Fetal brain: what's now and what's on the horizon?

To evaluate whether structured maternal lifestyle interventions based on Mediterranean diet or stress reduction could influence fetal-infant neurodevelopment, assessed by detailed fetal neurosonography (NSG) and infant Ages and Stages Questionnaires 3rd edition (ASQ).

This secondary analysis of a randomised clinical trial (2017-2020) included 1,221 singleton pregnancies at high risk for small for gestational age. Participants were randomised into three groups at 19-23 weeks of gestation: Mediterranean diet intervention, stress reduction program or usual care. A detailed NSG was performed on 881 participants at mean (SD) 33.4 (1.1) weeks' gestation. Neurosonographic measurements were done offline and linear regression analysis was adjusted by gestational weeks at the assessment, fetal sex, biparietal diameter or cephalic index to adjust head size. ASQ was performed on 276 infants at 12 months of corrected age and suboptimal score rates between groups were analysed by logistic regression adjusted by breast feeding and socioeconomic class.

Biparietal diameter were similar among study groups. Compared to usual care group, fetuses in the Mediterranean diet group had deeper insula [26.80 (1.68) vs. 26.63 (1.75) mm, p = 0.02] and longer corpus callosum [42.98 (2.44) vs. 42.62 (2.27) mm, p = 0.04], with lower rate of suboptimal score infants in ASQ problem-solving domain (6.2% vs. 16.3%, p = 0.03). Similarly, the Stress reduction group fetuses had deeper insula [26.90 (1.75) vs. 26.63 (1.75) mm, p = 0.04] and lower rates of suboptimal score infants in ASQ fine motor domain (4.3% vs. 12.8%, p = 0.04), compared to usual care group fetuses.

Structured maternal lifestyle intervention during pregnancy may positively influence offspring's neurodevelopment, already detected during fetal life. OP04.02: AI accuracy in detecting periventricular and subependymal cysts in fetal ultrasound images A. Khalil4, R. Besson1, N. Matevski1, G. Corda6, V. Debavelaere1, N. Fries7, J. Stirnemann3, Y. Oyelese5, Y. Ville2 1Sonio, Paris, France; 2Paris Descartes University, Paris, France; 3Obstetrics, Paris Descartes University, Necker-Enfants Malades Hospital, Paris, France; 4Fetal Medicine Unit, St George's Hospital Medical School, University of London, London, United Kingdom; 5Obstetrics and Gynecology, Beth Israel Deaconess Medical Center, Boston, MA, USA; 6Sonio, London, United Kingdom; 7Imagyn'Echo, Montpellier, France

Periventricular and subependymal cysts are frequent, often overlooked sonographic findings whose clinical significance when isolated is uncertain, but which may be one of the rare signs of CMV infection. An AI system capable of automatically recognising these signs could serve as a safety net, prompting practitioners to conduct further investigations when necessary.

This study evaluated an AI model (not yet approved) for detecting subtle signs such as periventricular or subependymal cysts. Trained on 17,236 images from 69 centres, including 390 images of brain cysts from 56 CMV-seroconverted pregnancies, the model was tested at an experienced screening centre with a comprehensive record of brain malformations, including 7 CMV cases that were used for training the AI. Experts assessed the clinical relevance of AI-identified cysts.

The AI model identified axial brain views in 3,447 images (from 1,431 studies) that had not been seen during training, originating from the targeted screening centre. It flagged a potential brain cyst in just 22 images. Of these, experts classified 4 as false positives and 18 as true positives. Upon reviewing the full examinations, all detected cysts were associated with additional ultrasound findings (e.g., periventricular echogenicity, hyperechogenic bowels, brain calcifications), raising concerns about possible CMV infection. Additionally, 7 cases had confirmed congenital malformations (6 agenesis of corpus callosum, 1 neural tube defect).

