Andrea Miyahira: Hi. I'm Andrea Miyahira here with the Prostate Cancer Foundation. I'm here at ASCO 2025, joined by Dr. Emily Grist, a clinician scientist at the University College London and a 2021 PCF Young Investigator. Dr. Grist, thanks for joining us.

Emily Grist: Hi. Thank you for having me. And I'm excited to talk through the data and our work.

Andrea Miyahira: Thank you so much. So you presented an oral abstract on a study using STAMPEDE data to look at survival in patients that received docetaxel versus abiraterone. So tell us about the goals and the rationale for that study.

Emily Grist: The standard of care treatment now for patients with advanced prostate cancer is ADT and an androgen receptor pathway inhibitor or ARPI. And we have also the option of docetaxel. But as a medical oncologist in clinic, we don't have that many tools to identify who will benefit from docetaxel added in as a triplet therapy, we use patient factors, clinical factors, including volume.

And then in ESMO last year, I also presented data around the DECIPHER mRNA score, also having utility to identify patients who may be more sensitive to docetaxel with a high score in the prostate tumor, indicating more sensitivity. Really, the objective of our study was to identify prognostic biomarkers so patients who are more likely to have shorter survival on ADT and ARPI and also identify a predictive biomarker to identify patients who may do better with the addition of docetaxel.

And we used the STAMPEDE framework to address this question because we had a cohort of patients randomized to ADT versus ADT and abiraterone or ADT versus ADT and docetaxel. And that was a huge cohort of over 3,909 patients, both with high-risk localized and metastatic disease.

So we initiated a massive multi-institutional effort. I'm talking over 100 trial sites in the UK and 95% of men very kindly donated their tissue to this research. We collected tissue from over 2,200 patients and all were cut for H&E sections. And we have a big digital repository now that the community can use.

And we collaborated with Veracyte and used their Affymetrix microarray, and then generated transcriptome wide expression profiles for over 1,500 of the cohort. We then identified classifiers, transcriptome classifiers we were interested in and we thought would be relevant in prostate cancer.

I presented the data for DECIPHER last year, and we pre-specified testing that as a predictive biomarker based on data from CHAARTED. And then today I'm presenting data around a classifier associated with PTEN inactivity. And we also cut sections for PTEN IHC.

Andrea Miyahira: OK. Thank you. And what were the major findings and conclusions?

Emily Grist: So the PTEN mRNA inactivity score is continuous based on mRNA, and it was used in prostate cancer and trained to predict PTEN genomic loss. So we looked at that and also looked at PTEN IHC scoring. We used a validated assay and collaborated with the Johns Hopkins team. So each section was scored by two expert pathologists.

And our first question was how does this score associate with PTEN protein. We found that a high PTEN inactivity score significantly associates with homogeneous and heterogeneous PTEN protein loss. Although I must say the relationship with homogeneous protein loss was much clearer.

And then what we did was we looked at the score across non-metastatic metastatic disease and found that the score was not significantly different, but much higher than low and intermediate risk disease, which isn't a massive surprise we know PTEN genomic loss is higher in more advanced disease.

And then we looked at the associations with clinical outcome. And our first finding was that a high PTEN inactivity score is associated with shorter survival on ADT and ADT and abiraterone. But we did not see that in the ADT and docetaxel cohort, which suggested to us that there was a differential treatment effect.

So the PTEN inactivity score is bimodal. So we picked a cutoff and high score patients were classified as PTEN inactive. And that was equally split between metastatic and non-metastatic disease and conversely low score PTEN active.

And we tested this formally in an exploratory analysis. So this wasn't pre-specified like DECIPHER, this was exploratory. And we saw really clear evidence that to patients with tumors that are PTEN inactive have greater benefit with docetaxel than those that do not. And there was a significant biomarker treatment interaction. So this really is strong evidence with a p-value of 0.038. So we saw a really strong signal there.

And then we went one step further and we looked in smaller subgroups split by disease volume. And we found patients, the direction of effect was the same irrespective of whether you had high volume or low volume disease. PTEN inactive tumors were more sensitive to docetaxel. But we did not see this in abiraterone. Treatment effect was uniform.

