But when we lose PTEN, when this is deficient, it creates an alternative proliferative drive for these tumors and it's really complementary to the AR pathway. So, in these tumors, when we block AR and androgen signaling, this PI3 kinase AKT pathway kicks in. It's unregulated and can lead to early disease progression. And so, the idea of targeting this is associated with a poor prognosis. This idea of targeting this with AKT inhibitors has been established as an earlier drug, ipatasertib that was studied in the metastatic castration resistance setting in all comers. And we'll talk about that a little bit later.
But this drug, capivasertib, I'm going to call it capi, is a potent selective inhibitor for all three AKT isoforms and was specifically developed for this setting, this what we're now terming, and I'll be the first to say it, androgen receptor modulation sensitive metastatic prostate cancer. Thank you very much. Okay, so this was the CAPItello-281 study designed. This was really a Herculean effort. PTEN is not the most common defect associated with prostate cancer early on. It gains in its frequency as this disease progresses. And to look at this, we were specifically looking by immunohistochemistry at tissues for 90% or greater loss of PTEN.
So, specifically focusing on the population we thought was most likely to be driven by this AKT pathway. And to do this, we screened over 6,000 patients submitting tumor tissue and centrally reviewed for PTEN status. And we found about 25% met this criteria for PTEN deficiency. Of those, about a thousand were actually enrolled in the study and randomized one-to-one to either capi given 400 milligrams BID, four days on, three days off. So, it's built-in breaks along with abiraterone and ADT, standard dosing versus placebo plus ADT and standard dosing. And these patients were followed for disease progression. Investigator assessed RPFS was the primary endpoint.
We also looked at overall survival and a number of secondary endpoints shown here. Took three and a half years to accrue this study, 6,000 patients. It's started unfortunately right after the pandemic, July of 2020, but we got it done. This was a massive effort by all investigators around the globe. Our primary RPFS data cut off was we had a press release about a year ago, and we have about 18 months on follow-up and we have a final OS that's planned after 522 deaths. So, these are the baseline characteristics and they're basically similar across the board. A couple of things I'll point out here that I think are important. One, we had a lot of bone mets. We had over 90% of patients had bone metastasis.
We had about 20% visceral disease and we had a very high rate of Gleason eight or above. It was almost 80% of patients with Gleason eight, nine, or 10 cancer. So, a pretty high risk and not as much high volume, but certainly high risk disease. This was the primary endpoint. This was the RPFS investigator assessed. And a couple of things to point out here. First, this is a positive endpoint. We hit our benchmark, our hazard ratio of 0.81, our P value of 0.03, statistically significant improvement in the RPFS. You can see our median improvement in RPFS of about seven and a half months. But look at our control arm, our control arm of abiraterone and ADT, 25.7 months.
You just heard somewhere, I think they mentioned around 36 to maybe even 40 months median RPFS and some of these other studies. So, we are seeing a decidedly different and much worse prognosis associated with our standard of care ARPI and ADT therapy in this PTEN selected population. If you look at our subgroup analysis in general, everything favors the capi arm some more or less. But basically we're seeing across the board, a consistency across all subgroups. And we had a number of key secondary endpoints. Overall survival was our first key secondary endpoint, and that has not been reached yet. It's only been 26% events. Remember, we only have 18 months follow up at this point in time, but we already have 26% deaths in this study.
There's a slight hazard ratio favoring capi, but it's very early. This is not statistically significant. And because of that in the hierarchical statistical methodology here, we stopped there, but we just from an exploratory standpoint, looked at a number of other key secondary endpoints shown here. And again, they generally are in favor of the capi arm. We're going to highlight just a few of these on their individual Kaplan-Meier curves, because I think there's a lesson to be learned here. This is the first one. This is the interim overall survival. And as I said, it's immature. We need longer follow up. There's no doubt about it and we'll get that. That's planned to do.
We've got a number of events to get to, but already we're seeing in our control arm 138 deaths. This is symptomatic skeletal event survival. Now this includes deaths, as well as our symptomatic skeletal events, our pathologic fractures, cord compressions, surgery, radiation. These are life-altering events. These are highly symptomatic life-altering events. And we're seeing with that and with these early deaths, a little bit more separation than we did with just OS alone. And so, you can see the hazard ratio here drops down to 0.82. Again, exploratory early, but we're seeing a little bit greater trend. I'll point out here in our placebo arm, we have 176 events at this point in time.
