Biologic Rationale for and Current Research Database with Immune Checkpoint Inhibition in Gynecologic Cancers

DR POWELL: We’re going to review the biologic rationale for the use of these drugs. I know we’ve already touched on a few key points here. We’ll talk about the FDA-approved indications that we have in the MSI-positive space for basically all tissue types. And then in cervical cancer talk a little bit about the CPS scoring: who to test, how to test them — it did look like there was some room for improvement. And some of the immunotherapy completed trials as far as single agents,. And then we’re going to talk about some of the ongoing and approved trials.

It’s really quite a fascinating story, as many of you know, from 2003 when we started to learn how the immune system was interacting with ovarian cancer and noting those patients that had high intratumoral T cells were doing better than those patients that didn’t. We knew there’s a signal there. How do we capitalize on this has really been a goal over these last 2 decades to get a better understanding of the way tumors interact with the immune system.

This is the more simplified Chen and Mellman diagram, with our understanding of how the cell recognizes a cancer, how the cancer hides from the immune system. And I wanted to point out a few key features, and again, the simplified diagram is better. They have an updated one that is so complicated, but it tells us so many things going on.

We talked about the effect of HPV. We do know it’s very immune stimulating. These cancers, typically — about half the cancer is immune cells when we look at them under the microscope. And we know anti-PD-1 has a signal, and we’ll be talking more about that. What we don’t entirely know is, if we give chemo with checkpoint inhibitors, is that going to improve the antigen presentation or is that going to actually harm things?

We also have some signals that PARP inhibitors may be helpful. And we’ll be touching on some of these issues.

And when you look at total tumor mutational burden across the genome, and these are mutations per megabase of DNA, you see how our gyn cancers kind of stack up to — bladder cancer probably the biggest one on the market, has 5 drugs I think approved in that space for IO therapy. You can see a large portion at the top of the graph here, and our blue one, with endometrial cancer, showing many of our tumors have that high mutational burden. Those are our MSI-positive tumors. Those are our POLE mutation tumors. Within ovarian cancer, it’s not quite as promising — you see less in that space above the theoretic bar of 11 or 12 mutations per megabase.

What about adding chemotherapy? I think this is probably our hottest area moving forward. This is some mouse model data. In this mouse model, at least the combination of checkpoint inhibitor with paclitaxel seemed to improve tumor control and with the appropriate biologic endpoints.

And if we look at this timeline just from 2011 with that first approval in melanoma, moving across here, now all these are new approvals in multiple sites, with our all-comer approval with MSI-positive in 2017, with basically all our tumors, and then looking in 2018 with our cervical cancer approval.

And we’re going to talk a little bit about what the label is saying. Pembrolizumab is approved in the treatment of recurrent or metastatic cervical cancer with disease progression on or after chemotherapy whose tumors express PD-L1 with a CPS, which is a combined pathologic score, greater than or equal to 1%.

We note that chemotherapy is vague. So does chemoradiation count?

And then the other indication that we’re going to talk about is pembrolizumab in adult and pediatric patients with unresectable or metastatic MSI high. Also, we say defective mismatch repair solid tumors that have progressed following prior treatments. And this term “who’ve had no satisfactory alternative treatments” really fits us in gyn cancers because, really, we’re not satisfied with how our patients are doing.

Talking a little bit more about: What is MSI? So microsatellite instability, can be called MSI-positive, mismatch repair deficient, defective — you’ll see these terms used interchangeably.

We also know that some endometrial cancers also have homologous recombination defects, so we’re learning more and more about how ovarian cancers, endometrial cancers and even cervical cancers can behave similarly. We’re understanding better biology, but we do know about 20% of our endometrial cancers will be classified as MSI high. This also should be a signal: Is there familial syndrome? We need to think about Lynch syndrome in these patients. And I think we all know that story quite well.

When we look at the ovarian subtype, and I think this is probably data that many of you haven’t seen, are high grade serous. Yes it’s less than 1%, but it’s still worth testing. If you find it, these patients still seem to benefit. But it’s our endometrioid subtype, as many of you would guess, would be the one with a 7% MSI-positive rate.

When you look at cervical cancer, a little more older data sets, but maybe as many as 5% — somewhere between 1% and 5%. So it’s worth looking. You can find these patients. They seem to benefit for a long time. So now I test all these patients, both with an IHC platform and using NGS platform, because there is not overlapping things you find.

One other key thing is, at relapse there’s about 7% more MSI found than on the primary tumor. So it may be worth rebiopsying and checking for these signals.

When we look across the trials in endometrial cancer, Dr Tater’s initial presentation showed a 56% response rate. And these trials partly what led to the label indication we just talked about. If you look at those that had PD-L1-positive tumors, nonselected for MSI, that was a 13% response rate.

