The main finding of this study was that the definition of ROI width significantly influence measurements of 4ChLS in the fetal heart late in pregnancy. The reduction in strain by increasing ROI width was consistent through midwall, multilayer and segmental analyses (Tables 1 and 2).
Our findings are similar to previous studies in adults. Spriestersbach et al.  examined 20 healthy subjects, and Stoebe et al.  examined 30 healthy adults and 15 patients with left ventricular dysfunction at different ROI widths, both found strain to decrease with increasing width. When comparing manually adjusted ROI width with default semi-automatic software selected ROI, Otterstad et al.  reported that increasing ROI width lead to lower strain values.
A recently published review by van Oostrum et al.  included 23 studies on STE in fetal hearts. The studies were contradictory concerning development of strain throughout gestation. This was explained by different study design and inconsistent results from varying tracking software and different ultrasound devices. However, another possible explanation for the inconsistencies may be different image acquisition and processing settings. In the present study we show that different ROI widths significantly influences 4ChLS, independent of layer and segment assessed.
At 8 weeks of gestation the structure of the fetal heart is comparable to that of the adult. The contractile fibres of the myocardium are oriented in a helical pattern, in the subendocardium in a right handed helix and in the subepicardium in a left handed helix, whereas the fibres in the midmyocardium are aligned circumferentially . In a complex interplay the myocardial contraction result in a radial thickening and longitudinal shortening with higher strain in the subendocardium compared to the mid- and subepicardium , as well as increased strain moving from the basis to the apex . In this study we confirm this also in the fetal heart. With increased ROI-width we show decreased strain both between layers, segments and in 4ChLS which is explained through involvement of myocardial fibres with less deformation.
Echocardiography during fetal life is challenging compared to the post-natal period. The thickness of the maternal abdominal wall and/or unfavourable fetal position or movements may hamper ultrasound examination of the fetal heart. As the beam must traverse through both maternal and fetal tissue to reach the fetal heart, low-frequency probes must be used, which result in reduced spatial resolution. Considering that the fetal heart by 37 weeks of gestation has a mean thickness of the septum and free wall is by average 3.2 mm , these factors pose challenges in fetal echocardiography and in speckle tracking echocardiography in particular .
In this study, we assessed multi-layer strain. Previous studies have shown its feasibility in older patients in the case of myocardial toxicity , aortic stenosis , and ischemic cardiomyopathy [33, 34]. However, the accuracy of multi-layer strain measurements remains disputed .
Although we see that ROI width affects strain values, it is not possible to determine which ROI settings yield the most accurate strain values as this would require a reference method to serve as a source of comparison. Cardiac MRI could have been an option, but the use in fetuses is still hampered by challenges in spatial and temporal resolution .
The ROI widths narrow and narrow* resulted in different numerical strain values in spite of ROI-widths being the same. Although this difference was not statistically significant (p = 0.425), nor of clinical relevance, it is nonetheless an interesting finding as it highlights that strain not only depends on the ROI-width, but also the initial ROI-width.
The most important strengths of the study were that analyses were carried out by one experienced cardiologist, and with excellent intraobserver variability, and that all the ultrasonographic examinations were performed at the same gestational age (37 weeks of gestation).
There are limitations to this study. We evaluated strain using equipment and software from a single vendor. Inter-vendor discrepancy in strain remains an on-going challenge for STE , although efforts are being made at achieving some form of standardization across vendors . Strain was assessed in the four-chamber view only, not from three apical views, as recommended for adults in the joint consensus paper by the European Association of Cardiovascular Imaging and the American Society of Echocardiography . As previously noted, echocardiographic evaluation of the fetal heart remains challenging, particularly for the 2-chamber and 3-chamber views, due to loss of the endocardial borders of the septal and lateral walls .
In conclusion, this is to the best of our knowledge the first study to examine the effect of different widths of the ROI when assessing fetal strain late in pregnancy. Strain assessment is feasible and highly reproducible in this age group. We found that the setting of ROI width significantly affected STE derived left ventricular longitudinal strain values in fetuses. Although the effect was greatest in the epicardial layer, a reduction in strain was consistent through midwall, multilayer, and segmental strain across the entire myocardium. Fetal cardiologists and obstetricians should be aware of the effect of different ROI widths as a source of variation in strain measurements when assessing strain. A standardization of the ROI setting would strengthen the method as a routine measure of ventricular function in the fetus