Current status and prospects of prenatal ultrasound diagnosis of congenital heart disease in fetuses: a narrative review

Introduction

Congenital heart disease (CHD), commonly referred to as congenital heart defect, is one of the most prevalent congenital structural heart defects, typically involving malformations of the heart valves and walls (1). Types of CHD include some simpler conditions, such as atrial septal defects and ventricular septal defects (VSDs), as well as more complex cardiac anomalies. Fetal congenital heart disease (FCHD) is a leading cause of death during the perinatal period and in children under 5 years old (2). According to epidemiological data, approximately 1.35 million newborns are diagnosed with CHD globally each year, with an incidence rate of 8.22 per 1,000 live births, making it a significant global health issue (3). Of these, about 50–60% of cases are detected prenatally through local and national screening programs (4).

According to national data, CHD affects approximately 8.98 per 1,000 live births in China (5). Within this spectrum, severe forms of CHD—those necessitating intervention or carrying a risk of neonatal death—account for an estimated 1.46 per 1,000 births (5). Furthermore, the overall prevalence of CHD has been increasing annually (2). A study on congenital heart defects indicated that VSD (6) accounting for 17.04% of nationwide screening for fetal diseases during pregnancy. The second most common is Tetralogy of Fallot (TOF), comprising 9.72% of pregnancy cases. Severe congenital heart defects (CHD) refer to those defects that may be fatal or require intervention or long-term follow-up during infancy (7). In early pregnancy, if the ultrasound measurement exceeds 3.5 mm nuchal translucency (NT), a suspicion of fetal heart defect may arise, but its sensitivity for detecting severe CHD is relatively low (8). Without timely treatment, infants with severe CHD may experience significant cyanosis, hypoxemia, or cardiogenic shock, which can ultimately lead to morbidity or mortality after birth (9). Furthermore, if offspring are diagnosed with CHD, it can trigger negative emotions (10) and psychological distress (11) in parents, subsequently affecting the cognitive and socio-emotional development of the newborn and imposing psychological and economic burdens on the family.

The prevention and treatment of CHD are divided into three levels: primary prevention of pathogenic factors, secondary prevention through prenatal diagnosis and in-utero intervention, and tertiary prevention through postnatal treatment. Research and clinical practices both domestically and internationally indicate that interventions at these three levels are effective and feasible in improving outcomes and alleviating the burden of the disease (12). We present this article in accordance with the Narrative Review reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-798/rc).

Methods

This study employed a systematic methodology to conduct a narrative review. The literature search was commenced in May 2025, covering the PubMed, Web of Science databases for publications and CNKI (China National Knowledge Infrastructure) from January 2019 to May 2025, aiming to encompass key developments over the recent five years. The search strategy was constructed based on the PICO (Population, Intervention, Comparison, Outcome) framework, with the core search string being: (“fetal echocardiography” OR “prenatal ultrasound”) AND (“congenital heart disease” OR “CHD”) AND (“screening” OR “diagnosis”). Supplementary searches were also conducted for specific technologies (e.g., “artificial intelligence”, “speckle tracking”, “three-dimensional ultrasound”).

The inclusion criteria were: studies involving echocardiographic techniques, protocols, or clinical research related to the screening or diagnosis of fetal CHD (including original research, reviews, meta-analyses, and authoritative guidelines). The exclusion criteria were: editorials, case reports, articles for which the full text was unavailable, and publications in languages other than Chinese or English.

The literature screening process was completed independently by two researchers. First, duplicate records were removed. Subsequently, titles and abstracts were reviewed, followed by a full-text assessment of potentially relevant articles. Any disagreements were resolved through discussion or by consulting a third researcher. Due to the heterogeneity among the included studies in terms of design, population, and outcome measures, a quantitative meta-analysis was not performed. Instead, a thematic synthesis and critical analysis were conducted. A detailed description of the search strategy is provided in Table 1.

The cornerstone of fetal CHD diagnosis

Fetal echocardiography (FE) serves as an indispensable cornerstone for the prenatal diagnosis of CHD, profoundly reshaping perinatal management strategies. Although the modern diagnostic framework has integrated genetic screening and advanced visualization tools for surgical planning [e.g., three-dimensional (3D) printing] (13), echocardiography, by virtue of its non-invasive, real-time, and reproducible nature, remains the preferred and primary imaging modality for achieving precise anatomical delineation of the fetal heart (14,15). As a core component of contemporary prenatal care, it is recommended for all pregnant individuals in the mid-trimester, while a comprehensive examination is mandatory for those with identified risk factors (16,17). The fundamental basis of its diagnostic accuracy lies in the standardized acquisition and interpretation of a series of standard cardiac views.

