RPS 912 - Insights into foetal imaging

Author Block: W. H. E. Hamed, G. Kendall, L. Dyet, L. Srinivasan, D. Peebles, A. David, M. Sokolska, K. P. Baruteau; London/UK
Purpose: Validates 3DSVR for detecting anatomical and structural pathologies in fetal MRI and assesses quality improvement in a cohort of antenatal ventriculomegaly confirmed by neonatal MRI.
Methods or Background: Detecting subtle anatomical abnormalities in fetal brain MRI is challenging due to motion artefacts and the limited spatial resolution of 2D slices. Recently, slice-to-volume reconstruction (SVR) software has been developed to realign multiple 2D stacks into a high-resolution 3D volume (3DSVR), enabling better visualisation through multiplanar reconstruction. However, clinical validation of 3D-SVR is limited due to lack of ground truth data.

A retrospective cross-sectional study was conducted on pregnancies with ventriculomegaly. Inclusion criteria included fetal and neonatal MRI performed with standard protocols and 3DSVR. The median gestational age at fetal MRI was 28weeks (range 21-33w), and at neonatal MRI, 1week (range 1d-4w).

Ventriculomegaly causes were assessed on fetal 2DT2w-HASTE and 3DSVR and confirmed with 2DT2w-TSE on neonatal scans. Eleven brain structures were scored on a 3-point visibility scale, and image quality was rated based on signal-to-noise ratio (SNR) and motion artefacts. Statistical analysis was performed using the Wilcoxon signed-rank test.
Results or Findings: Of 20 subjects, eight had aqueduct stenosis identified on neonatal MRI. This was confirmed in 3/8 on 2D and 8/8 on 3DSVR. Fetal 3DSVR improved visibility scores in six of eleven structures, with significant differences in PLIC (0.65vs1.85, p<0.001), Sylvian aqueduct (1.05vs1.9, p<0.001), olfactory bulbs (0.9vs1.7, p<0.01), and grey-white matter contrast (0.9vs1.9, p<0.01). SNR improved in 35% of scans, and motion artefacts were reduced in 25%.
Conclusion: 3DSVR provides improved diagnoses of aqueduct stenosis, as evidenced by comparison with neonatal ground-truth scans. This is achieved by improving the visibility and overall quality of fetal brain MRI. Future work will validate 3DSVR in other pathologies.
Limitations: Not applicable.
Funding for this study: No funding was provided for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: The study is retrospective.

Author Block: M. Kaandorp¹, H. Asma-Ull², H. G. Kim², D. Agbelese¹, K. Payette¹, A. Jakab¹; ¹Zurich/CH, ²Seoul/KR
Purpose: Fetal MRI is increasingly used for the quantification of the developing brain through segmentation of various anatomical structures. One of the challenges is that current segmentation algorithms perform poorly in cases with ventriculomegaly due to limited high-quality annotations. Additionally, privacy concerns often restrict data sharing. We aimed to overcome these challenges through the generation of realistic synthetic pathological MRIs from manipulated healthy label images using generative AI.
Methods or Background: We trained a stable diffusion model for 3D brain MRI synthesis (Med-DDPM) using 727 fetal and preterm neonatal MRI-label image pairs. We generated 47 synthetic ventriculomegaly label images (S47-ventriculomegaly) from 33 healthy fetal MRIs (R33-healthy) by dilating ventricular labels. Combining label images from R33-healthy and S47-ventriculomegaly, we generated 80 synthetic MRIs (S80-ventriculomegaly) using Med-DDPM and visually assessed their quality. To evaluate segmentation performance, we trained two nnUNet models on S80-ventriculomegaly or R33-healthy MRI-label pairs and measured performance using Dice score on 40 test cases from FeTA2021 Challenge and 26 spina bifida severe ventriculomegaly cases (SPINABIFIDA). Performance was also assessed for 80 healthy and pathological MRI-label pairs (R80) and their synthetic equivalent (S80).
Results or Findings: Med-DDPM generated diverse, high-quality synthetic fetal MRIs. In segmentation tasks, S80-ventriculomegaly outperformed R33-healthy, achieving higher Dice scores in FeTA2021 (0.773 vs. 0.751) and SPINABIFIDA (0.759 vs. 0.722), especially for ventricles (0.780 vs. 0.672). S80 also surpassed R80 in SPINABIFIDA (0.822 vs. 0.815), with comparable performance in FeTA2021 (0.766 vs. 0.773).
Conclusion: This study demonstrates that generating realistic pathological fetal MRIs by manipulating labels from normally developing subjects can enhance data augmentation and data anonymization in prenatal imaging. This is an important step towards addressing privacy concerns while improving segmentation performance.
Limitations: Our method could be expanded to include more pathologies, increasing clinical applicability.
Funding for this study: Funding was provided by SNF grant: IZKSZ3_218590
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study has been approved by the ethics committee of Zurich Children's Hospital, decision number: 2022-01157)

