ProBio Trial Demonstrates Improved Outcomes Using Biomarker-Guided Therapy Selection for Metastatic Prostate Cancer - Henrik Grönberg
January 4, 2024
Zachary Klaassen interviews Henrik Grönberg about the ProBio trial, presented at ESMO 2023. The ProBio trial, designed in 2019, is a platform trial aimed at determining the optimal treatment sequence for metastatic prostate cancer and identifying predictive biomarkers for treatment efficacy. Unique to ProBio is its use of circulating tumor DNA as a biomarker for assigning patients to experimental treatments and its implementation of outcome adaptive randomizations. This approach adjusts the probability of treatment allocation based on patient responses, making the trial more efficient and responsive. The trial compares androgen receptor pathway inhibitors (ARSI) like abiraterone or enzalutamide against taxanes and a control arm, revealing superior progression-free and overall survival for ARSI. The trial design allows for the continuous addition of new drugs and combinations, adapting to emerging treatments in oncology. Dr. Grönberg emphasizes the importance of academic-driven trials like ProBio for advancing cancer treatment.
Biographies:
Henrik Grönberg, MD, PhD, Professor of Cancer Epidemiology, Karolinska Institutet, Capio St. Göran Hospital Prostate Cancer Center, Stockholm, Sweden
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Well Star MCG, Georgia Cancer Center, Augusta, GA
Biographies:
Henrik Grönberg, MD, PhD, Professor of Cancer Epidemiology, Karolinska Institutet, Capio St. Göran Hospital Prostate Cancer Center, Stockholm, Sweden
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Well Star MCG, Georgia Cancer Center, Augusta, GA
Read the Full Video Transcript
Zach Klaassen: Hello, my name is Dr. Zach Klaassen, I'm a urologic oncologist at the Georgia Cancer Center in Augusta, Georgia. And I'm joined by Professor Henrik Grönberg at the Karolinska Cancer Institute in Sweden, a medical oncologist. Thanks so much for joining us today.
Henrik Grönberg: Thank you very much.
Zach Klaassen: So, we're going to discuss your presentation at ESMO 2023, looking at the ProBio trial, which was one of the special presentations at the Congress. And just by way of background, what was the genesis for the ProBio trial, and importantly, the trial design itself?
Henrik Grönberg: I think one of the key aspects now, when we are having multiple treatments for metastatic prostate cancer, is to know who to treat with what drug, and what is the optimal sequence of actually treating patients with metastatic prostate cancer? And we also need treatment predictive biomarkers. A biomarker that can tell us before we start treatment if the patient could benefit more from one treatment than another.
One example of that is, of course, the BRCA1 and BRCA2 mutations and PARP inhibitors. So, this is why we designed the ProBio trial in 2019, four years ago, and the design is what you call a platform trial. In which, you have one common control arm, and then you have multiple experimental arms. We have five now in the castrate-resistant metastatic prostate cancer. So, this trial design makes it possible to compare multiple treatments with a common control arm.
There are two other, I think, unique aspects of ProBio1 is that we use circulating tumor DNA as the biomarker for deciding which patient will go to which experimental treatment. So, I think that's very important. It's not a retrospective analysis of biomarkers in a trial; it's actually prospectively using biomarkers in a trial.
Zach Klaassen: Yeah, absolutely. It sounds like it's similar to the STAMPEDE platform trial. Is that a similar design?
Henrik Grönberg: Yeah. The STAMPEDE trial, it's also a platform trial that has been extremely successful, that they've been running now for almost 20 years. A new aspect of ProBio is that we use biomarkers. That's something that the STAMPEDE hasn't been able to do. So this is one of the key things.
The other thing is that we use something that's called outcome adaptive randomizations. And that's something that's been used in covid research, for example, when you have a lot of multiple drugs that you need to very rapidly show which one is effective. But it's also being used in cardiovascular disease. But not in oncology before, and outcome adaptive randomization means that you take each patient and see what biomarker he has and what treatment he has received, and also shows, did he respond or not?