Now we turn to time to castration resistance. So, castration resistance includes radiographic progression-free survival, so the patients that have radiographic progression, as well as PSA progression. So, now we see this curve's getting even a little bit wider and a little bit earlier. Our median time to castration resistance and again, our control arm's 22 months, see a seven-and-a-half month difference. Again, similar to our RPFS population. And if you look again under placebo abi in that upper right, now we've got 230 events. It's almost half the patients here that we're seeing this time to castrate resistance. So, this is happening much earlier than what we're seeing in some of our other study populations.
Now, this is where it really caught my eye, and that is the time to PSA progression because it doesn't look that bad. Here, we see the time to PSA progression. Again, there's a separation here. Again, it's exploratory, but capi's on top. But I think the bigger point here is look at our placebo plus abi arm. There's 82 events I just showed you. There are 231 castration resistant events, but only 82 of them have PSA progression. The majority of patients are developing castration resistance without PSA progression. We don't typically see that and it's a real concern. It's really the first time we've looked prospectively at loss of PTEN in a large population like this. And it's telling us something about this biology.
It's different and it's concerning because disease progression without PSA is maybe a precursor for developing more neuroendocrine-type of phenotypes, androgen-indifferent disease, disease that our drugs don't work as well in. So, this is a population that is really, it's an unmet need. We don't have effective therapies and abiraterone, ADT is not cutting it, this population. Here's the other interesting finding. So, along the way, as the study was going along, again, that ipatasertib study, the other AKT inhibitor I mentioned being developed in metastatic androgen receptor, modulating resistant disease was being done.
And they looked at all comers patients just with metastatic disease, but they also broke it down by PTEN status. And when they did and they saw patients that had greater than 50% loss of PTEN, they started to see more statistically significant RPFS benefit to the drug. When you get down to 90 or a hundred percent, you start to see even greater differences, even greater benefit associated with the drug. So, we looked at our CAPItello subgroups, and this is a post-hoc analysis. So, again, hypothesis generating, but again, this was our original analysis right here, looking at greater than 90% cutoff. Now what I'm going to show you are the CAPItello alone, the capi alone arms.
So, this is actually four curves here. The dark one is the original all population here that was treated with capi, that arm. But then there's the 95% cutoff, the 99% cutoff, and the hundred percent loss of PTEN cutoff. And they're pretty much superimposable. Those are all in light below and you really almost can't distinguish them. So, now back to that 90% cutoff here, this is the capi arm and the placebo arm. And now, I'm going to show you what happens when we look at 95% cutoff with the control arm. And you see it drop down a little bit, and you can see over here, our hazard ratio drops down a little bit. You look at 90%, 99% cutoff, it drops down even more and you look at the hundred percent cutoff and it drops down even further.
So, again, we're seeing a story here that the greater the complete loss of PTEN, the more dependency that tumor has on it, the earlier disease progression that we're seeing. And it really fits with this dual pathway for disease proliferation and growth. If you look at overall survival, we see the same trend. I'm not going to walk you through all four of them. Let me just show you. This is the 90% and this is the hundred percent. And you can see in light there, you can see the gradations there, but it's pretty fine. Again, this is still immature data, right? But again, now we're starting to see hazard ratio down there at a hundred percent 0.77. We're seeing things shifting over.
And again, it's primarily due to the early deaths, the early progression in our control arm. So, that's this population. If you look at our other key secondary endpoints, they all follow the same way. I mean, the numbers get smaller. We don't lose much going to 95%. That's 80% of our population. But then the numbers really drop as we go down to 99 and a hundred percent, but the trends are still there. And then lastly, I just want to mention the safety data here because I think this is a concern for folks. This drug is already FDA-approved and widely used in breast cancer with hormonal therapy, but prostate cancer, this is a new agent for us. It is associated with known side effect profiles. I'll show you those in a minute.
But if you look here overall at our grade three toxicities, placebo and abiraterone was 40%, this was 67%. So, we're taking a jump up in our grade three toxicities, about 27%, about a quarter patients developing some grade three toxicities. But if you look about halfway down here, you'll see discontinuation of the capi/placebo, 18.3% versus 4.8%. It's not that huge. Most of the patients are able to tolerate this drug and tolerate it for a long period of time. If you look at dose reductions, that's more. So, as you go down, dose interruptions I think are 62%, and then dose reductions are 29%. So, again, 70% of patients are able to tolerate the dose full dose, but about 30% are dropping the dose. So, what's causing that?
