OC07.03: Automated cortical analysis from third‐dimensional second trimester ultrasound scans

To describe the expected cortical development from a large cohort of healthy fetuses between 18+0 and 26+6 gestational weeks, from 3D ultrasound (US) volumes using a novel, fast, automated method. Using a deep learning algorithm that anatomically guides the segmentation in regions of acoustic shadows, the cortical plate (CP) was extracted from 3D US scans acquired from the INTERGROWTH-21st Project. The CP was parcellated into 5 lobes (see figure 1a), to measure global and local cortical properties. We analysed 2188 scans (1703 women), in less than 2s per scan. Cortical volume (range = 5-55cm3), surface area (40-150cm2) and 3D Sylvian Fissure (SF) depth (5-12mm) measures were consistent with previous findings (see figure 1A). Average cortical depth (CD), thickness (T) and volume for each lobe showed that the insula deepened and thickened at the fastest rate, due to SF opercularisation (see figure 1B). Frontal, parietal, and temporal lobe volumes increased the most, closely aligning with the MRI literature. Figure 1C shows the average CD and T across the population on a surface template. We have demonstrated that it is possible to characterise the developing cortex in 3D from US scans, potentially paving the way for more neuroanatomical studies to be performed from this modality in the future. Further, our automated pipeline makes it possible to describe the development from large cohorts, enabling future cohort comparisons. 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.