Teaching fetal hearts – why is it so difficult?

The first person to perform a fetal sonogram on a pregnant woman is by far the most important observer of potential fetal and newborn problems. Ultrasound examination has the potential to determine abnormal placentation, viability, fetal age, fetal number, amniotic fluid index, fetal movement, and make the mother and the family happy. That is a lot to ask of anyone when there are potentially so many things that can go wrong with the pregnancy, including the fact that 1/4 will end in miscarriage, while realising that more than 95% of all pregnancies scanned at 20 weeks’ gestation will have normal outcomes [1]. Scanning the fetal heart with ultrasound is considered the most difficult form of any ultrasound endeavour: major anatomic and functional anomalies can be missed [2]. Nonetheless, we recommend that a basic evaluation of the fetal heart using a simple protocol be performed at every prenatal ultrasound scan, including follow-up and indicated growth studies.
The referral indication for fetal echocardiography with the highest yield of true congenital heart disease (CHD) in our database and in other published articles is that of a questionable or abnormal fetal anatomy scan. Abnormal 4-chamber views account for the majority of fetal CHDs detected [3]. However, the most missed versions of CHD on fetal anatomy studies are conotruncal malformations including tetralogy, truncus, and transposition variants and great vessel anomalies including aortic and pulmonary atresias, coarctations, vascular rings, and interrupted aortic arch. Small ventricular septal defects (VSDs) and mild semilunar valve stenoses that we scan because of heart murmurs in well baby units are almost never seen in fetal echocardiograms [4].
My first boss in this business of perinatal cardiology, Bill Rashkind, often told me to go figure it out, explain what you found, then do it, then teach it. That was his version of the see one – do one – teach one rule. We got to do a lot of innovative stuff in the animal labs and of course in the cath lab. I had worked a bit as an undergraduate in the basement of the med school where some of the first ultrasound transducers were being made, so he assigned me to get the new A-mode ultrasound devise, which we attached to the cath lab recorder for m-mode recordings. If we did find something unusual with the heart in a child with ultrasound, Dr. Rashkind would say, “Let’s prove it in the cath lab.” I believe he would now say, “Let’s go fix it in the cath lab!” We became an alpha site for several companies to test the early wobbler and rotational mechanical 2-dimensional transducers in the early 1970s, which we used in Sidney Friedman’s congenital heart disease clinics.
Ultrasound for pregnancies was becoming popular in the late 1970s, and so we were asked by the perinatologists (maternal-fetal) to check fetal hearts that had arrhythmias by auscultation, and what we could see with our fancier machines like hydrops formation. This led us to a research project in which we invited several thousand pregnant doctors, nurses, and other hospital workers as well as some mothers of our paediatric patients to allow us to evaluate the fetal hearts and cardiovascular systems in order to establish our own nomograms for a cross-sectional matrix of anatomy and physiology from late first trimester through term. One of the first volunteers had a fetus with a VSD and another had what looked to us like a tetralogy of Fallot. We added to our database and modified it as the imaging equipment got better.
As we reached the mid 1980s we were doing so many fetal heart studies that we had to move them to before and after normal hours as they took longer than the routine paediatric echo exams (and they were still free research-oriented studies). We presented our findings in Circulation in 1984 and at the Second World Congress of Paediatric Cardiology in1985. That same year, my next boss, Bill Norwood, asked why none of the babies requiring early surgery, including those with single ventricle, complete atrioventricular (AV) canal, and d-transposition, had not been identified in utero as so many of the mothers had had obstetric ultrasound. He sent me out to the major ob/gyn groups in the greater Philadelphia area to find out why not and to teach a basic form of fetal cardiac evaluation. Our findings and answers from the sonologists and sonographers were as follows:

1. Pregnancy exams were mostly for viability and to measure head circumference and biparietal dimensions, femur length, and abdominal circumference to establish the due date and number of gestations.
2. Most studies took about 15 minutes of scanning time.
3. Half that time was spent looking for the sex of the fetus.
4. The other half was showing baby hands and feet to the mother and her attendants.
5. The heart moves too fast to say anything about it other than “it seems to be working”.
6. “I can never find both of the outflow tracts of the heart” (even from those sonographers trained in adult cardiology).
7. The mother was obese (now one of the most common reasons for fetal echo).
8. There were twins or triplets, so the fetuses not in front could not be seen clearly.
9. They want us to make a faster throughput for our cases.
10. The obstetricians and/or radiologists do not want us to say anything about the heart.

