JCO Precision Oncology Conversations

JCO PO author Dr. Amar U. Kishan, Professor, Executive Vice Chair, and Chief of Genitourinary Oncology Service in the Department of Radiation Oncology at the University of California, Los Angeles, shares insights into his JCO PO article, “Transcriptomic Profiling of Primary Prostate Cancers and Nonlocalized Disease on Prostate-Specific Membrane Antigen Positron Emission Tomography/Computed Tomography: A Multicenter Retrospective Study.” Host Dr. Rafeh Naqash and Dr. Kishan discuss the relationship between Decipher genomic classifier scores and prostate-specific membrane antigen (PSMA) PET/CT-based metastatic spread.

TRANSCRIPT

Dr. Rafeh Naqash: Hello and welcome to JCO Precision Oncology Conversations, where we bring you engaging conversations with authors of clinically relevant and highly significant JCO articles. I'm your host, Dr. Rafeh Naqash, Assistant Professor at the OU Health Stephenson Cancer Center at the University of Oklahoma. Today we are joined by Dr. Amar Kishan, Executive Vice Chair of the Department of Radiation Oncology at the David Geffen School of Medicine at UCLA and UCLA Jonsson Comprehensive Cancer Center, and also the corresponding and senior author of the JCO Precision Oncology article entitled, “Transcriptomic Profiling of Primary Prostate Cancers and Non Localized Disease on Prostate-Specific Membrane Antigen (PSMA) Positron Emission Tomography/Computed Tomography: A Multicenter Retrospective Study.”

Dr. Kishan, welcome to our podcast and thank you for joining us today.

Dr. Amar Kishan: Thank you so much for that kind introduction and the invitation to be here today.

Dr. Rafeh Naqash: Well, it seems to me that there's a theme that people in the GU space, investigators in the GU space, are very interested in trying to understand risk predictors for prostate cancer. We had somebody, I believe from Huntsman Cancer Center a few months back on a previous podcast, where they were trying to do risk prediction modeling as well. Could you tell us why that's something that the GU community is very interested in? What's the background? Is it because there's no risk prediction approaches currently? And would this somehow influence management in the near future?

Dr. Amar Kishan: Yeah, that's a great question. So, I think this goes back to the point that we're in the era of precision medicine now, and many cancers have these molecular stratification scores and all that. Prostate cancer has lagged a little bit behind in that regard, despite the fact that it's such a common cancer that affects so many people across the country and across the world. So, we do have risk stratification schemes for prostate cancer. These are based off clinical and pathologic variables, like the level of PSA, the size of the tumor on digital rectal examination, now, we're incorporating MRI imaging as well, and then what the cancer looks like under the microscope, the Gleason score. And now there have been revisions to the Gleason score, but it's really kind of the architecture, what the biopsy looks like. And this was kind of developed many, many years ago by Donald Gleason, a pathologist at the VA. What we're not necessarily taking into account routinely is kind of the biology of the cancer per se. You know, what are the molecular drivers? How could that influence ultimate outcome? And that's very important because we have these risk groups, low risk, very low risk, favorable intermediate risk, unfavorable intermediate risk, high risk, very high risk. But within each of those groups, based on the clinical kind of pathological characteristics, there's a huge heterogeneity in outpatients too, and our treatments are effective, but they can be morbid. Putting someone on hormone therapy for an extended period of time has a lot of side effects. Dose escalating radiotherapy or doing surgery and then radiation afterwards, these are big things that have a big impact on the patient, and I think we really need better risk stratification tools to understand who needs intensification and who we can de-escalate treatment for.

Dr. Rafeh Naqash: I think those are absolutely valid points, perhaps not just for prostate cancer, more so for all cancers that we currently treat, especially in the current day and age, where we have a tendency to add more and more therapies, combination therapies for patients, and as you mentioned, risk stratification to help identify high risk versus low risk, where you can de intensify treatment, is of high value from a patient standpoint as well as from a financial toxicity standpoint.

So then, going to this next part of the approach that you used, and from what I understand in this paper, you had the radiological aspect, which is the PSMA PET, which we'll talk about. Then you had the genomic aspect, where you did some genomic risk-based stratification. Then you had the transcriptomic score based on the Decipher score. So, could you go into some of the details, first, for the PSMA PET, when is it used? What is the utilization? What is it based on, the science behind the PSMA PET? And then we can talk about some of the other genomic transcriptomic predictors that you use in this study.

