NCI Working Group on Biochemically Recurrent Prostate Cancer
Host Dr. Davide Soldato and guests Dr. David Einstein and Dr. Ravi Madan discuss JCO article, "National Cancer Institute's Working Group on Biochemically Recurrent Prostate Cancer: Clinical Trial Design Considerations," underscoring the need for a consensus on clinical trial designs implementing novel endpoints in this population and the importance of PSA doubling time as a prognostic factor, with an emphasis on treatment de-escalation to limit toxicity and improve patient outcomes. TRANSCRIPT The disclosures for guests on this podcast can be found in the show notes. Davide Soldato: Hello and welcome to JCO After Hours, the podcast where we sit down with authors from some of the latest articles published in the Journal of Clinical Oncology. I am your host, Dr. Davide Soldato, medical oncologist at Ospedale San Martino in Genoa, Italy. Today, we are joined by JCO authors Dr. David Einstein and Dr. Ravi Madan. Dr. Einstein is a medical oncologist specializing in genitourinary malignancy working at Beth Israel Deaconess Medical Center, part of the DFCI Cancer Center, and an assistant professor at Harvard Medical School. Dr. Madan is a senior clinician at the National Cancer Institute (NCI), where he focuses on conducting clinical research in prostate cancer, particularly in the field of immunotherapy. Today, we will be discussing the article titled, "National Cancer Institute's Working Group on Biochemically Recurrent Prostate Cancer: Clinical Trial Design Considerations." So, thank you for speaking with us, Dr. Einstein and Dr. Madan. David Einstein: Thanks for having us. This is a great pleasure. Ravi Madan: Appreciate being here. Davide Soldato: So, I just want to start from a very wide angle. And the main question is why did you feel that there was the need to convey a consensus and a working group to talk about this specific topic: biochemically recurrent prostate cancer? What has been the change in current clinical practice and in the trial design that we are seeing nowadays? And so, why was it necessary to convey such a consensus and provide considerations on novel clinical trials? David Einstein: Yeah, so I think it's very interesting, this disease state of biochemically recurrent prostate cancer. It's very different from other disease states in prostate cancer, and we felt that there was a real need to define those differences in clinical trials. Years ago, metastatic castration-resistant prostate cancer was the primary disease state that was explored, and over time, a lot of things shifted earlier to metastatic disease defined on a CAT scan and bone scan to an earlier disease state of metastatic castration-sensitive prostate cancer. And the clinical trial principles from late-stage could be applied to MCSPC as well. However, BCR is very different because the patients are very different. And for those reasons, there are unique considerations, especially in terms of toxicity and treatment intensity, that should be applied to biochemically recurrent prostate cancer as opposed to just using the principles that are used in other disease states. And for that reason, we thought it was very important to delineate some of these considerations in this paper with a group of experts. Davide Soldato: Thanks so much. So, one of the main changes that have been applied in recent years in clinical practice when looking at biochemically recurrent prostate cancer is the use of molecular imaging and particularly of PSMA PET. So, first of all, just a quick question: was the topic of the consensus related on which threshold of PSA to use to order a PET scan to evaluate this kind of patient? David Einstein: Yeah, thanks for that question. It's a super important one. The brief answer is that no, we did not address questions about exactly when clinicians would decide to order scans. We were more concerned with the results of those scans in how you define different disease states. But I think as a broader question, I think a lot of folks feel that finding things on a scan equates that with what we used to find on conventional scans. And fundamentally, we actually sought to redefine that disease space as something that's not equivalent to metastatic disease, and rather coined the term "PSMA-positive BCR" to indicate that traditional BCR prognostic criteria and factors still apply, and that these patients have a distinct natural history from those with more advanced metastatic disease. Ravi Madan: And if I may just add that the National Cancer Institute is running a trial where we're prospectively monitoring PSMA-positive BCR patients. And that data is clearly showing that, much like what we knew about BCR a decade ago, PSMA findings in BCR patients do not change the fact that overall, BCR is an indolent disease state. And the findings, which are usually comprised of five- to seven-millimeter lymph nodes, do not endanger patients or require immediate therapy. And so, while PSMA is a tool that we can be using in this disease state, it doesn't really change the principal approach to how we should manage these patients. And as Dr. Einstein alluded to, there is a drive to create a false equivalency between PSMA-positive BCR and metastatic castration-sensitive prostate cancer, but that is not supported by the data we're accumulating or any of the clinical data as it exists. Davide Soldato: One thing that it's very important and you mentioned in your answer to my question was actually the role of PET scan and conventional imaging, so CAT scan and bone scan that we have used for years to stage patients with metastatic prostate cancer. And you mentioned that there is a distinction among patients who have a positive PET scan and a BCR, and patients who have a positive conventional imaging. And yet, we know that sometimes the findings of the PET scan are not always so clear to interpret. So, I