.0006). <h3>Conclusions</h3> Neonatal MRI with manual ventricular segmentation quantifies preterm gross ex-utero left ventricular growth, highlighting differences from in-utero cardiac development. Increases in preterm LVM and EDV may represent pathological remodelling or physiological ex-utero adaptation. We have constructed provisional computational atlases that currently allow visual comparisons of size and shape, but which after further analysis will enable more sophisticated quantification and characterisation of preterm ventricular growth and remodelling.
BibTeX
@inproceedings{cox2014ps,
  author = {David J. Cox and Wenjia Bai and AN Price and A. David Edwards and AM Groves and Daniel Rueckert},
  title = {{PS-011 Quantitative Assessment Of Preterm Left Ventricular Anatomical Development And Remodelling Using Neonatal Cardiac Mri And Atlasing Techniques}},
  booktitle = {Archives},
  year = {2014},
  doi = {10.1136/archdischild-2014-307384.307},
}
Figure