And if you responded to that treatment, the chance of a similar patient coming in with the same biomarker signature to get that treatment increases. Vice versa, if he doesn't respond, the randomization probability could have been decreased. That means that this is a learning process as it goes. So this has been shown to be an extremely effective way of actually running clinical trials. Because if you have, for example, a hypothesis that was wrong from the beginning, you can correct for that during these trials.
Zach Klaassen: Right. Well, it's very unique, and mCRPC is the perfect space to do this, especially with the biomarker. The design is fantastic, so take us through the results of the first iteration of your presentation at ESMO 2023.
Henrik Grönberg: Yeah. About a year ago, we got the first signal that one of the arms was superior to the control arm. Then, of course, our data safety monitoring board told us that I needed to look at the data. And then we stopped that arm, together with another arm that was stopped for futility. And the data I presented at ESMO was comparing the control arm, which is the physician's choice without knowing the biomarkers, compared to the ARSI, the androgen receptor pathway inhibitors, abiraterone or enzalutamide, compared to taxanes, docetaxel.
And what we showed here at ESMO is that both for progression-free survival and overall survival, the androgen receptor pathway inhibitors were superior to taxanes and the control arm. Really, when it comes to progression-free survival, it was four more months compared to the two arms. But that actually translated to almost 17 or 18 months' overall survival benefit in the long run.
Zach Klaassen: How many patients in that first presentation were re-randomized, as you described?
Henrik Grönberg: The number of randomizations in this analysis is about 260 patients that are randomized into this. And you can look at that. Is that enough? Usually, we are used to having 1,000 patients, and 500 in each arm. When you have a platform trial like this, you have what we call “patient statistics,” and it's been shown they're much more effective than the frequentist statistics you have in a fixed control arm.
At least the, what do we call it? The studies that we've done before, that this is enough to actually show superiority. And of course, this trial is not meant to be the definitive answer, but it gives a very strong indication that this most likely is true. Of course, if this was a completely new hypothesis, that ARSIs were superior to taxanes, we must run a validation trial. But there are a lot of other indirect evidences that this is true. At least in my view, it is actually strengthening the analysis or the results we have.
Zach Klaassen: Absolutely. So, I think the ctDNA aspect and the trial design, where does the ProBio trial go from here? How do we think about future trial designs, based off of this trial design and what you guys showed?
Henrik Grönberg: I think there's two things that we do right now. We are adding new drugs into this trial, because we stopped the ARSI and the taxane arms. Now, we have, of course, a combination with the PARP inhibitor and one with the AKT inhibitor. We are looking into getting lutetium into the trial, and to adding new drugs and new combinations into the trials. And as long as there are new drugs, we will continue the trial.
The other thing that I think this trial for the first time shows is that this is most likely the way that we need to evaluate new drugs in oncology. And not only prostate cancer, but... For example, we go from myeloma, where you have 10 different drugs. And we actually really don't know which order or which one we should use, because it's very difficult to do comparative trials. And I think this is the way to do it.
Zach Klaassen: Yeah, absolutely. I think it's exciting to see that you can just keep adding in. Like you said, the PARP inhibitors, lutetium, and whatever comes down the road in the next three to five years. It gets put back into that trial design. Specific to the ctDNA, is there any derivations you guys will be looking towards in terms of measuring response and future re-randomization two or three times down the line?
Henrik Grönberg: Yeah, this is really interesting. I think we can use the ctDNA in two ways. Of course, we can get predictive biomarker signatures. That is obvious. The other thing that we are exploring right now is if we can use ctDNA as a treatment biomarker. We have preliminary data, and other results have shown that if you look at the ctDNA concentration after one and two months of treatment, it can be very predictive of treatment response. So, we are looking into... With the data from ProBio now, we could, as early as after one month, predict whether a patient could respond to treatment or not, and that could be completely a game changer for oncology.
Zach Klaassen: Absolutely.
Henrik Grönberg: You probably know, as an oncologist, that we are... It is a little bit of trial and error. Some of the patients really didn't respond that well, but we continued another two or three months because you never know. But, if you have a good biomarker already after one or two months, most likely it's of no use to continue.
Zach Klaassen: Move on to the next treatment, yeah. That's been a great discussion. Anything else we haven't touched on, and maybe a couple of take-home messages for our listeners?