For the atezolizumab, 13% response rate. Again, in line with the group for just unselected recurrent endometrial cancer patients. And then if you look at the pooled MSI cohort that was in the label for pembrolizumab, you see that 36% overall response rate and that nice range of durability of response. So that’s where we see the 17+ months’ durability, which I think has really excited us, although in the unselected population maybe not as good.

So TSR-042 is a novel anti-PD-1 agent looked at in an MSI-stable and MSI-high population. This is the high population that was presented at ESMO in 2018. And you can see again that very nice response rate, 52%, some CRs in here. And the median duration of response hadn’t been reached at the time of this publication. You can see a few of these patients were still ongoing treatment. So quite exciting. And you see the treatment-related adverse event rate, about 11%, which I think is in line with many of the other agents.

I think it has a fairly traditional dosing initially but then has a q6week dosing after a number of cycles. Actually, I think makes it compelling for our patients that may be traveling and can be monitored less frequently.

So in the immunotherapy space for cervical cancer, I know we’ve talked a little briefly about this, the IPI data showed us a response rate of about 8.8 with a disease control rate of about 32%, with manageable toxicities, the KEYNOTE-028, 12 and a half percent. The 158 cohort, 14%, with that range. Duration response, again this really nice, long duration response of 18-plus months.

There was the Checkmate trial looking at all our kind of HPV-driven lower genital tract cancers with vaginal and vulvar included in that, 20% for the whole population, but that cervical cancer population was 26% again with that kind of similar safety signal. So I know we talked about these rates aren’t as high as we’d like, these response rates, so there’s room for improvement, but when we find these long durations, we’re pretty happy about that.

We’ve seen a lot of this. As a single agent I would say we’re overall a little bit disappointed when we see response rates somewhere between 10% and 15%. The PFS is also relatively modest.

And this is just a landscape of ongoing Phase III trials in endometrial cancer, cervical cancer and ovarian cancer. It’s a very exciting time to be a practicing gyn oncologist or medical oncologist seeing gyn patients. We have a lot of new things to be offering our patients. As I mentioned, within the endometrial space we’ve never seen anything like this.

So just in summary, we should be looking for MSI defectiveness measured among all of our cancers, and I would say both IHC and next gen probably make sense. Consider rebiopsying also. There seems to be some data. CPS score for cervical cancer.

 Targeting Folate Receptor Alpha (FRα) in Patients with Gynecologic Cancers

DR BIRRER: So we’ll just talk briefly about targeting the folate receptor alpha in patients with gynecologic cancers, really focused mostly on ovary. So again, just the take-home message, folate receptor alpha is a member of a gene family that binds folic acid. It’s a big gene and it’s a big protein. It sits in the membrane. But it’s also secreted, so you can actually find folate receptor alpha floating around in the blood.

One common misconception is that the cell gets folate solely through this receptor. It doesn’t. And, actually, most folate in the cell comes through the reduced folate transporter.

We’re not really quite sure what the folate receptor alpha does for cancer cells. It’s only expressed in 2 normal tissues: Type II alveolar cells in the lungs and renal tubules, and that makes it a great target, because it’s differentially overexpressed in a number of cancers. And I’ll just put some IHC images there showing you the membrane staining.

So there’s a very rich history of trying to target folate receptor alpha, and I wanted to just go through it really quick so you can appreciate how much work has gone into this. The actually first approach was just a humanized monoclonal antibody, the folate receptor alpha, called farletuzumab. And it was felt to be able to block whatever this receptor was doing. It was active in ADCC assays. It was very active in xenografts. Looked like it had a great toxicity profile. And, unfortunately, had very low activity in clinical trial and was relatively rapidly abandoned.

The second version of this is vintafolide, which some of you may be familiar with, because it went into a very large randomized Phase III trial. Its former name was EC145, and that was actually the first antibody-drug conjugate, not really an antibody. It has, actually, folate bound to a vinca alkaloid by a linker. So it was delivered to the receptor because the folate would bind to it. It would be endocytosed, release the vinca alkaloid and the tumor cell would be killed.

In combination with this or in parallel with this, they actually developed an imaging molecule, which you can see on the right here. And that was able to bind to the surface of the cells and express folate receptor alpha and you could actually see it by a technetium scan. And that was used in the trials as identifying patients who would likely benefit from this.

The Phase II trial I’m not going to show you, but it was very encouraging. Showed a lot of activity and then went on then to a randomized Phase III trial, and the design is shown here. You got the folate-binding molecule with pegylated liposomal doxorubicin versus just PLD alone. And this is in platinum-resistant patients. Big study, 640 patients randomized, and they were looking for PFS.

And, unfortunately, this was the announcement, which was, the entire trial was stopped for futility in an interim analysis. So at that point, we’re 0 for 2 in targeting this receptor.