The only fetal cardiovascular standards for obstetrical ultrasound in the mid 1980s was for “positive cardiac activity”. The most ruled out fetal anomaly was for trisomy 21, which was done using ultrasound for amniocentesis (1/732 pregnancies). We implemented a program, along with maternal fetal practitioners, called “The 2-minute fetal heart study” (Figure 1). This included 6 standardised images of the fetal heart in the chest in order to prove normal 4-chamber views and outflow and arch views in transverse and sagittal planes. The heart was to be looked at for 2 minutes and, if possible, sweeps recorded as short videos, as well as obtaining m-modes of the heart rate. The purpose was to convince the sonographer of their confidence in declaring that the fetal heart’s position, size, function, and rhythm were normal. The right and left sided chamber, and valve and vessel sizes had to be approximately balanced or equal in proportion to be considered normal, while understanding that the right-sided structures might be a little larger than the left-sided structures. These lectures and courses always came with handouts, the most important of which were checklists and worksheets, which are still recommended for both screening and targeted heart studies after 45 years. We taught that if the sonographers performed 100 fetal ultrasound studies, they would find at least 4 cases of abnormalities of the heart, such as being in the wrong place in the chest, the heart taking more than 1/3 of the thoracic space in the 4-chamber view, the heart “rocked” in the chest rather than seeing the ventricles squeeze equally, and of course an unusual rhythm, which is still missed by experienced sonographers but not by auscultation at the non-stress test (NST). The bit about disproportion may be the hardest to teach. When teaching one person or 1000 people, we try to give visual reminders such as that the normal right ventricle (RV) and right ventricular outflow tract (RVOT) spirally cross the left ventricle (LV) and left ventricular outflow tract (LVOT), like when you clap your hands with the left hand on top. Try that right now! Your fingers are the outflow tracts. My favourite bit to teach is the direction of the outflow tracts: put you left hand on your right shoulder to represent the LVOT then your right hand on your left shoulder to represent the RVOT and show the normal crisscrossing. If you see parallel great vessels, especially in the sagittal views, you have transposition of the great arteries or maybe double outlet right ventricle (but remember that in a 4CV to 3VV transverse sweep you can also see only one great vessel at a time in d-transposition). If you can only find one great vessel, you have either pulmonary atresia or aortic atresia or truncus arteriosus. Then look at the colour flow during your 4CV to 3VV sweeps to make sure the flow directions are appropriate and not reversed in either the ductal arch or the aortic arch.
In the second half of the 1980s we published our findings that fetuses with surgical congenital heart disease did better when they had fetal echo diagnoses with early family and team preparation. We were asked by several of the local MFM, Ob/Gyn, and radiology organisations to do lectures and conferences on fetal heart in our tri-state area. We worked with the MFM ultrasound mavens of Drs. Bolognese, Librizzi, and Weiner and participated in their regular teaching conferences and weekend seminars. I must mention the help of Victoria Lee Vetter on fetal arrhythmias, Paul Weinberg on the anatomy of CHD, and Henry Wagner on physiology and newborn transitional changes from fetal life. Later, Jim Huhta developed a research group with young but now international stars like Saemundur Gudmundsson, Gerald Tulzer, and Maria Respondek, who not only helped with the teaching but introduced us to the many Europeans such as Torvid Kiserud and Lindsey Allen, who were quite a bit ahead of us in finding congenital heart disease in fetuses. We shifted over to a Perinatal Cardiology format aligned with maternal fetal medicine and did some good research with Juha Rasanen, again with pregnant volunteers, for fetal lung physiology. We started a quarterly national course at the request of Thomas Jefferson University through their internationally famous radiology ultrasound department under Barry Goldberg which continued until COVID-19 closed all courses in 2021. We also established a national course for those about to take the fetal echocardiography exam for ARDMS certification. Our courses were also designed for sonographers performing ultrasound and for MFM fellows and attendees responsible for ultrasound interpretation. We now give international webinars for sites that send cases for discussion. Most of this teaching was and continues to be pro bono.
OK, so why is fetal heart disease still missed? From my experience, missing diagnoses that have become devastating will lead to a closer look by the sonographer in the future. However, the sonographer may not get that feedback information or have enough experience. I would postulate the following reason for missing CHD:

1. The usual: maternal obesity, fetal lie, polyhydramnios and anhydramnios, gestational age – too early first trimester or too late third trimester, poor equipment or transducer selection, and improper system tuning [5].
2. Not following a checklist or worksheet for each anatomy study, especially after the sixth study of the day (see attached examples).
3. Sonographers may pay more attention to the fetal heart in pregnancies when there was prior CHD, but the sonographer may not have that information prior to the scan.
4. Multiples are difficult to capture all the suggested images in each fetus, although we are getting much better, especially in monochorionic pregnancies, at expecting twin to twin transfusion syndrome and TRAP but not at diagnosing twin anaemia-polycythaemia sequence (TAPS) [6].
5. Developmental structural CHD: Aortic Stenosis and Pulmonary Stenosis can look normal in the first and second trimester. Ebstein’s Anomaly of the tricuspid valve and tricuspid valve dysplasia can appear to be normal in an 18- to 22-week gestational study if colour Doppler is not used with the proper tuning [7].
6. Hydrops fetalis (congestive heart failure) from supraventricular tachycardia including atrial flutter almost always has a normal anatomy scan in the second trimester and may only be found after the mom complains of lack of fetal movement in the third trimester.
7. Volume-loaded fetal congestive heart failure from AV malformation or fistulae such as hepatic AV connections or extra hepatic absent ductus venosus, sacrococcygeal teratoma, or vein of Galen aneurysm are usually not identified until after 28 weeks’ gestation or after the canalicular stage of lung development.
8. Volume loaded fetal congestive heart failure from anaemia caused by parvovirus b19 or auto immunity is often missed if it occurs after 22 weeks’ gestation.
9. Pressure-loaded fetal heart from increased placental resistance due to maternal hypertension or preeclampsia may lead to fetal growth restriction and eventual heart failure after a normal anatomy scan.
10. Rhabdomyomas are rarely seen before 24 weeks’ gestation. If identified in the early second trimester, fetuses may succumb to fatal arrhythmias [8].
11. Late-onset diabetes can be missed during pregnancy, and therefore fetal diabetic cardiomyopathy. Remember that a large portion of still-born babies are from undiagnosed diabetic cardiomyopathy.
12. Mass effects on the heart such as diaphragmatic hernia or pulmonary CCAM or sequestration or bronchopulmonary cysts may lead to underdevelopment of the left side of the heart.
13. Very active fetuses (caused by maternal overuse of cardiac stimulants such as coffee, tea, or soda) often develop a redundancy of the septum primum, causing a windsock effect returning some ductus venosus flow back to the right atrium thus causing right sided structures greater than left-sided structure and a possible aortic isthmus narrowing (R > L disproportion).
14. If heart issues are not recognised, the sonographer does not dwell on additional imaging that they might get, for example, of a dilated cavum septum pellucidum. An absent CSP occurs 2.5/100,000 fetuses whereas CHD including arrhythmias and bicuspid aortic and pulmonary valves occurs in approximately 8/100 full term newborns [9].
15. Not enough ultrasound of the fetal heart is taught at ultrasound schools, and many sonographers and sonologists are undertrained and overworked.

