PSMA PET/CT Imaging for Staging High-risk Prostate Cancer Prior to Curative-intent Surgery or Radiotherapy (proPSMA) - Michael Hofman and Declan Murphy
Declan Murphy, MB, BCh, BaO, FRACS, FRCS Urol, Professor, Urologist & Director of GU Oncology, Peter MacCallum Cancer Centre, Associate Editor, BJUI, Honorary Clinical Professor, The University of Melbourne
Michael Hofman, MBBS (Hons), FRACP, FAANMS, Professor Michael Hofman is a nuclear medicine physician and physician-scientist. Peter MacCallum Cancer Center, Victoria, Australia
Alicia Morgans, MD, MPH Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.
Alicia Morgans: Hi, I am so excited to have here with me today, Dr. Declan Murphy and Dr. Michael Hofman, two professors from the Peter MacCallum Cancer Center in Melbourne, Australia, also with the University of Melbourne there in Australia. I am so excited to have you here to talk with me today about the ProPSMA trial and the results that are being presented at EAU 2020 as well as being published in the Lancet this year.
So Michael and Declan, can you tell me a little bit about why this trial, why now, what are the issues that we have as clinicians trying to understand and stage patients with high risk localized disease and why was this work so important?
Declan Murphy: Well, thank you Alicia and I'm good day from Melbourne. It's very nice to be connecting with you all today. The reason we did this study was, first of all, we know that patients with high-risk localized prostate cancer are at higher risk of biochemical occurrence or biochemical persistence following treatment of the primary, whether that's surgery or radiotherapy. And we all know that one of the reasons why patients have a higher rate of biochemical recurrence is that standard conventional imaging fails to identify the true state of disease in these patients.
We know that many of these patients will have lymph node disease or even distant disease that's not apparent on conventional imaging using CT bone scan and even MRI scan. So that's the background to the patient population that we studied in this study. And here in Australia and in a number of other countries around the world, we have had really quite a lot of encouraging experience using novel imaging, PSMA PET/CT to be precise over the past five or six years.
And we already know from retrospective or community experience actually in Australia that PSMA PET/CT appears to have better performance characteristics in staging these higher-risk patients. So with that in mind, we designed this study and before we go into the detail of that, I'll hand over to Michael who leads our PSMA PET/CT program here at Peter MacCallum Cancer Center to just explain a little about PSMA PET/CT.
Michael Hofman: Thank you. So PSMA is prostate-specific membrane antigen. This is a cell surface glycoprotein overexpressed in prostate cancer. And to do the scan, we injected intravenously a small radiolabeled peptide that binds to this receptor. It's then taken up into the prostate cancer cells and we put the patient on the PET scanner, a positron emission tomography scanner. And this enables us to image the whole body and see where that radiation has gone to.
And it's combined at the same time with a CT on the PET/CT scanner so that we can see the anatomic background. And this allows us to localize both the primary prostate cancer and any distant metastatic disease in a way that bone scan and CT have never really been able to do before. It's a very sensitive test. We can detect deposits down to around three millimeters in size. And with the CT we can localize very precisely where that abnormality is.
Alicia Morgans: Wonderful. So as we think about this, particularly from the United States where we don't have routine availability of things like PSMA PET/CTs. I always think about how hard it might be to replace our conventional imaging strategy. So our CTs and our bone scans with something new. There's going to have to be a bar that we meet that shows that a new imaging strategy is at least as good, if not better, than the strategy that we already have.
And it seems that this study was ... really one of its goals, was really to help to define that or to meet that bar, at least to help us understand if it could meet that bar. But I'd love to hear about the study and how did you design it? What were its goals and objectives and what were the arms in the trial?
Michael Hofman: Yeah so this is a multicenter, randomized study conducted at 10 large centers around Australia. These were centers that had nuclear medicine facilities that could do the gallium PSMA PET scans, but also multidisciplinary uro-oncology teams with both urologic surgeons and radiation oncologists treating these patients. And what we did is the primary aim was to see whether the PSMA PET/CT was more accurate than the combined findings of conventional staging. That being bone scan and CT and we randomized patients to receive either the PSMA PET or the current standard of care imaging, bone scan or CT.