Henrik Grönberg: I think the take-home message is that if I have a patient with metastatic castrate-resistant prostate cancer and who has not been exposed to an androgen receptor pathway inhibitor, I think you should most likely treat the patient with that first. The other thing is I think we need to have platform trials like this to evaluate multiple treatments at the same time. And the third thing is also that I think it's important that this is an academic-driven trial, with, of course, support from industry, because we need their drugs to try them. But I think it is very important that the academic trials are continuing, because I think we need both. Because these are trials that the industry will never do.
Zach Klaassen: Right. No, that's fantastic, and I think we look forward to more work coming out of the ProBio trial. And it's exciting to hear additional arms being added, so thank you so much for your expertise and for your discussion today.
Henrik Grönberg: Okay, thank you very much.
Zach Klaassen: Thank you.
Zach Klaassen: Hello, my name is Dr. Zach Klaassen, I'm a urologic oncologist at the Georgia Cancer Center in Augusta, Georgia. And I'm joined by Professor Henrik Grönberg at the Karolinska Cancer Institute in Sweden, a medical oncologist. Thanks so much for joining us today.
Henrik Grönberg: Thank you very much.
Zach Klaassen: So, we're going to discuss your presentation at ESMO 2023, looking at the ProBio trial, which was one of the special presentations at the Congress. And just by way of background, what was the genesis for the ProBio trial, and importantly, the trial design itself?
Henrik Grönberg: I think one of the key aspects now, when we are having multiple treatments for metastatic prostate cancer, is to know who to treat with what drug, and what is the optimal sequence of actually treating patients with metastatic prostate cancer? And we also need treatment predictive biomarkers. A biomarker that can tell us before we start treatment if the patient could benefit more from one treatment than another.
One example of that is, of course, the BRCA1 and BRCA2 mutations and PARP inhibitors. So, this is why we designed the ProBio trial in 2019, four years ago, and the design is what you call a platform trial. In which, you have one common control arm, and then you have multiple experimental arms. We have five now in the castrate-resistant metastatic prostate cancer. So, this trial design makes it possible to compare multiple treatments with a common control arm.
There are two other, I think, unique aspects of ProBio1 is that we use circulating tumor DNA as the biomarker for deciding which patient will go to which experimental treatment. So, I think that's very important. It's not a retrospective analysis of biomarkers in a trial; it's actually prospectively using biomarkers in a trial.
Zach Klaassen: Yeah, absolutely. It sounds like it's similar to the STAMPEDE platform trial. Is that a similar design?
Henrik Grönberg: Yeah. The STAMPEDE trial, it's also a platform trial that has been extremely successful, that they've been running now for almost 20 years. A new aspect of ProBio is that we use biomarkers. That's something that the STAMPEDE hasn't been able to do. So this is one of the key things.
The other thing is that we use something that's called outcome adaptive randomizations. And that's something that's been used in covid research, for example, when you have a lot of multiple drugs that you need to very rapidly show which one is effective. But it's also being used in cardiovascular disease. But not in oncology before, and outcome adaptive randomization means that you take each patient and see what biomarker he has and what treatment he has received, and also shows, did he respond or not?
And if you responded to that treatment, the chance of a similar patient coming in with the same biomarker signature to get that treatment increases. Vice versa, if he doesn't respond, the randomization probability could have been decreased. That means that this is a learning process as it goes. So this has been shown to be an extremely effective way of actually running clinical trials. Because if you have, for example, a hypothesis that was wrong from the beginning, you can correct for that during these trials.
Zach Klaassen: Right. Well, it's very unique, and mCRPC is the perfect space to do this, especially with the biomarker. The design is fantastic, so take us through the results of the first iteration of your presentation at ESMO 2023.
Henrik Grönberg: Yeah. About a year ago, we got the first signal that one of the arms was superior to the control arm. Then, of course, our data safety monitoring board told us that I needed to look at the data. And then we stopped that arm, together with another arm that was stopped for futility. And the data I presented at ESMO was comparing the control arm, which is the physician's choice without knowing the biomarkers, compared to the ARSI, the androgen receptor pathway inhibitors, abiraterone or enzalutamide, compared to taxanes, docetaxel.
