AI biomarkers are becoming incredibly important during the course of design and endpoint selection for cardiovascular clinical trials. One, because it allows you to understand, if I were to choose certain endpoints, could I potentially simulate what that trial would look like, how long that would take, what my event rate would be like, what my population size could look like? And so one example is we actually looked at the, you know, Novo Nordisk select trial and cardiovascular outcomes and thought, hey, this is a really huge trial that shows cardiovascular benefit of DLP1s, but is there a way that we potentially could have come to that conclusion sooner by leveraging an algorithm based on ECG? 

So we partnered with an algorithm developer who has a MACE prediction algorithm based on ECG and looked at all of these patients that were part of the trial, and looked to see if we had anticipated when there would have been MACE events earlier in the course of that trial and been able to kind of model out the event rate appropriately for that. So I think as you think about endpoints, know, unfortunately, in cardiovascular trials, a lot of the endpoints are related to, you know, MACE events, right? And so if there's a way to be able to predict those MACE events using something like an ECG, which is part and parcel of every single clinical trial, then you can really start to understand those populations better, you can start to like model out and predict those event rates and then really, really narrow in on the population that could be most beneficial for this trial.

At Dandelion, we built a data platform that really allows Clin Dev leaders to harness all of these data points in a way that makes it really easy to play around with different I/E criteria and see the impact that could have on their trial design.

If you look at the I/E criteria of trials, you can go to any trial on CT.gov and look at those I/E criteria; you're like, wow, there's a lot of criteria, there's a lot of data and most of that data is only captured in the course of clinical trials, we can actually on almost all occasions, match all of those I/E criteria directly to the data that we're seeing in our platform, which is super, super important because then it allows you to run simulations on, how would this trial change if I were to tweak some of these criteria? How can I optimize for population size, for event rate, for other things that I care about to make sure that I'm finding the population of patients that can benefit the most from these therapies? 

And so our platform is really built to harness the richest multimodal data set available, but really make it easy to use and quick to query, so that you can start to run simulations and model trials quickly, so that you can really inform the strategy that you pursue.