FDA ALERT: Efficacy Issue Identified in Some Patients Taking Pembrolizumab or Atezolizumab as Monotherapy to Treat Advanced Urothelial Cancer - Petros Grivas

June 21, 2018

(Length of Discussion: 16 min)

Petros Grivas and Alicia Morgans discuss the late-breaking FDA Alert announcing the label changes for Pembrolizumab and Atezolizumab as Monotherapy to Treat Urothelial Cancer with Low Expression of PD-L1.


Petros Grivas MD, Ph.D. a medical oncologist at Seattle Cancer Care Alliance with expertise in genitourinary cancers such as bladder cancer, prostate cancer, and testis cancer.  He is the Clinical Director, of the Genitourinary Cancers Program at the University of Washington Medicine and an Associate Professor of the Department of Medicine, Division of Oncology at the University of Washington School of Medicine.

Alicia Morgans, MD, MPH

Read the full video transcript

Dr. Alicia Morgans: Hi and thank you so much for joining us. My name is Alicia Morgans. I'm a medical oncologist at Northwestern University and joining me today is Dr. Petros Grivas. He's an associate professor at the University of Washington. He is also the clinical director of the GU cancers program at the University of Washington and the Seattle Care Alliance. Thank you so much for talking with us about this late breaking news, Dr. Grivas. Can you tell us a little bit about what came out just yesterday on June 20th from the FDA regarding testing and use of checkpoint inhibitor therapy?

Dr. Petros Grivas: Thank you so much for having me and this is a very relevant question since the FDA, yesterday on June 20th, 2018, changed the label for the use of pembrolizumab, the anti-PD-1 checkpoint inhibitor, and the atezolizumab, the anti-PD-L1 checkpoint inhibitor for use in patients with metastatic urothelial carcinoma or locally advanced urothelial carcinoma, specifically for the first line indication in patients who have not received chemotherapy for advanced urothelial cancer. 

The previous label was based on the Keynote 052 trial with pembrolizumab, and the IMvigor 210 trial with atezolizumab. These were both single-arm phase two studies that showed rapid and durable responses with the checkpoint inhibitors that led to the FDA approval, an accelerated approval by the FDA in the past. And those approvals for those agents regarded cisplatin-ineligible patients, those patients who cannot receive cisplatin for other medical reasons and the chemotherapy made for advanced disease in first-line settings.

And the approval was in all comers, meaning regardless of their expression or over-expression of PD-L1 in the tumor tissue.

Yesterday, the change of the label restricted the use of pembrolizumab and atezolizumab in this particular first-line cisplatin ineligible, unfit patient indication, only for those patients who have a positive expression or over-expression of PD-L1 as defined distinctly in two separate assays. 

Regarding pembrolizumab is the immunohistochemistry antibody 22C3 assay, and there is the so-called CPS score that has to be at least ten or more for PD-L1 to be considered positive and thus pembrolizumab to be indicated in that particular setting. 

Correspondingly, atezolizumab label in the first-line setting, cisplatin-ineligible patients was restricted to those who have positive PD-L1 as defined by distinct assay, the so-called VENTANA SP142 antibody assay that was used in the relevant trial and, in that particular study and assay, the PD-L1 expression had to be at least five percent or more in tumor infiltrating cells in the tumor tissue. The interesting point though was that the FDA kept an unrestricted label for the use of either atezolizumab or pembrolizumab in this particular first-line setting in patients who cannot get any platinum-based chemotherapy. So, patients who cannot get even carboplatin-based chemotherapy because they were not fit enough, these patients can still get either atezolizumab or pembrolizumab regardless of PD-L1 status. 

And these changes in the label by the FDA came a few weeks after a similar update from the EMA, the European Medical Agency, that restricted the used of atezolizumab and pembrolizumab in the first-line setting of cisplatin ineligible unfit patients in urothelial carcinoma settings, but the difference is that in Europe the restriction that the EMA placed was in every patient in the first-line setting and they did not separately allow use of atezo or pembro in patients who are not chemotherapy eligible, regardless of PD-L1 assay so even those who can authorize chemotherapy in Europe, they still have to get PD-L1 testing.

In contrast, the FDA made this separate distinction that patients who cannot get any chemo, including carboplatin, they can still get either checkpoint inhibitor regardless of PD-L1.  So, this is definitely practice changing with practice implications both in Europe and in US. For example, if someone cannot get cisplatin chemotherapy in the first-line setting but still can get carboplatin-based chemotherapy, has to be tested for PD-L1 in the tumor tissue and of course this create more questions in the clinical practice of oncologists' both in the academia and community.

For example, which lab is capable and has the capacity and the validation to do the assay. There are multiple commercially available and I would say well-known labs that can be used for the conduction of the assay, so-called residence labs that have to be, again, validated and consent so they can do the assay. 

There could be a pathology department in different cancer centers that also may be able to do the assay. They have the antibody and they are built to interpret the results. Of course, this has to be a clinically approved environment because a clinical decision is being made based on the assay.

The second question is, what is the turnaround time of the results, and what time it takes from the time the tumor tissue is sent to the lab until the result is reported, and of course costs covered by insurance, by payers of the assay. And since this is now an EMA and FDA indication with a label change, we assume that the payers might cover that particular test because it's relevant to clinical practice.

I think more information will come as we move forward, but definitely, the implications are imminent in the clinical practice and in advanced urothelial cancer. And moreover, in my opinion, it’s probably also relevant in the conduction of clinical trials in the research context where we have clinical trials that combine PD-1 or PD-L1 inhibitors with other agents, either immunotherapy agents or targeted therapies or vaccines or other different compounds.

