Susan Desmond-Hellmann
speaker
362 appearances
1 recordings
1 series
first heard Apr 2025
last heard Apr 2025
Susan Desmond-Hellmann’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
I'm not close enough to the robotics piece of it.
I think it is when you're trying to connect all the dots. That's the thing. What AI does so brilliantly is it just takes a lot of data and it comes out with observations. And if there are ways that that can assist at the bedside, that's a massive improvement, especially when people are changing, even me, University of Washington to UCSF. Yeah.
it's so hard to change caregivers, to change health systems. Those kinds of things can decrease workloads. But I also think it's the kinds of things where clinical observations could be AI-driven.
These guys, what they did is they made possible, and we talked about preclinical. This is pre-preclinical.
This is figuring out what you're going to do. If you can make figuring out what you're going to do much, much faster, which they did- You're going to have the opportunity. The way I think of it is you've got like a mountain of opportunity, but it's shown a light on just a limited number of things where you can see the opportunity and take advantage of the opportunity.
I think it's a start, but I think it's great that they were recognized.
So far.
So far.
Anything where AI can help us with outcome measures. I told you that my husband's an HIV doc. When we were both at Bristol Myers Squibb, I was doing two by two measurements of tumors on x-rays for Taxol. And he was looking at viral load. Viral load allowed us to have 20 HIV drugs in like five years. It was crazy how good it was. I want a viral load for everything.
We need a good biomarker for more things. And you were talking about all the different types of breast cancer. So think about what you just talked about with breast cancer, that you have ER positive, ER negative, HER2 positive, triple negative. There's all these. What if actually there's 15?
So then you're in 15 trials, but you only need 10 patients in each trial because it's so obvious you have the perfect remedy for each of those patients. I always think of it as switch is on, turn it off, and you see clinical benefit. Anything we do that sets up like that, especially if we can not just measure switch on, but switch off. That's why viral load is so powerful.
I have been pretty negative based on the data. I just have not seen the data that suggests to me that we're helping.
So can AI help us? Possibly. The problem is just really hard.
I think that appears to be the problem because if they did, I think it would work. So that's the most important problem. The other problem is something that I think we all tend to underestimate because I love the concept of prevention. And I think Make America Healthy Again, in part, is we'll go to preventive therapy and stop all these. And
I understand that in oncology, we've often celebrated tiny successes, but you can't have big successes before you have tiny successes. I don't think it's easy to do early detection. The only two things that are, well, now three, colonoscopy works for cervical cancer, a pap smear works. Even better, HBV vaccine is my ad. And now you can do a spiral CT for lung cancer.
I'm not even using one handful of fingers. And we've been trying to do early detection as long as I've been an oncologist.
You know what you just said? That's something that if someone wanted to start a company, they could simplify that and make something more turnkey for patients and physicians.
They shouldn't be on the list.
That's the neat thing is you can just take out the polyp. That's always been the beauty of colonoscopy.
I think it'd be interesting to look at protein.
Showing 341–360 of 362 · page 18 of 19
← Previous
Next →