And what we showed here at ESMO is that both for progression-free survival and overall survival, the androgen receptor pathway inhibitors were superior to taxanes and the control arm. Really, when it comes to progression-free survival, it was four more months compared to the two arms. But that actually translated to almost 17 or 18 months' overall survival benefit in the long run.
Zach Klaassen: How many patients in that first presentation were re-randomized, as you described?
Henrik Grönberg: The number of randomizations in this analysis is about 260 patients that are randomized into this. And you can look at that. Is that enough? Usually, we are used to having 1,000 patients, and 500 in each arm. When you have a platform trial like this, you have what we call “patient statistics,” and it's been shown they're much more effective than the frequentist statistics you have in a fixed control arm.
At least the, what do we call it? The studies that we've done before, that this is enough to actually show superiority. And of course, this trial is not meant to be the definitive answer, but it gives a very strong indication that this most likely is true. Of course, if this was a completely new hypothesis, that ARSIs were superior to taxanes, we must run a validation trial. But there are a lot of other indirect evidences that this is true. At least in my view, it is actually strengthening the analysis or the results we have.
Zach Klaassen: Absolutely. So, I think the ctDNA aspect and the trial design, where does the ProBio trial go from here? How do we think about future trial designs, based off of this trial design and what you guys showed?
Henrik Grönberg: I think there's two things that we do right now. We are adding new drugs into this trial, because we stopped the ARSI and the taxane arms. Now, we have, of course, a combination with the PARP inhibitor and one with the AKT inhibitor. We are looking into getting lutetium into the trial, and to adding new drugs and new combinations into the trials. And as long as there are new drugs, we will continue the trial.
The other thing that I think this trial for the first time shows is that this is most likely the way that we need to evaluate new drugs in oncology. And not only prostate cancer, but... For example, we go from myeloma, where you have 10 different drugs. And we actually really don't know which order or which one we should use, because it's very difficult to do comparative trials. And I think this is the way to do it.
Zach Klaassen: Yeah, absolutely. I think it's exciting to see that you can just keep adding in. Like you said, the PARP inhibitors, lutetium, and whatever comes down the road in the next three to five years. It gets put back into that trial design. Specific to the ctDNA, is there any derivations you guys will be looking towards in terms of measuring response and future re-randomization two or three times down the line?
Henrik Grönberg: Yeah, this is really interesting. I think we can use the ctDNA in two ways. Of course, we can get predictive biomarker signatures. That is obvious. The other thing that we are exploring right now is if we can use ctDNA as a treatment biomarker. We have preliminary data, and other results have shown that if you look at the ctDNA concentration after one and two months of treatment, it can be very predictive of treatment response. So, we are looking into... With the data from ProBio now, we could, as early as after one month, predict whether a patient could respond to treatment or not, and that could be completely a game changer for oncology.
Zach Klaassen: Absolutely.
Henrik Grönberg: You probably know, as an oncologist, that we are... It is a little bit of trial and error. Some of the patients really didn't respond that well, but we continued another two or three months because you never know. But, if you have a good biomarker already after one or two months, most likely it's of no use to continue.
Zach Klaassen: Move on to the next treatment, yeah. That's been a great discussion. Anything else we haven't touched on, and maybe a couple of take-home messages for our listeners?
Henrik Grönberg: I think the take-home message is that if I have a patient with metastatic castrate-resistant prostate cancer and who has not been exposed to an androgen receptor pathway inhibitor, I think you should most likely treat the patient with that first. The other thing is I think we need to have platform trials like this to evaluate multiple treatments at the same time. And the third thing is also that I think it's important that this is an academic-driven trial, with, of course, support from industry, because we need their drugs to try them. But I think it is very important that the academic trials are continuing, because I think we need both. Because these are trials that the industry will never do.
Zach Klaassen: Right. No, that's fantastic, and I think we look forward to more work coming out of the ProBio trial. And it's exciting to hear additional arms being added, so thank you so much for your expertise and for your discussion today.
Henrik Grönberg: Okay, thank you very much.
Zach Klaassen: Thank you.