just wanted to understand if the consensus reached an agreement as to when to use conventional imaging to potentially resolve some findings that we have on PET scan among thess patients with BCR? David Einstein: Yeah, I think there's a number of questions actually buried within that question. One of which is: does PSMA PET result in false positives? And the answer has definitely been yes. There's a known issue with false-positive rib lesions. And so, first and foremost, we need to be very careful in calling what truly is suspicious disease and what might actually not be cancer or might be something that is totally separate. So I think that's the first part of the answer to that question. The second is to what extent do we need to use paired PET and conventional imaging to define this disease state? In other words, do you have to have positive findings on one and negative findings on the other in order to enter this definition? The challenge there, as we discussed, is that logistically, oftentimes it's hard to get patients to do multiple sets of scans to actually create that definition. Sometimes it's difficult to get insurers to pay for such scans. And finally, it's hard to sometimes blind radiologists to the results of one scan in reading the other. So, we did have some deliberations about to what extent you could use some of the CAT scan portion of a PSMA PET in order to at least partially define that. We also talked about using bone scans to confirm any bone findings seen on PET. But I think another important part of this is not just the baseline imaging, but also what's going to be done serially on a study in order to define responses and progression. And that's sort of a whole separate conversation about to what extent you can interpret changes in serial PET. Ravi Madan: And just to pick up on the key factor here, I think that the PSMA PET in BCR is pretty good at defining lymph node disease, and that's actually predominantly 80 to 90 percent of the disease seen on these findings. It might be pretty good at also defining other soft tissue findings. The real issues come to bone findings. And one thing the group did not feel was appropriate was to just define only PSMA-positive bone findings confirmed on a CT bone window. There's not really great data on that, but the working group felt that, when in the rare situation, because it is relatively rare, a PSMA-positive finding is in a bone, a bone scan should be done. And it's worth noting that Phu Tran, who is a co-author and a co-leader of this working group, his group has already defined that underlying genomics of conventionally based lesions, such as bone scan, are more aggressive than findings on next-gen imaging, such as PSMA. So, there is also a genomic underlying rationale for defining the difference between what is seen on a PET scan in a bone and what is seen on a bone scan. Davide Soldato: Coming back to this issue of PET PSMA sometimes identifying very small lesions where we don't see any kind of correlates on conventional imaging or where we see only very little alteration on the bone scan or in the CT scan, was there any role that was imagined, for example, for MRI to distinguish this type of findings on the PET scan? Ravi Madan: So, I think that, again, what can be identified on a PSMA frequently cannot be seen on conventional imaging. We didn't feel that it was a requirement to get an MRI or a CT to necessarily confirm the PSMA findings. I think that generally, we have to realize that in this disease state, that questionable lesions are going to be seen on any imaging, including PSMA. We've actually probably put way too much faith in PSMA findings thus far, as Dr. Einstein alluded to with some of the false positives we're seeing. So, I think that these false positives are going to have to be baked into trials. And in terms of clinical practice, it highlights the need to again, not overreact to everything we see and not necessarily need to biopsy everything and put patients' health in jeopardy to delineate a disease that's indolent anyway. Davide Soldato: Thanks so much. That was very clear. So, basically, the main driver was really also the data showing that if we have a BCR, so a patient with a biochemically recurrent disease that is positive on the conventional imaging, this is usually associated with a different aggressiveness of the disease. But coming back to a comment that you made before, Dr. Madan, you said that even if we talk about PSMA-positive BCR, we are still talking about BCR and the same criteria should apply. So, what we have used for years in this space to actually try to stratify the prognosis of patients is the PSA doubling time, so how quickly the PSA rises over time. So, coming back to that comment, was the consensus on the PSA doubling time basically retained as what we were using before, so defining patients with a doubling time less than 12 months, 10 months, 9 months, as patients with a higher risk of progressing in terms of developing metastatic disease? Ravi Madan: Yes, so that's a very important point. And the working group defined high-risk BCR as a PSA doubling time less than six months. And this really comes from Johns Hopkins historical data, which shows that if your doubling time is three months or less, there's about a 67 percent chance of metastasis at five years. If it's between three and six months, it's 50 percent. And if it's over six months, if it's between six and nine months, it's roughly only 27 percent. There are trials that are accruing with eligibility criteria that they may describe as high-risk that are beyond six months, but the data as really it's been defined in the literature highlights that truly high-risk BCR is less than six months. And the working group had a consensus on that opinion, and that was our recommendation. David Einstein: And I think an important follow-on to that is that's regardless