Author Block: G. Biechele¹, B. Stos², D. Laux², S. Stöcklein¹, D. Grevent², L. J. Salomon²; ¹Munich/DE, ²Paris/FR
Purpose: Congenital heart defects (CHD) are common severe birth defects, resulting in a clinical need for precise prenatal assessment. Magnetic Resonance Imaging (MRI) of the fetal heart recently became feasible with the advent of novel gating techniques , but its feasibility has not yet been systematically evaluated in pregnancies with fetal CHD.
Methods or Background: This study yet evaluated 42 singleton pregnancies, consisting of 25 fetuses (60%) with diagnosis of CHD and 17 healthy control fetuses (40%). Diagnosis of CHD was refined by a cardiopediatrician. All MRI images were reviewed by a single blinded operator. Image quality of cardiac sequences was rated from 0 (not usable) to 5 (high image quality in multiple orientations). Cardiac anatomy was evaluated and correlation between MRI findings and US diagnosis was assessed.
Results or Findings: Fetal cardiac MRI was feasible in 40 patients (image quality score >0) due to extensive fetal movement in n=2 cases. Among, in n=13 (32.5%), n=21 (52.5%) and in n=6 (15.0%) cases, image quality was rated high (image quality score 5), medium (image quality score 3-4) and poor (image quality score 1-2), respectively. Among the 25 fetuses with CHD, fetal cardiac MRI was capable to fully confirm US-diagnosis in n=19 cases (76.0%) and to partly confirm US-diagnosis in n=6 (24.0%) cases, mostly due to incomplete recording of structures of interest. There was a substantial agreement between MRI and US (κ-value 0.70 with an accuracy of 0.85).
Conclusion: Our preliminary study suggests a similar diagnostic performance of MRI to US.
Limitations: With improvements in sequences, gating techniques and operator experience, MRI examination might soon become an essential part of the prenatal management of CHD, not only to identify associated abnormalities but also to reinforce the assessment of the heart itself.
Funding for this study: DFG Walter-Benjamin-Stipendium (BI 2563/1-1, G.B.)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Local ethic commitee number NCT 04142606

Author Block: I. C. Simcock¹, A. Lamouroux², S. C. Shelmerdine¹, C. Hutchinson¹, N. Sebire¹, O. Arthurs¹; ¹London/UK, ²Nimes/FR
Purpose: To identify the range and frequency of thoracic diagnoses at less-invasive autopsy, following a post-mortem micro-CT investigation for early gestation pregnancy loss.
Methods or Background: Micro-CT provides high-resolution imaging for early gestation fetuses (<300g), typically following miscarriage or termination of pregnancy, allowing parents a less invasive autopsy option.
We retrospectively analysed micro-CT diagnoses made on an unselected population of over 1190 early gestation fetuses between 2017 and 2024 at our tertiary referral institution.
Results or Findings: Thoracic abnormalities on micro-CT were identified in (147/1190; 12.3%), comprising 148 cardiac and 79 non-cardiac individual abnormalities.
The commonest cardiac abnormalities were septal defects (50/147; 34.0%), aortic abnormalities e.g. coarctation (23/147; 15.6%), congenital heart disease particularly tetralogy of Fallot (16/147; 10.9%) and ventricular abnormalities e.g. hypoplastic left heart (11/147; 7.5%).
Most common non-cardiac chest abnormalities were abnormal fluid accumulation (41/147; 27.9%), which included pleural effusion (34/147; 23.1%), hydrops (4/147; 2.7%), and pericardial effusion (3/147; 2%). A thoracic wall defect was observed in (18/147; 12.7%) of cases.
Conclusion: A range of thoracic abnormalities were made by micro-CT in our cohort, commonly from cardiac causes. This information is useful for parents regarding the likelihood of congenital abnormalities in subsequent pregnancies and provides an alternative to conventional invasive autopsy.
Limitations: Single centre data from a large specialist centre. Not all parents consented to invasive autopsy, so some diagnoses could not be histologically confirmed.
Funding for this study: ICS was funded by a National Institute for Health Research (NIHR) Clinical Doctoral Research Fellowship (ICA-CDRF-2017-03-53), Development and Skills Enhancement Award (NIHR302390) and a Research for Patient Benefit Award (NIHR206174), OJA was funded by a NIHR Career Development Fellowship (NIHR-CDF-2017-10-037) and SCS is supported by a NIHR Advanced Fellowship Award (NIHR-301322), and the work is funded by the Great Ormond Street Hospital Children’s Charity. AL was funded by two mobility funding from Nimes and Montpellier university hospital All research at Great Ormond Street Hospital NHS Foundation Trust and UCL Great Ormond Street Institute of Child Health is made possible by the NIHR Great Ormond Street Hospital Biomedical Research Centre. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health & Social Care.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: We obtained ethical approval from the UK NHS Health Research Authority (HRA) Ethics Committee (IRAS ID: 131395).