Screening begins with the acquisition of standard views. International consensus once established the four-chamber view, left ventricular outflow tract view, and right ventricular outflow tract view as the basic requirements (18). However, clinical practice confirms that relying solely on these “three views” systematically misses critical anomalies such as great vessel abnormalities. Consequently, more comprehensive screening protocols have been continuously validated and promoted. For instance, the study by Bak et al. (19) demonstrated that the “five-axial view” protocol, which incorporates the three-vessel and trachea view, performs better in detecting various malformations and offers good cost-effectiveness. The four-chamber view itself is crucial, capable of primarily screening approximately 48.2% of major cardiac anomalies, and its imaging quality directly affects the assessment of atrioventricular connections, chamber symmetry, and septal integrity (20). To enhance the standardization of view acquisition, artificial intelligence (AI) technology has been developed to automatically identify standard views and perform real-time quality control, becoming a vital tool in promoting screening standardization (21,22).

Two-dimensional (2D) ultrasound forms the bedrock for displaying cardiac morphology, capable of clearly presenting the anatomical relationships of chambers, valves, and great vessels to identify structural anomalies. Doppler ultrasound (including pulsed-wave and color Doppler) provides indispensable hemodynamic information, crucial for evaluating valve stenosis and regurgitation, and for determining abnormal vascular connections (7). These two modalities constitute the reliable foundation of global fetal cardiac screening practice. A comprehensive fetal echocardiogram is precisely based on the integrated application of these core imaging modalities, encompassing meticulous 2D imaging, Doppler assessment of valvular and vascular flows, and evaluation of rhythm and function (21,23-25).

The development of diagnostic technology is driving fetal cardiac assessment toward greater standardization, quantification, and intelligence. The optimal timing for basic screening is typically 18–22 weeks of gestation, whereas for high-risk pregnancies such as those with chromosomal abnormalities or increased NT, targeted fetal cardiac examination can be advanced to the early first trimester (12–14 weeks) to facilitate early diagnosis and planning (26-30). The application of AI in view recognition and quality control (22,31) is a prime example of how technological integration aids in enhancing screening standardization and efficiency. The key screening views of the fetal heart are illustrated in Figure 1.

Figure 1 Schematic diagram of key screening views of the fetal heart. (A) Four-chamber view. (B) Horizontal four-chamber view. (C) Left ventricular outflow tract view. (D) Right ventricular outflow tract view. (E) Three-vessel and trachea view. (F) Three-vessel view. AAD, aortic arch descending; AAO, ascending aorta; ARCH, aortic arch; DAO, descending aorta; FO, foramen ovale; LA, left atrium; LPA, left pulmonary artery; LV, left ventricle; MPA, main pulmonary artery; MV, mitral valve; PA, pulmonary artery; PV, pulmonary valve; RA, right atrium; RV, right ventricle; SV, systemic vein; SVC, superior vena cava; T, trachea; TV, tricuspid valve.

Technology-enhanced ultrasound

Three-dimensional/four-dimensional spatiotemporal image correlation (3D/4D STIC) echocardiography

Building upon traditional 2D echocardiography, FE combined with 3D/4D STIC technology has significantly enhanced the assessment of complex CHDs such as double outlet right ventricle (DORV). This technology not only provides more intuitive cardiac structural information but its derivatives and related techniques have also expanded diagnostic capabilities.

For instance, the SlowflowHD technique proposed by Turan et al. optimizes imaging sensitivity for the microvascular system, allowing clear visualization of low-velocity blood flow within fetal branching vascular beds. It features high frame rates, high resolution, and high sensitivity, demonstrating particularly high sensitivity and specificity in diagnosing congenital vascular rings (32,33). Research by Vaidyanathan et al. confirmed that combining 3D/4D STIC rendering techniques for prenatal diagnosis and anatomical analysis is feasible for isolated congenital coronary artery fistulas (34).

Furthermore, the integration of STIC technology with other ultrasound modalities further increases its diagnostic value. Studies show that incorporating STIC imaging into routine fetal CHD examinations can yield reliable, high-quality diagnostic results, contributing to improved diagnostic accuracy (35). The combination of high-definition flow (HD-flow) and STIC aids in the diagnosis of fetal interrupted aortic arch (IAA) (36). For functional assessment, STIC technology combined with speckle tracking echocardiography (STE) can be used to evaluate myocardial contractility in fetuses with a history of maternal diabetes (37).

Although STIC technology can provide clearer diagnostic images (38), its application in primary screening remains limited by requirements for fetal quiescence and lengthy post-processing times. Therefore, it currently serves primarily as a powerful adjunctive diagnostic tool for the in-depth evaluation of complex cases and the refinement of diagnostic details, offering unique value particularly in providing critical anatomical information for prenatal counseling and surgical decision-making.

