Is it heterogeneity? Is it radioresistance? Can we turn it around with Alphas? Hopefully we'll hear a bit more about that. And then can we overcome some of this resistance by combination strategies looking at different mechanisms of action, as well as different targets? I put this slide up, and in fact, this is not a promise, but it actually is coming very, very soon that we have a lot of therapies in the clinic at the minute. And so there is the ability to look at combination strategies that are rational. AR signaling, DNA repair pathways, chemotherapy, ADCs, immune checkpoints, perhaps the next-generation T-cell engagers. And then I think we do need to think about the stroma in this disease, which unfortunately we are very far behind. So firstly, leveraging androgen receptor signaling. This is an obvious pathway. And this is wonderful work done by my colleague, Louise Emmett, really showing that in the context of castration-resistant prostate cancer, androgen receptor pathway inhibitors really quite rapidly upregulate PSMA expression. And so by virtue of that, you can exploit that biology. You get greater PSMA expression and consequently a greater payload of radiation being delivered to the tumor. This is some single-cell work from my lab, really showing a dichotomy between AR expression and PSMA expression. And you can see across different sites in an individual patient, it is pretty heterogeneous.
And so, combining therapies are useful from the purpose of both targeting PSMA-positive and -negative disease and AR-positive and -negative, but at the same time, also with that upregulation of PSMA. To this end, Louise showed that when you give an ARPI first, you cause that upregulation of the PSMA expression, and then you can come in with the radiation payload, in this case, targeting PSMA. And then that translates into improvement in RPFS as well as overall survival, which is ultimately what we want. This forms the rationale for the PSMAddition study as well that was presented this morning. But I put up this slide to highlight the fact that disease is pretty heterogeneous. What we see here is in primary tissue, in the hormone-sensitive setting, in the metastatic hormone-sensitive disease, and then in the castrate-resistant using our rapid autopsy cohort, you can see that there's considerable overlap in many of these targets in the individual deposits of sites, but you can also see there is also considerable heterogeneity. I make the point of saying here that actually in the context of neuroendocrine, we don't see a lot of PSMA, but we do see DLL3. But it's interesting that we do see B7H3, which I think is quite an interesting target, particularly for the more aggressive phenotype.
So key lessons, co-targeting of AR signaling is definitely an effective strategy. And beyond ARPIs in the CRPC space, I would expect that we'll start to see combinations with AR degraders and RIPTACs, which look really promising. When considering this co-targeting strategy, some key considerations are the level of expression, heterogeneity across sites, does target expression change with the disease state, hormone-sensitive versus CRPC, and where to position these treatments within the therapeutic paradigm of this disease? Is the target expression being modulated by androgen receptor signaling most relevant for PSMA, perhaps less relevant for B7H3 and STEAP1 and STEAP2? But I remain to see some data in that space. And then what happens to the tumor microenvironment? And I think we are really falling short as a community, in the field really, because we don't really understand what these therapies do to the microenvironment. Especially when you start at the onset with these therapies, it completely modulates what else you can or can't do down the track. So first, then looking at DNA damage response and exploiting this, we know that lutetium PSMA delivers both single-strand and double-strand breaks, and perhaps if you block one of the DNA repair pathways, in this case with the PARP inhibitor, you could potentially convert some of those single strands into double strands, causing more cell death, and therefore more activity.
