SonoCNS™ appears to be a useful tool for assessing the fetal head and central nervous system structures. The results of our study indicate high reliability of the tool in terms of biometric measurements commonly used in most models for estimating fetal mass and size. However, measurements of Vp and CM were less favorable, with some cases not significantly different from 0.5, indicating poor performance. A characteristic feature of Vp and CM measurements is that they are fluid spaces whose boundaries in the ultrasound image are demarcated by fetal CNS structures. Also, both structures are not located in the basal plane (thalamus) identified by the operator, but in planes automatically delineated by the control system. During image analysis, it may become apparent that despite good visibility of intracranial structures in the thalamic plane, the system’s delineation of planes based on the acquired 3D volume in other planes does not provide clear boundaries for Vp and CM due to ultrasound artifacts. The positioning of these structures also makes them susceptible to obscuration by the acoustic shadow of the fetal skull bones. Both structures are also relatively small, typically up to 10 mm in normal cases. In contrast, other structures with better reproducibility and repeatability have larger sizes and are usually located in the peripheral parts of the fetal head.
In the literature, there have been few studies evaluating SonoCNS™ in terms of performance. The largest of these, published in 2021, involved 143 patients screened mid-trimester6. This study compared intraobserver variability in manual measurements and measurements using SonoCNS™. The results of that study differed significantly from ours, with only the HC and BPD parameters in the published study having an ICC greater than 0.8. In other cases the ICC was in the poor range (< 0.5). The cited study used a different protocol, involved more researchers (eight physicians), and was conducted in the United States with a study group characterized by a slightly higher BMI. These factors may contribute to the differences obtained; however, the trends in the results are similar to those in our work, with the lowest ICC achieved for measurements of CM, Vp and TCD.
In another study published by Gembicki et al9., the mean error for parameters ranged from 1.26 mm (standard deviation (SD) = 1.6) for BPD to 0.87 (4.22) mm for HC, and 0.55 (0.82), 0, 16 (0.82) for 0.16 (1.34) for TCD, and 0.13 (0.67) for CM. However, it is important to remember that such presentation of results does not reflect the relative (percentage) error in individual cases, and that measurements of TCD, CM and Vp have significantly smaller absolute values than HC and BPD. This makes these results difficult to compare with the findings from our research.
Another study compared the detection rate of CNS structures between the 18th and 34th week of pregnancy, which was 75%, and for pregnancies before the 28th week it was 85%. The differences in the percentage of manual and automatic identification were also greatest for intracranial structures10. These variations highlight the need for careful consideration and contextual understanding when implementing SonoCNS™ in clinical practice, especially for specific patient structures and demographics.
Accurate measurement repeatability is not equally important for all structures measured during a fetal head examination with SonoCNS™. The accuracy of the measurement seems to be more important for dimensions that have a strong correlation with the estimated fetal weight and are part of the algorithms used for its assessment (HC, BPD, TAD). Lower repeatability may be acceptable for structures where the absolute measurement value is not critical, but rather whether it falls below the deviation threshold (e.g. Vp < 10 mm). In this situation, the more important factor is whether automatic measurements result in false-positive findings within the obtained measurements. In our study, the percentage of such measurements was minimal: 0.2% for Vp and 1.8% for CM.
One problem that arises during the acquisition of ultrasound images are instances where the system fails to delineate specific measurements, which may be due to blurring of the boundaries of structures of interest. In our work we calculated the percentage of such situations. We observed that in some cases this was the result of an incorrect plane angle in cases of automatic delineation of the transthalamic plane of the fetal central nervous system. Such situations occurred under difficult research conditions where obtaining the original, perfect thalamic projection required for volume acquisition was challenging or partially obscured, requiring relocation of the probe to another bony window to visualize all structures of a given plane . Such situations may include the placement of the fetal head in the small pelvis or an acoustic shadow of a fetal limb, causing the projection obtained by the abdomen to be slightly oblique. The amount of adipose tissue can also influence repeatability6. We also observed that the repeatability of some fluid space measurements, such as Vp and CM, is greater in the third trimester of pregnancy. Intuitively, we expected the opposite result. We expected that the increased calcification of bones in the third trimester would result in acoustic shadows that would reduce the quality of imaging planes. However, repeatability was higher, probably due to better delineation of the above structures. We believe that bone calcification can still impact situations where automatic measurements of structures are impossible; 28 of 42 (66%) failures of automatic CM measurement occurred in the third trimester.
