Radiographic Progression-Free Survival Correlation with Time-to-Event Endpoints in the VISION Trial - Michael Morris

December 20, 2022

In this discussion, Michael Morris and Alicia Morgans discuss a post hoc analysis of the VISION Trial aimed at assessing earlier maturing time-to-event endpoints. Earlier time-to-event endpoints than overall survival (OS) will allow more timely regulatory approval and overall accelerated drug development. Morris and colleagues aimed to estimate correlations between time-to-event endpoints such as rPFS, time to first symptomatic skeletal event, and OS and provide further evidence for rPFS as an early endpoint for regulatory approval, in the context of radioligand therapy.


Michael Morris, MD, Medical Oncologist Clinical Director, Genitourinary Medical Oncology Service & Prostate Cancer Section Head, Division of Solid Tumor Oncology, Memorial Sloan Kettering Cancer Center, New York, New York, USA.

Alicia Morgans, MD, MPH, Genitourinary Medical Oncologist, Medical Director of Survivorship Program at Dana-Farber Cancer Institute, Boston, Massachusetts

Read the Full Video Transcript

Alicia Morgans: Hi, I'm so excited to be at ESMO 2022 with Dr. Michael Morris here to talk about advances in the VISION trial presented here at this meeting.

Michael Morris: Well, since VISION was published and originally presented, we've tried at each meeting to update the community on more data, deeper dives into the data. And so for this meeting, we looked at the association of some of the intermediate endpoints with overall survival. I wouldn't say that they will shock everybody. These are pretty conventional intermediate endpoints by now, and the associations are quite good, which everyone expected. To remind your viewers, all of the intermediate endpoints really were hit on VISION. PSAs declined. There were radiographic CRs and PRs, and there was a prolongation in radiographic progression and death, and then OS. We more formally analyzed the relationships between PSA kinetics and RPFS and OS. Unsurprisingly, the more your PSA declines, the more it's associated with a greater clinical benefit.

From an rPFS standpoint we actually did an exercise that I think is an important one. If you go back some years in the formation of prostate cancer working group two and working group three, we showed an association between a bone scan defined and cross-sectional imaging defined definition of progression and overall survival. And we used COUGAR-302, the pre chemotherapy registration trial for abiraterone as a initial study that we looked at. And then we looked at the same thing with PREVAIL, the enzalutamide pre chemotherapy study. There's an Alliance trial that we also did to look at that association. And the correlation there between rPFS and OS was around 0.7.

It was on that basis that regulatory agencies, particularly US regulatory agencies, then said the association is good enough that we will permit our PFS to be an endpoint on which we approve drugs. We don't have those data for radioligand therapy, but there are a lot of radioligand therapy trials that are coming up and are planned. And we would really like to be able to say that our PFS is a standalone endpoint. We took VISION and applied the same statistical tests basically that we had with the ARSIs to qualify our PFS as an approvable endpoint and found the correlation coefficient to be about the same, around 0.7. Our hope is that provides firm data so that rPFS will be used as a standalone endpoint for further approvals either of Pluvicto in earlier patient populations or other drugs that are just coming into phase three or have ongoing phase threes in which there's rPFS primary endpoint.

Just to remind your viewers, PSMA-4 has an rPFS primary endpoint and PSMA edition has an RPFS primary endpoint, so it's important to know what the data set from VISION is in terms of that correlation between rPFS and OS so that people can say, "Oh, this isn't some arbitrary endpoint and we don't actually know that there's a downstream overall survival benefit." They'll have these data and see that it's the same as it was with between rPFS and OS for the earlier trials using the ARSis.

Alicia Morgans: Thank you for walking through that. And I just want to emphasize how important it is to make sure that this is an endpoint that holds true in a radioligand therapy space. We have seen, for example, Cabozantinib and had very interesting effects on bone scans, for example.

Michael Morris: Yeah.

Alicia Morgans: And if we were trying to use in rPFS with that particular agent, that would be very challenging if somehow bone scans were incorporated into that endpoint. Ensuring that the truth remains the truth in this setting, I think is really critical. And I'm just curious, this is rPFS based on conventional imaging.

