Fetal Cardiac Screening: What Are We (and Our Guidelines) Doing Wrong?

Artificial Intelligence in Obstetric Anomaly Scan: Heart and Brain

BACKGROUND

The ultrasound scan represents the first tool that obstetricians use in fetal evaluation, but sometimes, it can be limited by mobility or fetal position, excessive thickness of the maternal abdominal wall, or the presence of post-surgical scars on the maternal abdominal wall. Artificial intelligence (AI) has already been effectively used to measure biometric parameters, automatically recognize standard planes of fetal ultrasound evaluation, and for disease diagnosis, which helps conventional imaging methods. The usage of information, ultrasound scan images, and a machine learning program create an algorithm capable of assisting healthcare providers by reducing the workload, reducing the duration of the examination, and increasing the correct diagnosis capability. The recent remarkable expansion in the use of electronic medical records and diagnostic imaging coincides with the enormous success of machine learning algorithms in image identification tasks.

OBJECTIVES

We aim to review the most relevant studies based on deep learning in ultrasound anomaly scan evaluation of the most complex fetal systems (heart and brain), which enclose the most frequent anomalies.

The Left Outflow Tract in Fetal Cardiac Screening Examination: Introduction of Quality Criteria Is Not Always Associated With an Improvement of Practice When Supervised by Humans

OBJECTIVES

Since 2016, the French CNEOF included the left ventricular outflow tract (LVOT) in the second and third trimester of pregnancy in addition to the four-chamber view and the parasagittal view of the right outflow tract. The objective of this study was to define quality criteria for fetal LVOT assessment and to perform a human audit of past and current practices, before and after the implementation of those quality criteria at a large scale.

METHODS

Seven quality criteria were investigated and rated from 0 to 1 during three periods of interest. Files were randomly selected from three centers, and average total and specific scores were calculated.

RESULTS

LVOT pictures were present in more than 94.3% of reports. The average quality score was 5.49/7 (95% confidence interval [CI]: 5.36-5.62), 5.91/7 (95% CI: 5.80-6.03), and 5.70/7 (95% CI: 5.58-5.82) for the three centers in the three periods of interest. There was no significant difference following the introduction of the quality criteria, 2017 versus 2020, P = .054.

CONCLUSION

Fetal LVOT images were present in most of ultrasound reports but the introduction of the proposed quality criteria under human supervision seems not associated with a significant change in practice.

Prenatal Detection of Congenital Heart Disease: Importance of Fetal Echocardiography Following Normal Fetal Cardiac Screening

Fetal echocardiography may be performed because of noncardiac indications (the pregnancy is identified as high risk for fetal cardiac disease), or because of fetal cardiac indications (abnormal fetal heart at the time of a screening ultrasound). Considering recent improvements in fetal cardiac screening over the past decade, the goal of this single institution study was to reconsider the importance of performing fetal echocardiography purely for screening (noncardiac) indications. We performed a retrospective analysis to review screening and fetal cardiac indications and fetal cardiac findings for fetal echocardiograms performed at UCLA between 2015 and 2019. Fetal heart disease was identified in 391 (15%) of 2592 pregnancies in this study. Among these 391 cases, 227 (58%) occurred in low-risk pregnancies (without screening indications). While 79% of the cases of fetal cardiac disease were referred with fetal cardiac indications, 21% of the cases were referred with exclusively screening indications. Fetal cardiac disease was discovered on fetal echocardiograms in 4% of pregnancies referred for exclusively screening indications, but the frequency of fetal cardiac disease following normal fetal cardiac screening has decreased from 6% in 2015 to 3% in 2019. In our population, we recommend continued referral for fetal echocardiography for pregnancies identified as high risk for CHD. However, as fetal cardiac screening continues to improve, referral for fetal echocardiography following normal fetal cardiac screening will have diminishing value and yield.

A 'holistic' sonographic view on congenital heart disease - How automatic reconstruction using fetal intelligent navigation echocardiography (FINE) eases the unveiling of abnormal cardiac anatomy part I: Right heart anomalies

Attempting a comprehensive examination of the fetal heart remains challenging for unexperienced operators as it emphasizes the acquisition and documentation of sequential cross-sectional and sagittal views and inevitably results in diminished detection rates of fetuses affected by congenital heart disease. The introduction of four-dimensional spatio-temporal image correlation (4D STIC) technology facilitated a volumetric approach for thorough cardiac anatomic evaluation by the acquisition of cardiac 4D datasets. By analyzing and re-arranging of numerous frames according to their temporal event within the heart cycle, STIC allows visualization of cardiac structures as an endless cine loop sequence of a complete single cardiac cycle in motion. However, post-analysis with manipulation and repeated slicing of the volume usually requires experience and in-depth anatomic knowledge, which limits the widespread application of this advanced technique in clinical care and unfortunately leads to the underestimation of its diagnostic value to date. Fetal intelligent navigation echocardiography (FINE), a novel method that automatically generates and displays nine standard fetal echocardiographic views in normal hearts, has shown to be able to overcome these limitations. Very recent data on the detection of congenital heart defects (CHDs) using the FINE method revealed a high sensitivity and specificity of 98% and 93%, respectively. In this two-part manuscript, we focused on the performance of FINE in delineating abnormal anatomy of typical right and left heart lesions and thereby emphasized the educational potential of this technology for more than just teaching purposes. We further discussed recent findings in a pathophysiological and/or functional context.