Dr. Amar Kishan: Sure. Absolutely. So, a PSMA PET is an advanced molecular imaging tool. PSMA stands for prostate specific membrane antigen. It's a membrane protein that is expressed on the surface of prostate cancer cells. It is expressed elsewhere in the body as well. The utilization of this for imaging has been a revolution in the staging of prostate cancer, both upfront and in the recurrent setting. We basically had fairly recent approval for PSMA PET being used more routinely in upfront staging and recurrent staging in 2022. Essentially, what this is it gives us an ability to detect whether prostate cancer has spread at a time of diagnosis or try to localize the recurrence. Now, no imaging test is perfect, of course, and a PET has a resolution of about 3 mm. There are questions about the sensitivity of the PET. You get it on a patient with high-risk disease, the PET is negative; you do surgery, there are positive lymph nodes. That can happen, but it's far superior to the tools that we have had before. For instance, beforehand, all we would have is a contrast enhanced CT, bone scan, and MRI. And the sensitivity of those is far below that of a PSMA PET. And that has actually been shown in a randomized trial called the ProPSMA trial out of Australia, where they compared conventional upfront imaging versus PSMA upfront imaging with a crossover design, and there was better detection of disease with the PSMA PET. So that's been a revolution in how we stage prostate cancer.

But I'm sure many of your listeners and others are aware of the concerns. When you get a new test and you're detecting disease that's extra prostatic, for instance, are you seeing truly significant new disease that we do need to change our management for, or are we just seeing stuff that wasn't there before that actually wouldn't impact anything? And what I mean by that is, let's say you're seeing things that would never have made a difference to the patient, but now you're saying they have metastatic disease. You're changing their entire treatment paradigm, all kinds of things like that. There's implications to this that hasn't been fully fleshed out. But very recently, like we're talking in July of 2024, essentially, there was a Lancet Oncology paper that looked at the long-term prognosis of patients who had extra prostatic disease on PSMA PET, judged by something called a PROMISE score, kind of gives a quantification on the volume of disease, the brightness of disease, and they correlated that with long term outcomes. And that was really the first time that we have long term follow up data that this extra prostatic disease on PSMA PET actually is prognostically important. So, we're getting there. I mean, now that it's approved and, in some sense, the cat is out of the bag, patients are coming in asking for a PSMA PET, etc. I'm sure everyone has experienced that, but I think we now do have good evidence that it actually is prognostically important as well.

Dr. Rafeh Naqash: Thank you for that explanation. And again, to put this into context for things that I've seen and that might also help the listeners in other tumors, so, for example, melanoma surveillance tends to be or while on treatment, patients tend to have more PET scans than what you see, maybe in individuals with lung cancer, where you get a baseline PET and then you have follow up CT scan based imaging is that something that you guys have shifted from in the prostate cancer space with the approval for PSMA PET, where follow up imaging, whether patient is on treatment or surveillance imaging, is PSMA PET based?

Dr. Amar Kishan: Yeah, that's a good question. I think there's actually less robust data to support it as a means of treatment response. But in terms of evaluating a recurrence, then, yes, that has become kind of a standard tool. It's very complicated because all of the metrics that we have for, say, a treatment failing are based on conventionally detected metastases or something that shows up on a CT or bone scan. So, again, that question arises if someone is on systemic therapy and then you see something on a PSMA PET, are you going to abandon the therapy that you're on? It technically would be earlier than you would otherwise have done that, or what are you going to do? So, that hasn't been fully fleshed out, but it is used in that circumstance. So, I'd say less for treatment monitoring and more for evaluation of suspected recurrence.

Dr. Rafeh Naqash: Understood. And I'm guessing, as a futuristic approach, somebody out there may perhaps do a trial using PSMA PET based imaging to decide whether treatment change needs to be made or does not need to be made.

Dr. Amar Kishan: Yeah. It is being incorporated into trials as we speak, I think.

Dr. Rafeh Naqash: Now, going to the second part of this paper is the Decipher score. Could you explain what the score is, what its components are, how it's calculated? Is it DNA, is it RNA, is it both combined? Is it tissue based; is it blood based?

