Giving birth is one of the riskiest activities a woman can engage in. Although developed nations have much lower maternal mortality rates, there are still countless things that can go wrong. The World Health Organization (WHO) estimates that about 295,000 women died in 2017 alone due to complications from pregnancy and childbirth.
Fortunately, skilled care can help save lives by preventing those complications from occurring or facilitating early interventions.
Research from the Perinatal Institute has shown that tracking fetal growth in utero is an essential component of good prenatal care. It can help prevent stillbirths, lets physicians identify growth restrictions, and increases the odds of a successful pregnancy.
Samsung Medison and Intel are working together on a smart, artificial intelligence (AI) driven approach to fetal growth tracking. Intel’s hardware and OpenVINO toolkit power Samsung’s BiometryAssist and LaborAssist programs. Together, they are used to automate and simplify fetal measurements and automatically estimate the fetal angle of progression (AoP) during labor.
I had a chance to speak with Dan Monaghan, senior sales director at Samsung USA about the tech. You’ll find the interview below after a recap of this innovative approach to maternal health.
Samsung’s BiometryAssist tool can be used to automate and standardize fetal measurements in just 85 milliseconds. With a single click, providers are able to get a 97 percent accurate reading. Typically, this process is much more involved and takes a lot of time. By automating fetal measuring, physicians are able to spend more time talking with their patients.
Meanwhile, the BiometryAssist system’s ability to standardize fetal measurements is a welcome answer to a historically difficult problem. It helps physicians quickly verify the measurements while providing care to a large number of patients.
Without a tech-forward approach, this wouldn’t be possible. The power of AI has a huge impact on this application. Won-Chul Bang, corporate vice president and head of product strategy at Samsung Medison says, “Samsung is working to improve the efficiency of new diagnostic features, as well as healthcare services, and the Intel Distribution of OpenVINO library and OpenCV library have been great allies in reaching these goals.”
Intel’s OpenVINO toolkit helps businesses develop apps and platforms that emulate human vision. It uses Intel hardware and neural networks to deliver high-level performance regardless of the end application.
For something like fetal measuring, emulating vision is critical. Ultrasound images allow providers to visualize the size and growth patterns of the fetus in utero. BiometryAssist is able to do the same thing thanks to the OpenVINO toolkit.
Helping in Labor
While BiometryAssist is a helpful tool that can be used throughout the pregnancy, Samsung Medison and Intel also developed a solution that’s designed specifically for the labor process.
The fetal angle of progression is a very helpful measurement during labor. It can be used to predict outcomes, keep the woman informed on how labor is progressing, and help doctors determine the best approach.
The LaborAssist tool helps physicians estimate the fetal AoP in real-time throughout the labor process. This is crucial in many ways.
For instance, slowing labor can be a sign that something is wrong and can result in issues for the baby. Without real-time info, many patients are forced into a cesarean section delivery that isn’t actually necessary. This can lead to complications for the baby and a lengthy recovery for the mother. The accurate, real-time data provided by LaborAssist helps physicians determine the best delivery method and can help cut back on unnecessary C-sections.
Professor Min Jeong Oh from the Korea University Guro Hospital in Seoul says, “LaborAssist provides automatic measurement of the angle of progression as well as information pertaining to fetal head direction and estimated head station… It is expected to be of great assistance in the assessment of labor progression and decision-making for delivery.”
Making an Impact
Currently, BiometryAssist and LaborAssist are in use in 80 countries around the world—including the United States, Korea, Italy, France, Brazil, and Russia. The tools received Class 2 clearance from the U.S. Food and Drug Administration (FDA) this year.
Moving forward, Intel and Samsung Medison plan to continue their collaboration to develop state-of-the-art ultrasound solutions by leveraging the power of advanced technologies like AI. It will be interesting to see how these systems impact the healthcare field in the coming years.
Keep reading to learn more about these innovative technologies from my conversation with Samsung’s Dan Monaghan.
The Burn-In (TBI): You mention that this smart workflow helps speed up the fetal ultrasound process while also providing accurate results. How exactly does Intel’s AI play a role in this?
Dan Monaghan (DM): Samsung selected the Intel Distribution of OpenVINO toolkit as a software toolkit that can run on the existing hardware platform, without requiring modifications or upgrades. OpenVINO enabled Samsung to speed up the inference.
TBI: Going off of that last question, how was the AI trained? What sort of data set was used?
DM: In Biometry Assist, 12,000 images have been loaded into the deep learning reference data set. The technology also improves over time, as more images and results are added to the database.
TBI: As with any AI application, it’s nice to know that a safeguard exists. What happens if a physician disagrees with the automated system based on their own observations of the ultrasound?
DM: These solutions are there to assist physicians, relieving them of a tedious task while providing a standard for the measurement. An OBGYN can conduct her own fetal measurements if she feels it’s necessary, but we are hearing from physicians around the country that BiometryAssist and LaborAssist are valuable in helping them provide care to pregnant women.
TBI: What sort of traction is this system gaining in the U.S.? Do you have an estimate of how many facilities are using it?
DM: In the US alone we have close to 1,200 systems currently in use with these technologies and that number is increasing every week.