STE

Over recent decades, Doppler echocardiography has served as a key tool for evaluating fetal cardiac structure and hemodynamics by providing information on blood flow velocity and direction throughout the cardiac cycle (39). However, its utility is limited by angle dependence, Nyquist limits, and relatively low temporal resolution. STE has emerged as an innovative technique that overcomes these constraints. Since its first application in detecting adult cardiac defects in 2004, STE has gained broad acceptance in pediatric and adult cardiology for ventricular function assessment due to its angle independence, sufficient frame rates, strong reproducibility, and ability to simultaneously evaluate ventricular size, shape, and contractility (40).

In fetal cardiology, STE has demonstrated significant diagnostic value. For instance, it can improve or provide incremental value for the detection of suspected CoA during the final prenatal check (41). Furthermore, in fetuses with congenital heart defect, STE reveals reduced right ventricular functional parameters such as fractional area change (FAC), longitudinal fractional shortening (LFS), and transverse fractional shortening (TFS) (40). Advanced derivatives like Blood Speckle Tracking (BST) combine high-frame-rate ultrasound with angle-independent tissue motion analysis to enable quantitative vortex mapping of myocardial speckle patterns, offering detailed hemodynamic insights (42).

Research applications of STE continue to expand. It is used to investigate cardiac function and myocardial stress in fetuses with aortic coarctation (43), assess critical aortic stenosis (44), and evaluate blood flow patterns in the ascending aorta (AAo) of fetuses with TOF in relation to cardiac geometry (45). Additionally, STE supports early screening of cardiac dysfunction in fetuses of mothers with gestational diabetes mellitus (46) and in cases of hypertensive disorders during pregnancy (47,48). Novel technologies such as FetalHQ further extend the utility of speckle tracking by enabling detailed measurement of atrial size, shape, and contractility in the fetal heart (49,50).

STE and its evolving methodologies provide a robust, non-invasive framework for comprehensive assessment of fetal cardiac structure and function, proving especially valuable in complex CHD and conditions associated with altered fetal cardiac load.

Application of AI in FE

Deep learning-based FE

The application of deep learning in FE is advancing rapidly along a clear trajectory: from assisting imaging to enabling automated diagnosis, and ultimately progressing toward clinical validation and integration.

Its development begins with the optimization of image acquisition and quality control. Early research primarily utilized AI models to assist sonographers in rapidly locating standard cardiac views—for example, through integrated frameworks combining recognition, localization, and quality assessment (such as MCLN/ICLN) or dedicated recognition networks (such as FHUSP-NET)—to enhance the standardization and image quality of examinations (21,51,52).

Building on the foundation of standardized imaging, research focus quickly shifted to the automated detection and differentiation of CHD. Multiple studies have validated its feasibility: convolutional neural networks (CNNs) have achieved automatic classification of FE (53); the DGACNN model attained an 85% accuracy rate in identifying fetal heart disease (54); and more complex models have demonstrated performance comparable to clinical experts [with area under the curve (AUC) as high as 0.99] in distinguishing between normal and complex CHD (55). For the differentiation of specific conditions, such as VSD and TOF, specially designed fine-grained image classification models (e.g., NTS-Net and WSDAN) have exhibited excellent performance (56). Furthermore, the DDCHD-DenseNet model, employing a two-stage deep transfer learning approach, demonstrated screening performance from aortic arch views on par with that of senior sonographers (57).

To address the challenge of scarce high-quality annotated data, techniques such as generative adversarial networks (GANs) have been utilized to synthesize images for dataset augmentation (58). Simultaneously, the technological frontier is expanding from structural recognition to intelligent quantification and prediction of cardiac function. Examples include the development of a Fetal Heart Rhythm Intelligent Quantification System (HR-IQS) for automated index extraction (59) and the construction of the first machine learning prediction model (CR-FGR) that quantifies fetal cardiac remodeling as a direct biomarker (60). At the image analysis level, a YOLOX-based instance segmentation network (IS-YOLOX) has achieved precise localization and segmentation of cardiac anatomical structures (61).

The clinical practicality and generalizability of these AI tools are being validated. Research indicates that AI-based software can effectively assist clinicians in identifying suspicious signs of CHD during prenatal ultrasound (62), and its screening performance in low-risk populations within community healthcare settings can match or even surpass that of specialist physicians (63,64). This highlights its significant potential as a standardized initial screening tool for optimizing resource allocation and improving the overall detection rate of CHD (65,66).

In summary, by providing automated and standardized analytical tools, deep learning not only enhances the workflow efficiency and quality of fetal cardiac ultrasound but is also driving the field toward greater efficiency and precision through the realization of high-accuracy, reproducible quantitative diagnosis and risk prediction.

Fetal intelligent navigation echocardiography (FINE)

Assessing the fetal cardiac axis is a crucial part of FE that can improve the detection rate of CHD. FINE (67) is a relatively advanced technology: the software generates a virtual map of STIC volume data sets, forming nine standard fetal echocardiographic views and detecting the cardiac axis. Combining this result with a semi-automated method for assessing the cardiac axis may further enhance the detection rate of fetal CHD. This technology has been proven effective for prenatal screening of complete transposition of the great arteries in fetuses (68).