Dr. Amar Kishan: Yeah. So, the Decipher is also an approved test now, was approved in 2018. What it is, essentially, and how it's derived is based on the idea originally that patients might have a recurrence after surgery for prostate cancer. And it's just a PSA recurrence. It's this way. It's literally what we call a biochemical recurrence. That patient might not have any problems, whereas other patients with a recurrence might go on to develop metastatic disease. And we didn't have a good way of determining which patient is which. Get back to that prognostic problem that we have. So, some investigators, they looked at men that had radical prostatectomy from 1987 to 2001 at the Mayo Clinic that had archived tissue. They looked at FFPE, or basically paraffin embedded tissue. They extracted the RNA and then did a microarray analysis and looked at transcriptomic signatures and wanted to see, could this discern the patients who had mets, who had clinically significant recurrences from those that didn't? And out of that exercise came the Decipher Genomic Classifier, which basically is based on 22 genes. These are involved with cell proliferation, etc., but it's an RNA-based, tissue-based assay. So, if you wanted to order a Decipher on somebody, you would need to use a biopsy or prostatectomy specimen to do so. Essentially, that the samples, they would take the highest grade, highest Gleason grade specimen, send it to their lab. Their main lab is in California. The company is called Veracyte. And then they will do this RNA express analysis with a microarray and then return a score. The score is 0 to 1. Basically, 0 is the lowest, one is the highest, and it is a way of prognosticating the risk of metastasis. Originally, when you get a Decipher report, it actually will tell you the 5 and 10-year risks of distant metastasis, and we'll quantify that.

Dr. Rafeh Naqash: And you said this is approved or has been approved in 2018. So, is this insurance reimbursable at this point?

Dr. Amar Kishan: Most insurances do, not all, and the criteria for getting it can vary, so we can talk about it, but it was initially developed in this post-op setting. On the basis of a significant amount of validation studies, it has been moved to being used in the upfront setting as well. So, if you look at some of the ongoing NRG trials, for instance, they are stratifying patients based off the upfront Decipher score. And this is based off of validation studies that have been conducted looking at past RTOG trials and other trials. That said, sometimes it is not approved by commercial insurances in the upfront setting, because that wasn't where it was initially validated and derived. But honestly, here in 2024, that's very uncommon. It's much more common that it's approved.

Dr. Rafeh Naqash: Understood. And in your practice, or the medical oncologist practice at your institution or other institutions, is this something that is commonly used for some sort of treatment decision making that you've seen?

Dr. Amar Kishan: Yeah. So, as a radiation oncologist, I do think it's a useful test, because my approach is, if we're talking about adding hormone therapy, for instance, which is oftentimes dominating the conversation, we know that it offers a relative benefit to a lot of patients. We've published on this; others have published on it. Let's say it reduces the chance of metastasis by about 40%. 10-year risk of metastasis has a ratio of 0.6. So, 40% reduction. But if your risk of metastasis is 2%, that benefit is not that much in absolute terms. And we don't historically have a great way of saying, what is your absolute risk of metastasis? And I think Decipher is one tool that does tell us that - it literally gives it on the report. Now, is that a holy grail? Is it 100% accurate? Nothing is 100% accurate. But it does give us some quantification. Then I can go back to the patient and say, yes, you will get a benefit from adding hormone therapy, but you're talking about going from 2% to 1%, and so they can decide if that's worth it to them. Conversely, it could be a situation where they really don't want hormone therapy, but it comes back that their risk of metastasis is 20%, and then there's actually a big absolute benefit. So that's how I use it as a radiation oncologist, and we would use it upfront. Now surgeons, and if I was consulting on a post operative patient, maybe it plays more of a role. And do we need to do post operative radiotherapy on this patient, or do we need to add hormone therapy in the postoperative situation?

From the medical oncology perspective, there are emerging data that may be useful in the choice of systemic therapy for metastatic disease, but that is a little bit earlier in the investigational stage, I would say. So, when I'm working with medical oncologists, it's often still in this localized setting, and typically, do we add hormone therapy or not, and that type of thing.

Dr. Rafeh Naqash: Understood. And from a reporting standpoint, so the Decipher score, I'm guessing it's some sort of a report that comes back to the ordering physician and you basically see the score, it gives you a potential recurrence free survival percentage or a metastasis percentage of what is your risk for having metastasis in the next five years - is that how they generally do it? Because I've personally never seen one, so I'm just curious.

Dr. Amar Kishan: Yeah, essentially, it comes back with a score, a numerical score, again, from 0 to 1, and it will basically give you the five-year risk of distant metastasis. The ten-year risk of distant metastasis. You can request an extended report that provides additional, not as well supported signatures that are out there, like ADT response signature, etc. But those maybe may have been published, but are not clinically validated as much, but the actual Decipher report, which goes to patients too, just has this kind of 5,10-year risk of distant metastasis. They have some estimations on prostate cancer specific mortality as well.

Dr. Rafeh Naqash: Sure. Now, the third part of this project, and correct me if I'm wrong, the grid database of the 265 genomic signature score. From what I understood, this is a different component than the Decipher score. Is that a fair statement?

