The Impact of Unequivocal Clinical Progression on Survival Outcomes in Metastatic Castration-Resistant Prostate Cancer - Arpit Rao
April 5, 2021
Arpit Rao, MD, MBBS, Assistant Professor of Medicine, Division of Hematology, Oncology, and Transplantation, University of Minnesota, Minneapolis, Minnesota, USA.
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. My name is Alicia Morgans, and I'm a GU medical oncologist and Associate Professor of Medicine at Northwestern University. I am so excited to have here with me today, Dr. Arpit Rao, who's an Assistant Professor of Medicine and a GU medical oncologist at the University of Minnesota. Thank you so much for being here with me, Dr. Rao.
Arpit Rao: Thanks, Dr. Morgan. It's great to be back.
Alicia Morgans: Wonderful. So I want to just speak with you a little bit, Arpit, about a recent publication that you had, I think a really clinically pertinent one, on the concept of unequivocal clinical progression that is within some data that was analyzed for COUGAR-302 looking at abiraterone in the metastatic CRPC setting. Can you tell us a little bit about the study?
Arpit Rao: Yeah, absolutely. So, I want to set a stage for many of our colleagues who don't routinely do prostate cancer. There are certain unique parts of management that you and I are very familiar with. For example, prostate cancer is bone-predominant or bone-exclusive metastatic in a majority of patients. And then there's significant heterogeneity in the biology.
We also have trouble, and we quite often struggle, with the low sensitivity of our conventional imaging, which is CAT scan or bone scans, and, especially, things that happen during the treatment course, like false-positive findings, like bone flare phenomena, that are rather unique to prostate cancer care and make longitudinal radiographic response assessment a little bit challenging.
And then, PSA, which is the next best thing in clinical practice, again, it wasn't developed to be a longitudinal marker for response assessment. So, it, as is, is difficult, but then it becomes less reliable in mCRPC. And, so, I've seen, and I'm sure you have too, a good number of patients, or a significant proportion of patients, who get switched based on PSA progression, and then they run out of many of the effective treatment options.
And, so, the counterpart to that is, that there are men who start developing symptoms, more pain, subtle things, decline in performance status, or have one or two new lesions that are causing imminent risk for cord compression and things like that, needing surgery or radiation. And those are clinically important events. We know they are relevant. The problem is that the trial progression assessments don't really capture them. That was the clinical issue that we went after with this paper.
Alicia Morgans: I completely agree with you and really think that that's such an important entity. In fact, on other clinical trials, things like the CARD trial, for example, there was a portion of patients, I think over half of those patients, who actually had what sounds like unequivocal clinical progression or progression on pain as entry criteria for the trial. So it is so clinically meaningful, and certainly meaningful even as we do our studies. So, can you tell us a little bit about the population included? What were you looking for and what did you find?
Arpit Rao: Yeah, so we took this really amazing, rigorous data set from the COUGAR-302 study and did a post-hoc analysis. So, just as a reminder, 302 was a phase three randomized study, abiraterone versus prednisone, in men with pre-docetaxel metastatic CRPC. And, so, we took that data set. And the one unique part, or which was unique back then, is that patients were allowed to be on treatment even after serologic, so PSA, or radiographic progression, on the 302 study.
So that allowed for a design where we could actually track the biology and natural trajectory of these patients after a radiographic progression, while they were still on abi or prednisone. And, so, that's really what we tapped into.
The way UCP was defined was another important piece of this trial. And, so, there were three criteria: either ECOG performance status of 3 or greater, so if you decline that much, or the need to start chemotherapy, radiation, or surgery for a cancer-related complication, or cancer pain that required chronic opiate analgesia, which, as you obviously know, that doesn't happen in most of the men who have mCRPC. That really is a major shift in their overall clinical profile, when they start needing that chronic opiate.
So, this definition's directly clinically relevant for people like us and clinicians out in the community, because these are the kind of triggers that they need to be looking out for.
Alicia Morgans: So, how common did you find this unequivocal clinical progression within that COUGAR-302 data set, which, as you mentioned, is a Phase III trial population. So maybe sometimes, in some settings, more healthy than some of our real-world populations. What was the prevalence in that population?