Okay, these are again, volcano plots here, and you can see things shifted over in favor of capi. In terms of toxicities, really the top five. The top three, diarrhea, hyperglycemia, and rash, these are known toxicities associated with it. They happen early. They happen within the first couple of months. And remember, this is a 4-day on, 3-day off. So, the diarrhea tends to get better over those three days on, and it's off again. So, it's not as constantly chronic for patients, but we do sometimes have to hold a dose, sometimes have to stop. Rash is probably the one we stop the most for. Again, it's not huge amounts of grade three rash here. As you can see, it's about 10% grade three rash.
So, it's mostly grade one, two, but it is something to know about. It's manageable in the vast majority of these patients. And then you can see the hyperglycemia in there as well. Again, can you manage hyperglycemia? Yeah, we have drugs for that. Can we manage non-PSA producing disease progression in this population? Not so easily. So, these are the trade-offs they're known, but I think for a lot of our patients, they're manageable. Anemia, hypokalemia, these are probably the exacerbations. Anemia something I think, again, we know how to manage mostly low grade. The hypokalemia is really an exacerbation of that. Abiraterone probably from the diarrhea, and again, mostly low grade. And then there's the rest of them here.
So, I think in conclusions for this study, we would say that PTEN deficient metastatic androgen receptor modulation sensitive disease has a poor prognosis. And I would say that not only that, but PTEN is a driver, AKT is a driver of disease progression in patients with loss of PTEN. And I think that's an important new finding that is established with this large prospective study. The study met its primary endpoint. It's statistically significant RPFS benefit with capi plus abiraterone, median improvement in the RPFS from 25.7 months to 33.2 months. And the hazard ratio and confidence intervals are shown there. These events, this is really consistent across the secondary endpoints.
They're exploratory, but they're trending all in the same way, including the clinically relevant predefined subgroups. The post-hoc analysis is interesting. It really shows the greater and greater dependency on loss of PTEN when we see greater levels of this. And this is something that's easily quantifiable in our pathology. Immunohistochemistry for PTEN has been well-established. It's routinely done in breast cancer now. It could be routinely done in prostate cancer. I would venture say it should be routinely done in prostate cancer, knowing the clinical significance it carries. And I think the most common grade three AEs of rash, hyperglycemia, diarrhea, these are known, expected, manageable.
Yes, we do have to stop drug in some patients, that's absolutely true, but I think the majority of patients that can get through with this drug. So, I think this combination, I think it represents really a potential first in class targeted treatment for patients with a defined PTEN deficient metastatic disease. It's really our only opportunity to use this agent if it's approved in this setting. I think it's an important unmet need. And even if these results are modest in your mind, they're positive for a patient population that we ultimately really need more drugs for. Just like to acknowledge individually each one, no, I'm not going to go through this, but there's a number of investigators all over the globe doing this really was a global study.
You can see that in light blue, the number of countries represented here, over 6,000 participants. Incredible effort. Thank you. Thank you to the families, the patients, the caregivers, and our whole staff and everything, our steering committee that made this happen. Lastly, this has been published now. It's available in Annals of Oncology. There's a QR code you can scan here for that. We have the supplementary data, plain language summary, and an animal video if you want to see it. No, it's animated video, sorry, associated with this that you can download with this QR code. Thank you all very much. Okay, I think you were up first. Yes.
Cornelia Ding: Hi, I'm Cornelia Ding, UCSF Pathology.
Daniel George: Yes.
Cornelia Ding: Just from pathology's working perspective, I wonder if you could comment on, for example, if my clinician asked me to do a PTEN apply for this medication, but the patient doesn't have a metastasis biopsy, should we recommend our clinician get the metastasis focused to do the PTEN? Or do you think the primary prostate staining will be good enough?
Daniel George: Yes, it is a great question. I think these were vastly done off of the primary biopsy. So, you can absolutely use the primary biopsy from that assessment. And I think it's relevant even in our high risk localized disease, it has prognostic significance. So, I think you can justify it in a lot of situations like that.
Cornelia Ding: And should I comment on the percentage of a cell staining PTEN moving forward? It sounds like there's some significance for treatment. Also comparing to other way like NGS and FISH, is there any benefit, pros and cons between these methods for PTEN status?
Daniel George: Great question. So, yes, I think any more information you give us around the percentage of loss of PTEN I think may be helpful. Again, these are exploratory analyses, but we've now got two separate studies showing that trend. So, I think it would be particularly helpful. In terms of NGS, I think there is concordance with NGS and high loss of PTEN. So, we're going to be looking at that in this study as well going and sequencing. It's just going to take some time. But I think immunohistochemistry is really our gold standard because there are other ways of knocking out PTEN besides just genetic loss. So, I think, and the scalability of this across pathology is a lot easier.