Dr. Amar Kishan: Yeah. No, that's exactly correct. And that was an exploratory part of this analysis, to be honest. Basically, I think our main focus in the paper was those advances that we've talked about PSMA and Decipher, those happened concurrently. People started developing PSMA PET, people started developing Decipher. And so, what we wanted to understand was, if you have a patient that has extra prosthetic disease on PSMA PET, are those biologically more aggressive cancers, is their Decipher score going to be higher? What can we learn about the biology of this? And we were the first, to my knowledge, where we actually had a large data set of patients that actually received PSMA PETs and Decipher. And that's kind of the gist of the paper. We have patients in the upfront setting, patients in the post radical prostatectomy setting, and we're essentially showing that there is this correlation. In the upfront setting, the odds of extra prosthetic disease are higher for higher Decipher scores, which is kind of maybe validating that this biology is capturing something that's akin to this ability to spread.

And in the post-op setting, because we have time to failure, technically, we can calculate a hazard ratio rather than odds ratio. So, we have a hazard ratio that's significantly associated with an increased risk of spread for patients with higher Decipher. The grid portion, which is the genomic resource information database, was more of an exploratory part where I mentioned the Decipher score is based off this microarray, they're looking at 1.4 million transcripts. Only 22 are part of the Decipher, but you can request the rest of the signature data as well. And so, we wanted to look at other pathways, other signatures that have been published, like looking at DNA repair, neuroendocrine pathway, just to see if we could see any correlations there that's not necessarily as clinically actionable. These are more exploratory. But again, we were trying to just look at whether patients who had non localized disease on their PSMA PET, whether their primary had more aggressive biology. We did see that. So that's kind of loosely speaking things like PTEN loss, androgen receptor, DNA repair, metabolism, neuroendocrine signaling, which are thought to be portenders of aggressive disease. Those pathways were upregulated at the RNA level in patients who had non-localized disease. And that's kind of the take home from that. But I wouldn't say any of that is clinically actionable at this point. It's more kind of defining biology.

Dr. Rafeh Naqash: Some of the interesting correlations that you make here, at least in the figures that we see, you're looking at different local occurrences, nodal metastases, M1A and M1B disease. And one thing that I'm a little curious about is the Decipher score seems to be lower in pelvic nodal metastasis, that is, PSMA PET positive versus local recurrence, which has a slightly higher Decipher score. Is that just because of a sample size difference, or is there a biologically different explanation for that?

Dr. Amar Kishan: Yeah, that's a good point. I would assume that's probably because of a sample size in this case, and it's a little bit complicated. It wasn't statistically different. And it was 0.76 on average for patients with local recurrence and 0.7 for patients with a pelvic nodal metastasis. Well, what I think is interesting is we can maybe think that in this post-op setting the time to failure could have been long in some of these cases. So, it is conceivable that an isolated nodal recurrence 10 years after the surgery, for instance, is not as aggressive a cancer as a local recurrence in a short time after the surgery. And that's not taken into account when you're just looking at median scores like we are in this fox and whiskers plot. But overall, I think what it's suggesting is that there are patients who have more indolent disease. That's actually pretty widespread there. There are pretty indolent cases that have these nodal metastases. So just because you have a nodal metastasis doesn't mean it's an incredibly aggressive cancer, biologically.

Dr. Rafeh Naqash: Now, the exploratory component, as you mentioned, is the grid part where you do look at TP53, which is a cell cycle gene, and higher TP53 associated with worse recurrences, from what I understand. Do you see that just from a cell cycle standpoint? Because from what I, again, see in the paper, there's a couple of other cell cycle related signatures that you're using. Is that just a surrogate for potential Gleason score? Have you guys done any correlations where higher Gleason score is associated with maybe higher cell cycle checkpoint, pathway related alterations and replication stress and DNA damage and perhaps more aggressive cancers?

Dr. Amar Kishan: Yeah, that's a great question. We haven't done that in this paper, but it has been published before that there is this correlation loosely between grade and some of these parameters - so repair, metabolism, androgen receptor signaling. However, it's a very great point that you bring up, which is that it's pretty heterogeneous and that's why we need something like this as opposed to Gleason score. So, you can have Gleason 10 cancer. I mean, that would be pretty uncommon. But within the Gleason 9, at least, which we have published on and looked at, there's a heterogeneity. There are some that are biologically not that aggressive. And the converse Gleason 7, you can have some that are actually biologically aggressive. That's why it may be useful to move away from just the pathological architecture and get a little bit more into some of these pathways.

Dr. Rafeh Naqash: What's the next step here? I know this perhaps isn't ready for primetime. How would you try to emphasize the message in a way that makes it interesting and clinically applicable for your colleagues in the GU community?