of PET findings, right? And so, we present a couple of case studies of patients with positive PET findings who have a long doubling time, in whom the disease is in fact indolent, as you would have expected from a traditional BCR prognostic standpoint. Obviously, there are patients in whom they have fast doubling times, and even if they do not have PET findings, that doesn't make them not high-risk. Ravi Madan: And just to follow up that point, I will let you know a little bit of a free preview that my colleague Melissa Abel from the NCI will be presenting PSMA findings in the context of PSA doubling time at ASCO GU if that data is accepted. Davide Soldato: Looking forward for those data because I think that they're going to clarify a lot of the findings that we have in this specific population. And coming back to one of the points that we made before, so PET PSMA has a very high ability to discriminate also a very low burden of disease, which we currently refer to as oligometastatic biochemically recurrent prostate cancer, which is not entirely defined as an entity. But what we are seeing both in some clinical trials, which use mainly conventional imaging, but also what we're starting to see in clinical practice, is that frequently we use the metastasis-directed therapy to treat these patients. So, just a little bit of a comment on the use of this type of strategy in clinical practice and if the panel thought of including this as, for example, a stratification criteria or mandated in the design of novel clinical trials in the field of BCR? David Einstein: Yeah, I think that's an incredibly important point. You know, fundamentally, there's a lot of heterogeneity in practice where some folks are using local salvage approaches, some are using systemic therapies, in some cases surveillance may be reasonable, or some combination of these different strategies. We certainly have phase two data from multiple trials suggesting that met-directed therapy may help buy patients time off of treatment until subsequent treatments are started. And that in and of itself may be an important goal that we can come back to in discussing novel endpoints. I think what our panel acknowledged was that, in some sense, the clinical practice has gotten even farther ahead than where the data are, and this is being offered pretty routinely to patients in practice. And so, what became clear was that we, in developing clinical trials, cannot forbid investigators from doing something that would be within their usual standard of care, even if it might not be supported by the most robust data. But at minimum, it definitely should be used as a stratification factor, or in some trial designs, you can do met-directed therapy after a primary endpoint is assessed. And that offers a compromise between testing, say, the effect of a systemic therapy but also not excluding patients and investigators from doing what they would have done had they not been on a study. Ravi Madan: And I would just like to follow up your phrasing in the question of "oligometastatic prostate cancer." We have a figure in the paper and it highlights the fact that, unfortunately, that term in prostate cancer is imaging agnostic. And we've already discussed in this podcast, as well as in the paper, that imaging used to define a metastatic lesion, whether it's PSMA or conventional imaging, carries with it a different clinical weight and a different prognosis. So, we feel in the working group, that the correct term for this disease state of PSMA-positive BCR is just that: PSMA-positive BCR. We also have to realize that when we talk about oligometastatic disease, while it's imaging agnostic, it seems to be numerically based, whether it's five or three or 10 depending on the trial. But PSMA-positive BCR does not have a limit in terms of the number of lesions. And so again, we just feel that there is an important need to delineate what we're seeing in this disease state, which again is PSMA-positive BCR, and that should be differentiated frankly from oligometastatic disease defined on other imaging platforms. David Einstein: Right, and that also makes clear that patients can have polyfocal disease on PET that still is not what we would consider metastatic, but goes beyond the traditional definition of oligometastatic. So, in other words, just because someone has PET-detected disease only, that does not automatically equate with oligometastatic. Davide Soldato: Thanks so much. So, you were speaking a little bit, Dr. Einstein, about the different types of treatment that we can propose or not propose to this patient because you mentioned, for example, that in clinical practice MDT, so metastasis-directed therapy, is becoming more and more used. For these patients, we can potentially use systemic treatments, which include androgen deprivation therapy, which can be given continuously or in an intermittent fashion. And recently, we can also use novel systemic therapies, for example, enzalutamide, to treat this type of patient. So, given that the point of the consensus was really to provide consideration for novel clinical trials in this space, what was the opinion on the panel regarding the control arm? So, if we're looking at a novel therapy in the BCR space, does the control arm need to include a therapy or not? And if so, which therapy? David Einstein: Yeah, this is a super important question and one that's subject to a lot of discussion, especially in light of recent data from EMBARK. What we came to a consensus around was the fact that neither MDT nor systemic therapy should be required as a control arm on BCR trials. And we can talk about a number of reasons for that. There's also the pragmatics of what investigators might actually accrue patients to and what they would consider their standard of care, and that's important to factor in, too. I think that one of the major goals of our working