Cancer AI could support research, drug development, and care

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 9:37

   i know oftentimes especially here on mts we talk about cool models what they're capable of all the things that they'll solve but then when it comes to actually getting them implemented getting doctors use it, healthier use it, getting the everyday person to reach these benefits. It's not this like immediate societal impact. And so what does it actually take for maybe something like Holland S1 to get implemented at scale to start changing how we actually think about cancer research? Yeah, that's a good question. I mean, I think like, yeah, there's a variety of ways that Holland S1 can be used. There are ways that it could potentially be used right now. For example, in cancer research, you can, you know, again, all of these cancer patients have pathology slides.

2. 10:17

   And so, you know, on the research side, you can try to apply Holland S1 to like huge archives of slides and extract novel insights about what, you know, like maybe different types of, you know, patients have different types of biology, things like this. So it can be very useful on the research side. And, you know, we, you know, we started to explore this a little bit with, with, with our key studies that we've published on our website. And then on, then of course, on top of that then there's a lot of interesting applications in life sciences and pharma companies as well, especially as they are, you know, looking at, you know, the treatment response of their patients to their drugs, right? And can those treatment responses be predicted? We can utilize the features that Holon S1 produces to be able to do this, and we can do that in a very interpretable

3. 11:01

   way because we're actually looking at the exact biological pathways. That's what's being predicted by halon S1. So that's another thing where it's like there, you know, it can help with the development of, of these drugs. And as these drugs are moving towards the clinic, it can help on that side. Then also it can help on, on the, you know, on, you know, like actual care of patients in terms of, you know, developing different tests for, you know, diagnostic tests for, you know, tests that, you know, again, help, help patients select different treatments. Because again, if you can predict the treatment response using the features from Hall and S1, then you can tell a patient, well, you may respond to this treatment versus you may not respond to this treatment, and that's how you can decide that. But of course, these sorts of things require a lot

4. 11:45

   of validation. You need to have, you know, like these sorts of prospective clinical trials, for example, in order to ensure that, you know, this model or this approach is accurate in a very large scale way and in the general. like in the general population. So that's like one of our next focus is like, you know, actually to conduct more sort of both retrospective and eventually also prospective trials for what we're doing with Holiness 1 and so that it can hopefully then finally reach the patient in the end. So that's the eventual end goal, I guess, of what we're doing is, yeah, we actually do want to