1 in 10 births in the U.S. are premature. Here's how AI could help doctors predict it

Last year, 1 in 10 U.S. babies was born before 37 weeks of pregnancy, which is considered preterm. That’s one of the highest premature birth rates among developed nations, according to the March of Dimes. We hear from parents of preterm babies about their experiences, and Ali Rogin speaks with an entrepreneur who’s using AI to help doctors predict when preterm births are likely.

Read the Full Transcript

Notice: Transcripts are machine and human generated and lightly edited for accuracy. They may contain errors.

John Yang:

Last year, one in 10 U.S. babies was born before 37 weeks of pregnancy, which is considered preterm. The March of Dimes says that's one of the highest premature birth rates among developed nations.

Preterm births can have long term health effects and are hard to predict. Ali Rogin spoke with an entrepreneur who's using AI to help doctors know when it's likely. But first, March of Dimes helped connect Ali with parents of preterm babies to hear about their experiences.

Erica Shoemate:

My name is Erica Shoemate and I am Mia's mom.

Roxanne Romeo:

Hi. My name is Roxanne Romeo and I'm Zara's mom.

Lucero Payano:

My name is Lucero Payano and I am Zael's mother.

Ruben Contreras:

My name is Ruben Contreras and I'm Zael's dad.

Ashley O'Neil:

My name is Ashley o' Neil and I am Collin's mom.

Emma Lee Greenley:

My name is Emma Lee Greenley and I'm Lois's mom.

Erica Shoemate:

My baby was born at 24 weeks, 2 days. He she weighed 1 pound 7 ounces or the best that they could measure so quick. She was intubated immediately.

Roxanne Romeo:

Zara was born at 26 weeks and four days. She weighed 2 pounds, 6.7 ounces and she spent 93 days in the NICU.

Lucero Payano:

Zael was 25 weeks and two days when I delivered him.

Erica Shoemate:

We learned that my daughter actually has Ellis van Crevel syndrome, which has less than 300 cases reported in the world, but it's highly concentrated in the Amish Dutch population. It's very rare in general, but then to be a family of color, we only know two other families of color that have been directly impacted in the world.

Ashley O'Neil:

I went into labor with Collins at 25 weeks gestation at a small community hospital that didn't have the resources to properly care for a child. He was born at roughly a pound and a half.

Roxanne Romeo:

Every day I would have to call in the morning to talk to one of the NICU nurses to find out how many episodes she had the night before, which is when a baby stops breathing. And to even imagine your child not breathing is just crazy.

Ashley O'Neil:

He fought. He was a feisty little thing, repeatedly pulled out all of his tubes. But he came home 183 days later.

Emma Lee Greenley:

Lois is five years old now. She is in preschool. She's doing very well.

Erica Shoemate:

She just turned 8 on September 27th. It has been an honor, a privilege to be this little rock star's mom.

Roxanne Romeo:

She is a spunky, sassy little toddler. She's gonna be three in December.

Lucero Payano:

He is, thank God, doing amazing. When I tell people that he's a 25-weeker, their jaws drop.

Ashley O'Neil:

Collin is six now. He is in the first grade. He loves school. He is non-speaking, uses a communication device. He does boss me around, mom, iPad, please. He has the best smile that lights up a room. He loves playing tricks on his teachers. If they ask him a question, he'll hover over the right answer, giggle and hit the wrong answer and giggle some more.

Roxanne Romeo:

I just encourage people to learn more about prematurity, especially if you know you're pregnant or if you plan to have a child. And I'm not just talking about women. Fathers need to be in the know as well because they are definitely part of this journey.

Erica Shoemate:

To anyone that is navigating this thing that you never sign up for as a parent, just know that you can get through it. There is life and love on the other side. And also to allow yourself the moments to grief, to grieve through the grief of the pregnancy you didn't have. The parenthood experience that you are not having or will not have, and to also not allow yourself to get in the business of comparing your child to someone else's who is not having the same experience as you.

Ashley O'Neil:

I share my story a lot just to tell other people that like I know how dark it can get. But if you just keep going, eventually you'll find the light. It may not be as bright as you imagined or even the same colors you imagine your light would look like, but it can still beautiful nonetheless.

Ali Rogin:

We now turn to Robert Bunn, an entrepreneur who knows this issue firsthand. He and his wife have four daughters, but his wife experienced nine miscarriages along the way. Her experience inspired him to work on technology that can help predict whether a pregnancy will end in a premature birth and help a medical team prepare for that outcome. Robert, thank you so much for being here with us.

Robert Bunn, Founder, Ultrasound AI:

Thank you. It's a pleasure to be here.

Ali Rogin:

Tell me a little bit more about what you and your wife went through to build your family and how that inspired you to begin working in this field.

Robert Bunn:

Sure. We were beginning to start our family. My wife suffered numerous miscarriages, and I was beginning to wonder why this could happen over and over and nothing could be done about it.

After doing a bit of research, because I have no medical background, I had to do some research to understand the problem. What I realized is that the reason there's so many premature births is because typically they're a surprise. And when they happen, it's too late to do anything about that time. And doctors typically don't have any forewarning.

And I so thought maybe if I could create an AI that could predict that a premature birth is impending, then doctors would have a fighting chance to prevent that, and they have time to do something about it.

Ali Rogin:

How does this technology know what to look for and flag issues, potential issues?

Robert Bunn:

I eventually learned that the solution is to obtain millions of ultrasound images, and then I have the outcomes of what happened to those babies in those images. Then what I did is I created an AI that could figure out how to map what was the pixels in the images to what actually happened. And I had to build a supercomputer from parts in my basement to get the computer power necessary. And it took months to train this AI.

But basically what the AI eventually learned how to do was to find patterns in these images that then correlated to the actual outcomes, which opened up a whole lot of new discoveries, as well as the capability to predict the delivery date of these babies, which then, of course, the doctors could then act upon and determine if that was going to work or if they needed to do something.

Ali Rogin:

Why is it traditionally so difficult to screen for prematurity?

Robert Bunn:

One is because ultrasound images are not -- are hard to read. They're very noisy. It's very hard to see things in them. There are some things that doctors know that typically lead to premature birth, but those things typically only happen to a small percentage of women. But the vast majority of women is just you have a history of premature birth or history of miscarriages. So therefore we'll monitor you closely in the future.

But it's not -- the accuracy is not as good as it could be. And I think that's the problem I needed to solve is to give doctors very precise delivery dates of when these babies were going to be delivered so then they could then act confidently to treat those women.

Ali Rogin:

I understand your technology is already being used in countries in South America. What have the results been like there?

Robert Bunn:

They've been in trial use in both Brazil and Chile and the doctors are extremely happy. Probably the biggest thing that makes them happy is the fact that it's able to identify the women who are on their first child who may have a premature birth so they can make sure those women get the care they need they would not have gotten otherwise.

Then there's also the case where there's high risk pregnancies and they want to keep the babies in as long as possible because that reduces time in NICU and they have a good idea of when the baby's going to spontaneously deliver and a spontaneous deliver may not be the best choice in that case. And they can do a C section right before then. So it helps them time the deliveries that need to be much better and have better outcomes.

Ali Rogin:

And then what's your outlook for becoming available to people here in the US.

Robert Bunn:

So we've recently submitted our final review for the FDA and we hope to have approval here very soon, maybe within the next few months. Nothing life is uncertain, but we feel very good about this technology and that its ability to begin saving the lives of babies in America here too.

Ali Rogin:

Robert Bunn with Ultrasound AI. Thank you so much.

Robert Bunn:

Thank you. It's a pleasure being here.

Listen to this Segment