group was outlining what kinds of trials we would like to see in the future and where the limitations of the current data stand. For example, EMBARK proposes a strategy of a single treatment discontinuation and resumption at a predefined threshold indefinitely. That's probably not how most people are practicing. Most folks are probably using some version of intermittent therapy as they would have before this trial, but we actually don't have any data supporting that. Moreover, we don't have data comparing different intermittent strategies to one another. We don't know what the right thresholds are, we don't know how much time we buy patients off treatment, and we don't know to what extent MDT modifies that. And so, those are all really important questions to be asking in future versions of these trials. I'd say my second point would be that a lot of drug development is happening with novel therapies that are not hormonal, trying to bring them into this space. And when you think about trying to compare one of those types of therapies to a hormonal therapy on short-term endpoints, the hormonal therapy is always going to win. Hormonal therapy is almost universally effective, it will bring down PSAs, and it will prolong, quote-unquote, "progression." The downside of that is that hormonal therapy doesn't actually modify the disease, it suppresses it, and it tends to have fairly transient effects once you remove it. And so, part of our goal was in trying to figure out some novel endpoints that would allow these novel types of therapies to be examined head-to-head against a more traditional type of hormonal therapy and have some measurement of some of the more long-term impacts. Davide Soldato: So, jumping right into the endpoints, because this is a very relevant and I think very well-constructed part of the paper that you published. Because in the past we have used some of these endpoints, for example, metastasis-free survival, as potentially a proxy for long-term outcomes. But is this the right endpoint to be using right now, especially considering that frequently this outcome is measured using conventional imaging, but we are including in these trials patients who are actually negative on conventional imaging but have a positive PSMA when they enter this type of trial? David Einstein: Yeah, there's a number of challenges with those types of endpoints. One of which is, as you say, we're changing the goalposts a little bit on how we're calling progression. We still don't exactly understand what progression on PET means, and so that's something that is challenging. That said, we're also cognizant of the fact that many times investigators are likely to get PET scans in the setting of rising PSA, and that's going to affect any endpoint that relies purely on conventional imaging. So, there's some tension there between these two different sets of goalposts. One thing that we emphasize is that not only are there some challenges in defining those, but also there're challenges in what matters to a patient. So, if a progression event occurs in the form of a single lesion on a PET scan or even a conventional image, that might be relevant for a clinical trial but might be less relevant for a patient. In other words, that's something that, in the real world, an investigator might use serial rounds of metastasis-directed therapy or intermittent therapy to treat in a way that doesn't have any clinical consequences for the patient necessarily. In other words, they're asymptomatic, it's not the equivalent of a metastatic castration-resistant disease progressing. And so, we also need to be cognizant of the fact that if we choose a single endpoint like PFS, that there's going to be many different versions of progression, some of which probably matter clinically more than others, and some of which are more salvageable by local therapies than others. Ravi Madan: So I think the working group really thoughtfully looked at the different options and underscored perhaps strengths and weaknesses, and I think that's presented as you mentioned in the paper. But I think it's also going to depend on the modality, the approach of the therapeutic intervention. In some cases if it's hormone-based, then maybe PSA is providing some early metrics, maybe metastasis-free survival is more relevant in a continuous therapy, but intermittent therapies might have a different approach. There's emerging immunotherapy strategies, radiopharmaceutical strategies, they might have some more novel strategies as well. I think we have to be open-minded here, but we also have to be very clear: we do not know what progression is on a PSMA scan. Just new lesions may not carry the clinical significance that we think, and we may not know what threshold that ultimately becomes clinically relevant is. So, I do think that there was some caution issued by the working group about using PSMA as an endpoint because we still do not have the data to understand what that modality is telling us. Again, I'm optimistic that the National Cancer Institute's prospective data set that we've been collecting, which has over 130 patients now, will provide some insights in the months and years ahead. Davide Soldato: So, just to ask the question very abruptly, what would you feel like the best endpoint for this type of trials is? I understand that is a little bit related to the type of treatments that we're going to use, whether it's intermittent, whether it's continuous, but do we have something that can encapsulate all of the discussion that we have up until this point? David Einstein: Yeah, so that's a perfect segue to the idea of novel endpoints, which we feel are very important to develop in these novel disease spaces. So, one thing that we discussed was an endpoint called treatment-free survival, which conceptually you can think of as exactly what it sounds like, but statistically you actually