Michael Morris: Yes.

Alicia Morgans: And how do you see PSMA imaging coming into play with this endpoint?

Michael Morris:  To answer your first question, this is rPFS based on standard imaging because VISION only had a pre-treatment screening PSMA scan as opposed to serial post-treatment PSMA scans. This is basically confirmation of our old criteria applied to a new drug class. But you raise a very important question, which is, can we do better? And what are the challenges that are posed by PSMA PET for a time to event endpoint?

Obviously PSMA PET is a lot trickier because especially with anything that implicates the androgen receptor in terms of a therapeutic, because that will alter PSMA expression and alter potentially a post-treatment PSMA scan. We face that same challenge though with standard conventional imaging with bone scan flare. We probably, in looking at PSMA based imaging to document progression or a beneficial change, may need to wait to understand beyond the first post treatment scan what's actually going on because PSMA expression could go up or it can go down, and tumor volume is also going either up or down, so we probably will need to establish a rule set similar to working group three in that you don't make your assessment based on just the first scan.

But I think that PSMA PET also offers a unique opportunity to get beyond a time to event endpoint because response is accessible on a PSMA pet, which it hadn't been on cross-sectional imaging for bone disease or even bone scintigraphy for bone disease. Whereas PSMA PET does have the opportunity to have a quantifiable biomarker that's response. Now, response occurs much earlier than progression, especially for patients who have castration sensitive disease or otherwise early disease. And so you could get a readout on a drug's efficacy by looking at response rather than time to progression.

The question isn't just what will rPFS look like for PSMA PET, but more optimistically, what will response look like? And then we would be able to cut the time that it takes to develop a drug, not just by months for castration resistant disease, but potentially by years for castration sensitive patients and other early patients. PSMA PET is huge focus of prostate cancer working group four in which many, many people are working on, and along with the regulatory agencies to try and figure this out so that we could build these credentialing correlative components of ongoing and upcoming phase three trials in order to qualify a response parameter for future therapeutic drug trials.

Alicia Morgans: Yeah, it will be so interesting with PSMA expression being dynamic-

Michael Morris: Yeah.

Alicia Morgans:... with the therapy and understanding that to also then ensure that down downward expression and decreased PSMA avidity that actually is associated with whatever your end point is, of course of interest.

Michael Morris: I think we're up to it though. We had the same challenge with bone scintigraphy and we found a way through it that put forth a hypothesis and we built the data around the hypothesis in order to prove or disprove an association and a definition and a means of collecting data, we're going to have to do that same thing. The main challenge I think from a feasibility standpoint of doing that is that when we were qualifying working group two and three bone scans were standard of care in assessing post treatment effects. In every clinical trial you would've bone scans just built into the study.

PSMA based imaging is not, and so we're going to have to figure out how to have the financial aspects of the sponsors of various trials work together to build that database because it will be to every sponsor's advantage to have a nice early endpoint in future trials, but there has to be some investment by the community up front to build in the serial PSMA scans to gather the data. Hopefully that can be an aggregated effort, so no one sponsor is on the hook for that. But we need a plan at least to see how many patients that we need in order to do that, and we're developing that plan now with the regulatory agencies.

Alicia Morgans: Well, that's fantastic. And I'm really glad that you and the team are on the case defining the problem and really putting forth the effort to figure out, how do we walk through this and get where we need to be. Because to your point, this is where we need to go, and I do think that we are up for the challenge. Any final words on these additional analyses within VISION very helpful work?

Michael Morris: Sure. I think that the next step would be taking all of these intermediate endpoints and synthesizing them into a unified model. Whether that is as a nomogram or some other expression as a multi-variable model. We've looked at each of these individually, now we need to put them all together and see how in aggregate they predict for OS. And then hopefully we'll have a multi-variable way of predicting for an individual patient as to whether they're going to be a good candidate for a drug like Lutetium or perhaps whether they should go on to something else.

Alicia Morgans: Wonderful. Well, thank you so much for taking the time to talk with us about these presentations at ESMO 2022. It has been a pleasure.

Michael Morris: Thank you so much for inviting me, Alicia.