Then there was a selective crossover because we did not want to deny patients with a normal CT and bone scan access to this new technology. So patients with fewer than three unequivocal distant metastases crossed over and had the other imaging arm. And then importantly, we followed patients up carefully for six months and at six months patients selectively had repeat imaging and then the principal investigators at each site used all the information up until that six-month time point, including the repeat imaging to define the ground truth.
That may have been histopathologic evaluation through a biopsy of bone metastasis or it might've been the fact that a bone metastasis went from being absent on CT to a sclerotic lesion at six months combined with appropriate changes in PSA. So we had a predefined criteria system that enabled us to define the ground truth and this enabled us to very cleanly define the accuracy of the staging investigations and accurately compare at bone scan to CT.
We also had a number of key secondary outcome measures, the main one being management impact. The referring physicians were asked to record their management intent prior to randomization, then after first-line imaging, after second-line imaging so we could see whether these results resulted in a change in patient management. We also looked at the difference in equivocal results between the two arms. These are the number of uncertain findings as the result of a scan report.
We looked at the radiation dose of the two imaging and modalities between the two arms. And we also looked at findings in the patients that crossed over to second-line and that gave us an assessment of the incremental accuracy of both conventional imaging or PSMA PET after first-line imaging.
Alicia Morgans: Wonderful. So can you tell me and just remind us specifically who were these patients? We talked about high-risk patients at the start, but I think this was high-risk patients. And how many did you end up enrolling in each of your arms?
Declan Murphy: So the patients we selected for this study were newly diagnosed patients with high-risk features. So predominantly, this is around what we consider to be traditional high-risk features by D'Amico or NCCN or EAU criteria. So PSA greater than 20 or clinically T3 cancer or ISUP grade group, four or five, which are the classic high risks. But we also chose to include ISUP group three so that's old fashioned Gleason 4+3=7. Because we're aware within that higher intermediate risk category, these men have a higher chance of biochemical recurrence. So we thought it would be of value to include these higher intermediate risks as well as the classic high risks.
So these were newly diagnosed patients. They were allowed to have PSA levels and up to 100 and in other words, they were being considered for curative-intent treatment either with surgery or radiotherapy. And as Michael pointed out, management intent was a key secondary endpoint of the study. So once we diagnosed these patients that had their biopsy, we hadn't done any staging, we then recorded the management intent, whether that may be surgery or radiotherapy or so on. And then we submit them to the study and I'll hand over to Michael and he'll describe at the study schema, which we can show to you as well on the screens.
Michael Hofman: Yeah, so this was a 300 patient study. We randomized 302 patients, 152 in the conventional arm and the 150 in the PSMA PET/CT arm. They were followed up at six months. We had very few drop-offs. We had two patients in the conventional arm and three patients in the PSMA arm. And that means we were able to analyze the primary endpoint in 150 patients in the conventional arm and 145 patients in the PSMA PET arm and the results, in summary, the reference standard showed that 30% of these men had nodal or distant metastatic disease. And what we showed is that PSMA PET/CT had a 27% greater accuracy than conventional imaging. So that's an accuracy of 92% for PSMA PET compared to 65 for bone scan CT. And this is one scan, the PSMA PET compared to two scans or two visits, the CT and the bone scan. So we must remember that the advantage of a PSMA PET is that it's a one-stop-shop.
And we also broke that down into nodal or distant metastatic subgroups. So end-stage or end-stage and PSMA PET was as similarly superior. And we also performed a sensitivity analysis where equivocal lesions were considered positive for metastatic disease rather than at negative. And the superiority of PSMA PET persisted in that sensitivity analysis as well. So I think the take-home message here is that the primary endpoint of accuracy showed a large difference in favor of PSMA PET/CT.