Limitations

Although this study followed a systematic literature search protocol and aimed to provide a structured analysis of technological evolution, as a narrative review, its strength of evidence is inherently lower than that of a systematic review or meta-analysis. Firstly, despite rigorous efforts in the process of literature screening and synthesis, there remains an inherent risk of selection and interpretive bias. Secondly, this field (particularly AI in cardiac imaging) is rapidly evolving, meaning the latest technological breakthroughs and clinical evidence may not be fully captured. Finally, the significant heterogeneity among the included original studies—in terms of design, population characteristics, and outcome measures—limits the feasibility of performing quantitative pooled analyses or direct comparisons. It is also important to note that the diagnostic accuracy of FE itself is constrained by factors such as operator expertise, equipment, and fetal conditions; a single examination cannot rule out all congenital heart defect. Therefore, it is emphasized that prenatal screening must be embedded within a sequential clinical pathway that includes standardized protocols, ongoing training (69), and postnatal screening (70).

Current landscape, challenges, and future directions

Current status and challenges in diagnostic efficacy

FE serves as a cornerstone tool for the prenatal screening of CHD, yet its diagnostic performance exhibits significant heterogeneity, with detection rates ranging approximately from 60% to 90%. This variability primarily stems from operator experience, limitations in the screening views used, and the diagnostic difficulty associated with certain disease subtypes (71). Diagnostic errors can lead to serious consequences. Although its high diagnostic accuracy (exceeding 90%) has been confirmed, accessibility in rural areas remains constrained by economic and logistical factors (71). Furthermore, the geographic diversity of the served population—spanning urban and rural areas as well as different states—places greater demands on the technology’s general applicability (72).

Quality assurance: standardization and comprehensive verification

Establishing a systematic pathway for quality assurance is crucial for enhancing and stabilizing diagnostic quality. The fundamental foundation lies in cultivating a workforce of sonographers capable of performing standardized examinations through nationally certified training centers. In clinical practice, for cases with screening abnormalities or complex presentations, a multimodal comprehensive diagnostic strategy should be employed for verification and precise management. This involves the integration and cross-referencing of FE with fetal cardiac magnetic resonance imaging (MRI), postnatal imaging studies, and genetic analysis (73). It is essential to acknowledge the inherent limitations of the technique: the uniqueness of fetal circulation and physiological changes pose difficulties in diagnosing certain anomalies, and a single examination cannot definitively rule out all abnormalities. Consequently, even with a normal prenatal ultrasound result, neonatal cardiac screening remains an indispensable safety net.

Future research directions

Future research should converge on several closely interconnected fronts. The primary task is to enhance the diagnostic accuracy for cases flagged as abnormal during mid-trimester screening and to strengthen the consistency of operations and reporting across different healthcare institutions. Secondly, there is a need to further clarify the complementary roles of FE and routine obstetric ultrasound within the screening pathway and to deepen the capability for risk stratification utilizing quantitative FE parameters (74). Thirdly, rigorously designed multicenter real-world studies are necessary to validate the clinical utility, cost-effectiveness, and generalizability of AI-assisted tools and the tiered screening framework they support. Finally, the advancement of all technological applications relies on the construction of a foundational research and development infrastructure. This entails establishing standardized, high-quality, and compliant shared imaging databases to train more robust and generalizable AI models, and actively exploring multimodal predictive models that integrate imaging features, biomarkers, and genetic information. A conceptual framework for a future-tiered intelligent screening ecosystem for fetal CHD is presented in Figure 2.

Figure 2 Conceptual framework of a future-tiered intelligent screening ecosystem for fetal CHD. 3D/4D STIC, three-dimensional/four-dimensional spatiotemporal image correlation; AI, artificial intelligence; CHD, congenital heart disease.

Conclusions

FE is an indispensable core tool for the prenatal diagnosis of CHD. Its developmental trajectory is evolving from a qualitative assessment reliant on individual expertise towards a new phase of technology-enabled, increasingly standardized, and quantitative practice. The current system still faces systemic challenges in diagnostic consistency, detection of complex malformations, and equitable access to medical resources. The key to a successful future prenatal cardiac care system lies in constructing a tiered and integrated intelligent ecosystem. This ecosystem would utilize AI and standardized training to ensure homogeneous quality in primary screening, while leveraging tele-collaboration platforms and multimodal comprehensive diagnostic pathways to accurately direct complex cases to regional advanced diagnosis and intervention centers. This systemic evolution represents not merely an iteration of ultrasound technology but a profound transformation in healthcare delivery models. It holds crucial practical significance for comprehensively improving the prenatal detection rate of CHD, optimizing the perinatal management cascade, and ultimately enhancing the long-term prognosis of affected children.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-798/rc