And interestingly, we did not see any prediction using PTEN protein IHC assays. So if you were PTEN protein loss, we did not see that you were more likely to benefit from docetaxel chemotherapy. It was just if you were PTEN inactive defined by the classifier.

Andrea Miyahira: OK. That's really interesting that you have this classifier that PTEN inactivation by a signature actually corresponds with a better response to docetaxel. Do you have any insights as to what the mechanisms that are driving that?

Emily Grist: I think it's a really exciting novel result. And the answer at the minute is it's not clear, why. I think it's interesting that PTEN protein loss was not predictive, but in our hands there was a much stronger association between homogeneous PTEN protein loss and PTEN inactivity by the classifier. Whereas heterogeneous PTEN protein loss was quite split actually between PTEN inactive and PTEN active transcriptionally.

So potentially the signature is telling us something about dysregulation of downstream transcriptional pathways that the IHC assay isn't. There isn't much preclinical data to explain why we would observe this treatment interaction. There is some data that PTEN interacts with cell microtubules and that is the mechanism of action also of docetaxel. I think we're going to be able to understand the biology a bit better with more targeted approaches now we have this result.

Andrea Miyahira: OK. Thank you. And now we have several different treatment options for patients with metastatic hormone sensitive prostate cancer including various doublets and also triplet therapy. So how do you think this biomarker and also integration with other biomarkers, how should we approach using these to select treatments for patients?

Emily Grist: So I think it's really exciting. We've got two transcriptome biomarkers now that have utility in this disease setting. Although I should add DECIPHER again, we pre-specified. So we generated level 1 B evidence. And our PTEN result was an exploratory analysis. So this will need further prospective validation.

But given we have the DECIPHER result as well we looked at the overlap between the two signatures, there is no correlation between the PTEN inactivity score and DECIPHER. They do not overlap, and only 28% of tumors were classified as high DECIPHER and PTEN inactive.

So we split into smaller subgroups and did an exploratory analysis to look at the effect with DECIPHER high and PTEN inactive. We found that tumors classifying as both 45% reduction of hazards of death with docetaxel. So I think really composite biomarkers now we've got a really strong rationale for prospectively testing this in future trials. So that we'll be looking at how we go about that as our next steps.

It's really hard because we don't have other previous trials that directly randomized to docetaxel with long-term outcome data and transcriptional data. We have CHAARTED. So that has a biomarker cohort of 160. And we did look within that cohort. And we did again see the same direction of effect with PTEN inactive patients having significantly more benefit from docetaxel chemotherapy.

But I think we do have to look at the ENZAMET trial and some of the triplet therapy trials and see if we can generate further data and perhaps perform a meta-analysis in patients that receive docetaxel as well as an RP and ADT.

Andrea Miyahira: Thank you. And do you have any final take home messages for clinicians and also for researchers?

Emily Grist: So to summarize, PTEN inactivity associated with PTEN protein loss and patients with PTEN inactive tumors have shorter survival on ADT and abiraterone and have more docetaxel sensitive tumors.

It's possible that tumors that are high DECIPHER and PTEN inactive are the most sensitive to docetaxel, and they should be tested in prospective trials as a composite biomarker. I think from ASCO and ESMO, we're presenting data that really shows strong data for using transcriptional biomarkers for stratifying patients to treatments. And there's clear data to show utility level 1 B evidence for DECIPHER. Again, PTEN needs to be tested prospectively.

And so then my take home message is and I really want to thank STAMPEDE participants and all their families and more than 11,000 patients now and friends that supported them, and all the site staff. You've got to remember, STAMPEDE ran over 100 sites with 3,000 staff members, and without which we wouldn't have this trial or this translational data.

And also funders, particularly PC UK, John Black Charitable Foundation, and the Prostate Cancer Foundation, who have funded me to conduct this work. Wouldn't have been possible without you, so thank you.

Andrea Miyahira: Well, thank you so much, Dr. Grist, for sharing this study with us. It really teaches us a lot about how powerful correlative research can be in teaching us something new, and also helping to guide treatment for patients. So thank you.

Emily Grist: Thank you.