Alicia Morgans: So really interesting. And Declan, as an urologic oncologist, I'd love to hear your thoughts on those patients who ended up going forward with surgery. Because presumably some of them would have. Were there patients who had nodal positivity, for example, at the time of surgery that were missed by the PSMA or what was the false positive and false negative rate at the time of surgery? Because I think that this is something that people talk about, "Oh, PSMA could have a lot of false positives or certainly miss things if they're too small." So what was your experience?
Declan Murphy: Yeah, so yes, a significant proportion of these patients did go on and have surgery. So, therefore, the radical prostatectomy and lymph nodes become the ground truth, if you like, for those patients. But let me tell you, as we've seen in other retrospective series and in big systematic reviews that we've done on experience of PSMA PET/CT. We see first of all that the specificity is extremely high, well over 90% in other words, if you see high avidity in a lesion, a lymph node for example, on PSMA PET/CT, it's almost always right.
And that's a really important point. I think that ProPSMA will help emphasize previous data predominantly retrospective open to selection bias, but in this prospective study, we show with very tight control of multicenter on the PET imaging, we show very high specificity. If these things light up, it's almost always right.
On the other hand, on the sensitivity side, we always know that every imaging modality, even advanced molecular imaging like PSMA PET/CT will always be incapable of detecting tiny lesions below a certain threshold. But as Michael mentioned at the start, we do see extraordinary sensitivity once lesions get to about three millimeters in size or thereabouts compared to, for example, as we know with conventional imaging with CT and we cannot define a node as being positive until it gets to 10 millimeters in size.
You know that is the threshold because of the lack of specificity in conventional imaging. But what we see is once you're down to, once you're at about three millimeters or above, you will see avidity in many lesions, below that you will not. So a PSMA PET/CT from a surgeon's perspective will never replace a lymph node dissection if you really value seeing whether these nodes are positive or not.
But as we will see in the management impact from this study, this imaging has moved the conversation forward for a lot of clinicians managing these patients. So what do you do in your high-risk patient who has a negative PSMA PET/CT for nodes, do we still need to go and do a node dissection? Well, if you really value the staging, many will still do a node dissection, but we also know that lymph node dissection does not improve survival in high-risk prostate cancer and those lead to complications in a proportion of patients, again by current EAU guidelines.
And I think what we see in Australia and other countries where there's a lot of PSMA PET/CT is people will counsel their patients and say, "Well, look your PET/CT is negative for the nodes." If we did a node dissection, a proportion of patients will still have microscopic disease, but it's not going to improve your survival and it does lead to an increase in complications. And I believe that we will see less lymph node dissection in patients who have negative PET scans and whether that's right or wrong.
And one of the other aspects about novel imaging for staging these patients, Alicia, is that you do see lymph nodes in places that are outside the standard template for an extended pelvic lymph node dissection. And this explains to me why these patients do not have a survival benefit. It's because we are seeing lymph nodes in the presacral spaces, the mesorectal spaces, left supraclavicular lymph nodes and so on without necessarily having nodes in the classic internal-external iliac regions there for example.
So I think the management impact is very significant as Michael will come to in a second when we speak about this, but an obvious flow on from that, which I'm sure we'll discuss is does this improve outcomes for our patients? It certainly has higher accuracy as we've just demonstrated. It has a high management impact. But this study is not designed to show whether that will improve longer-term outcomes for our patients.
Alicia Morgans: And maybe the next study will be an adaptive surgical approach around the PSMA findings. And you can see if you do a dissection in the mesorectal area or if you address more distant disease either by not operating on them or by using SBRT for something that perhaps you can change outcomes. But the first step may be in identifying the extent of disease and its location. So certainly this seems to be a step in the right direction.
And I'd love to hear from Michael, from a nuclear medicine perspective. This was a group of patients who all had a PSA that was greater than or equal to 20 nanograms per milliliter or as I understand, although that, please correct me if I'm wrong. If patients have high-risk features by other criteria, perhaps they didn't need to have a PSA that high. But if you're trying to apply this data in a clinical space, you may have a PSA that's lower. What is your PSA threshold for thinking about using PSMA PET in this setting?
