3D ultrasound assessment of the central sulcus in very preterm infants: feasibility and reproducibility of opening metrics study

Quantitative studies of cortical development in preterm infants are needed. Until now, magnetic resonance imaging (MRI) has been the only technique capable of providing these 3D measurements. However, bedside ultrasound may offer a more accessible, repeatable alternative for both clinical and research applications. To assess the feasibility and reproducibility of three-dimensional transfontanellar ultrasound (3D US) for the semi-automated quantification of central sulcus (CS) morphology in very preterm infants, we performed a prospective study acquiring 109 serial 3D US volumes from 13 infants born < 32 weeks' gestation. Twelve CS morphometric parameters were extracted with a semi-automated pipeline; reliability was assessed with intra-class correlation coefficients (ICC), and maturational trends were described using group comparisons across PMA categories and complemented with linear mixed-effects models reporting fixed-effect estimates and variance components. Single-sweep ICCs ranged from 0.15 (length) to 0.71 (mean opening), while averaging three sweeps improved reproducibility to 0.34-0.88 (highest ICC 0.88, 95% CI 0.74-0.95). CS length increased from 25.9 ± 9.8 mm at < 28 weeks to 34.3 ± 8.9 mm at 36 weeks post-menstrual age (PMA) (p = 0.006), and mean sulcal opening from 4.0 ± 1.8 mm to 5.7 ± 1.9 mm (p = 0.023). In mixed-effects models, CS length increased with PMA (β = 0.96 mm/week; p = 0.008), and maximum opening also increased (β = 0.27 mm/week ; p = 0.002).The right hemisphere consistently showed larger openings than the left (5.4 ± 1.3 mm vs. 3.2 ± 1.5 mm; p < 0.0001). 3D US-based CS metrics thus capture lateralised, maturational changes and, when multiple sweeps are averaged, achieve good reproducibility (ICC up to 0.88), supporting their use for serial monitoring of cortical folding in the neonatal intensive care unit (NICU).