3-D Ultrasound for In Vivo Imaging

Ultrasound stands out for its ability to provide detailed structural information from the first trimester, coupled with its affordability, widespread availability and real-time examination capabilities. While traditional 2-D ultrasound is commonly used, there is a growing shift toward 3-D ultrasound for anomaly evaluation, volumetric measurements, and surface anatomy visualization, offering the advantage of revisiting stored 3-D volumes. Recent advancements in 3-D-ultrasound technology, including innovative visualization software, have greatly enhanced the clarity and lifelikeness of images, aiding clinicians, students and parents [Reference Shah, Al-Memar, de Bakker, Fourie, Lees and Bourne3]. Additionally, these advancements include see-through 3-D rendering applications that enhance the contrast of internal structures, further improving the utility and accuracy of ultrasound in early pregnancy [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2, Reference Shah, Al-Memar, de Bakker, Fourie, Lees and Bourne3].

Micro-CT Imaging for Ex Vivo Imaging

Microfocus computed tomography (micro-CT), an evolving tool in the biomedical field, offers high-resolution imaging without material disruption, making it useful in various nonmedical fields like precision engineering and geosciences [Reference Docter, Dawood, Jacobs, Hagoort, Oostra and van den Hoff4]. Similar to conventional CT in its use of X-ray attenuation, micro-CT typically involves a fixed radiation source and a rotating sample platform. The application of iodine-based staining, known as diceCT (diffusible iodine-based contrast-enhanced CT), enhances soft-tissue imaging and is increasingly used for studying ex vivo human embryonic and fetal anatomy, having been widely employed in animal research [Reference Dawood, Hagoort, Siadari, Ruijter, Gunst and Lobe5]. The versatility of diceCT is a key advantage, effective across various developmental stages, from 6 to 24 weeks of gestation [Reference Simcock, Shelmerdine, Hutchinson, Sebire and Arthurs6, Reference Shelmerdine, Simcock, Hutchinson, Guy, Ashworth and Seibre7, Reference Lamouroux, Cardoso, Bottero, Gallo, Duraes and Salerno8]. This method provides detailed anatomical insights and has been adopted by multiple research groups as a virtual autopsy tool, particularly beneficial for parents who decline invasive autopsy following fetal loss [Reference Kang, Cos, Guizani, Cannie, Segers and Jani9]. It has been evaluated in prospective studies comparing it to conventional autopsy for gestations up to 20 weeks [Reference Hutchinson, Kang, Shelmerdine, Segers, Lombardi and Cannie10]. Additionally, research demonstrates the feasibility of studying human fetal anatomy as early as eight weeks’ gestation, facilitating virtual autopsies in the early first trimester [Reference Lamouroux, Cardoso, Bottero, Gallo, Duraes and Salerno8]. By obtaining precise phenotypic data, especially from miscarriage specimens, this approach supports reverse phenotyping, enabling the selection of variants of interest in genome-wide analyses [Reference Lamouroux, Cardoso, Bottero, Gallo, Duraes and Salerno8]. This offers the potential for genetic counseling to bereaved parents, adding a layer of support during a difficult time. Additionally, its minimally invasive nature allows for its use on rare or invaluable samples, including those with unique abnormalities or museum specimens [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2].

The Use of AI in Ultrasound

Artificial intelligence (AI), particularly deep learning (DL), has made significant strides in the last decade, showing great promise in medical fields that rely heavily on imaging. Artificial intelligence’s application in fetal imaging has been the subject of recent reviews and various studies, focusing on uses like fetal diagnosis and estimating gestational age. Most of these studies emphasize implementing DL in 2-D ultrasound, such as automatic detection of 2-D planes, probe-motion tracking and enhancing biometric measurements such as head circumference [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2].

Artificial intelligence is expected to further enhance embryonic and fetal imaging, particularly in complex classifications and diagnosing congenital defects. We anticipate that AI, trained on models like those from the 3D Embryo Atlas, could assist sonographers in analyzing 2-D and 3-D fetal images. This would involve comparing real-time scans with existing models to identify deviations in fetal anatomy, potentially indicating birth defects. The integration of AI in prenatal imaging promises to improve the detection of congenital abnormalities, leading to better prenatal diagnosis and enhanced postnatal care [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2].

Synchrotron Imaging

Synchrotron radiation-based X-ray phase-contrast tomography (sCT) is emerging as a promising technique for ex vivo embryonic and fetal imaging. This method is based on the principle that both amplitude and phase of an X-ray beam are altered as they pass through an object [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2].

The key advantage of sCT lies in its sensitivity to X-ray phase shifts, which provides clear visualization of soft tissues at a micrometer resolution, without the need for contrast agents or staining, positions it as an invaluable tool in biomedical research. This technique offers detailed insights into embryonic and fetal structures, potentially revolutionizing our understanding of early human development [Reference Dawood, Buijtendijk, Shah, Smit, Jacobs and Hagoort2].