So, what can we do about it? Education must be ongoing, but how do we get it and how can it be provided? I would begin by recommending that all pregnant women receive basic fetal cardiovascular anatomy studies at the their routine anatomy studies in the third trimester after 29 weeks’ gestation [10]. Again, I will postulate the following:

• Teach the following image adjustments: Auto-Opt with ZST – Gain/DGC: 2D Gain/DGC; Display Zoom (pan/zoom on frozen image up to 4× mag); Greyscale Map; B Mode Tints; Dynamic Range; Persistence; Edge Enhancement/Smoothing; Up/Down Invert; Left/Right Invert; Frame rate; Power; 2D Gain/DGC; Contrast Gain; Depth; Frequency; Display Zoom (pan/zoom on frozen image up to 4× mag); Dynamic Range; Edge Enhancement/Smoothing; Persistence; Maps; L/R Invert; U/D Invert; Auto-Opt with ZST – Gain/DGC; Tints; Compounding; Harmonics.
• Annotation and arrows.
• Teach with a model of the heart.

1. Give feedback immediately to cases referred for potential anomalies, especially structural and functional CHD, and insure that the performer of the first study gets the feedback. Do it in a careful “this is what we believe we learned” format.
2. Be willing to share what you have found to be anomalous: for the fetal heart, actively teach and support the acquisition of images that will prove normalcy, and take a shot at making CHD diagnoses when the sonographers find something wrong. Stress that anything even suggesting an abnormality should prompt a dedicated or targeted study.
3. Establish an on-going case review for all sonographers and sonologists at least bi-weekly.
4. Take advantage of on-line information, including free webinars. My website, www. perinatalechoimaging.com has over 100 international seminars on congenital heart disease as well as many published papers and videos.
5. Build a library of standard ultrasound techniques and great images for your team and continually up-date it.
6. Set up a weekend course where competent speakers can teach how to perform both screening and targeted fetal heart studies as well as technical diagnoses and outcomes, but dwell on the hands-on training.
7. Encourage the building of a database of interesting cases.
8. Encourage the building of a video library.
9. Encourage research into anatomy, physiology, and outcomes as a team effort.
10. Encourage the use of video clips for sweeps of the entire fetal anatomy, not just the heart. I believe that this will encourage sonographers and sonologist to get better images and make better diagnoses.

Finally, there is a new wind blowing that may help practitioners of fetal ultrasound. Many of the newer pieces of equipment have forms of computer learning modules that are the forefront of artificial intelligence (AI), including generative heart imaging. AI prompts on some of the newer ultrasound machines will inform the user which views should be recorded with drawings or pre-recorded images of the proper views of the heart. Some of these AI programs may guide the newer practitioner to produce protocol imagery. I hope these programs will not replace any part of the team effort we have established with paediatric cardiologists and surgeons and high-risk maternal fetal medicine specialists and sonographers. Make sure that the AI does not to give automatic readings of any abnormal 4-chamber view but reasonably offers differentials for diagnoses.

A protocol for cardiovascular imaging during the fetal echocardiogram (Figure 2)

The basic fetal heart screening ultrasound during the routine anatomy scan provides a set of data that leads to confidence in stating that the heart is within normal limits for specific parameters. Therefore, standard images and video clips of observed real-time ultrasound sweeps through the abdomen and thorax from all available planes in any fetal lie should be obtained to begin the study prior to making measurements required for the fetal echocardiogram. Use a worksheet or a checklist for every study performed. The fetal echocardiogram is designed to identify abnormalities and diagnose CHD and congestive heart failure (CHF). Remember that as many forms of CHD and CHF are progressive over gestation; all follow-up studies should be comprehensive and complete heart studies. Likewise, you cannot take too many video clips or still images.
Any abnormal finding that may lead to fetal congestive heart failure or in utero treatment or require newborn surgery requires follow-up with a perinatal or paediatric cardiologist.

