ProPSMA Versus PreRP: Value of the Prospective Literature “Presentation” - Thomas Hope

February 8, 2024

At the 2024 UCSF-UCLA PSMA Conference, Thomas Hope explores the impact of PSMA PET imaging on prostate cancer care, emphasizing the ProPSMA trial's comparison of PSMA PET to conventional imaging and the PreRP trial's pursuit of FDA approval, highlighting their roles in enhancing diagnosis and treatment. Dr. Hope concludes with a patient case to illustrate how these studies, through their design, findings, and clinical relevance, collectively advance prostate cancer management by improving detection and informing treatment strategies.

UCLAUCSF video logo stroke


Thomas Hope, MD, Professor of Radiology, UCSF Vice Chair, Clinical Operations and Strategy, Director of Molecular Therapy, Chief of Nuclear Medicine, SFVAMC, San Francisco, CA

Read the Full Video Transcript

Thomas Hope: Okay, here we go. PSMA in the initial staging setting. I think what I want to do in this talk is just compare and contrast ProPSMA, and I called it PreRP because that's our initial staging trial, but really the whole category of registrational studies that were performed in the US in the initial staging setting, and think about what are the differences and what's the value of both of them.

I'm going to start off talking about ProPSMA. And this is, I think, one of the most impressive diagnostic imaging studies to date in this field. It's a randomized study. Over 300 patients were randomized. 10 centers. I don't really go into this, but the Australian group did a really remarkable job harmonizing all of their PET/CTs, harmonizing everything. The SUVs were relatable and everything across all the sites. Just a really impressive study across multiple centers.

And what they were looking at here is actually the difference between conventional imaging and PSMA PET. A very different question than what all the other trials were looking at. And probably the main reason for this is, this was not done to get a drug approved. This was done because it's an academic question. This is actually answering probably the most important clinical academic question of, is this or not better than conventional imaging.

When I say the word conventional imaging here, we're talking about a bone scan or a CT scan. Each patient would be randomized to get either CT and bone scan. Half the patients would receive that first and then the other half got a PSMA PET. And then there was the option for crossover, so depending on the results of those imaging studies, you could get conventional imaging or a PSMA PET. And then they did a six-month follow-up in those patients to determine what was true positive, false positive, et cetera.

They had a composite endpoint to determine the sensitivity and specificity in essence. But the minority of patients actually had pathology available to determine sensitivity and specificity. I think maybe the most important thing here is this bottom right hand. This got into Lancet. For our field to get into Lancet, that's pretty unusual, and I think that's a huge statement for our field. And really a statement about the value of trials like this.

This took a lot of work, a lot of planning, and that work and planning really brings us out of our... When we're in nuclear medicine, usually we're in journals that have impact factors of 10, something like that. This gets us up there and people start reading it. And this really brought PSMA PET into the mainstream. This is the trial that really got us out of our retrospective trial data and was really critical for our field.

These are the results here. This is sensitivity and specificity, again, looking at that, compared to the composite endpoint. On the far right there, the red dots are the sensitivity for PSMA PET. The blue dots are the sensitivity for bone scan. And you can see obviously a marked improvement in the sensitivity of PSMA PET compared to bone scan. And that's obviously the main point of this.

In the middle column, you see the specificity. And the important thing here is oftentimes when you improve sensitivity, you actually trade off specificity, sort of a false positive. If you start over-calling things, you might increase your sensitivity, but you'll lose on the specificity side. But here in this trial, you can see actually the specificity is still, if equal to, slightly improved with PSMA PET. Not only is there a marked improvement in sensitivity, but there's also an improvement in specificity.

That's, I think, really important. This is true across different locations of disease. You have pelvic nodal disease and distant metastases. And I think this is where this trial is really critically important is the look at distant metastases. All of the initial staging studies are focused on pelvic nodes because that's what we're going to compare to at the time of surgery. This study is looking much more at the change in management, which is my next slide, and also metastatic disease outside of the pelvis.

The next slide here is the impact on the change in management. I think it's actually pretty interesting here. If you look at the top part there, so this is any metastatic disease. Blue, again, is PSMA PET. Red is conventional imaging. There is a markedly higher change in management if there's metastatic disease, but in pelvic nodes, it's not statistically different, the change in management, if you notice that there are pelvic nodes. And that's because pelvic nodes we're going to hear about and there's going to be a debate about this is, how does that actually alter what you do? Do you not do a radical prostatectomy if the pelvic nodes are there or not? There's actually much less change in management depending on whether or not the patient had pelvic nodes.

Actually, you can see this big change in the change in management, particularly in the patients who have metastatic disease. And you can see here in the bottom part, they broke this up between if it's a first-line imaging versus second-line. Remember how you could cross over? If you crossed over and got a PSMA PET, there was very little change in management. If you cross over to get a bone scan, but a markedly higher rate of change in management, and that's still in patients who had negative conventional imaging upfront. You can see that even a crossover you have a marked improvement in change in management.

