Dx Dialogues: predicting HIV treatment failure with AI

I'm looking forward to using these technologies and seeing them used to help us really fill in those gaps.

[RELAXING MUSIC] As we learn more and more about the mutations that

occur to HIV, even wild-type HIV, but certainly people who are exposed to HIV therapies,

we could see that could be a complex mix of mutations of the virus that may reduce susceptibility of it

to different medicines. This is where I think newer technologies and AI can take large sets of data and help us understand more

of a profile of an individual virus and what medications may work best against it.

And that could help us craft better regimens. The other thing about this is that sometimes people acquire resistance when they're not taking HIV medicines the way

that they should, or they may have acquired a virus that is already resistant. And so as we think about, well, what do we do after a regimen doesn't work any longer?

When a medicine fails, a medical regimen fails somebody, how do we develop a new regimen that is likely to work?

And in some cases, this can be very complex. In the old days, we used to have to memorize a lot of these mutations.

It's gotten more complicated. And now we have databases and tablets where we look up tables. I think that we won't have to do that anymore.

We could push a button. We can ask a prompt, and we will get a profile of here's the best match for this individual virus

that we know about and what mutations it harbors that will help us understand what's most likely to work. [AUDIO LOGO]