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-798/prf

Funding: This study was supported by the Tianshan Talent Training Program (the Third Batch of Medical and Health Leading Talents) of Xinjiang Uygur Autonomous Region (No. TSYC202401A107) and the Science and Technology Innovation Team (Tianshan Innovation Team) Project of Xinjiang Uygur Autonomous Region (No. 2022TSYCTD0016).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-798/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cao J, Wu Q, Huang Y, et al. The role of DNA methylation in syndromic and non-syndromic congenital heart disease. Clin Epigenetics 2021;13:93. [Crossref] [PubMed]
2. Zhang Y, Zhang W, Xu H, et al. Epidemiological Aspects, Prenatal Screening and Diagnosis of Congenital Heart Defects in Beijing. Front Cardiovasc Med 2021;8:777899. [Crossref] [PubMed]
3. Zhang S, Wang L, Yang T, et al. Parental alcohol consumption and the risk of congenital heart diseases in offspring: An updated systematic review and meta-analysis. Eur J Prev Cardiol 2020;27:410-21. [Crossref] [PubMed]
4. Pasupathy D, Denbow ML, Rutherford MA, et al. The Combined Use of Ultrasound and Fetal Magnetic Resonance Imaging for a Comprehensive Fetal Neurological Assessment in Fetal Congenital Cardiac Defects: Scientific Impact Paper No. 60. BJOG 2019;126:e142-51. [Crossref] [PubMed]
5. Zhao QM, Liu F, Wu L, et al. Prevalence of Congenital Heart Disease at Live Birth in China. J Pediatr 2019;204:53-8. [Crossref] [PubMed]
6. Chen X, Zhao S, Dong X, et al. Incidence, distribution, disease spectrum, and genetic deficits of congenital heart defects in China: implementation of prenatal ultrasound screening identified 18,171 affected fetuses from 2,452,249 pregnancies. Cell Biosci 2023;13:229. [Crossref] [PubMed]
7. Alanen J, Korpimaki T, Kouru H, et al. First trimester combined screening biochemistry in detection of congenital heart defects. J Matern Fetal Neonatal Med 2019;32:3272-7. [Crossref] [PubMed]
8. Alanen J, Leskinen M, Sairanen M, et al. Fetal nuchal translucency in severe congenital heart defects: experiences in Northern Finland. J Matern Fetal Neonatal Med 2019;32:1454-60. [Crossref] [PubMed]
9. Sun HY. Prenatal diagnosis of congenital heart defects: echocardiography. Transl Pediatr 2021;10:2210-24. [Crossref] [PubMed]
10. Wu Y, Kapse K, Jacobs M, et al. Association of Maternal Psychological Distress With In Utero Brain Development in Fetuses With Congenital Heart Disease. JAMA Pediatr 2020;174:e195316. [Crossref] [PubMed]
11. Harris KW, Brelsford KM, Kavanaugh-McHugh A, et al. Uncertainty of Prenatally Diagnosed Congenital Heart Disease: A Qualitative Study. JAMA Netw Open 2020;3:e204082. [Crossref] [PubMed]
12. Zhang Y, Wang J, Zhao J, et al. Current status and challenges in prenatal and neonatal screening, diagnosis, and management of congenital heart disease in China. Lancet Child Adolesc Health 2023;7:479-89. [Crossref] [PubMed]
13. Madeddu P. Looking at a baby's heart through the lens of the mother's blood. EMBO Mol Med 2023;15:e18680. [Crossref] [PubMed]
14. Yang S, Qin G, He G, et al. Evaluation of first-trimester ultrasound screening strategy for fetal congenital heart disease. Ultrasound Obstet Gynecol 2025;65:478-86. [Crossref] [PubMed]
15. Cawyer CR, Kuper SG, Ausbeck E, et al. The added value of screening fetal echocardiography after normal cardiac views on a detailed ultrasound. Prenat Diagn 2019;39:1148-54. [Crossref] [PubMed]
16. Sillesen AS, Raja AA, Pihl C, et al. Copenhagen Baby Heart Study: a population study of newborns with prenatal inclusion. Eur J Epidemiol 2019;34:79-90. [Crossref] [PubMed]
17. Chen J, Wang J, Sun H, et al. Fetal cardiac tumor: echocardiography, clinical outcome and genetic analysis in 53 cases. Ultrasound Obstet Gynecol 2019;54:103-9. [Crossref] [PubMed]
18. Practice Bulletin No. 175: Ultrasound in Pregnancy. Obstet Gynecol 2016;128:e241-56. [Crossref] [PubMed]
19. Bak GS, Shaffer BL, Madriago E, et al. Detection of fetal cardiac anomalies: cost-effectiveness of increased number of cardiac views. Ultrasound Obstet Gynecol 2020;55:758-67. [Crossref] [PubMed]