This is the data that really is the foundation of being able to state just straight up, PSMA PET is better than conventional imaging. And really critical for us moving PSMA PET before conventional imaging. And actually now in all of our guidelines to say, not just should it go in front of conventional imaging, but we should not get conventional imaging at all. We don't require negative conventional imaging in our guidelines. This should just replace conventional imaging altogether in patients at initial staging. Really, a critical set of data for our understanding of PSMA PET.

I think the other thing I want to point out here is not many patients underwent radical prostatectomy. There were 126 men in that population who underwent radical prostatectomy. Only 83 of them actually had pelvic nodal dissection. This was all done off-trial. There was no requirement to do a pelvic nodal dissection or anything like that. This is just looking at patients over time, and only 14 of the patients had pelvic nodes at pathology. The vast minority had actually pelvic nodes, which doesn't really give you any statistical power between the population to understand sensitivity and specificity with a histopathology background.

That's not the goal of this trial. Again, the goal is the change in management in comparison to conventional imaging. And those are oddly perfectly evenly distributed between both cohorts. Only seven patients who got a PSMA PET had pelvic nodes at pathology.

Okay, so this is the preRP study, and this is very similar to the other studies out there, which we'll talk about in a minute. We had very similar inclusion criteria. Not identical. ProPSMA required a higher PSA to get into the trial. We were including unfavorable intermediate risk and high risk, very high-risk patients as well. There was no randomization. Everyone got a PSMA PET. And then we had a surgical cohort. And you can see actually in our trial, not a lot of patients overall got a radical prostatectomy. We enrolled 764 patients and only 277 of them underwent surgery. And that does reflect where this trial came from.

Back in 2015, when we started designing this trial, it was more of an access trial. No one had access to it, and we had to run clinical trials to give access to patients. People would say that they were going to get surgery in order to get onto the study, but if they had mets or they already knew they had mets, they were just trying to get the PSMA PET for improved imaging. And a lot of the patients, and most patients who didn't undergo surgery, underwent radiation treatment, so people were getting the PSMA PET to plan for their radiation therapy.

But we had 277 patients in this cohort who underwent radical prostatectomy, and 40 of them were N1 positive, which is consistent with the literature out there. This was our trial, and the primary endpoint here is to look at sensitivity and specificity. You have to remember when this was coming through, we were pairing this with the BCR trial. We saw the OSPREY trial that was going on, but we actually did something very different. I don't know if any of you remember the original implementation of the OSPREY trial before it was changed. The primary endpoint for the initial staging cohort was the detection of tumor in the prostate. Not at all a clinically relevant endpoint. That's when we changed over to being sensitivity and specificity in pelvic nodes.

And then you had the BCR component of that trial, which required you to have a metastatic lesion that was biopsiable. They had an average PSA in the first BCR cohort of CONDOR of 10. Whereas here, we weren't requiring that. We wanted to look in lower PSA populations, and we had to create this composite endpoint similar to what ProPSMA did as a validation of metastatic disease because the vast majority of patients would have a pathology in the BCR cohort. This was really our path foundation of PSMA PET. Do we see stuff and is that stuff real compared to pathology? We were using this initial staging cohort to prove pathologically that what we're seeing is real, if that makes sense.

This is really the main cohort of the data here, and we had three blinded readers of note. ProPSMA didn't have blinded readers. It wasn't a reader variability type of study. It was more looking at change in management in sensitivity and specificity. Here we had three blind readers, and you can see our sensitivity and specificity there in the fourth and third bottom rows there about 0.4.

Now note that our sensitivity did not meet our primary endpoint. Our primary endpoint was created... Actually, this whole trial came from many of the speakers here. Wolfgang Fendler, Matthias Eiber were at UCLA with Ken Harriman. And when we were designing the endpoints, we actually designed it based on Matthias Eiber's paper. And he had, I think, a 68% sensitivity in your first paper right out of the tomb. And that's a really high number. And we were like, oh, we're going to hit that. And clearly, that was not true, but the FDA still approved this based on this because it still is impactful on patient management. You'll see as time goes on, these numbers are actually probably right. When you get blinded readers, you don't know the outcomes, this is the actual sensitivity of this imaging study.

Now, a couple of things we followed up on. One thing to note is that in patients who have lower uptake in the primary tumor, they tend to have a lower sensitivity for pelvic nodes. Not surprising. Lower uptake in the primary tumor, you're going to have lower uptake in your nodal metastases, and you can see our sensitivity falls as the uptake goes down. If your primary tumor has uptake less than or equal to the blood pool, the sensitivity falls down to 25%. If it's above that of the prod, it's 47%. Okay? A big impact on the uptake in the primary tumor based on the sensitivity for pelvic nodal metastasis.

Now, the other thing I would say is it's not always related, the uptake in the primary tumor to the uptake in the metastases. In this patient, we have pretty low uptake in the primary tumor of six, but you can see the bones are really hot, and the pelvic nodes actually have low uptake. What's really interesting is there are often differences between the primary tumor and the metastases. And I think as Dr. Fendler mentioned, sometimes you should pay attention to that, is it heterogeneous or not. But in the initial staging setting, it's just for different reasons. If disease is in the bone or in the nodes, you can have different uptake than you get in the primary tumor. And this was a patient we had just this week that I threw in there.

