Dr. Claire de la Calle

Now, in those patients, it does seem to predict outcomes after active surveillance.

But I think, like I said,

We need to do these studies in more unselected active surveillance patients to understand better what is the role.

And then we also need to do modeling studies where we add these genomic classifiers to our complex clinical models that include all the information we already have, and especially crib reform.

and introductory because there are a lot of people out there that think that maybe these genomic classifiers are essentially just capturing that.

So if these tests are obviously very expensive or if they don't have added value over what we already know, then that doesn't argue to using them.

Now, in actual practice, and even in my practice, I have to recognize I use these classifiers and I do notice that they change management.

I have certainly seen patients with

Great group two disease, small volume pattern four, and a really, really high decipher score.

How can one ignore that?

It makes me really wonder if the underlying biology of the cancer is much worse at the molecular level compared to what is visible on H and E. So do I use them?

Yes.

How much they should be used?

We still need to answer that.

I also, I do like to stress to the patient that these genomic classifiers are only based on one biopsy core.

And we don't even know which biopsy core to use.

Maybe it's the targeted biopsy core, but maybe not.

And we know that prostate cancer is very heterogeneous within a single prostate and even within a single tumor.

And it would be, I think, too expensive to basically do the testing on every single core.

Oftentimes, you don't even have enough cancer in every single core to do that.