OC11.07: Automated grading of fetal cortical development using deep learning algorithms: a preliminary three‐dimensional ultrasound study in healthy fetuses
Ultrasound in Obstetrics and Gynecology, vol. 58, pp. 35
Abstract
Grading of fetal cortical development using manual scoring systems exhibits significant intra- and inter-observer variability; this can influence the accuracy of the estimated brain maturation. We developed a deep learning algorithm to automatically evaluate gyral development. Bounding boxes centered around the Sylvian fissure (SF), parieto-occipital fissure (POF) and calcarine sulcus (CLC) were manually extracted from 811 fetal brain three-dimensional (3D) transabdominal volumes of 533 healthy fetuses (GA 19-30 weeks) collected in the INTERGROWTH-21st study (IG-21). A neural network was trained on 85% of the dataset to estimate the gestational age of each fissure and evaluated on the final 15%. Our network estimated the GA of the SF, POF, CLC with a mean error of 3.4, 5.0 and 4.8 days, respectively, with predictions generated in less than one second. This preliminary experiment shows our novel pipeline is able to predict the GA of three fissures to a high standard, which has the potential to eliminate user-dependent fissure grading and allow more in-depth exploration of cortical development. This proposed technique enables objective evaluation of cortical maturation and thus can be used for (early) detection of abnormal cortical development. We aim to extend this to new datasets to ensure it is generalisable. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Authors 7
-
Affiliation as printed
IBME University of Oxford Oxford United Kingdom
IBME, University of Oxford, Oxford, United Kingdom
-
Affiliation as printed
IBME University of Oxford Oxford United Kingdom
IBME, University of Oxford, Oxford, United Kingdom
-
Moska Aliasi Aachen
Leiden University · Leiden University Medical Center
Affiliation as printed
Leiden University Medical Centre Leiden Netherlands
Leiden University Medical Centre, Leiden, Netherlands
-
Affiliation as printed
IBME University of Oxford Oxford United Kingdom
IBME, University of Oxford, Oxford, United Kingdom
-
Monique C. Haak Aachen
Leiden University · Leiden University Medical Center
Affiliation as printed
Leiden University Medical Centre Leiden Netherlands
Leiden University Medical Centre, Leiden, Netherlands
-
Affiliation as printed
Nuffield Department of Women's and Reproductive Health University of Oxford Oxford United Kingdom
Nuffield Department of Women's and Reproductive Health, University of Oxford, Oxford, United Kingdom
-
Affiliation as printed
IBME University of Oxford Oxford United Kingdom
Nuffield Department of Women's and Reproductive Health, University of Oxford, Oxford, United Kingdom
Cited by 0 stored of 0
No patents citing this paper on Lens.org (checked 2026-10-11).