Normative volumetric growth modeling of the whole fetal body, placenta, and amniotic fluid for three-dimensional T2-weighted magnetic resonance imaging.
Authors
Affiliations (7)
Affiliations (7)
- Research Department of Early Life Imaging, School of Biomedical Engineering and Imaging Sciences, King's College London, Westminster Bridge Road, London, SE1 7EH, UK. [email protected].
- Research Department of Early Life Imaging, School of Biomedical Engineering and Imaging Sciences, King's College London, Westminster Bridge Road, London, SE1 7EH, UK.
- Department of Women and Children's Health, King's College London, London, UK.
- Biomedical Image Technologies, ETSI Telecomunicacion, Universidad Politecnica de Madrid and CIBER-BBN, Madrid, Spain.
- Research Department of Imaging Physics and Engineering, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK.
- Institute of Information Processing, Leibniz University Hannover, Hanover, Germany.
- Fetal Medicine Unit, Guy's and St Thomas' NHS Foundation Trust, London, UK.
Abstract
Magnetic resonance imaging (MRI)-based volumetry of the fetus, placenta, and amniotic fluid is clinically valuable but rarely used due to labor-intensive manual segmentation of motion-corrupted two-dimensional (2-D) stacks. Existing deep learning approaches are typically limited to single structures and 2-D data, while no robust automated solution exists for whole-uterus volumetry in reconstructed three-dimensional (3-D) MRI, and normative reference ranges are lacking. To develop an automated pipeline for whole-uterus volumetry in 3-D T2-weighted fetal MRI and derive normative growth models for fetal, placental, and amniotic fluid volumes. Motion-corrupted T2-weighted stacks (0.55-3-T field strength) were reconstructed into 3-D isotropic images using deformable slice-to-volume reconstruction, followed by automated segmentation with a 3-D U-Net. The method was applied to 357 normal-control datasets with confirmed term birth (16-41 weeks gestational age range) to derive quadratic normative growth curves. Performance and clinical utility were further evaluated on 43 independent datasets. Segmentation was highly accurate (Dice: fetus 0.997, placenta 0.995, amniotic fluid 0.998) with low volume errors (<1%) and minimal manual refinement required in <25% of cases. In the control cohort, fetal and placental volumes increased with gestational age (P<0.001), while amniotic fluid followed a quadratic trend. Longitudinal growth rates were 146.6 cc/week (fetus) and 38.8 cc/week (placenta). Preterm pregnancies showed significantly lower fetal and placental volumes (P<0.001) and reduced amniotic fluid (P<0.01). This work presents the first automated pipeline for simultaneous whole-uterus volumetry in 3-D fetal MRI and establishes normative growth models across gestation. The approach enables accurate, standardized volumetric assessment and provides a practical tool for detecting abnormal growth patterns in both normal and high-risk pregnancies.