How AI-enabled, handheld ultrasound could help save moms' and babies' lives
The Butterfly handheld ultrasounds are also in use in a Texas hospital system, helping patients when it comes to IVs.
Artificial intelligence is giving an assist to pregnant women and their care teams to help those moms deliver healthy babies.
Ultrasounds are a routine part of prenatal care, but the Gates Foundation says that two-thirds of women around the world lack access to them. So, the group has backed the development of a handheld, AI-enabled ultrasound, called Butterfly, in hopes of changing that.
The Butterfly devices were created to help solve the problem of access to diagnostic imaging, but also to be a low-cost, portable and durable option for clinicians so they can get answers right away when treating their patients.
Some of the Butterfly probes are about the size of a men's electric shaver.
Perhaps among the most important benefits, it can help health care workers discover high-risk conditions early during pregnancy - potentially saving a mother and baby's lives.
"Butterfly is the world's only patented ultrasound on a microchip. So inside the probe head, there are micro conductors that create sound waves. The probe connects to a phone or tablet where a clinician can see an image on the screen. The sound waves travel through the body, through the probe, and then return information back to your phone or tablet, so that the clinician can see the screen, look at the image, and make interpretations right at the bedside very quickly," explained Dr. Sachita Shah, Sr. Director of Global Health at the Butterfly Network.
Butterfly is also unique, Shah said, because of the embedded AI to help make calculations easier on the actual ultrasound screen versus a provider having to do the calculations manually.
That's something that would be useful, for example, before a catheterization, where health care workers need to know bladder volume.
AI is also incorporated into an education app called ScanLab.
"It has AI embedded such that a provider can be a new learner for ultrasound, and go through a walkthrough of how to scan the heart, or how to scan the lung, and then see their own images with AI labeling for all of the anatomy," she explained.
That's an important piece because ultrasound is still very operator dependent, Shah added. "Providers need to be trained how to move the probe, how to make the images, and then also how to interpret the images."
Ultrasound is the ideal form of imaging because it has no radiation, Shah told ABC13, adding that the World Health Organization (WHO) recommends every pregnant woman get an ultrasound scan before 24 weeks of pregnancy to determine high-risk conditions of the placenta or if there are multiple babies.
But in so-called obstetric deserts, ultrasound may not always be accessible in a timely manner.
Butterfly was created to address maternal care access globally, but you don't have to look far to see or understand that there's a need for equity and accessibility right here in Texas.
A March of Dimes report found that roughly 46% of counties in the state are defined as maternity care deserts. That includes Chambers, Liberty, Waller and Austin Counties.
"This means that a little less than half of our counties do not have a birthing hospital. They don't have a birth center as well as they do not have an obstetric provider meaning an OB/GYN family medicine doctor who delivers babies, or a certified nurse, midwife," said Alicia Lee, the director of the Maternal Child Health Collective Impact at March of Dimes.
March of Dimes has been working with Stanford University to study how AI can help predict the health of premature babies and diseases related to pre-term birth - in some cases before the baby arrives.
"This paves the way for some very targeted medical care for our pregnant women," Lee said.
Butterfly isn't currently used in those southeast Texas-area obstetric deserts, but they are in use in the Lone Star State and for reasons outside of maternity care.
The nursing team within the Baylor, Scott and White hospital system in Texas found that when using the Butterfly ultrasound, it made placing IVs much easier, safer and less painful for the patient.
"In the early part of their deployment, they used Butterfly with the nursing team to study how it impacted patients getting IVs placed and what they found was a drastic reduction in the missed IVs," Shah told ABC13.
But why does having a smaller, point of care ultrasound matter?
As Shah explains, ultrasound traditionally involves a radiologist in a reading room, a sonographer in a sonography suite completing the scan, and a clinician at the bedside to order a test.
"So there's this big lag time, and there's also a physical space issue, where the patient has to leave the clinician taking care of them, go to see a sonographer who makes pictures," Shah said. "The pictures get read by the radiologist, and then that feeds back to the clinician. And so there's a big-time lag in getting that information."
Point of care, which was first invented by trauma surgeons and emergency doctors who needed a quick way to determine where a critically injured patient was bleeding, is performed by the clinicians themselves. It's designed to answer a simple "yes" or "no" question, especially when time is critical.
It's one of the aspects that made it attractive to the Gates Foundation, who Shah said was an early investor and helped the company to deploy ultrasound probes in sub-Saharan Africa in 2022.
Another important part is the cost.
Shah says one Butterfly device offered is in the $2,000 range, compared to the types of ultrasound machines you might find in a hospital system that can cost between five to six figures.
"The other piece is the environmental sustainability piece. The probe is tiny and portable. If it needs to be shipped back to be fixed, it can be very easily. Much of the ultrasound equipment that exists today is too large, and then when it breaks, it ends up in what they call medical graveyards," Shah said.
Shah adds AI in health care is part of the next frontier.
"Really smart app developers around the world are building AI for diseases that we don't even use ultrasound for right now," she said. "When we think about medicine, and how it could change, and how to empower clinicians that might not have as advanced training, but have a whole boatload of patients that are really sick to take care of, and they need an affordable, really easy way to do it, that's why I'm here."