And the question is whether those trials that are being already conducted, already launched, or are in the design phase, they will require to have a PD-L1 assay as a selection or certification factor. This I think remains to be seen.

If you ask my personal bias, if we talk about the single-arm study without randomization, if you have a PD-1 or PD-L1 inhibitor and you combine it with a compound X and this is in a clinical research setting, personally I do not think that is necessary to select patients by PD-L1 because the combination has a premise and sometimes that may overcome PD-L1 status assumptions. However, if you have a randomized trial that randomizes patients in the cisplatin unfit first-line setting, again the cisplatin-ineligible patients in the first-line setting, then you might actually need to select the patients based on PD-L1 status because there is a fifty percent chance in a one-to-one randomization design that these patients may get single agent checkpoint inhibitor. 

So, I think the implications are definitely clear, both in clinical practice and also in clinical trials. So very interesting evolving landscape in urothelial cancer over the last few years and, in my opinion, again combination statuses are being tested in clinical trials and the question will be, in the future, whether the PD-L1 assay will remain relevant or not in the context of combination therapies that are being tested.

Dr. Alicia Morgans: So, fantastic description, and I really appreciate you referencing the EMA's recommendations or guidance as well, which is different than the FDA. At least in the United States, for patients who are not eligible for chemotherapy with a platinum-based chemotherapy, they do not need to undergo testing with either assay but can proceed with simply undergoing treatment with a PD-1 or a PD-L1 agent, pembrolizumab or atezolizumab.

For patients in Europe, that does not seem to be the case. That if they do not stain high even if they are not chemo eligible then I'm not sure how things will play out there. But, at least in the United States, for non-chemo eligible patients, they can proceed with treatment, which I think is really important to emphasize -- treatment with checkpoint inhibitor therapy.

The other thing that I would just love to hear your thoughts on in the last minute or so that we have is really, how do we operationalize this? And I know that you reference CLIA labs and sending things off but, you know, for practitioners who treat predominantly urothelial cancer with these drugs, we have not had a companion diagnostic in the past. And now we really do for these non-cisplatin eligible patients who have to think about testing.

Do you test all of your patients? What's the turnaround time in your clinic? Or do you, for sick patients, proceed directly to carboplatin-based therapy for the cisplatin-ineligible patients?

Dr. Petros Grivas: That's a great question and I will tell you my personal bias here. I think at this point of time, and this is, of course, all news and just happened within a few hours.  So, at this point of time, the two assays that I quoted, the Dako 22c3 and the VENTANA SP142 antibodies, are not deemed companion diagnostics yet.  I think the FDA changed the label, but in my understanding, I don't think that the assays are automatically FDA approved as a companion diagnostic. 

Dr. Alicia Morgans: No, I agree. And I apologize, I don't mean to say that they are. What I meant to say is we have not had to have any companion diagnostic in the past and, no, they're not considered companion at this point, though if we need them in order to get the drugs, I'm not sure if that label will change. So, I apologize, I don't mean to make implications. They are not considered companion diagnostics.

Dr. Petros Grivas: I agree with you. I think your point is very, very valid and very well-taken. And they're actually complimentary diagnostics. That's how I think the FDA made them before. But, as you said correctly, they were not companions. So, this raises all these questions about the applicability and the operationalization of those results and those updates.

So, I think, number one, we need to have a better sense of which labs, reference labs, can provide this assay and what will be the turnaround time for each of these two assays. It's an important note for providers that if they want to use pembrolizumab, they have to use the correct assay corresponding to that and vice versa for atezolizumab. They have to use the corresponding assay that is relevant to atezolizumab based on the relevant studies that were done, so they cannot mix the assays because the assays are different and they were tested in the context of its agents.

Now, in our institution we're actually discussing the operational aspect of it and how we're going to move forward. We're in discussion with pathology and also with laboratories exactly for that aspect.

In terms of clinical practice and selection of agents, personally, if someone can get carboplatin-based chemotherapy and I don't have the assay available right now, but again I may have it in the near future, a common practice is that someone can get chemotherapy with carboplatin gemcitabine and then there are two maintenance trials out there that are still ongoing actively.  One with avelumab versus best supportive care, and the other with pembrolizumab. Both those clinical trials are enrolling patients in a switch-maintenance manner.

After people get platinum-based chemotherapy, cisplatin-gemcitabine or carboplatin gemcitabine after four to six cycles, if they have no progression they can go on checkpoint inhibitors again versus control. And these trials are very interesting and relevant. It could potentially be relevant to clinical practice down the road, based on the results. 

So, a strategy, if someone selects to start with chemotherapy, is to think about those two trials. If not, I think the question will be, if someone cannot get chemotherapy at all, they can use atezolizumab or pembrolizumab, regardless of the PD-L1 assay. They have to document that in their medical records, and of course if someone selects one of those checkpoint inhibitors in patients who still can get carboplatin, they have to use the assay and I think each institution has to operationalize this in an optimal way so we can minimize turnaround time and select which lab, you know, whether they can do the assay quickly, either reference lab or a local lab, pathology lab.

Dr. Alicia Morgans: Great. Well, I truly appreciate your expertise, and your guidance as we continue to operationalize and implement the changes brought about first by the warning essentially by the FDA and now by this new guidance, and I really appreciate your time.

Dr. Petros Grivas: Thank you so much and I appreciate your excellent questions that are really, really practice-relevant and have significant impact and very, very thoughtful for our community and academic colleagues.

Dr. Alicia Morgans: Thank you.