Arpit Rao: Yeah, so I was actually surprised that when the numbers came back. We were estimating maybe 20%, 25%, but the numbers came back, and 38% of the patients who had disease progression, had unequivocal clinical progression only. So, they didn't have radiographic progression, just UCP. And then an additional 17%, so now we're reaching the halfway mark, had radiographic and UCP together.
Alicia Morgans: That's a large number of patients. And that actually, from my perspective, speaks to what I see in clinic when we get to this patient population, although we have to remember like you said, this is a pre-docetaxel population. So, that's, I think, very impressive.
Now, did UCP tend to correlate with cancer-type outcomes, things like survival and progression?
Arpit Rao: Yeah. So that was the other big question, right, is this is happening. It's happening very commonly, and it's irrelevant. Now, does it have any prognostic value? And, in fact, it did.
So, we looked at the correlation between the occurrence of UCP and overall survival, and there was a clear drop in the median overall survival when patients have UCP. To give you a sense of that, in the UCP-only arm, the median overall survival was 25.7 months. And in the radiographic-only progression arm, the median overall survival was 33 months. So, that's a 40% higher chance of death, a higher likelihood of death in UCP patients, all else being equal.
Alicia Morgans: Wow. Well, can you predict who these patients will be? Or is this something that you just have to monitor and see who those patients are going to be during practice?
Arpit Rao: Yeah. So there are predictors. We looked at a variety of baseline clinical pathologic variables, and one of them turned out to be the PIM risk group, which Dr. Charles Ryan had previously reported some data on, and Dr. Halabi.
So, I think that is a composite metric that was derived from the 302 data set, that correlated very strongly with UCP. The caveat, obviously is, that we got to investigate that relationship further in other studies and other settings as well.
Alicia Morgans: Absolutely. So, as you think about this data, and you think about your clinical practice, how do you counsel patients, and what do you do with this information when you see an unequivocal clinical progression in your clinic?
Arpit Rao: Yeah. So, I use this information in a number of different ways. I will go back and say, one of the things that I was interested in is when we give patients the magnitative treatment benefit, so when you're talking abiraterone, you say median survival, median rPFS is, whatever, 16 months. Those 16 months don't include these patients with more aggressive disease trajectory because they were censored from the rPFS analysis. They never had the radiographic-progression event, so they were removed.
And, so, when you put those patients back in, we actually did that, and we created a composite, progression-free survival endpoint called crPFS, clinical radiographic progression-free survival. Again, that was, in many ways, it was a true magnitude of benefit for the entire population. And that crPFS was, again, much lower at 13.3 months, as opposed to rPFS, which was 16.5 months in the abiraterone arm, again, statistically significant. And, again, the UCP was having a deleterious impact on, not just abi arm, but the prednisone arm. So, it was independent. It was happening to everybody.
So, I use that information. I tell patients that, "Hey, if you have the certain risk factor, you may have an increased chance of UCP, and if we get there, it would indicate that your disease biology may be a little bit more aggressive, that you may have less benefit from either of these treatments. And you may have a shortened survival."
The other thing we factor in is treatment and sensitivity. So, we now know, okay, time to progression is one metric. But what happened before then? Did they actually have a response in PSA?
And what we found is, patients who did poorly on both of those metrics, so did not have PSA decline or PSA response, and then had UCP, actually got the worst deck of cards, essentially. And their median survival was 22.6 months, so very different. And if you didn't have either of them, and you had PSA response and you didn't have UCP, well, your median survival was 40.6 months in that group. So, a huge difference in outcomes.
Alicia Morgans: It really is huge, and it does help patients, as they're trying to think through what to expect. It also raises the question as to whether patients who are progressing clinically, so clearly progressing clinically, should be censored from a radiographic progression-free survival, or if we should have a progression-free survival that really incorporates that clinical progression as part of it.
This is really fascinating work, and I congratulate you and the team for putting it together and for putting it into context for us, as we think about how we use data that we analyze, even data from trials that have happened a few years ago and reported a few years ago, they can still inform our practice today, not just in terms of providing us therapeutics, but helping us better understand how we predict these things for our patients. I appreciate your time and your expertise. Thanks so much, Dr. Rao.
Arpit Rao: Thanks, Alicia. Great to chat on this topic.