Dr. Josh Ofman discusses Grail's AI technology that detects multiple cancer types from blood samples by analyzing DNA patterns. The approach moves beyond screening for one cancer at a time, using genomics and machine learning to identify signals across different tumor types.
So the way traditionally in healthcare, we have screened for cancer is by looking at one cancer at a time. So for women, we look for breast cancer, cervical cancer, colon cancer, and if she happens to be a smoker, also lung cancer. And we're looking for those single cancers. So we call those single cancer screening tests. Right. A pap smear or a psa. Well, Grail was invented and kind of founded some really breakthrough technology that looks at DNA that is being shed by cancer cells into your blood. And we figured out a way to combine genomics and artificial intelligence and machine learning to read a pattern in that DNA that you see very commonly in many different kinds of cancer and importantly, that you almost never see in patients without cancer. So so far, this pattern recognition technology has enabled us to find well over 50 different types of cancer.
Wow.
And we're grouping them so we know it's well over that we know, you know, probably 60 to 70 already. And because we're not looking for any particular cancer, we're looking at whatever cancer the individual may have because there's a shared signal in that DNA that we're reading. And. And then we can localize that signal to where in the body that signal comes from to tell the doctor where to go look to find that cancer, if it exists.
I've read that there isn't a lot of genetic overlap. You take two different pituitary tumors, and they could have entirely different genetics in them. How do you know what cancer DNA looks like, given that a lot of cancer looks like it might not be DNA caused?
You know, when most people are talking about DNA, they're talking about mutations, changes in the code of the DNA, and.
