Christopher Sweeney: Thanks, Tanya. And it's an honor to be talking about some results that we're presenting at ASCO 2026 about a genomic study.
Tanya Dorff: Yeah. Because you work in a genomics institute, perhaps that's very appropriate, but more so because you led this study ENZAMET, from which we learned that intensified therapy with enzalutamide improved overall survival in metastatic hormone-sensitive prostate cancer compared to traditional doublet with nonsteroidal like bicalutamide. Those were really practice-changing findings. Now you're presenting data looking at a genomic classifier called Decipher. Taking a step back, what is the Decipher genomic classifier score?
Christopher Sweeney: It's a great question. As we know, tumors have variable RNA and DNA quality, and variations in those genes can predict how patients do with a given therapy. The RNA profile that comes from the formalin-fixed paraffin-embedded tissue is sent to a company called Veracyte, and they actually run an assay called Decipher. And there are 22 genes that they measure, but it comes off of a 25,000 gene transcriptome, but they've gone in and they've identified 22 genes from that 25,000 genes that have the ability to explain a lot of the biology and clinical outcomes of patients with prostate cancer. The 22 genes, we don't know its proprietary gene set, so we don't know the weightings of the genes and how they come to it. That's locked. But the reassuring thing about that gene set is that many, many studies have shown that it identifies men more likely to progress to metastatic prostate cancer if they have localized disease.
And also, we're now learning that men who have a higher Decipher score have a poorer prognosis for metastatic hormone-sensitive prostate cancer. So predicting relapse, but also in patients who have metastatic disease. Even just analyzing the 22 genes in their primary, even though they may relapse years later and have metachronous metastatic hormone-sensitive or have de novo, the information from those 22 genes does have information on how patients respond to hormonal therapy for metastatic hormone-sensitive prostate cancer.
Along the way, we've tested this gene set first in the CHAARTED study, and we saw that patients who got ADT alone did poorer with a high Decipher score, and then looked like they had some benefit when they got docetaxel [inaudible 00:03:08], starting to suggest that they had a predictive ability to identify men who benefit from adding docetaxel to ADT. And then, it was tested again and then validated in the STAMPEDE study with all the right power to say it is predictive for who benefits from adding docetaxel to ADT. So, it's a gene set tested in CHAARTED, validated in STAMPEDE when added to ADT.
Tanya Dorff: That's fascinating. We traditionally thought about Decipher as something we used in the localized prostate cancer setting that really predicted for more aggressive behavior, which made us maybe want to utilize additional layers of therapy such as radiation or hormone therapy. What made you think it could be predictive in this more advanced setting? And specifically, it seems like you applied the question to chemotherapy.
Christopher Sweeney: Absolutely. You're absolutely right. We, on the medical oncology community, see the urologist and the radiation oncologist using it and helping them make decisions on who to get what type of treatment in localized setting. It was a bit of a leap of faith to say, does it have a role in metastatic disease? Gerhardt Attard and myself, back in 2019, actually wrote a grant saying, let's look at a gene expression profile and see if that can help us work out the outcomes of patients based on their genes in their primaries. We chose the Decipher assay because it's 25,000 genes, clear lab, reproducible, and a reliable clinically implementable assay. People said, "Well, why don't you do RNA-seq?" There's challenges with doing RNA-seq off of paraffin-embedded tissue, but this is reliable even in archived old tissue to give information. That's why we chose the assay.
And in that assay, there were multiple classifiers. And we looked at multiple different ones, and that was in the Annals of Oncology. And a colleague of mine, and they submit was the first author. In a biomarker study, we do the strategy of testing many different assay classifiers. And then the leading ones, promising ones, we then validate in the next cohort. And that's why we then now tested it with a priori statistical analysis plan with a cut point that we defined before we looked at any of the ENZAMET data.
Tanya Dorff: And I feel like we should jump to the results. I have some questions on the methods and integration with clinical risk predictive factors. But tell me, what were the top-line results when you applied this Decipher score to the ENZAMET population?
Christopher Sweeney: It basically shows that patients who have a tumor score, a Decipher score greater than 0.85 have a clearer benefit of getting the benefit of adding docetaxel to an ADT ARPI backbone and receptor pathway inhibitor of which there is abiraterone, enzalutamide, apalutamide, and darolutamide. And in the ENZAMET study, we had ADT and enzalutamide, and patients were selected for docetaxel. We can see, even though patients who had poorer risk disease were more likely to get docetaxel for all the right reasons because it was the standard of care for patients with poor-prognosis disease, that was their backbone, as opposed to some patients who just got ADT with or without enzalutamide. Some patients got ADT enzalutamide with or without docetaxel.