Our AI system improves the screening rate of periventricular and subependymal cysts with remarkable specificity, flagging only 0.006% of images. Importantly, 82% of alerts were deemed justified by experts, demonstrating clinical relevance. Despite these promising results, further testing is necessary due to the uncertain and subjective nature of this ultrasound finding. OP04.03: Sacral lypomyelocele with a dural defect a case associated with a tail-like cutaneous appendage M. Izquierdo1 1Prenatal Diagnosis, VITHAS Valencia 9 de Octubre, Vithas Hospitales, Valencia, Spain Among the closed neural tube defects are lipomyeloceles. This entity falls within congenital spinal lipomatous malformations, a very heterogeneous group of pathologies with different prognoses depending on dural involvement and location. Supporting information can be found in the online version of this abstract OP04.04: From womb to wisdom: a familial journey of atretic cephaloceles across generations L. Kaur1, M. Preet2 1Prime Diagnostic Centre, Chandigarh, India; 2Fetal Medicine, Prime Diagnostic Centre, Chandigarh, India Supporting information can be found in the online version of this abstract OP04.05: From diagnosis to triumph: a fetal subdural hematoma success story M. Preet1, L. Kaur1 1Fetal Medicine, Prime Diagnostic Centre, Chandigarh, India Supporting information can be found in the online version of this abstract OP04.06: Longitudinal assessment of cerebellar growth based on head circumference ratio in a large cohort of normally grown fetuses M. Sanz-Cortes4, Z. Chen1, W. Lee4, L.F. Goncalves2,5, J. Espinoza3, L.M. Mack6, J. Gleason1, D. He1,7, K. Grantz1 1Division of Population Health Research, Division of Intramural Research, NICHD, NIH, National Institutes of Health, Bethesda, MD, USA; 2Radiology, Phoenix Children's Hospital, Phoenix, AZ, USA; 3Obstetrics and Gynecology, University of Texas Health Science Center at Houston, Houston, TX, USA; 4Obstetrics and Gynecology, Baylor College of Medicine, Houston, TX, USA; 5Child Health and Radiology, University of Arizona College of Medicine, Phoenix, AZ, USA; 6General Electric Company, Cincinnati, OH, USA; 7Prospective Group, Inc., Fairfax, VA, USA

Understanding the relationship between fetal cerebellar and head growth can help in the differential diagnosis of a small cerebellar size. There is a gap in knowledge related to the normal growth ratio between cerebellar and head size. The objective of this study is to create a fetal growth standard for transcerebellar diameter (TCD) to head circumference (HC) ratio.

As part of the NICHD Fetal 2D Study (2009–2013), we included 1714 singleton fetuses with no known fetal anomalies and normal neonatal outcomes from pregnancies without obstetric complications. TCD (mm) and HC (mm) were measured at up to five scans between 15 to 40 weeks' (wks) gestation following a standardised protocol. All cases had at least one TCD measurement. We used linear mixed models with cubic splines to flexibly model curves (5th, 50th, 95th percentiles) for TCD and TCD/HC between 15–40 wks.

The TCD 50th percentile was similar to the gestational week from 15 to 21 wks (e.g. 17.9 mm at 18 wks), within one to two weeks from 22 to 24 wks (e.g. 24.8 mm at 23 wks), and then accelerated through the end of pregnancy where it was 55.1 mm at 40 wks (figure 1A). The 50th percentile TCD/HC ratio was 0.12 to 0.13 from 15 to 30 weeks, and then accelerated through the end of pregnancy where it was 0.17 at 40 wks (figure 1B).

Supporting information can be found in the online version of this abstract OP04.07: Imaging features of prenatal intracranial hemorrhage S. Shinar1, P. Tripathy4, D. Chitayat3, E. Miller2 1Obstetrics and Gynecology, Sinai Health, Toronto, ON, Canada; 2Radiology, Hospital for Sick Children, Toronto, ON, Canada; 3Prenatal Diagnosis and Medical Genetics, Sinai Health, Toronto, ON, Canada; 4University of Toronto, Toronto, ON, Canada

To analyse the imaging features of prenatally diagnosed intracranial hemorrhage (ICH) across different etiologies.

- COL4A1/2 variants; - Fetal/Neonatal Alloimmune Thrombocytopenia (FNAIT); - infections; - other; and - unknown. Imaging findings on US and MRI were analysed to identify common distinct patterns within each group.

Seventy cases were identified, with 42 meeting inclusion criteria. Most severe injuries were observed in COL4A1/2 cases (n = 9, 21.4%), all presenting with intraventricular hemorrhage (IVH), six extending into extra-axial spaces. White matter injury was present in all, with 67% showing extensive damage. Porencephaly and bilateral schizencephaly were frequent. FNAIT accounted for four (9.5%) cases, with three showing IVH and one parenchymal hemorrhage, all characterised by focal bleeds. Four cases (9.5%) were infectious (3 parvovirus, 1 CMV), with parvovirus cases complicated by cerebellar hemorrhage. Other causes (n = 3, 7.1%) included maternal disease (severe hypoxia), fetal coagulopathy, and a NOTCH3 variant. In 22 cases (52.4%), the cause remained unknown; six of these had identified genetic variants, but none had an established link to hemorrhage. Unknown cases predominantly showed IVH with periventricular leukomalacia resembling IVH of prematurity.