20. Gottschalk I, Strizek B, Jehle C, et al. Prenatal Diagnosis and Postnatal Outcome of Fetuses with Pulmonary Atresia and Ventricular Septal Defect. Ultraschall Med 2020;41:514-25. [Crossref] [PubMed]
21. Zhang Y, Zhu H, Cheng J, et al. Improving the Quality of Fetal Heart Ultrasound Imaging With Multihead Enhanced Self-Attention and Contrastive Learning. IEEE J Biomed Health Inform 2023;27:5518-29. [Crossref] [PubMed]
22. Qiao S, Pang S, Luo G, et al. FLDS: An Intelligent Feature Learning Detection System for Visualizing Medical Images Supporting Fetal Four-Chamber Views. IEEE J Biomed Health Inform 2022;26:4814-25. [Crossref] [PubMed]
23. Mebius MJ, Clur SAB, Vink AS, et al. Growth patterns and cerebroplacental hemodynamics in fetuses with congenital heart disease. Ultrasound Obstet Gynecol 2019;53:769-78. [Crossref] [PubMed]
24. Ling W, Wu Q, Guo S, et al. Four-section approach of fetal congenital heart disease at 11-13(+6) weeks. Front Cardiovasc Med 2023;10:1206042. [Crossref] [PubMed]
25. Moon-Grady AJ, Donofrio MT, Gelehrter S, et al. Guidelines and Recommendations for Performance of the Fetal Echocardiogram: An Update from the American Society of Echocardiography. J Am Soc Echocardiogr 2023;36:679-723. [Crossref] [PubMed]
26. Byrne JJ, Morgan JL, Twickler DM, et al. Utility of follow-up standard sonography for fetal anomaly detection. Am J Obstet Gynecol 2020;222:615.e1-9. [Crossref] [PubMed]
27. Minnella GP, Crupano FM, Syngelaki A, et al. Diagnosis of major heart defects by routine first-trimester ultrasound examination: association with increased nuchal translucency, tricuspid regurgitation and abnormal flow in ductus venosus. Ultrasound Obstet Gynecol 2020;55:637-44. [Crossref] [PubMed]
28. Karim JN, Bradburn E, Roberts N, et al. First-trimester ultrasound detection of fetal heart anomalies: systematic review and meta-analysis. Ultrasound Obstet Gynecol 2022;59:11-25. [Crossref] [PubMed]
29. Zidere V, Vigneswaran TV, Syngelaki A, et al. Reference Ranges for Pulsed-Wave Doppler of the Fetal Cardiac Inflow and Outflow Tracts from 13 to 36 Weeks' Gestation. J Am Soc Echocardiogr 2021;34:1007-1016.e10. [Crossref] [PubMed]
30. Expert Panel on GYN and OB Imaging. ACR Appropriateness Criteria® Nuchal Translucency Evaluation at 11 to 14 Weeks of Gestation. J Am Coll Radiol 2020;17:S447-58. [Crossref] [PubMed]
31. Pu B, Lu Y, Chen J, et al. MobileUNet-FPN: A Semantic Segmentation Model for Fetal Ultrasound Four-Chamber Segmentation in Edge Computing Environments. IEEE J Biomed Health Inform 2022;26:5540-50. [Crossref] [PubMed]
32. Karmegaraj B, Kumar S, Srimurugan B, et al. 3D/4D spatiotemporal image correlation (STIC) fetal echocardiography provides incremental benefit over 2D fetal echocardiography in predicting postnatal surgical approach in double-outlet right ventricle. Ultrasound Obstet Gynecol 2021;57:423-30. [Crossref] [PubMed]
33. Turan S, Goetzinger KR. First-trimester fetal heart evaluation: time to move forward. Ultrasound Obstet Gynecol 2021;57:677-80. [Crossref] [PubMed]
34. Vaidyanathan B, Verma R, Kappanayil M, et al. Prenatal Diagnosis of a Large Left Coronary Artery to Right Ventricle Fistula With 3-Dimensional/4-Dimensional Spatiotemporal Image Correlation Rendering Followed by Successful Neonatal Transcatheter Closure. Circ Cardiovasc Imaging 2022;15:e014247. [Crossref] [PubMed]
35. Wang Y, Wang L, Zang H, et al. Comparison of Different Ultrasonic Screening Methods and Analysis of High Risk Factors for Fetal Cardiac Malformation in Second Trimester of Pregnancy. Pediatr Cardiol 2025;46:769-77. [Crossref] [PubMed]
36. Li TG, Wu WR, Su XR, et al. Prenatal diagnosis of interrupted aortic arch using high-definition flow render mode and spatiotemporal image correlation. Echocardiography 2024;41:e15828. [Crossref] [PubMed]
37. Oliveira LB, Peixoto AB, Bravo-Valenzuela NJ, et al. Evaluation of Fetal Myocardial Contractility of Pre-Existing Diabetes Mellitus by Three-Dimensional Ultrasonography Using STIC and Vocal Methods. J Clin Ultrasound 2025;53:1078-86. [Crossref] [PubMed]