1. Obtain a video of the lower uterine segment with colour through sweeping caudad to cephalad in both transverse and sagittal planes. This is to establish fetal lie, general anatomy, placental position, and relative amniotic fluid volume.
2. Abdominal to thoracic sweep to establish situs. Can be side by side still frames of stomach and heart positions. The stomach should be seen on the left side of the fetus and just below the heart and diaphragm. The abdominal aorta should be on the left of the spine and the inferior vena cava on the right of the spine anterior to the level of the aorta. The still images show the stomach and then the heart on the same side with identification of the right and left sides of the fetus.
3. Slow sweeps through the lower abdomen to include the bladder, umbilical vessels, cord insertion site, intestines, liver, stomach, kidneys, and diaphragm. Repeat with colour Doppler to see both IAUAs, IAUV, portal system, DV and hepatic veins, and IVC. Repeat in sagittal. The IVC should be on the right of the spine, and the abdominal and thoracic aorta should be on the left.
4. Take several clips of the 4VC with and without colour in the transverse views. Twist the transducer or move to get the 4CV with the IVS perpendicular to the ultrasound beam. Repeat with colour Doppler to prove no obvious VSD and normal RA to LA blood flow. Capture the pulmonary veins in colour, using a low-frequency scale with a small colour box.
5. Sweep cephalad to obtain views of the LVOT and RVOT. Prove with colour Doppler that there is normal crisscrossing of the flow patterns and no aliasing into the ascending aorta or main pulmonary artery of the colour using a high-velocity scale. Repeat these sweeps in the sagittal view to show the ductal and aortic arches. Colour Doppler is used to show vessel and chamber size and direction of blood flow.
6. In the transverse view, show the 3VV + trachea with and without colour Doppler. There should be approximately equal in size and colour flow in the MPA and As Ao. The “V” view shows the descending aorta on the left, whereas a “U” view shows a right descending aorta with a possible vascular ring.

All these sweeps of the fetal heart require the observation of potential changes from the normal heart in position, size, function, and rhythm without chamber or vessel size disproportion. Always narrow the window, place the thorax in the centre that of that magnified image, then use a narrower than taller high-resolution box to get the optimal image filling the screen. Use harmonics and adjust overall gain. Measurements for comprehensive fetal cardiovascular ultrasound:

1. Standard cardiac dimensions to include: CC/TC ratio, Cardiac IVS to Sternum to spine angle, and the annuli of all 4 valves at their hinge points. Dimensions of VSDs, ASDs, and oddities should also be included.
2. M-mode for shortening fractions that evaluate LV function, and the myocardial thicknesses measure in diastole and systole. An isolated thickened IVS will suggest diabetic cardiomyopathy, although either or both ventricular free walls may also be involved. If the RV wall and IVS thickening suggests RVOT obstruction while IVS and LV wall thickening suggests LVOT obstruction, including coarctation. M-mode through the atrial and ventricular myocardial walls or through a semilunar valve and atrial wall can be used to show AV conduction, rate, and ectopy or type of tachycardia or type of bradycardia.
3. The Myocardial Performance (Tei) Index is performed to detect potential heart failure or rhythm disturbances from normal. The PR interval as well as the A period, B period, and isovolumetric and isovolumetric relaxation period and ejection times can be measured in any fetus at any age. Also measured with this acquisition are the E and A waves of the mitral and D pulsed Doppler flow patterns and maximum velocities immediately above the aortic and pulmonic annuli. The semilunar flow patterns with the vessel diameters and heart rate are the determinants of either the right or left cardiac outputs. Tissue pulsed Doppler may be performed at the outer annuli of the AV valves and at the IVS just below the central semilunar valve. Some laboratories will include ventricular strain measurements in cases of pressure and volume loading.
4. Standard vascular pulsed Doppler with colour Doppler measurements are taken as close to a 0 angle of insonation as possible. These include the FLUA, FLUV, IAUV, DV, DA, and MCA. Any regurgitant or reversed flow patterns should be Doppler imaged and recorded. Sweep #1 of the abdominal cross-sectional view through the thorax to the fetal head to establish situs should be repeated with colour Doppler to ascertain any vascular abnormalities.