The range of examination time for automatic measurements was low and depended only on the time to acquire the thalamic projection and set the volume acquisition gate. For manual measurements, this time was more varied due to the need to visualize three planes: transventricular, transthalamic, and transcereberral. Appropriate measurements had to be made on each plane obtained. Our study included only fetuses with normal intracranial anatomy; we did not test the system in cases where an anomaly was identified. However, the system may be useful in patients in whom abnormal central nervous system anatomy is associated with abnormal dimensions of structures, such as microcephaly, ventriculomegaly, or Dandy-Walker malformation. However, due to the low prevalence of some of these complications, it can be challenging to verify the performance of the system in practice.
Despite the need for improvements in measuring intracranial structures, it is important to recognize a significant benefit of SonoCNS™ software beyond biometric measurements. The system delineates all planes of the fetal central nervous system, recommended in the 20 + 2 methodology by the International Society of Ultrasound in Obstetrics and Gynecology11,12 and additionally outlines the fetal profile, the assessment of which is also recommended during examination of fetal anatomy. This allows the assessment of suitable structures in all areas and also the profile. Under good examination conditions, the corpus callosum and central nervous system structures of the fetus can also be visualized in this plane. This may be useful for novice sonographers who have difficulty visualizing this plane due to the need for proper probe rotation. Therefore, this tool may be particularly useful in the hands of sonographers with appropriate theoretical background but lower technical skills, potentially shortening the learning curve.
Undoubtedly, AI-based systems are the future of fetal screening studies. Current literature indicates that such algorithms can locate suitable planes and classify sections as abnormal, even in real time13. Modern real-time software is also able to recognize standard levels, store them in the device’s memory and check whether they meet standard diagnostic levels (e.g. SonoLyst, Voluson, GE)14. Such algorithms are also being developed for diagnostics in the first trimester of pregnancy15where AI is used for automatic measurements of non-hereditary markers of chromosomal abnormalities in the first trimester (such as nuchal translucency and the presence of the nasal bone). Descriptions of such software exist in the literature16. Software capable of identifying all necessary structures of the fetal head between 10 and 14 weeks of gestation, such as the thalami, midbrain, palate, 4th ventricle, cisterna magna, nuchal translucency (NT), nasal tip, nasal skin and the nasal bone, is also available. described17. Based on brain images, software can also estimate gestational age with high accuracy18,19. Recognize facial expressions of the fetus, such as blinking, mouthing, a face without any expression, frowning and yawning20. Although SonoCNS™ is available on devices from one manufacturer, other ultrasound device manufacturers also offer software that can perform similar functions, such as 5DCNS+™ (Samsung Medison, Seoul, Republic of Korea), Smart Planes CNS™ (Mindray, Shenzhen, China) .
The limitation of our study is the lack of analysis taking into account possible confounding factors such as subcutaneous fat thickness and fetal positioning. Our goal was to reflect the general population without overcomplicating the results. The examination is also limited to specific periods during pregnancy, but this is consistent with the recommended times for screening studies of fetal anatomy. The software also allows for marker adjustment before measurements are confirmed or for manual marking on areas defined by the system when automatic measurements are not taking place; However, in our research protocol we allowed the software to run without external control. Furthermore, patients were recruited into the study only if the participating operator was available in the prenatal testing laboratory. Nevertheless, we believe that the study population represents the general population attending screening examinations.