38. Zhang WD, Ma B, Qi PA, et al. Anomalous drainage of the ductus venosus into the coronary sinus: prenatal ultrasound diagnosis utilizing two-dimensional and three-dimensional imaging techniques and differential diagnosis. BMC Pregnancy Childbirth 2025;25:437. [Crossref] [PubMed]
39. Soveral I, Crispi F, Guirado L, et al. Fetal cardiac filling and ejection time fractions by pulsed-wave Doppler: reference ranges and potential clinical application. Ultrasound Obstet Gynecol 2021;58:83-91. [Crossref] [PubMed]
40. DeVore GR, Satou G, Sklansky M. Using speckle-tracking echocardiography to assess fetal myocardial deformation: are we there yet? Yes we are! Ultrasound Obstet Gynecol 2019;54:703-4. [Crossref] [PubMed]
41. DeVore GR, Haxel C, Satou G, et al. Improved detection of coarctation of the aorta using speckle-tracking analysis of fetal heart on last examination prior to delivery. Ultrasound Obstet Gynecol 2021;57:282-91. [Crossref] [PubMed]
42. Liu Y, Zhou D, Yang Y, et al. Vortex formation of cardiac bilateral ventricles in fetuses with tetralogy of fallot assessed by blood speckle-tracking echocardiography. BMC Pregnancy Childbirth 2025;25:1279. [Crossref] [PubMed]
43. Chen G, Xiang W, Zhang Y, et al. Abnormalities in Cardiac Function, Anatomy, and Myocardial Strain in Coarctation of the Aorta. Echocardiography 2025;42:e70328. [Crossref] [PubMed]
44. Murlewska J, Witkowski S, Strzelecka I, et al. Fetal Speckle Tracking Technology for Critical Aortic Stenosis: Advancing Through Innovation. Diagnostics (Basel) 2025;15:2591. [Crossref] [PubMed]
45. Zhou D, Liu Y, Xu G, et al. Abnormal flow patterns in ascending aorta in fetuses with tetralogy of Fallot: evaluation by blood speckle-tracking echocardiography. Ultrasound Obstet Gynecol 2025;66:614-20. [Crossref] [PubMed]
46. Tsokkou S, Konstantinidis I, Matsas A, et al. Doppler Evaluation of Fetal Cardiac Function in Gestational Diabetes Mellitus: A Scoping Review Providing Insights into Hemodynamic and Structural Alterations. J Clin Med 2025;14:5622. [Crossref] [PubMed]
47. Zhang Y, Zhang L, Sun C, et al. The effects of hypertensive disorders of pregnancy on fetal myocardial remodelling and cardiac function assessed by two-dimensional speckle tracking imaging. J Matern Fetal Neonatal Med 2025;38:2531146. [Crossref] [PubMed]
48. Peng Y, Feng W, Yu CX, et al. Fetal heart quantification ultrasound technology for the quantitative analysis of fetal cardiac morphology and function in hypertensive disorders of pregnancy. Quant Imaging Med Surg 2025;15:3517-31. [Crossref] [PubMed]
49. DeVore GR. Measuring Atrial Size, Shape, and Contractility of the Fetal Heart Using FetalHQ: A New Technique Using Speckle Tracking Analysis. J Ultrasound Med 2024;43:2399-410. [Crossref] [PubMed]
50. Sharma S, Bennasar M, Yadav M. Evaluation of Fetal Heart Using Fetal Heart Quantification (fetalHQ) Technique in an Indian Cohort. J Clin Ultrasound 2025;53:271-85. [Crossref] [PubMed]
51. Wu H, Wu B, Lai F, et al. Application of Artificial Intelligence in Anatomical Structure Recognition of Standard Section of Fetal Heart. Comput Math Methods Med 2023;2023:5650378. [Crossref] [PubMed]
52. Li F, Li P, Wu X, et al. FHUSP-NET: A Multi-task model for fetal heart ultrasound standard plane recognition and key anatomical structures detection. Comput Biol Med 2024;168:107741. [Crossref] [PubMed]
53. Shi J, Bai Y, Hou Q, et al. Automated and accurate view classification of fetal echocardiography using convolutional neural networks. Quant Imaging Med Surg 2025;15:12582-92. [Crossref] [PubMed]
54. Gong Y, Zhang Y, Zhu H, et al. Fetal Congenital Heart Disease Echocardiogram Screening Based on DGACNN: Adversarial One-Class Classification Combined with Video Transfer Learning. IEEE Trans Med Imaging 2020;39:1206-22. [Crossref] [PubMed]
55. Arnaout R, Curran L, Zhao Y, et al. An ensemble of neural networks provides expert-level prenatal detection of complex congenital heart disease. Nat Med 2021;27:882-91. [Crossref] [PubMed]