Author Block: A. Verma, A. Malik; New Delhi/IN
Purpose: To evaluate fetal lung maturity using multiparametric ultrasound in pregnant women with preterm premature rupture of membranes (PPROM), and to determine the accuracy of ultrasound parameters in predicting neonatal respiratory distress in this population. This study highlights the role of multiparametric ultrasound in predicting neonatal respiratory distress.
Methods or Background: We conducted an 18 month observational prospective cohort study on 81 women with singleton pregnancies under 37 weeks gestation, complicated by PPROM with fetuses between the 10th and 90th weight percentiles. Fetal biometric measurements included biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL). We also assessed distal femoral and proximal tibial epiphyseal ossification centers, placental maturity (Grannum system), and fetal main pulmonary artery (MPA) Doppler flow measuring resistive index (RI), pulsatility index (PI), acceleration time, and ejection time. These parameters were correlated with neonatal respiratory distress to predict fetal lung maturity in PPROM cases.
Results or Findings: BPD demonstrated highest diagnostic accuracy (81.48%) for predicting neonatal respiratory distress syndrome (RDS). The appearance of proximal tibial epiphyses showed remarkable sensitivity (92.31%) in predicting RDS. Among the Doppler parameters, fetal MPA resistive index (RI) showed the highest diagnostic accuracy (88.89%). Notably, combined fetal main pulmonary artery indices yielded the highest diagnostic accuracy (95.06%), followed by combined fetal biometry (91.36%).
Conclusion: Fetal lung maturity, a key determinant for neonatal respiratory distress can be assessed by grayscale ultrasound and Doppler parameters. Multiparametric ultrasonographic assessment is a promising tool in prediction of neonatal RDS in women with PPROM.
Limitations: The study's generalizability is constrained by its limited sample size. Additionally, assessment of the fetal main pulmonary artery via ultrasonography is subject to inter-observer variability and may be affected by inherent artifacts in ultrasound and Doppler imaging techniques.
Funding for this study: The cost of this study was covered by VMMC and Safdarjung Hospital, which provided essential resources and infrastructure to ensure its completion.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by Institutional Ethics Committee, VMMC and Safdarjung Hospital, New Delhi (06/2022/CC-265)

Author Block: A. Omar, A. Mohamed Tharwat, M. Aboelnasr, H. Abo-Ali Hamza, S. A. Hassanein, W. Gaber Eldamaty; Shebin El Kom/EG
Purpose: Fetal lung maturity assessment is the most critical factor for identifying the optimal delivery time. A non-invasive sonographic technique is necessary to evaluate fetal lung development. we aimed to predict maturity of fetal lung utilizing 3D lung volume ultrasound, lung to liver intensity ratio, and pulmonary artery doppler indices measurement
Methods or Background: A prospective observational study was conducted on 200 pregnant females with gestational age of 32 to 40 weeks age underwent 3D ultrasound (3DUS) for determination the fetal lung volume (FLV) and fetal lung-to-liver intensity ratio (FLLIR) (tissue histogram) with doppler examination of the main pulmonary artery (MPA) for the following parameters; acceleration-time to ejection-time ratio (At/Et), pulsatility index (PI) and resistive index (RI) during a period of one week from delivery and comparing the results to the neonatal outcome.
Results or Findings: Of 200 fetuses investigated; 113 cases (56.5%) were found to have respiratory distress syndrome. The MPA RI and PI were significantly greater in fetuses with respiratory distress syndrome comparing with those without (2.6± 0.3 and 0.9±0.05 vs. 1.9±0.3 and 0.8±0.2, respectively with p-value < 0.001 for both). MPA At/Et was significantly lesser for fetuses with RDS than fetuses without RDS (0.2±0.1 vs. 0.3±0.1 respectively, p-value less than 0.001). FLLIR was significantly lesser in RDS +ve group comparing with RDS -ve group (0.9±0.2 versus 1.3±0.3 respectively, p-value < 0.001) and FLV was significantly smaller in fetuses with respiratory distress syndrome comparing with those with no (31.5±2.5 vs. 38.1±2.8; p-value < 0.001).
Conclusion: The utilization of main pulmonary artery Doppler indices, together with mean fetal lung volume and FLLIR for assessing fetal lung is a quick non-invasive accurate technique for estimation of neonatal lung maturity and respiratory distress syndrome
Limitations: A uni-center study
Funding for this study: No funding was present.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Menoufia faculty of medicine research ethics committee under code no: (4/2022OBSG35