An ensemble of neural networks provides expert-level prenatal detection of complex congenital heart disease

Congenital heart disease (CHD) is the most common birth defect. Fetal screening ultrasound provides five views of the heart that together can detect 90% of complex CHD, but in practice, sensitivity is as low as 30%. Here, using 107,823 images from 1,326 retrospective echocardiograms and screening ultrasounds from 18- to 24-week fetuses, we trained an ensemble of neural networks to identify recommended cardiac views and distinguish between normal hearts and complex CHD. We also used segmentation models to calculate standard fetal cardiothoracic measurements. In an internal test set of 4,108 fetal surveys (0.9% CHD, >4.4 million images), the model achieved an area under the curve (AUC) of 0.99, 95% sensitivity (95% confidence interval (CI), 84-99%), 96% specificity (95% CI, 95-97%) and 100% negative predictive value in distinguishing normal from abnormal hearts. Model sensitivity was comparable to that of clinicians and remained robust on outside-hospital and lower-quality images. The model's decisions were based on clinically relevant features. Cardiac measurements correlated with reported measures for normal and abnormal hearts. Applied to guideline-recommended imaging, ensemble learning models could significantly improve detection of fetal CHD, a critical and global diagnostic challenge.

Discordant congenital heart defects in monochorionic twins: Risk factors and proposed pathophysiology

A six-fold increase in congenital heart defects (CHD) exists among monochorionic (MC) twins compared to singleton or dichorionic twin pregnancies. Though MC twins share an identical genotype, discordant phenotypes related to CHD and other malformations have been described, with reported rates of concordance for various congenital anomalies at less than 20%. Our objective was to characterize the frequency and spectrum of CHD in a contemporary cohort of MC twins, coupled with genetic and clinical variables to provide insight into risk factors and pathophysiology of discordant CHD in MC twins. Retrospective analysis of all twins receiving prenatal fetal echocardiography at a single institution from January 2010 –March 2020 (N = 163) yielded 23 MC twin pairs (46 neonates) with CHD (n = 5 concordant CHD, n = 18 discordant CHD). The most common lesions were septal defects (60% and 45.5% in concordant and discordant cohorts, respectively) and right heart lesions (40% and 18.2% in concordant and discordant cohorts, respectively). Diagnostic genetic testing was abnormal for 20% of the concordant and 5.6% of the discordant pairs, with no difference in rate of abnormal genetic results between the groups (p = 0.395). No significant association was found between clinical risk factors and development of discordant CHD (p>0.05). This data demonstrates the possibility of environmental and epigenetic influences versus genotypic factors in the development of discordant CHD in monochorionic twins.

Improving fetal congenital heart disease screening using a checklist-based approach

To determine if using a checklist of specific ultrasound image criteria to screen the fetal heart improves the cardiac exam completion rate, defined as the ability to classify the heart as normal or abnormal. This is a retrospective cohort study of patients with singleton pregnancies who underwent a fetal anatomy survey between 18 and 28 weeks' gestation. A checklist was used from 1 September 2015 to 31 March 2016 to categorize exams as complete-normal, complete-abnormal, or incomplete. Performance was compared with a 7-month period prior to checklist introduction (1 December 2014 to 30 June 2015). Checklist utilization improved the cardiac exam completion rate by 8.9%. With the checklist, 1083 of 1202 exams (90.1%) were completed compared to 987 of 1193 (82.7%) pre-checklist, P < .001. We did not detect a change in cases classified as abnormal and referred for echocardiography: 25 (2.1%) with the checklist and 16 (1.3%) pre-checklist, P = .16. We did not detect more congenital heart disease (CHD), 12 (1.0%) with checklist screening, 5 (0.4%) pre-checklist, P = .14. Critical CHD was not missed in either group. Using the checklist improved the cardiac exam completion rate. There was no change in congenital heart disease detection.

[The left outflow tract: Normal features]

Congenital heart disease remains an important cause of perinatal morbidity and mortality. Screening for these is based on a good knowledge of normal fetal heart anatomy and ultrasound views to be performed. After recommending the use of the four chambers and the right outflow tract views in 2005, CNEOF proposes recently adding the left outflow tract assessment. The use of this one should sensitize the operator to the notion of sweeping essential to obtain all these views and make it possible to improve the detection of the conal VSD and conotruncal pathologies.

Detailed Fetal Anatomic Ultrasound Examination: Effect of the 2014 Consensus Report on a Tertiary Referral Center

This study evaluates the impact of extended cardiac views on examination time, repeat imaging, and anomaly detection before and after implementation of 76811 guidelines (American Institute of Ultrasound in Medicine Consensus 2014). It is a retrospective study of singleton pregnancies undergoing detailed ultrasound imaging at 18 weeks' gestation or greater before and after the protocol change in an academic, tertiary care fetal center. Views required prior to 2014: 4-chamber, left outflow tract, right outflow tract. Additional views required after 2014: bicaval, aortic arch, 3-vessel, and 3-vessel trachea. Fetuses with known anomalies were excluded. Rates of detection of congenital heart disease (CHD), examination completion, repeat examination recommendation, fetal echocardiogram recommendation, completion by body mass index, and cardiac examination time were determined. Six hundred twenty-four subjects were included, 217 before and 407 after protocol change. Views obtained were as stated in the American Institute of Ultrasound in Medicine/Society for Maternal-Fetal Medicine consensus. Detection of CHD was not improved. Examination times increased by 20% (6.4 vs 7.7 minutes, P < 0.05). Number of incomplete studies increased by 130% (11% to 26%, P < 0.05). Twice as many patients were referred for repeat examination (6% vs 13%, P < 0.05). Completion rates were negatively correlated with body mass index. Recommendations for fetal echocardiogram were unchanged (5% vs 6%, P = 0.6). Additional imaging did not increase detection rate of CHD (3% vs 2%, P = 0.3). Extended cardiac views resulted in increased examination time, more incomplete examinations, and more repeat examinations without changing detection rates of CHD.