The atria should be approximately equal in size. The flap of the foramen ovale (septum primum) should move into the left atrial chamber [11]. The ventricles should be approximately equal in size with the tricuspid valve attaching to the septum slightly apical relative to the attachment of the mitral valve. The right ventricular apex should contain the moderator band. Both ventricular chambers should reach the apex of the heart and should squeeze equally. The still image shows the 4 chambers of the heart as about 1/3 the area of the chest with the cardiac axis being approximately 45° to the left of the sternal spine axis [12].
Sweeping toward the fetal neck, the aortic valve and ascending aorta are seen arising centrally from the 4-chamber view with the ascending aorta aiming toward the right shoulder. Pulmonary veins should be seen in this view as attached to the posterior wall of the left atrium. The still images show the LVOT and the pulmonary veins.
Continuing the sweep toward the fetal neck; the pulmonic valve and the main pulmonary artery are seen arising at a 90° angle anterior and to the left of the ascending aorta and aiming toward the right shoulder. The bifurcation of the pulmonary artery and ductus arteriosus completes this view. The still image shows the RVOT with the MPA leading to the bifurcation of the RPA and ductus arteriosus.
Continuing the sweep toward the fetal neck with a medial twist of the transducer, the short-axis views of the main pulmonary artery, the ascending aorta, and the superior vena cava are seen with their relationships to the trachea and oesophagus. The main pulmonary artery diameter is slightly greater than the aorta, which is slightly greater than the superior vena cava. The still images show the cross sections of the MPA, ascending aorta, and SVC and the convergent views of the transverse aortic arch and aortic isthmus meeting the ductus arteriosus at the left sided descending aorta. This convergent view should show equal-sized isthmus and ductus joining in a “V” formation. A “U” formation with the trachea in the middle indicates a right aortic arch and right descending aorta with a vascular ring.
This sweep shows the superior vena cava, right atrium, and inferior vena cava on the right and the aortic arch centrally and the ductal arch to the left. Both the superior vena cava (SVC) and the inferior vena cava (IVC) drain from a posterior position to a medial connection to the right atrium. The hepatic veins ascend from an anterior position connecting to the right atrial (RA) junction with the IVC. The 3 sagittal images show the SVC and IVC, the aortic arch with brachiocephalic vessels, and the ductal arch connecting to the descending aorta.
Taken as a whole, these images show normalcy in the following fashion:

1. Position: The axis of the fetal heart, made up of the interventricular and interatrial septa, should be seen in coming from the centre of the chest into the left hemithorax at a 45° angle to the sternum to spine axis in a transthoracic 4-chamber view.
2. Size: In the transthoracic 4-chamber view, the heart should occupy approximately 1/3 of the chest or be seen so that 3 hearts could fit into the thorax.
3. Function: In the transthoracic 4-chamber view, both ventricles should squeeze vigorously and equally toward the ventricular septum, such that the heart does not appear to “rock” within the chest.
4. Rhythm: While observing the heart, there should be a relatively steady rate (average 142 bpm) within the range of 120-180 beats per minute without early beats, blocked beats, or periods of bradycardia or tachycardia.
5. Proportion: pairs of cardiac chamber structures should be seen as approximately the same size, right to left, including the atria, ventricles, AV valves, semilunar valves, great arteries, and the SVC and IVC.

Disclosures

Institutional review board statement: None.
Financial support and sponsorship: None.
The author declares no conflict of interest.

References