Automated Fetal Cardiac Function in Babies Affected by Heart Diseases

Investigation Plan

Participating patients will be offered a fetal cardiac function echocardiography between 27+6 and 29+6 weeks of gestation and another between 34+6 and 36+6 weeks.

At the stage patients will come in to carry out fetal cardiac function echocardiography they will already know whether the baby is affected by congenital heart disease since malformation screening is carried out between 18 and 20 gestational weeks and will have already been undertaken. In the remote case of detecting an undiagnosed fetal malformation during one of the research scans, patients will be reassured, and appointments arranged at Maternal and Fetal medicine (MFM) clinic for further consultation along with notification of to their own treating doctor.

Every participant will be assigned a study number following recruiting order of participation. De-identification will be undertaken at image acquisition by the doctors prior to image review by the engineers. Researchers will access an online patients' form system, collect the assigned patients' number and add the necessary outcomes information. Ultrasound images will be saved locally in the ultrasound machine by the same assigned patient's study number and uploaded to a specific Microsoft (OneDrive) folder created in Sydney. This data will be stored for a minimum of 5 years after publication, then the data record will be securely destroyed in accordance with the University of New South Wales (UNSW) Records Disposal processes.

The research ultrasound scans will be performed by a trained sonographer using either a clinical ultrasound system or a research ultrasound system that is Therapeutic Goods Administration (TGA) approved for research imaging. The machine used clinically is made by one of the common commercial manufacturers (in this case General Electric E10) and is an identical model to that used in the Department of Maternal and Fetal Medicine. The machine used for research porpoises is called Vantage 256 and it is manufactured by a private company in the United States (Verasonics). It uses the same fundamental electronic circuitry and transducer design as conventional commercial machines, and in fact uses the same transducers as the commercial machines. However, the way that the ultrasound is delivered differs. Instead of transmitting beams as used in conventional ultrasound machines, Verasonics scans the area of interest by unfocused waves, allowing high quality images with a limited number of compounded plane-waves, and reduced acquisition time.

Women will be placed in a semi-recumbent position, as standard for pregnancy ultrasounds. After routine biometry, the research fetal cardiac function ultrasound will be carried out.

Each fetal cardiac function examination will include the following parameters:

1. Fetal biometric parameters (biparietal diameter, head circumference, abdominal circumference, femur diaphysis length)
2. Standard fetal Doppler parameters (umbilical artery, medial cerebral artery, ductus venosus)
3. Fetal cardiac heart rate
4. Presence of pericardial effusion or hydrops
5. Cardiac morphometry - all measurements carried out at the end of diastole, with the exception of atrial dimensions measured in systole (at their maximum extension)

• Heart/thorax area measurement
• 4 heart chambers measurements (apical/basal 4-chambers view in 2D)
• Atrial and ventricular areas (apical/basal 4-chambers view in 2D)
• Ventricular and atrial sphericity calculation
• Inter-ventricular septum and myocardial walls thickness measurement (transversal 4-chambers view in 2D or M-mode)

1. Cardiac contractility

• Spatio-temporal image correlation (STIC) M-Mode stroke volume, ejection fraction and shortening fraction
• Automated STIC Mitral Annular Plane Systolic Excursion (MAPSE), Tricuspid Annular Plane Systolic Excursion (TAPSE), Septal Annular Plane Systolic Excursion (SAPSE)
• Automated Pulsed Wave Doppler (PWD) Left and Right modified myocardial performance index (Mod-MPI)

1. Atrioventricular valves' function evaluation

• Cine-loop evaluation of correct opening and closing
• Anterograde Colour Doppler without regurgitation
• Pulsed Doppler evaluation of flow velocity (monophasic or biphasic)
• Left and right E/A ratio calculation
• If any regurgitation: peak velocity and duration quantified

1. Aorta outflow evaluation

• Aorta artery measurement (at the level of valvular ring in systole)
• Aortic flow evaluation (Colour Doppler evaluation of systolic peak velocity)

1. Pulmonary outflow evaluation

• Pulmonary artery measurement (at the level of valvular ring in systole)
• Pulmonary flow evaluation (Colour Doppler evaluation of systolic peak velocity)

1. V-sign evaluation

• Confirmation of anterograde flow in the entire length of the arteries
• Pulsatility index of aortic isthmus and ductus arteriosus

All fetal morphometric and functional cardiac parameters will be normalised to Z-score by gestational age where possible. Fetal cardiac volumes and 2D images with inadequate quality due to fetal movements, presence of acoustic shadows of fetal ribs or spine, and maternal breathing will be excluded. If hydrops develops, cardiovascular profile score will be added to the routine cardiac function exploration.

The study population will be followed up until delivery and discharge of both mother and neonate. Patients' information will be collected anonymously. Each patient's history will be evaluated and information about previous pregnancies (maternal or fetal diseases during pregnancy) and outcomes (type of delivery, maternal and neonatal conditions at birth, long-term outcome of the pregnancy) will be collected. Furthermore, investigators plan to collect information about the current pregnancy (maternal and fetal observations during pregnancy) and outcomes (type of delivery, maternal and neonatal conditions at birth and up to hospital discharge of both).