Author Block: M. Schwarz, V. Schmidbauer, N. M. Nowak, P. Kienast, D. Bettelheim, J. Binder, T. Reiberger, D. Prayer, G. Kasprian; Vienna/AT
Purpose: This study aimed to investigate the effects of IUTs on MRI findings in the fetal liver and spleen.
Methods or Background: Intrauterine transfusions (IUTs) are a life-saving treatment for fetal anemia. However, with each transfusion iron bypasses uptake regulation through the placenta and accumulates in fetal organs. Unlike other imaging modalities, fetal magnetic resonance
imaging (MRI) is capable of non-invasively assessing fetal liver disease and/or organ iron overload. This study aimed to investigate the effects of IUTs on MRI findings in the fetal liver and spleen.
For this retrospective study, we included eight fetuses undergoing IUT and prenatal MRI from 2014 to 2023. The fetuses were gestational age-matched with a cohort that received fetal MRI for other indications, but no IUTs. Signal intensity (SI) and
volumetric analyses of the liver and the spleen were performed.
Results or Findings: Fetuses receiving transfusions had significantly larger volumes of both liver (p=0.003) and spleen (p=0.029). T1 SI inversely correlated with the number of IUTs (Pearson’s r=-0.43, p = 0.099). This effect regressed over time (r=0.69, p=0.057). T2 SI did not correlate significantly with transfusion frequency but showed a strong positive correlation with the number of days between IUT and MRI (r=0.91, p=0.002). For splenic SI measures, similar effects were observed regarding T1 SI reduction per received transfusion (r=-0.36, p=0.167) and recovery of T2 SI after IUT (r=0.88, p=0.004).
Conclusion: This is the first study to report the effects of IUTs on MRI data of fetal livers and spleens. We observed considerable dose- and time-dependent SI alterations of the liver and spleen following IUT. Furthermore, fetal hepatosplenomegaly can be expected following IUT.
Limitations: Two fetuses have had parvovirus B19 infection, which may cause hepato-/splenomegaly, but may also mandate the need for IUT.
Funding for this study: Medical University of Vienna
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: In concordance with the principles of the Declaration of Helsinki, approved by the ethics committee of the Medical University of Vienna.

Author Block: Z. Zhu, C. Yan, Z. Lin, B. Zhang; Nanjing/CN
Purpose: To assess the potential of high dielectric constant (HDC) pad in increasing image quality and decreasing specific absorption rate (SAR) in 3.0T fetal MRI.
Methods or Background: This prospective single-center observational study included 168 pregnant participants taking 3.0T fetal MRI scanning with and without HDC pad between 1 May 2021 and 31 November 2023. Quantitative Image-quality analysis included signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Qualitative analysis was performed by three radiologists with four-point scale to evaluate overall image quality, dielectric artifact and diagnostic confidence. Whole-body total SAR was also compared. Correlation between image quality variable changes and participant clinical characteristics was evaluated using spearman correlation.
Results or Findings: 128 participants (mean gestational age 30.25±3.53 weeks, range 22-37 weeks) undertook balanced steady state free precession (bSSFP) sequence and 40 participants (mean gestational age 30.38±3.50 weeks, range 23-37 weeks) undertook single-shot fast spin-echo (SSFSE) sequence. With HDC pad, SNR and CNR was significantly higher (41.45% increase in SNR, 54.05% increase in CNR on bSSFP, p<0.001; 258.76% increase in SNR, 459.55% increase in CNR on SSFSE, p<0.001). Overall qualitative image quality, dielectric artifact and diagnostic confidence improved significantly (p<0.001). Adding HDC pad significantly reduced Whole-body total SAR (32.60% on bSSFP, p<0.001; 15.40% on SSFSE, p=0.005). There was no significant correlation between image quality variable changes and participant clinical characteristics (p>0.05).
Conclusion: In a clinical setting, adding HDC pad can increase overall quantitative and qualitative image quality while reducing dielectric artifact and SAR.
Limitations: Due to technical limitations, we could not conduct further analysis to compare the radiofrequency power deposited on uterus, whole fetal body and fetal brain separately. We only conducted experiments on magnetic resonance machines produced by United Imaging and didn’t include machines from other manufactures.
Funding for this study: National Science and Technology Innovation 2030 (2022ZD0211800)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethic committe of Nanjing Drum Tower Hospital