GWEN IFILL: Now: a new way of quickly diagnosing genetic diseases in newborns and the potential impact.
Margaret Warner has the story.
MARGARET WARNER: When a critically ill infant is rushed to the neonatal intensive care unit, it can take weeks for doctors to determine which genetic mutation may be endangering the baby's life.
But, today, researchers announced they have developed a new way of testing an infant's DNA that can scan for hundreds of genetic disorders and get results in just two days. The research was published in the journal "Science Translational Medicine."
The paper reported the tests of just six newborns in neonatal units, but the implications could be widespread. Roughly 20 percent of infant deaths in the U.S. are caused by inherited genetic conditions, according to the study.
And Dr. Stephen Kingsmore led the research team at Children's Mercy Hospitals in Kansas City. He's the director for the Center for Pediatric Genomic Medicine there.
And, Dr. Kingsmore, welcome, and thank you for being with us.
First of all...
DR. STEPHEN KINGSMORE, Children's Mercy Hospitals and Clinics: Thank you very much.
MARGARET WARNER: ... how big a breakthrough is this?
DR. STEPHEN KINGSMORE: This is a big breakthrough.
We have been working toward this goal for a couple of years now. There has been a big gap between the knowledge that we have of genetic diseases, about 3,500 of them, and the ability for doctors to identify which of these was a problem in any given child with an illness.
MARGARET WARNER: And up until now, how much have you been able to diagnose the DNA abnormalities? How quickly? I mean, I said it can take weeks and weeks, but what's the process that's making it so slow now?
DR. STEPHEN KINGSMORE: Well, typically, the way that this has been tested is for a doctor to pick the leading candidate gene or part of the DNA code and to look at just that.
It's kind of like fishing with a single fishing line. It can take months. Sometimes, it takes five years to make a diagnosis. And during that period, the family's waiting and probably the child is getting worse.
MARGARET WARNER: So, is it more important to have quick diagnosis in genetic diseases than in ordinary ones?
DR. STEPHEN KINGSMORE: Well, this is true of all diseases, but genetic diseases are a particular problem for physicians just because of the staggering number of them that present in children, about 3,500 diseases.
I mean, it doesn't really matter how expert a physician you are. You just can't comprehend each of those, and then tie them to specific children. Very many of them have very similar presentations, and so it's bewildering and often takes months or years to untangle.
MARGARET WARNER: So, in the simplest terms, how does your new process work? I gather it's in two stages.
DR. STEPHEN KINGSMORE: Yes, that's right.
So, we took this process, which takes months, and we collapsed the first stage, which is to decode the 3.2 billion letters of an individual's DNA language. That took about a day, and then another day for us to go from those three billion letters down to what we believe to be causing the particular child's illness. So, the first half was technical.
MARGARET WARNER: OK, and so...
DR. STEPHEN KINGSMORE: I'm sorry.
And the second half was very computational.
MARGARET WARNER: Well, from the description, it sounds as if the doctor basically punches -- I mean, uses software to say what they're just seeing as a doctor. He's not eating. His eyes are big, whatever.
DR. STEPHEN KINGSMORE: Yes.
MARGARET WARNER: And then you sort of narrow down how that matches up with...
DR. STEPHEN KINGSMORE: Yes, that's correct.
MARGARET WARNER: ... certain genetic mutations, and then you test his blood or her blood.
DR. STEPHEN KINGSMORE: That's exactly right.
So, a big problem here has been doctors don't actually know what precise tests to order. They have heard of some of the diseases, but many others, they haven't. And so we built a software system with clickable buttons where the doctor really enters, as you said, what he is seeing in a given patient, and that targets the analysis phase just to the parts of the vast genome that are on target for that particular child.
MARGARET WARNER: Now, I did notice that, sadly, almost all the babies in your study have died.
DR. STEPHEN KINGSMORE: Yes.
MARGARET WARNER: Even if a diagnosis is made of one of these rare genetic diseases, how often is there actually effective treatment?
DR. STEPHEN KINGSMORE: Well, there are treatments available for about 500 of the 3,500 diseases. Not all of these treatments are truly effective, but they all have some benefit.
And so we're looking to pick out the subset for which there is a treatment and to get that in a very timely manner, particular in babies who are born acutely ill. But even for those for whom there isn't a treatment available, getting a definitive diagnosis, and doing that quickly, makes a huge, huge impact on families who have this shock of a horribly ill baby.
MARGARET WARNER: So, how soon do you expect this to be really available clinically and at what cost?
DR. STEPHEN KINGSMORE: We estimate that the cost today is about $13,500 per baby. That's really the reason why we felt it was a very good place to start with newborns in an intensive care unit, because hospital stays are very expensive in these children.
When will it be available? Well, we're working now to put this into practice in our hospital here in Missouri, and by the end of the year, we project that we will be able to offer this test, albeit as a research test, where we will have to talk to parents and get their permission to run the test.
Hopefully, next year, we're able to broaden that out and offer it to other children in other hospitals.
MARGARET WARNER: And, finally, just an ethical question. What do the doctors do if they pick up a marker in this baby that predisposes him or her to a disease much later in life, let's say breast cancer? Do they tell the parents?
DR. STEPHEN KINGSMORE: Yes.
MARGARET WARNER: Not?
DR. STEPHEN KINGSMORE: That's a really good question.
And you're right. If you decode an entire genome, you find things that you wanted to see and possibly things that you didn't really want to see. And that's some of the beauty of what we have described today, is that we are targeting -- using this software system, just the regions of the genome which are relevant to a given baby.
And so we don't actually see the regions which are off-target. And this gets around this, this phenomenal issue by giving relevant information, but by not seeing the information that's irrelevant to what's going on in a given baby at a given time.
MARGARET WARNER: Very interesting.
Well, Dr. Stephen Kingsmore, thank you very much.