Study Procedure Risks

There is no increased risk related to participating in this study. The study uses conventional ultrasound machinery as used in routine fetal evaluation, with no alteration in power output (as defined by Thermal Index or Mechanical Index). The Verasonics research ultrasound system is also comparable to conventional ultrasounds in terms of risks.

Approximately 20-30 minutes scan duration is anticipated for acquisition of the necessary research data which is in keeping with standard ultrasound examinations, and the As Low As Reasonably Achievable (ALARA) principle conventionally applied in fetal imaging.

Sample Size Calculation

The primary outcome of the study is the difference in the mean of automated fetal functional cardiac parameters between CHD cases and controls. This will be analysed using a two-sample t-test.

To estimate the sample size, the most commonly applied fetal functional parameter was used, specifically the left ventricle MPI (LV-MPI) as a proxy of all the automated fetal cardiac parameters.

Due to the rarity of congenital heart disease, the sample size calculation was performed based on recruiting two controls for each case. Pooled across cases with isolated pulmonary valve stenosis (n = 16) and controls (n = 48), a previous work observed a standard deviation of 0.098 in LV-MPI measurements. Using this observed pooled Standard Deviation (SD), a total sample of 381 pregnancies (127 CHD + 254 controls) with completed measurements is required to achieve at least 80% power to detect a difference of 0.03 in mean LV-MPI, with a two-sided type I error rate of 5%.

Investigators acknowledge that some pilot data may be required to evaluate the limited number of pathological cases and therefore some approximations are necessary e.g. for standard deviation within the population. For this reason, investigators have aimed to recruit a larger number of participants (approximately 30%), 165 CHD and 330 Controls, allowing also for some patient exclusions due to patient drop out, difficulties in scanning due to fetal movements etc, and incomplete data sets.

The aim would be to recruit sufficient cases to be able to estimate if there is significant difference in terms of fetal cardiac function parameters between affected and not affected fetuses to inform further research.

Data Analysis Plan

Raw (radio-frequency) ultrasound data generated using the Verasonics will allow the researchers to analyse the signal/image processing that takes place prior to display, enabling refinement of this imaging technique.

Image analysis will be carried out first manually through optical evaluation and then through the use of mathematical algorithms which will recognise and analyse only high-quality images. This could be a limitation because automatically only high-quality images will be included (which is not representative of real clinical work) but also guarantees that parameters are collected only from almost perfect research material (showing true differences if they exist).

Ultrasound images will be analysed and cardiac function parameters interpreted by South Eastern Sydney Local Health District (SESLHD)/UNSW researchers based at the Royal Hospital for Women (RHW), Randwick. Images will be analysed by a team of fetal medicine doctors at Royal Hospital for Women and UNSW engineers to assure that algorithms are correctly applied to calculate fetal cardiac function parameters.

Comparisons of interest between cases and controls in baseline characteristics will be performed using two-sample t-tests, Wilcoxon rank-sum tests or Pearson Chi-squared tests, as appropriate.

Secondary outcomes comparing cases and controls at a single time point will be analysed in a similar way to the primary outcome, subject to checks of assumptions. Analyses of changes in fetal function parameters over time (i.e. between the 27+6-29+6 and 34+6-36+6 week scans), and the comparison of these changes between cases and controls will employ generalised linear mixed models, as appropriate for the parameter.

Logistic regression will be used to estimate the association between fetal cardiac parameters and the incidence of hydrops. Receiver-operating-characteristics-curve analysis will be carried out to assess functional cardiac parameters compared to the routinely used cardiovascular profile score to predict cardiac failure in fetuses with congenital heart disease.

Statistical analysis will be performed using SPSS version 22.0 (SPSS Inc., Chicago, Illinois (IL), USA).

For those cases without complete data acquisition (i.e. intending but not undertaking a second scan), analysis will take place for only the isolated value and not for any temporal change. Their single gestational data set of ultrasound measurements will still be included in analysis but excluded from any analysis of sequential change.

Data Safety and Monitoring Board

To assure high quality data collection, images will be collected at each participating centre by experienced fetal medicine doctors who have practiced fetal medicine for over 10 years. Each image will be stored securely. Data will be anonymised and monitored by researchers at SESLHD/UNSW. Researchers at UNSW will also review and complete data collection in case of missing data.

Our team from RHW will supervise and guarantee the quality of the data. Each participating centre will have a site researcher, a fetal cardiologist who will discuss and sign consent form, collect ultrasound images and relevant patients' data. The data collected will be uploaded and securely stored onto the UNSW platform.

Once ethical approval for each overseas center will be granted, UNSW team will download data from all centers, merge the information in a unique database and analyze them with the help of a statistician. UNSW team will be responsible for images analysis and data analysis. All the participants will then collaborate in drafting manuscripts for publication.

Outcome data will be accessed only by authorised researchers using an encrypted code for data protection. Final drafts for publication will be reviewed by all the authors from each research site.

Dissemination of results and publication policy

Results of the study will be published in peer-reviewed scientific journals, presentations at conferences or other professional forums. In any publication, patient privacy will be protected and presented in a de-identified manner.