It gave us the opportunity to look at that score with the current standard of care of ADT plus an ARPI, in this case, enzalutamide. Because the previous work we did with docetaxel was just with an ADT, aloan backbone. But there's been no study of ADT, docetaxel, enzalutamide, or ARPI, versus ADT and enzalutamide, ARPI comparison. This is the only study where we actually currently have the contemporaneously enrolled cohort of patients who've got ADT and enzalutamide with or without docetaxel. And we can see the score greater than 0.85 identifies men more likely to have benefit from docetaxel.
Tanya Dorff: That's super helpful. I think we're always grappling with, when do we use triplet therapy? We don't have that pure study of ADT AR pathway inhibitor plus/minus docetaxel. Obviously, the ASPIRE trial and triple switch are trying to answer that to some extent. Now we're saying we can take someone whose clinical characteristics might make us want to consider docetaxel, and this Decipher cut point 0.85 might help push us in one direction or the other.
Christopher Sweeney: I think that's a very, very fair assessment. Number one, get the patients on the ADT and the RP that you can access. That's the baseline therapy, because that improves the prognosis of all patients in all prognostic groups. But it is a conundrum who to give the docetaxel to. And so, you're absolutely right. The next question we should also ask, can the patient tolerate the docetaxel? Are they fit for it? And then, send the test if you think they are and you need help deciding whether to add docetaxel. Characteristically, we've been using docetaxel for men who have the most rapidly progressive poor-prognosis disease like synchronous high-volume, we feel more comfortable. But even in that group, some patients do poorly with ADT and enzalutamide, and some patients do great with ADT and ARPI, but we don't know who they are even if they're synchronous high volume.
This actually gives more comfort of saying, yes, this patient is more likely to have a benefit in terms of a longer survival than patients who just get the hormonal therapy treatment alone. Conversely, it also gives you comfort of actually saying maybe this patient doesn't need chemotherapy. And so, you can have the test and you can have a more informed decision process for the patient in front of you.
Tanya Dorff: And you mentioned you were looking at other biomarkers. One thing that comes to mind with AMPLITUDE and TALAPRO-3, we're now going to be looking at DNA and trying to identify patients who might benefit from PARP inhibitor therapy. How do you think the Decipher story will interact with that?
Christopher Sweeney: Two things is, yes, PARP inhibitors are a very important therapy for men with metastatic prostate cancer, and they have to have the right biomarker. And the number of patients who are eligible for a PARP inhibitor with the BRCA1 or 2 mutation where the signal is strongest is probably up to 10%. The other gene mutation's unclear, but that may take you up to 20%, but there's 80% of patients who aren't eligible for any version of a PARP inhibitor. But I limit PARP inhibitors personally from the data to patients with a BRCA mutation. There's 90% of patients who need intensification who don't have X, the benefit from a PARP inhibitor.
The other thing is I would still propose, and they did this in the Niraparib study with the AMPLITUDE, they did allow prior docetaxel. A patient who we have a stronger evidence of benefit from docetaxel, I'd say still give the docetaxel for the patients with the high score. But if they have the BRCA mutation after basically intensified induction therapy, then give them a maintenance therapy with an ADT and an ARPI plus a PARP inhibitor after you got rid of the AR-negative disease that's independent of PARP inhibition independent of hormonal therapy. And then, hopefully you've got a greater disease population that is hormonal therapy dependent and PARP inhibition responsive and get a longer outcome by induction followed by a maintenance phase is how I think a patient with a high Decipher score should be treated with a BRCA mutation.
Tanya Dorff: Given the design of ENZAMET, in which docetaxel use was left to the discretion of the treating physician, there must have been some complexity in the analysis. Why don't you talk a little bit about that?
Christopher Sweeney: I would actually upgrade some complexity to a lot of complexity. What's clear is that people talk about the chemotherapy sniff test, who do they think should get chemotherapy? And in ENZAMET it was, as you said, at the discretion of a patient and a physician based on their clinical characteristics. The patients who got docetaxel had much greater rate of poorer prognosis, high volume, metastatic hormone-sensitive. We actually even see they have a higher rate of a score greater than 0.85. They had a much greater rate of a poorer prognosis with the hormonal therapy.
The patients the physicians added the docetaxel. And so we had to do something called propensity scoring, which accounted for those variables. Not completely, but once you do account for those variables, we can actually start to see that patients who got ADT, docetaxel, enzalutamide did better than their counterparts with a score greater than 0.85 who got ADT, enzalutamide alone on a hazard ratio for overall survival of 0.75. Almost like a randomized trial, but nowhere close or good enough, but approximates what you would see if it was a direct randomization. The propensity score tries to get it like equal randomization. We didn't get there, but we got closer to it. And it is a very particular analysis called propensity scoring that helped us get there.
Tanya Dorff: Well, that's fascinating. And it's so encouraging to see us finally moving closer towards personalized medicine and prostate cancer. Thank you so much for being here today.
Christopher Sweeney: Thank you for the opportunity to share some new data, Tanya.