Advancements in artificial intelligence (AI) have revolutionized the field of obstetrics and perinatology, introducing groundbreaking tools like SonoCNS™ that aid in the assessment of fetal anatomy and biometry. This cutting-edge software automatically delineates appropriate planes and measures key structures within the fetal central nervous system (CNS), offering a new level of precision and efficiency in prenatal care. In a comprehensive study, researchers have now evaluated the repeatability and reproducibility of these AI-acquired fetal brain measurements, shedding light on the potential and limitations of this transformative technology.
Revolutionizing Fetal Brain Imaging
The SonoCNS™ software, developed by GE Healthcare, is a prime example of the integration of Click Here
</b>, <b>occipitofrontal diameter (OFD)</b>, <b>head circumference (HC)</b>, <b>cisterna magna (CM)</b>, <b>transverse cerebellar diameter (TCD)</b>, and <b>posterior horn of the lateral ventricle diameter (Vp)</b>. This comprehensive assessment of fetal CNS structures can provide valuable insights into the development and potential abnormalities of the fetal brain.</p>
<p><figure class="wp-block-image size-large"><img decoding="async" src="https://famefreq.s3.us-east-2.amazonaws.com/1729765579_41598_2024_77313_Fig2_HTML.png" alt="figure 2" class="wp-image-2" /><figcaption class="wp-element-caption">Fig. 2</figcaption></figure>
</p>
<p><h3>Evaluating the Reliability of AI-Powered Measurements</h3>
<p>In the study, researchers aimed to assess the repeatability and reproducibility of these AI-acquired fetal brain measurements across a diverse population of 381 pregnant women, with 270 in the second trimester and 111 in the third trimester. The researchers compared the manual measurements performed by experienced sonographers with the automated measurements obtained using the SonoCNS™ software, calculating <b>intraclass correlation coefficients (ICCs)</b> to determine the level of agreement.</p>
<p>The results were promising, with the ICC values for BPD, HC, and OFD ranging from “good” to “excellent” reproducibility in both the general population and the second and third trimester subgroups. This suggests that the AI-powered software can reliably measure these key biometric parameters, which are crucial for assessing fetal growth and development.</p>
</p>
<p><h3>Navigating the Challenges of Fluid-Filled Structures</h3>
<p>However, the study also revealed some limitations of the SonoCNS™ software. Measurements of the <b>cisterna magna (CM)</b> and <b>posterior horn of the lateral ventricle (Vp)</b> showed “moderate” to “poor” reproducibility, indicating a need for further refinement in the AI’s ability to accurately delineate and measure these fluid-filled intracranial structures.</p>
<p>The researchers hypothesize that the challenges in measuring these structures may be due to their small size and susceptibility to ultrasonographic artifacts, as well as the software’s reliance on automatically delineated planes that may not always provide clear boundaries for these delicate features. Interestingly, the reproducibility of these measurements was found to be better in the third trimester, potentially due to improved visualization as the fetal bones become more calcified.</p>
</p>
<p><h3>Balancing Efficiency and Precision</h3>
<p>Despite these limitations, the SonoCNS™ software offers a significant advantage in terms of efficiency. The study found that the automated measurements took an average of just 14 seconds, compared to 63 seconds for manual measurements. This time-saving feature can be particularly beneficial for busy clinicians, potentially reducing examination times and allowing for more thorough assessments.</p>
<p>Moreover, the software’s ability to delineate the appropriate planes for the fetal CNS, as recommended by the International Society of Ultrasound in Obstetrics and Gynecology, can be a valuable tool for less experienced sonographers. By guiding them through the process, the software can help shorten the learning curve and improve the consistency of fetal brain examinations.</p>
<p><h3>Embracing the Future of Fetal Imaging</h3>
<p>The findings of this study highlight the tremendous potential of AI-powered tools like SonoCNS™ in the field of obstetrics and perinatology. While further refinements may be needed to improve the reliability of measurements for specific intracranial structures, the overall performance of the software in assessing fetal biometry is highly promising.</p>
<p>As AI continues to advance, we can expect to see even more sophisticated tools that can not only perform precise measurements but also aid in the detection and diagnosis of fetal abnormalities in real-time. This integration of AI with medical imaging holds the promise of enhancing prenatal care, improving outcomes for both mothers and their unborn children.</p>
<p>Author credit: This article is based on research by J. Mlodawski, A. Zmelonek-Znamirowska, M. Mlodawska, K. Detka, K. Białek, G. Swiercz.</p>
<hr />
<p>For More Related Articles <a href=){kind=link}