Michael Hofman: Yeah, so in this study, you did not have to have a PSA greater than 20 which was one of the measures of high-risk disease. But you might just have had Gleason grade group three, four or five or clinical T stage three or greater.
Alicia Morgans: Very good. But if you were applying this data, I usually think about this cutoff of around two or sometimes people say down to one or 1.5. Where is your cutoff when you're thinking about a PSMA PET? For those of us who don't have access to these scanners at this point, where do you see the highest accuracy? And I think we all recognize and you should tell us that this is a curve, right? So there's no absolute value of PSA that's going to be the cutoff above which or below which will have an accurate scan. But where do you think about things being generally more accurate versus less?
Michael Hofman: Yeah, look, it's a great question. In this study, most of the patients are actually selected on the basis of histopathology, so the Gleason grade group and PSA, therefore, did not play a part. So we have not actually teased out some of these subgroup analyses, but I think your PSA question is a very interesting one to go back and look at the data and see if there's a threshold below, which even despite having a Gleason grade group three perhaps if your PSA is 1.5, do you need a PSMA PET? These are some further analyses that we can look at.
Declan Murphy: But I think we could say to clinicians out there, Alicia, that if you're looking at this data and trying to think how does it apply to my practice or how might it apply if I have access to PSMA PET/CT? I would make two points.
First of all, they had these inclusion criteria. That's what we use in practice nowadays. So if you have these high-risk or higher intermediate-risk, in other words, that grade group three at least in 4+3 at diagnosis, then what this trial shows is first of all, and the accuracy is much higher and if you have a PSMA PET/CT, so that's our routine staging, we might as well get that accurate information.
And the second point, which will I hope have global implications is that we see this trial as being a vehicle to help get PSMA PET/CT standardized around the world and reimbursed around the world. Unlike a lot of the oncology trials that readout as positive endpoints, the survival benefit, et cetera. But sometimes the drug is so expensive in countries it's really not going to be impactful.
PSMA PET/CT, at least in our experience is not an expensive imaging tool. And we see this as a platform for people all around the world to work with their local nuclear medicine people with their regulators, with their funding bodies to say, "Hey, come on, the cost, we'll show this in this study of doing a PSMA PET/CT" for us is about equivalent to doing a CT and a bone scan in our environment.
So why would we do that when we have higher accuracy and, and perhaps Michael, we'll talk about the equivocal studies and the radiation dose, which were important secondary endpoints, which I think feed into an argument that we would like to see carried on around the world about making this scan available on a widespread basis?
Michael Hofman: Yeah, and I think it's worth adding that a key feature of this study is that we assist PSMA PET as a replacement for CT and bone scan. So many imaging studies for new tests look at the additional value of adding a new test. This study was saying, "Well, if you were scheduled for a CT and bone scan" and not everyone needs imaging. If you're low risk, you don't need imaging. But if you are a patient that needs a CT and bone scan by current guidelines, then you are better off having a PSMA PET as singles test rather than two tests." Instead, you no longer need the CT and the bone scan.
And just want to go back and focus on some of the secondary objectives of the study. So the primary objective being accuracy, but then management impact. We looked at management impact and we defined a medium or high impact as a change in therapeutic modality. So let's say surgery to hormone therapy or a change within the therapy. So for example, Declan's looked at the scan and is now going to extend his pelvic nodal dissection a little bit or the radiation oncologist is going to change his radiation field. And that's encompassed a significant management change. And that occurred in 15% of patients who underwent a CT in bone scan and 27% in patients with a PSMA PET/CT. So that's almost like doubling of the high to medium management impact of the study.
Another key secondary objective was looking at equivocal or uncertain findings between the imaging arms because no one likes to get a bone scan or CT scan back saying that there is an equivocal abnormality, it could be a metastasis, but we really don't know. And the conventional arm had equivocal findings in 23% so almost a quarter of scans versus PSMA PET that had equivocal findings in 7%, so that's a significant reduction and that's attributed to the high tumor contrast.