56. Yu X, Ma L, Wang H, et al. Deep learning-based differentiation of ventricular septal defect from tetralogy of Fallot in fetal echocardiography images. Technol Health Care 2024;32:457-64. [Crossref] [PubMed]
57. Tang J, Liang Y, Jiang Y, et al. A multicenter study on two-stage transfer learning model for duct-dependent CHDs screening in fetal echocardiography. NPJ Digit Med 2023;6:143. [Crossref] [PubMed]
58. Qiao S, Pan S, Luo G, et al. A Pseudo-Siamese Feature Fusion Generative Adversarial Network for Synthesizing High-Quality Fetal Four-Chamber Views. IEEE J Biomed Health Inform 2023;27:1193-204. [Crossref] [PubMed]
59. Yang X, Huang X, Wei C, et al. An intelligent quantification system for fetal heart rhythm assessment: A multicenter prospective study. Heart Rhythm 2024;21:600-9. [Crossref] [PubMed]
60. Zhu C, Li J, Ren M, et al. Machine learning-enhanced prediction of fetal growth restriction using fetal cardiac remodeling parameters. BMC Med 2025;23:634. [Crossref] [PubMed]
61. Lu Y, Li K, Pu B, et al. A YOLOX-Based Deep Instance Segmentation Neural Network for Cardiac Anatomical Structures in Fetal Ultrasound Images. IEEE/ACM Trans Comput Biol Bioinform 2024;21:1007-18. [Crossref] [PubMed]
62. Lam-Rachlin J, Punn R, Behera SK, et al. Use of Artificial Intelligence-Based Software to Aid in the Identification of Ultrasound Findings Associated With Fetal Congenital Heart Defects. Obstet Gynecol 2026;147:108-17. [Crossref] [PubMed]
63. Athalye C, van Nisselrooij A, Rizvi S, et al. Deep-learning model for prenatal congenital heart disease screening generalizes to community setting and outperforms clinical detection. Ultrasound Obstet Gynecol 2024;63:44-52. [Crossref] [PubMed]
64. Patey O, Hernandez-Cruz N, D'Alberti E, et al. Prenatal detection of congenital heart defects using the deep learning-based image and video analysis: protocol for Clinical Artificial Intelligence in Fetal Echocardiography (CAIFE), an international multicentre multidisciplinary study. BMJ Open 2025;15:e101263. [Crossref] [PubMed]
65. Amedro P, Vincenti M. Response to the letter about the article "Prognosis of severe congenital heart diseases: Do we overestimate the impact of prenatal diagnosis?". Arch Cardiovasc Dis 2019;112:365-6. [Crossref] [PubMed]
66. Martins JG, Horgan R, Sinkovskaya E. Enhancing Fetal Cardiac Imaging With Artificial Intelligence: A Review of the Current Evidence and Future Directions. Clin Obstet Gynecol 2026;69:62-9. [Crossref] [PubMed]
67. Weichert A, Gembicki M, Weichert J, et al. Semi-Automatic Measurement of Fetal Cardiac Axis in Fetuses with Congenital Heart Disease (CHD) with Fetal Intelligent Navigation Echocardiography (FINE). J Clin Med 2023;12:6371. [Crossref] [PubMed]
68. Veronese P, Guariento A, Cattapan C, et al. Prenatal Diagnosis and Fetopsy Validation of Complete Atrioventricular Septal Defects Using the Fetal Intelligent Navigation Echocardiography Method. Diagnostics (Basel) 2023;13:456. [Crossref] [PubMed]
69. van Nisselrooij AEL, Teunissen AKK, Clur SA, et al. Why are congenital heart defects being missed? Ultrasound Obstet Gynecol 2020;55:747-57. [Crossref] [PubMed]
70. Lannering K, Kazamia K, Bergman G, et al. Screening for Critical Congenital Heart Defects in Sweden. Pediatrics 2023;152:e2023061949. [Crossref] [PubMed]
71. Bachnas MA, Andonotopo W, Pribadi A, et al. Fetal cardiac diagnostics in Indonesia: a study of screening and echocardiography. J Perinat Med 2025;53:561-75. [Crossref] [PubMed]
72. Chambers HN, Caris E, Conwell J, et al. Suboptimal Imaging on Obstetric Ultrasound Should Prompt Early Referral for Fetal Echocardiography. Pediatr Cardiol 2025;46:778-84. [Crossref] [PubMed]
73. Haxel CS, Johnson JN, Hintz S, et al. Care of the Fetus With Congenital Cardiovascular Disease: From Diagnosis to Delivery. Pediatrics 2022;150:e2022056415C. [Crossref] [PubMed]
74. Simpson J, Hornberger LK. What Does Fetal Echocardiography Add Beyond the Anomaly Scan? Circ Cardiovasc Imaging 2022;15:e014168. [Crossref] [PubMed]