There could only be a minimal amount of pretreatment. No more than 45 days of ADT or ARPI was allowed prior to randomization. So we're trying to get these patients upfront, but it's impossible, of course, to get them at the very, very beginning. So many of the patients had already started ADT, and some had started an ARPI as well. One-to-one randomization between PSMA-617 lutetium, 7.4 gigabecquerels, Q6 weeks times six. In combination with ADT and an ARPI of choice versus an ADT and ARPI of choice. RPFS primary with a crossover. Very important crossover allowed here after the blinded independent central review indicated that an RPFS event had occurred. Obviously, OS and other endpoints are going to be followed. Primary endpoint, of course, RPFS, key secondary OS, and a whole slew of secondary endpoints, some of which I'll cover, some of which I will not. The data here is from the second interim RPFS. Primary efficacy, 74% information fraction for the RPFS. The OS, only 47% of the information fraction. There's subsequent analysis to follow.
That's really critical. This is a relatively immature view of the data. We're all going to agree on that. There are not a lot of events that have occurred yet, and there are many, many more events that are yet to occur that are going to be critical. With regard to patient disposition, you can look on the left-hand side and see the ARPIs that were assigned by the investigator. Abiraterone, apalutamide, enzalutamide, darolutamide, and a little bit from China on rezvilutamide. Probably mispronounced that one. Nevertheless, a total of 1,500 patients were screened. About 1,420 patients had a PSMA PET. 87% or 1,232 were qualified by the PSMA PET. Interestingly, same as the VISION trial, about 13% were excluded. Randomization of 1,144. Then if you look down a little bit further, you can see the radiographic progressive disease by the blinded independent central review is only 19.6% in the lutetium arm, and in the hormonal therapy alone arm is 26.6%. Very immature data in many ways. If we look at the number of cycles, the vast majority of patients, 85% have received six cycles.
Critically important to crossover, which I mentioned before, but I'll mention again, this is potential confounding on the overall survival, because we all know that this lutetium compound is active after one ARPI, two ARPIs, an ARPI and a taxane, or an ARPI and two taxanes. This is an active compound regardless of when you give it in the conventional cycle. Critically important to understand. Baseline characteristics, there are a couple that I might want to point out. In terms of high tumor volume using CHAARTED criteria, about two thirds of the patients were high tumor volume. De novo was about 50%, so 50% recurrent. Remember, this is occurring over 19 countries, 176 sites, or really broad sampling. I don't think there are any of the baseline characteristics that are unbalanced between the arms, which is always important. Primary endpoint RPFS was met. You can see the hazard ratio of 0.72 confidence intervals that do not cross one. Please remember this is not a final RPFS. This is occurring with about three quarters of the information fraction that was planned. But regardless, we have more to understand about RPFS. I do think that the splitting of the curves looks nice. I'm anticipating that this will improve a little bit over time. I don't think it'll get worse.
Of course, I could be wrong. RPFS by subgroups were basically balanced in a nice way, between the individual subgroups. We don't really see outliers in terms of baseline PSA or anything else. You might be able to look down and see that most of the confidence intervals are to the left of the 1.0 line, but nevertheless, there's overlap. These are underpowered subsets. We're all going to know that when you start shrinking down the number of patients. I was a little bit gratified to see that the tumor volume, both high and low, turned out to have essentially the same hazard ratio. I think that's good. I think there's a lot of good here. We're not really seeing outliers in any particular group at this particular time, which is early. OS, obviously immature. Let's look at the number of events. 85 events in the lutetium arm, the hormone therapy-only arm is only 99. That's 17.3%. So we're missing a huge amount of data when it comes to the overall survival. Nevertheless, we have a positive trend at 0.84. Obviously, non-statistically significant, low number of events, but at least the trend is going the right way.
By the way, a very significant number of deaths have occurred after the crossover. And by the way, there's some really bad actors out there who are going to progress and they don't even care if you give the ADT or ARPIs. There's some people who just blow through these hormonal therapies, regardless of them being the best that we have today. Secondary endpoints I thought were interesting. If we look at the complete response by objective criteria using RECIST, we have 57 versus 42%. Time to PSA progression, the hazard ratio of 0.42, that's strongly positive. PSA nadir of less than 0.2, which I think is a pretty good prognostic finding. 87% in the lutetium arm, 74.9 for the hormone only arm. Time to symptomatic skeletal events, very few of these events occurred. Hazard ratio 0.89 there. Time to metastatic CRPC has a ratio of 0.7, not overlapping one. PFS by investigator assessment, not the central review, was 0.64. So a little bit better by the investigators, and we've seen some of these discrepancies before, but nevertheless, I think we're seeing some pretty positive trends here. Overall incidence of adverse events, of course, are going to be a little bit high. At the bottom of this slide, I think there's some important things about leading to discontinuations of PSMA lutetium. About 8% of the patients stopped. That's in marked contrast to some of the other trials that you might have seen along the way.
Capivasertib has a much higher discontinuation rate. Dose reductions only 3.9% dose delays only 12%. So this is very, very, very good in my opinion. Safety findings, dry mouth, fatigue, nausea, hot flashes, of course. There is nausea and vomiting more prevalent in the lutetium arm. You see it here. Health-related quality of life. This is from Mike's presentation, ASCO GU. Thank you, Mike. As it turns out, the hazard ratios are going the wrong way for the FACT-P total score in EQ-5D-5L, but not in a manner that is statistically significant. What Mike presented was a longitudinal assessment of the changes from baseline. And here you began to look at the FACT-P and various subsets, and you don't really see much differences. And by the way, if you go to the EQ-5D-5L, you really don't see that much. So this time to deterioration gives a little bit of a biased picture, in my opinion, with regards to the longitudinal assessment, which I think is a bit more accurate. I don't think the quality of life is really a detriment here. So conclusions. First of all, I want to make it clear that in my mind, this is an immature result. It's favorable. It's positive. The RPFS is positive. And we have a really good control arm. We have ADT and ARPI with abiraterone and apalutamide and enzalutamide.
That's really important. This is a good trial design. We have a positive trend for OS and no really unexpected toxicities. I think we're off to a good start and I hope we have a good finish. Thank you very much for the opportunity to be here today.