Dr. Amar Kishan: Yeah. I think for me, what I would try to emphasize here and what I think is the main takeaway is this is kind of a validation that having extra prostatic disease on PSMA PET is likely suggestive of a more aggressive disease biology. And I think what this stresses to me is the importance of getting a PSMA PET, particularly in patients with high-risk prostate cancer. This isn't always happening. And I think if we see things on a PSMA PET, we really need to consider systemic therapy intensification. And what do I mean by that as a practical point? You have a high-risk prostate cancer patient. You get a PSMA PET, you see an isolated pelvic lymph node. If we believe the results of the study, that's a more aggressive biology likely. Whether we have the Decipher or whether we have genomic signatures, which we may or may not have, maybe that patient should get treated with something like an androgen receptor signaling inhibitor in addition to ADT, more akin to a clinically node positive case. So, intensify the systemic therapy, more aggressive disease. That's how I would incorporate it practically into my practice, that really what we're seeing on the PSMA PET is real. It's a reflection of biology that's aggressive. It's not just some Will Rogers effect where you're upstaging stuff needlessly. I think this is telling us some true biology. So that's kind of what my takeaway would be.

I think future areas of investigation would be, honestly, to try to have a better idea of what's going on in these metastases. So, if you could design a study potentially, where your biopsy some of these and actually do sequencing and understand a little bit more of that. And so, we're looking into stuff like that. But my takeaway for like the everyday clinician would be to try to get a PSMA PET, if you can, and to intensify therapy on the basis of that, or at least consider it, discuss it in a multidisciplinary setting.

Dr. Rafeh Naqash: And I'm guessing somebody out there, perhaps even you, are thinking or planning on doing a ctDNA MRD based correlation here, since that's up and coming in this space.

Dr. Amar Kishan: That is up and coming, I think one of the challenges in prostate cancer is the amount of ctDNA can be low. But yes, you're right, that's certainly things that a lot of us are looking at, too.

Dr. Rafeh Naqash: Excellent. Well, thank you for the science discussion, Dr. Kishan, could you tell us a little bit about yourself, your career trajectory, where you started, what you're doing, and perhaps some advice for early career junior investigators, trainees, things that might have worked for you, that could also work for them as they are progressing in their careers.

Dr. Amar Kishan: Sure. So, yeah, I'm a radiation oncologist at UCLA. I run the prostate cancer radiation program. Clinically. I'm also heavily involved in our research enterprise, so I kind of oversee the clinical and translational research aspect. That's what I do currently. So, I did my residency in radiation oncology at UCLA. Just on a personal note, my wife is from LA, her parents live in LA. We really wanted to stay in LA, so I was fortunate to be able to join the faculty here. I always liked GU oncology, so that was kind of a natural thing for me to kind of go into this position here and try to build the GU program. I've been very fortunate to have great collaborators. My message to students and trainees is to try to reach outside your department for mentorship as well. It's important to have people inside your department who can mentor you. But as a radiation oncologist, I work so closely with urology, so closely with medical oncology that I'm very fortunate to have individuals in those departments who have a vested interest in me and my success as well. I like working with them. It's important to be a team player. If they need help, you help them. If you need help, you ask for help from them. So, I think that's the single biggest thing that I would say to any trainee is don't be intimidated. Please reach outside of your department. Lots of people are willing to help and provide mentorship, and it's helpful to have that perspective. We are in a very multidisciplinary environment and era of practicing medicine.

Dr. Rafeh Naqash: Well, thank you again for those personal insights and especially for submitting your work to JCO PO. And we hope to see more of this work perhaps in the subsequent sessions for JCO PO, and maybe we'll bring you back again. And at that point, the Decipher and the PSMA PET scan will have more data, more implementation in the clinically relevant real-world setting.

Dr. Amar Kishan: Thank you very much. And if I could just give one quick shout out. The first author of this work, which I presented, was Dr. John Nikitas, who is a trainee that works with me here at UCLA a PGY5 resident. So, I do want to give credit to him as well.

Dr. Rafeh Naqash: And John, if you're listening to this hopefully, it's always great to get a shout out from your mentor. Thank you both again for putting in the work and effort to submit this manuscript.

Thank you for listening to JCO Precision Oncology Conversations. Don't forget to give us a rating or review and be sure to subscribe so you never miss an 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 opinions of ASCO. The mention of any product, service, organization, activity, or therapy should not be construed as an ASCO endorsement.

Disclosures

Dr. Kishan

Honoraria Company: Varian Medical Systems, Boston Scientific, Janssen Oncology

Consulting or Advisory Role Company: Janssen, Boston Scientific, Lantheus

Research Funding Company: Janssen , Point Biopharma