Transforming Ultrasound Technologies
Recent research has determined that advanced AI systems tailored for ultrasound image interpretation are capable of significantly improving the identification of high-risk pregnancies. Prenatal care is now more effective than ever because the AI systems are able to detect up to 35% more complicated cases as compared to healthcare professionals who do not use AI technology to assist them. This improvement for identifying high-risk pregnancies can prevent tremendously damaging results like premature births or other severe birth complications.
Prenatal Emerges from University Collaboration
The results Prenatal claims are possible come from the company’s innovative technologies, which are a spin-off from a collaboration project with the Technical University of Denmark (DTU) and the University of Copenhagen. The interdisciplinary team consisting of veterinarians, computer scientists, and physicians supported each other under the Cooperative Technical University Hospital of Greater Copenhagen (TUH). This team has developed groundbreaking innovations in women’s healthcare.
Extracting Concealed Information from Ultrasound Pictures
The strength of this AI technology lies with processing the ultrasound images because they contain the volume of data that is not easily perceptible to the human eye. This is a vital aspect that Professor Aasa Feragen from DTU, also a co-founder of Prenaital, elaborates on. “Ultrasound images contain large amounts of data that the human eye cannot detect but which can be used to identify high-risk pregnancies,” she articulates while stressing the overwhelming information concealed beneath the visual data.
Determining Significant Fetal Elements
Professor Feragen goes on to describe more of the types of data the AI models can work with. Such as the fine features of a fetus’s brain, measurement of the fat percentage, and intricate matrix of tissues. These factors, according to Feragen, stand a chance of being in the risk-setting pathway prognosis during a scan and are possible defining factors on how a fetus might progress and the risks associated with the development.
Exceeding Boundaries of Conventional Assessment Techniques
In contrast, the fetal growth evaluation in practice today is assessed via relatively rudimentary image-based ultrasound measurements of the head circumference, abdominal circumference, and femur length. While these measurements do provide some data on growth, Professor Feragen points out that the AI model has much greater potential. “It can utilize all the information in the image,” she implies, which suggests a far more accurate evaluation than just a few parameters would assess AI’s potential.
Research and Development Stage and Product Portfolio
As of now, Prenatal’s AI models are in a critical phase of development, focusing on the capabilities of the models prior to routine clinical use. The company’s preliminary offering centered on ultrasound examination quality assurance AI and is currently undergoing a market access regulatory process. This phase is particularly important to ascertain that the technology used is safe and effective within the required thresholds for medical devices.
First AI Model Expected Soon
Prenatal expects to commercially offer its first specific AI model, designed for growth scanning, by 2026. It was the technology capable of detecting up to 35% of all fetuses potentially at risk of abnormal growth. Its launch is anticipated to empower practitioners to identify growth-related problems immensely earlier than possible today.
Collaboration with Healthcare Professionals
The entire workforce of frontline healthcare practitioners gave impetus to refine these models. Rigshospitalet’s sonographers, midwives, and doctors contributed actively by defining what their daily routines’ burdens and gaps needed to be solved most urgently with technology. This participation ensured that the AI technology in development attained sophistication but also tackled pertinent clinical problems.
Training on Comprehensive Danish Data Sets
In order to develop these advanced AI models, an extensive dataset was necessary. At DTU, the AI algorithms were meticulously trained using a dataset consisting of more than 10,000 images obtained from ultrasound scans conducted at different hospitals across Denmark. This dataset, along with other data, is essential in the development of AI models that are trustworthy and performant given the variability present in clinical ultrasound data.
Dramatic Increase in Scan Utility
The expectation is that implementing this AI technology will considerably increase the functionality of ultrasound scans conducted on pregnant women during the course of their pregnancy. Routine scans, as Prenatal’s technology aims to improve, currently only detect about 50% of high-risk pregnancies and are limited in their ability to identify many high-risk pregnancies during the early stages, which is a major gap in early detection.
Tackling the Issue of Premature Births and Their Costs
The advancement of the Capital Region of Copenhagen is a case where the new ultrasound technologies offered by AI Prenatal can be implemented. Monitoring the region’s pregnant population of roughly 22,000 women gives a little over 1,500 premature births annually. The societal cost of these premature births is staggering, reaching DKK 800 million per year in the Capital Region alone. Also, almost all (over 80%) of the high-risk cases do not get diagnosed on time, which means that the doctors cannot start the necessary preventive treatment. The new AI systems, however, can be game-changing in preemptive diagnosis and can take appropriate actions in real-time to solve such problems.
Conclusion: A Transformative Advance for Prenatal Care
To summarize, AI technology designed by Prenatal offers an unprecedented opportunity because it can diagnose significantly more high-risk pregnancies than previously detected. It has great potential to improve maternal and fetal welfare through providing more accurate diagnoses. Reducing the risks of prematurity and complications is extremely critical from a public health perspective.