We looked at testing out this hypothesis in the context of a LuPARP study, and essentially showed that the PSA-50 response was 69%, so a bit better, but it was the PSA-90 that is most promising at 52%, so a deepening of the responses. In the subset of patients receiving treatment at the day minus four, so making sure that that PARP was at maximum effect when the lutetium was delivered, we can see that the PSA-90 is in the range of 52, 55%, and in fact, all the patients with RECIST measurable disease had a complete or a partial response. How does this translate into PFS? RPFS was 12 months was 52%, and the median RPFS was 12.8 months. It translates into improvement of overall survival, 33 months for the combination strategy. Yes, it comes at a cost of additional myelosuppression, but actually manageable with the intermittent dose schedule. So my key lessons from targeting the DNA repair pathway is you can target the DNA repair pathway to achieve radiosensitization. It's viable. We clearly see efficacy from the combination. Most of the benefit is clearly being delivered by the radioligand, but it can be deepened, and it translates into a deeper, greater proportion of PSA 90s that translates into PFS and OS benefits. However, does it overcome radioresistance? The answer to that is no. Full blockade of DDR I think is important on day one if you want to really exploit that DNA breaks, and continuous dosing in the context of DNA repair response pathway agents may not be tenable. And certainly, we've done the experiment and showed that that's the case. Potential overlapping toxicities will need to be managed, so intermittent dosing schedules are important. And two concepts to consider, we are starting to molecularly stratify the disease.
I would almost argue that in the context of a DNA repair defect, why not have a induction strategy followed by a maintenance strategy? Then moving on to leveraging immunotherapy, both conventional ICI and perhaps T-cell engagers, why do we keep coming back to immunotherapy? Because it's been such a game changer in other diseases. And I think as clinicians, we'd like to see this sort of promise read out in prostate cancer. But of course all of you know, it's been a challenge. Radiation is thought to modulate the immune microenvironment. There's certainly a lot of preclinical and some clinical data to support that. But the fact of the matter is radiation can be both immunomodulatory in a positive or a negative way. And we've got a lot to learn from a point of view of how it does it in the setting of this form of radiation, because the reality is different doses, schedules, models, and associated immune checkpoints are relevant. So we undertook a phase 1b clinical trial looking at multiple doses of lutetium in conjunction with the backbone of pembrolizumab, reporting a PSA-50 response rate of 76% and objective response rate of 70%. And here's some PSA PFS, median of 11.2 months, 12 months RPFS of 40%, and 24 months RPFS of 16%. And OS, likewise, 84% at 12 months and 49% at two years. Landmark endpoints are important in the context of immunotherapy.
Then taking that experiment further, looking at the role of CTLA-4, targeting of CTLA-4 in conjunction with lutetium PSMA, we looked at IPI and NIVO in one arm in conjunction with lutetium versus lutetium alone. Again, showing an improvement in the landmark endpoints, PSA PFS at 12 months, 31% versus 17%. At two years, 15% versus 3.3%. And likewise, that's also reflected in the RPFS at two years, 18% versus 7.7%. Intriguingly, 16% of patients had PSAs of less than 0.2, and approximately 16% of patients were able to stop treatment and remain having suspended treatment at two years, and all of those patients fell into the combination group. This is the overall survival data. It's not been presented before, and you can see for the combination arm, 46% versus 31%. But as I said, there is a subset that remain responding. I think what's more exciting, and you'll hear more about this is the rationale of combining with T-cell engagers, maybe radioligands could result in effective debulking of the tumor, which is important for T-cell engagement efficacy, perhaps immunogenic cell killing and release the tumor antigens across sites, leverage different mechanism of action and targets, which could be highly advantageous, potential for greater retention of T-cell engagers potentially, no overlapping hematological toxicity, but many unknowns around scheduling impact on the tumor microenvironment.
Key lessons, most of the benefit from combination strategies, again, delivered by the radioligand therapy, IO toxicities to be anticipated, but no obvious overlapping toxicities. Is alpha better than beta? I hope our next speaker will tell us this. And issues around timing of agents really need to be considered, maybe concurrent with ICIs. Don't know for T-cell engagers. There appears to be a combination effect with improved RPFS in OS landmarks, however, likely to be a subset of patient. Contribution of parts, CTLA-4 probably contributes a little bit, and future efforts really looking at radioligands in combination with T-cell engagers. And then chemotherapy, this is a upfront study previously presented looking at the combination of docetaxel, lutetium to debulk followed by docetaxel showing clearly a benefit, and further studies looking at cabazitaxel and lutetium due to be presented at ESMO, I think. And then I asked the audience a question about the next-generation cytotoxins, ADCs, and quite interestingly targeting B7H3. So again, coming back to that rationale of targeting different surface markers. I might end there.