Supporting information can be found in the online version of this abstract OP04.08: Segmental spinal dysgenesis: a rare prenatal diagnosis M. Pérez Cruz3,4, M. Rebollo1, E. Ferriols2, J. Sabria3, J. Martinez3,5, L.P. Lopez1,4, M. Gomez Chiari1, E. Eixarch3,5 1Hospital Sant Joan de Deu, Barcelona, Spain; 2Obstetrics and Gynecology, Hospital del Mar, Barcelona, Spain; 3Fetal Medicine, BCNatal Barcelona Center for Maternal Fetal and Neonatal Medicine (Hospital Sant Joan de Deu and Hospital Clinic), University of Barcelona, Esplugues, (Barcelona), Spain; 4Institut de Recerca Sant Joan de Deu, Barcelona, Spain; 5Obstetrics and Gynecology, Hospital Clinic de Barcelona, Barcelona, Spain Segmental spinal dysgenesis (SSD) is a rare congenital condition characterised by agenesis or dysgenesis of the thoracolumbar or lumbar spine, congenital kyphosis, and abnormal or absent spinal cord at the defect level. A key feature is intact vertebrae above and below the defect, with a displaced lower spinal cord. We present a rare prenatal case of SSD in a pregnant woman who sought a 2nd opinion after a 20w US showing a mid-lumbar vertebrae deviation, without skin discontinuity or open NTD. Amniocentesis was normal. MRI at 25w showed posterior angulation of the thoracolumbar vertebrae and thinning of the spinal cord from T10 to T12-L1, suggesting a lumbar segmentation anomaly. At 32 w, follow-up NSG was similar. At 33 w, CT revealed T12 vertebral dysplasia, L1 posterior dislocation, and focal stenosis at T12-L1. Supporting information can be found in the online version of this abstract OP04.09: Proximal atrium assessment: addressing the need for reliable evaluation N. Prat4, N. Dominguez1,4, C. Guerrero Lopez2,4, N. Encabo4, M. Illa4,3, N. Masoller4, E. Gratacos4,3, E. Eixarch4,3, M. Pérez Cruz4,5 1Obstetrics and Gynecology, University Hospital of Ferrol, Ferrol, A Coruna, Spain; 2Obstetrics and Gynecology, Hospital Clinic de Barcelona, Barcelona, Spain; 3Institut d'Investigacions Biomediques August Pi i Sunyer, Barcelona, Spain; 4Materno-Fetal Medicine, BCNatal Fetal Medicine Research Center, Hospital Clinic and Hospital Sant Joan de Deu, University of Barcelona, Barcelona, Spain; 5Institut de Recerca Sant Joan de Deu, Barcelona, Spain

Measuring the proximal left ventricle (LV) remains challenging, potentially leading to the underdiagnosis of unilateral ventricular anomalies. This study evaluates the feasibility of measuring the proximal LV in both axial and coronal planes.

Prospective study including 343 fetal NSG in cephalic presentation from low-risk singleton pregnancies with normal growth and CNS structures, between 20 and 39 weeks of gestation being part of the ongoing Growing Brains study. In the axial plane, both the distal LV (ISUOG guidelines) and proximal LV (Yagel and Valsky's, 2021) were measured (figure 1a, 1b). LV widths were also taken in the coronal transventricular plane, aligned with axial landmarks at the PO sulcus and calcarine division (figure 1c).

The proximal LV measurement was successfully obtained in 96% of cases in the axial plane and 81% in the coronal plane (p < 0.005). The axial and coronal values for the proximal (4.7 ± 1.5 mm vs. 5.8 ± 1.3 mm, p = 0.000) and distal (4.7 ± 1.5 mm vs 5.7 ± 1.5 mm, p = 0.000) LV measurements showed an overestimation in the coronal plane, with a mean difference of 1.65 mm ± 1.25. The correlation and agreement between axial and coronal measurements were mild for the proximal (r = 0.31, ICC = 0.33) and distal (r = 0.24, ICC = 0.28), with a satisfactory Bland–Altman plot.

Supporting information can be found in the online version of this abstract