So, at least at this point in time, I think immunohistochemistry is probably the easiest way to adapt it.
Cornelia Ding: Thank you.
Daniel George: Sure.
Michael Morris: Great talk, Dan. Really nice job. I think that in the back of everyone's mind, while we're listening to both Oliver's presentation and yours is chemotherapy, that's the unspoken control arm really that's included in none of these trials. But I guess I would pose to you, given the tox profile, let's say, I'm a PTEN deficient PSMA avid patient, which therapy would you give me recognizing that everything is preliminary, we don't have OS?
Daniel George: Yeah, yeah. So, I think this is a great question because I think this is the limitations of these phase three studies, right? And to me, this is where we need, this is why we have medical affairs to do the studies beyond FDA approval to understand, do you combine those agents? Because we have a study of docetaxel and capi and castrate-resistant disease combining docetaxel and capivasertib. It was a negative study, but the safety is there. So, you can combine those, but that's an unanswered question. Should you sequence them? Should you do six cycles of chemotherapy and then add it? So, you do three doses of Pluvicto and then switch? There's a lot of different ways to address this. And I think to me, it's both.
Yeah, how you do that, I'm not sure. But at the end of the day, this is a bad disease. When I see this, I know this is an accelerated natural history. This is a patient that's going to run into complications sooner. I don't think I can look and say, "You're going to go 10 years, no problem." I don't want to congratulate him on the low PSA response, but I don't feel safe with that. So, these are the kind of patients that if they can, I'm going to want to try to intensify however you want to do that and we just need to do the studies to figure out the best way.
Michael Morris: Thanks, Dan. Congrats.
Daniel George: Yeah, yes.
Speaker 4: Very interesting talk. I have two questions, but hopefully they're both quick. So, first of all, in the patients that were PTEN null or PTEN deficient in the primary tumors, does PTEN get reactivated in the mets? And also, what's the status of AKT activation in both?
Daniel George: Great questions. And I think those are the limitations of what we have for data on. We did not include patients that had AKT mutations and activation, but one could argue that biology should apply here just as well. So, I mean, I think it's less common. PTEN is the most common way that this pathway is altered. But there are PI3 kinase alterations, AKT alterations. To me, that's an extrapolation. But we do that in clinical medicine all the time and I think those would be reasonable. In terms of reactivation the pathway in metastasis and stuff, that's probably a good thing in the sense that maybe those are a little less aggressive, but I wouldn't count on it.
So, to me, if I see this loss in the primary, I'm going to stick with that until proven otherwise.
Speaker 4: And then one other question is, and you're probably going to say no, not yet, but do you have any molecular insights as to the patients who are responding and not responding?
Daniel George: It's a great question because we know there's more to loss than just PTEN, right? And what other features could be there? So, I think with this next-gen sequencing data that we'll get on these samples, we'll have a lot more insights. So, you're absolutely right not yet.
Cora Sternberg: Dan?
Daniel George: Yes. Oh, Cora.
Cora Sternberg: Great data.
Daniel George: Yeah, thank you.
Cora Sternberg: Great, great, great talk. I participated in the ipatasertib trial, where we used PTEN as a cutoff immunochemistry of 50%, and it wasn't a negative trial. One patient did die due to hyperglycemia. And then when we went back again and looked at NGS, those patients that had PTEN loss by NGS was more though all of those patients actually were the ones who benefited. I'm just wondering, now that we know how serious this is, we be recommending that PTEN be monitored and reported either by immunochemistry or by NGS in all patients with metastatic hormone-sensitive prostate cancer?
Daniel George: Absolutely. In fact, our chair of pathology, Zhao Zhang is here and I'm going to grab him and say, "Look, we're going to start this on Monday because I think we need to know this," right? It's going to change how I counsel that patient, how often I scan them when their PSA is undetectable, what I'm looking for, and even when available, what we can intensify with. So, you're absolutely right. I mean, I think we're not going to get more evidence than this. This is about as good as you can get for saying this is an important driver of disease progression.
Cora Sternberg: Did you mention anything about high volume and low volume by the old CHAARTED criteria? Do they have more high volume or I'm not sure that I heard that.
Daniel George: It's interesting. So, this study, and to Mike's point earlier, I mean, as this trial evolved, we had the PEACE-1 and the ARASENS data come out. So, we lost a little bit of equipoise in our high volume patients to say, we can put them on abiraterone and ADT. That's enough. Some of those patients, we preferred to do chemotherapy with. So, we probably preferentially didn't selected out some high volume patients, so it was like 60/40. But I suspect that's probably a sign of the times than the real biology.