have to do some work to get there. And so essentially, you imagine a series of Kaplan-Meier curves overlaid: one about overall survival, one time to next therapy, one time on initial therapy. You can actually then take the area under those curves or between those curves and essentially sum it up using restricted mean survival time analysis. And that can give you a guide about the longitudinal experience of a patient: time spent on treatment versus off treatment; time spent with toxicity versus without toxicity. And importantly, each one of those time-to-event metrics can be adjusted depending on exactly what the protocol is and what is allowed or not allowed and what's prespecified as far as initiation of subsequent therapies. So, we felt that this was a really important endpoint to develop in this disease space because it can really capture that longitudinal aspect. It can really reward treatments that are effective in getting durable responses and getting patients off of therapy, because unfortunately, PFS-based endpoints generally reward more or longer systemic therapy versus shorter or no systemic therapy, and that's sort of an artificial bias in the way those endpoints are constructed. So, I think that there are challenges of course in implementing any new endpoint, and some of the things that are really critical are collecting data about toxicity and about subsequent therapies beyond what a typical trial might collect. But I think in this kind of disease space, that longitudinal aspect is critical because these are really patients who are going to be going through multiple rounds of therapy, going to be going on and off treatments, they're going to be using combinations of local and systemic therapies. And so, any one single endpoint is going to be limited, but I think that really highlights the limitations of using PFS-based endpoints in this space. Ravi Madan: I also think that in the concept of treatment-free survival lies one of the more powerful and, honestly, I was surprised by this, that it was so universally accepted, recommendations from the committee. And that was that the general approach to trials in this space should be a de-escalation of the EMBARK strategy as it's laid out with relatively continuous therapy with one pause. And so, I think again, buried in all of this highlights the need for novel endpoints like treatment-free survival. We get to the fact that these are patients who are not at near-term clinical risk from symptoms of their disease, so de-escalating therapies does not put them at risk. And if you look at, for example, lower-volume metastatic castration-sensitive prostate cancer, it's become realized that we need to de-escalate, and there are now trials being done to look at that. Historically, we know that BCR is an indolent disease process for the vast majority of patients who are not at near-term risk from clinical deterioration. So, therefore, we shouldn't wait a decade into abundant BCR trials to de-escalate. The de-escalation strategy should be from the outset. And that was something the committee really actually universally agreed on. David Einstein: And that de-escalation can really take multiple forms. That could be different strategies for intermittent therapy, different start-stop strategies. It could also mean actually intensifying in the short-term with the goal long-term de-intensification, kind of analogous to kidney cancer where we might use dual checkpoint inhibitors up front with some higher upfront toxicity but with the hope of actually long-term benefit and actually being able to come off treatment and stay in remission. Those kinds of trade-offs are the types of things that are challenging to talk about. There's not a one-size-fits-all answer for every patient. And so, that's why some of these endpoints like treatment-free survival would be really helpful in actually quantifying those trade-offs and allowing each patient to make decisions that are concordant with their own wishes. Davide Soldato: Thanks so much. That was very clear, especially on the part of de-escalation, because, as you were mentioning, I think that we are globally talking about a situation, a clinical situation, where the prognosis can be very good and patients can stay off treatment for a very long period of time without compromising long-term outcomes. And I think that well-constructed de-escalation trials, as you were mentioning and as the consensus endorsed, are really needed in this space also to limit toxicity. This brings us to the end of this episode. So, I would like to thank again Dr. Einstein and Dr. Madan for joining us today. David Einstein: We really appreciate the time and the thought, and I think that even starting these types of discussions is critical. Even just recognizing that this is a unique space is the beginning of the conversation. Ravi Madan: Yeah, and I want to thank JCO for giving us this forum and the opportunity to publish these results and all the expert prostate cancer investigators who were part of this committee. We produced some good thoughts for the future. Davide Soldato: We appreciate you sharing more on your JCO article titled, "National Cancer Institute's Working Group on Biochemically Recurrent Prostate Cancer: Clinical Trial Design Considerations." If you enjoy our show, please leave us a rating and review and be sure to come back for another episode. You can find all ASCO shows at asco.org/podcasts. The purpose of this podcast is to educate and to inform. This is not a substitute for professional medical care and is not intended for use in the diagnosis or treatment of individual conditions. Guests on this podcast express their own opinions, experience, and conclusions. Guest statements on the podcast do not express the opinion of ASCO. The mention of any product, service, organization, activity, or therapy should not be construed as an ASCO endorsement.
