How do we use a lot of this new technology and understanding to speed up clinical trial design and clinical trial timelines?

We recreated a recent cardiovascular trial that was looking at the effect of GLP-1s on major adverse cardiac event risk. That trial, on average, ran for about 3.3 years before you were able to see the impact of the treatment on the reduction of major adverse cardiac event risk. From AI, we were actually able to understand the change in risk after 1.7 years. So there's a huge difference, basically half the length of the trial using a different biomarker.

The other biggest aspect was it's very hard to understand who has a propensity to have a major adverse cardiac event. And that was who they needed to recruit for the trial. they actually recruited people that had already had one. And so really what the drug was doing was proving that it would stop you having a secondary event.

What we were able to do by recreating the trial was, and this new biomarker that we used based on an ECG that analyzed that risk, was we could find out if the drug helped reduce event risk for the primary prevention, i.e. people that had never had a major adverse cardiac event. Well, that was very much, I think it was 12 to 15 times larger in terms of population. So now, suddenly, not only are you faster using these new AI-derived biomarkers, but now you're actually opening up the potential group for treatment and identifying them way earlier and at a far better stage than you ever could before. So there are huge opportunities in terms of both clinical trials and then subsequently the treatment of those populations using a lot of these new