King's College London

Research portal

Texture-Based Analysis of Fetal Organs in Fetal Growth Restriction

Research output: Contribution to journalArticlepeer-review

Aya Mutaz Zeidan, Paula Ramirez Gilliland, Ashay Patel, Zhanchong Ou, Dimitra Flouri, Nada Mufti, Kasia Maksym, Rosalind Aughwane, Sébastien Ourselin, Anna L. David, Andrew Melbourne

Original languageEnglish
Pages (from-to)253-262
Number of pages10
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
DOIs
Published2021
Event3rd International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, UNSURE 2021, and the 6th International Workshop on Perinatal, Preterm and Paediatric Image Analysis, PIPPI 2021, held in conjunction with the 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 - Virtual, Online
Duration: 1 Oct 20211 Oct 2021

Bibliographical note

Funding Information: This research was supported by the Wellcome Trust (210182/Z/18/Z and Wellcome Trust/EPSRC NS/A000027/1) and the Radiological Research Trust. The funders had no direction in the study design, data collection, data analysis, manuscript preparation or publication decision. Publisher Copyright: © 2021, Springer Nature Switzerland AG.

King's Authors

Abstract

Fetal growth restriction (FGR) is common, affecting around 10% of all pregnancies. Growth restricted fetuses fail to achieve their genetically predetermined size and often weigh <10th centile for gestation. However, even appropriately grown fetuses can be affected, with the diagnosis of FGR missed before birth. Babies with FGR have a higher rate of stillbirth, neonatal morbidity such as breathing problems, and neurodevelopmental delay. FGR is usually due to placental insufficiency leading to poor placental perfusion and fetal hypoxia. MRI is increasingly used to image the fetus and placenta. Here we explore the use of novel multi-compartment Intravoxel Incoherent Motion Model (IVIM)-based models for MRI fetal and placental analysis, to improve understanding of FGR and quantify abnormalities and biomarkers in fetal organs. In 12 normally grown and 12 FGR gestational-age matched pregnancies (Median 28+ 4 wks±3+ 3 wks) we acquired T2 relaxometry and diffusion MRI datasets. Decreased perfusion, pseudo-diffusion coefficient, and fetal blood T2 values in the placenta and fetal liver were significant features distinguishing between FGR and normal controls (p-value <0.05). This may be related to the preferential shunting of fetal blood away from the fetal liver to the fetal brain that occurs in placental insufficiency. These features were used to predict FGR diagnosis and gestational age at delivery using simple machine learning models. Texture analysis was explored to compare Haralick features between control and FGR fetuses, with the placenta and liver yielding the most significant differences between the groups. This project provides insights into the effect of FGR on fetal organs emphasizing the significant impact on the fetal liver and placenta, and the potential of an automated approach to diagnosis by leveraging simple machine learning models.

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454