These tumors light up very brightly on a PSMA PET scan. It's easy for the readers to look at these scans. We also looked at reporter agreement, so all the PSMA PET scans were read a second time by a core imaging laboratory of experts and we compared it to the local review at the 10 sites around Australia. And it's important that PSMA PET was found to have a very high reporter agreement, a Kappa score of .88 so experts were agreeing amongst each other. This is really important because it's very hard to have an accurate test if five people look at the scan and come up with different conclusions.
And the last key secondary endpoint that I'd like to mention was radiation dose. Now because these are often well men that are going to be cured and we don't want to expose them to large amounts of radiation from diagnostic imaging and PET scanning might sound a little bit scary to people out there who aren't familiar with the technology, but in fact, the PSMA PET scan approach exposed men to half the radiation of a CT and bone scan. So that was eight millisieverts that's for a PSMA PET/CT compared to 19 millisieverts with a combination of a CT and bone scan.
Alicia Morgans: So when we think about it from a nuclear medicine perspective in terms of rolling out technology, there's high agreement between the local readers and these experts suggesting that with some training this is something that can be disseminated or implemented at sites around the world with really good accuracy. And for the patients to understand the importance of radiation. They all know it anyway, but they would actually receive half of the dose of radiation with this assessment as compared to the combination of bone scans and CTs that we actually re routinely use.
That's very, very critically important information. So as we wrap this up, and this is again, phenomenal work, gentlemen and to your entire teams. Because this has been something that I think the world has been waiting for, to think about trying to broaden the spectrum of applicability of PSMA PET. What would your take-home message be to audiences around the world as they're trying to think about how do we get this technology? How do we use your data in our day to day practices?
Declan Murphy: Well, thank you. Alicia. I'll give you my perspective on it as someone who's worked closely with Michael and Rod Hicks and his other colleagues in here and at other centers is I think there's a great opportunity space for clinicians like me and you working in prostate cancer to buddy up with talented nuclear medicine physicians all around the world. And often there's a bit of a disconnect there. Before all this came along, nuc med didn't come into our tumor board into our multidisciplinary team meetings.
But when all the PSMA PET started arriving, we started saying, "Hey folks, come and sit in our meeting. We'd all learn from each other." And now we have this really productive environment that's leading to lots of trials and also trials in therapies like lutetium PSMA. So I say to people out there, if you don't have a close working relationship with the nuc meds in your hospital, you should go down and find where they live. A knock on their door and say, "Hello, I'm a prostate cancer clinician, are you interested in PSMA PET?" And they're going to go, "Oh yes, I've just been to a conference and everyone's talking about PSMA PET" or "Do you have any prostate cancer patients?".
Because that community is very interested just as we are as prostate cancer clinicians and my messages to encourage people to go and work with each other. Encourage yourselves to share the tumor boards together or have virtual meetings together because I think there's a great opportunity space as this trial shows that for us to hopefully make a difference for our patients by collaborating more with our nuc med physicians.
Michael Hofman: Yeah. I share that key message as part of this study we set up a Tencent network around Australia with really multidisciplinary expertise. That network's now being used for several further studies both in the PSMA PET imaging space, but also in lutetium PSMA therapy space. And maybe I'll leave you with the key message of the trial, which is really that the data that we've produced here as support PSMA PET/CT as a replacement to the current standard of CT and bone scan, and we think this is beneficial for patients as a one-stop-shop investigation and also improving that patient navigation through the imaging pathway with fewer hospital visits and less radiation dose, more accurate imaging.
Alicia Morgans: Wonderful. Well, I sincerely appreciate your time. I know that everyone viewing and listening will be on the lookout for that Lancet paper, which is coming out in March, which is really going to detail the findings from this ProPSMA Phase III trial that looked at PSMA PET/CT as a replacement for CT and bone scan in the high-risk localized prostate cancer population. And certainly, I'm sure that they will be on the lookout for your presentation at EAU.
Thank you both so much for your time and again, congratulations on this excellent work.
Declan Murphy: Thank you.
Michael Hofman: Thank you so much.