Now, what's interesting is the initial staging studies then went on, and the design that we came up with through the SNMMI and in collaboration with UCLA then led to changes in design in the OSPREY trial. They changed the primary endpoint of that trial and redefined and did a post hoc analysis of it, and then obviously from there on, then we have LIGHTHOUSE. And these are all somewhat similarly designed studies. But there are some really important things here, which is the inclusion of intermediate unfavorable risk prostate cancer.

The OSPREY trial did not include intermediate unfavorable, whereas LIGHTHOUSE and the UCSF trial did include unfavorable intermediate. And that might impact your sensitivity. If your population has a lower chance of having pelvic nodes, then you're going to lower your sensitivity overall. Your sensitivity is going to be dependent on the patient population you're imaging.
Now, I really did appreciate that OSPREY and our trial had the exact same sensitivity. It was very nice. It made it very easy to say there's no difference. It gets more complicated. Unfortunately, the LIGHTHOUSE trial was lower. That's probably nothing to do with the fact it's better or worse. It's just they're all the same. But the trial design is different, which is leading to a different sensitivity number. I just wish it had come out as 40, then it would make everyone's life easier, and we could just say it's all 40 and walk away with it. But unfortunately, that wasn't the case in the LIGHTHOUSE trial. From my perspective, all of these should be considered equal in terms of the detection of metastatic disease, both pelvic nodal and distant metastasis.

Okay, now, what are the main issues here I think that are important? One, intent. Right? Why was our trial different? Our trial is different because we tried to get a drug approved. When you go to the FDA, you can't go with change in management as your primary endpoint. You can't go with metastatic disease in terms of sensitivity and specificity the composite endpoint, so we had to do a correlation to pathology. And that's really why those last three trials were designed. And so that's why we focused on sensitivity.

I think the funding was also a really important thing for our trial in particular. It was, I call it crowdfunding. This was a fully unfunded trial. No funding actually at all whatsoever for our initial trials. And it all came from patients who would pay out of pocket for the dose. Somewhere around $1,000 for the PSMA radiopharmaceutical was changed over time based on how many doses we made. It was a self-funded study, and that's why we didn't do it as a proprietary agent. We did it generically, and so other companies could reference it, etc. And now it's obviously available with multiple versions on the market.

We also were focusing, obviously, on the surgical cohort. Right? Because we wanted the pathology to get it approved, and that was also with blinded reads. We had to have the blinded reads by the FDA because if you don't do blinded reads, you end up with higher sensitivity. All literature out there that doesn't have blinded reads has higher sensitivity. And this is true across all imaging agents. There are a lot of people out there who will say, oh, this agent has higher sensitivity, look at this retrospective data. It's higher than it is in the CONDOR study. That's not true. They didn't do blinded reads. You cannot compare sensitivity without doing blinded reads, and I think that's really important.

Okay, and then just to highlight inter-reader agreement. Ours is a little lower, and the reason why it's lower is when you have small pelvic nodes, there's lower inter-reader agreement for small pelvic nodes versus metastatic disease, which is easier to detect.
One other thing here is specificity. I just want to throw this in there. We published this paper last year. Bone scans have an incredibly low specificity. If you take patients' initial staging with a positive bone scan, only 43% of them actually have a positive PSMA PET. Okay? This just goes down to the point of we should get rid of bone scans in terms of localizing disease in patients' initial staging. Key takeaway point.

And the last slide here, I just want to take this case. This is a patient at initial staging. Right? This patient would've undergone a radical prostatectomy before PSMA PET. A poster child for why PSMA PETs are important. We see supraclavicular nodal disease, mediastinal nodal disease. Surgery would've been futile on this patient. Okay? Now, what's important in this? Frankly, both are. ProPSMA and RP. Right?

ProPSMA is the one that's going to tell you what's relevant. Does it change management? Do we see metastatic disease? Is that metastatic disease real or not? ProPSMA is the one that's telling us about the clinical impact of the imaging. But the other three trials are critically important because we wouldn't have any PSMA PET imaging agents if we didn't do those registrational trials. You had to do the registrational trials to get the drug approved. The availability of the drug having two different branded PSMA 11 PET radiotracers, and the subsequent companies that use our trial design to get drugs approved were really critical to getting these agents available in the United States.

Both of these trials were critically important in helping patients like this get their correct treatment to properly and optimally treat these patients. ProPSMA shows us a comparison to conventional imaging and the change in management on PSMA PET. The preRP study was really focused on sensitivity and specificity and is what led to drug approval was focused on FDA endpoints to get these drugs FDA approved. Lots of people were involved in this over time, particularly the UCLA cohort. Really the greatest collaboration of my career was collaborating on these trials to get PSMA PET approved.

Thank you very much.