But that gave rise to, I think, the most encouraging molecule, and I’ll just go through it quickly. It is an antibody-drug conjugate. Antibody-drug conjugates are becoming, I think, into vogue, although there are still challenges with it. The antibody part will target the receptor or the tumor-specific antigen. They may also have some intrinsic activity, meaning it may block some sort of signaling.

Then on the other end of the antibody is the payload. And the key for an antibody-drug conjugate is, the payload has to be exceedingly active, because once it gets inside the cell, it kills the cell. There is still residual payload that’s active, and it is released from the dying cell, and you have what’s called a bystander effect.

And then finally there’s a linker, and the linker is where all the work has been done. So this linker has to be tight enough that you don’t have free drug in the blood but loose enough that it lets go of the payload in the cell.

And this is the drug — we’re talking about mirvetuximab soravtansine. I always tell my patients, if you can’t pronounce the drug, you can’t get it. And as stated before, it’s a humanized antibody to folate receptor alpha. It binds actually quite tightly. The linker is a proprietary linker, which has all those properties I just described.

And then the payload is DM4, which you’re familiar with, I’m sure. Very, very potent. When DM4 was used as a free agent it, essentially was too toxic to use.

The Phase I/Phase II trials are described here. There’s a dose-escalation trial, and then there were multiple cohorts. And, actually, over 200 patients were dosed with this as a single agent. There were cohorts that looked at this eye toxicity a little bit, and there was even a cohort of relapsed endometrial cancer.

This is a good summary, I think, of thatdata, and it was presented both at ASCO 2016 and 2017. You can see that the overall response rate for all patients is somewhere around 26% to 30%. If you then actually look at platinum-resistant ovarian cancer with 1 to 3 priors, so you’re not looking at tumors that had been really beat up too much, the response rate’s remarkably consistent, on the range between 40% and 50%. The spider plots are on the other side.

There was not a lot of difference in terms of the spider plots in terms of medium versus high folate receptor alpha. If you go into low folate receptor alpha, the response rates certainly did drop off. So if it’s going down and it’s staying down, it’s good.

So you can see, the vast majority of those patients, with the exception of a few who showed progression early on, have all gotten some sort of response. And it’s relatively durable. Spider plots are great, because it gives you response and durability all in one plot.

This is the safety profile. And we see some GI toxicities. Frankly, in my experience, I’ve never taken a patient off of mirv because of GI toxicity. But it’s there. And there’s some nausea. And then there’s this blurred vision. The blurred vision was discovered fairly early on. I always tell the story that it was my patient. We discussed her in our tumor board, our conference. Everyone said, “Yes, she’s doing great.” I got back to the office, and she had called and left a message and said, “Dr Birrer, I’m blind.” And at that point, I said, “Ooh.”

It turned out she wasn’t blind. She had blurred vision. And they develop blurred vision because they develop pseudocysts in the cornea, and you get so many of them it actually — it interferes with vision. It is 100% reversible and actually can be prevented with steroid eye drops.

And so you’ll see that even on these top 4 toxicities, almost all of it’s Grade 1/Grade2. There’s very little Grade 3 toxicity with this drug.

And the point I would finally make is, you don’t even see myelosuppression here on the list. So that’s one of the real advantages of this drug.

And that brings up all sorts of interesting issues we can talk about later on, which is, could this effectively be brought up earlier in the treatment of ovarian cancer as a substitute for taxanes? No alopecia. No neuropathy. And we’re talking about that.

So this, all this data got a lot of us excited, and it led on to the design of a randomized Phase III trial. This just shows you the folate receptor alpha cutoffs that we used. The FORWARD I is the randomized Phase III trial. The FORWARD II, being led by my colleague, sometimes friend, Dave O’Malley. So the slightly different criteria. But we use medium and high for FORWARD I. In retrospect, that may have been a bit of a mistake, but one never knows.

So this is the design for FORWARD I. You can see patients had to have folate receptor alpha staining in their tumor to get on. And again, we define that as medium or high. It was done centrally with VENTANA. And they were platinum resistant with 1 to 3 priors.

It got randomized to single-agent mirv, 6 mg/kg, q3week and then investigator’s choice — weekly paclitaxel, PLD or topo. I led this trial with Katie Moore.

It accrued extremely fast. Closed. And this was the patient population. Again, just defining that greater than 50% of the cells with greater than 2+ intensity by staining.

It closed and we analyzed it, and these are the top-line results. And I regret to say it did not meet its primary endpoint of PFS in either the entire study population or a prespecified subset of patients with high folate receptor alpha.

Now, there’s a caveat here in the sense that the way the trial was designed, when we went to the really high expression, we actually needed a p-value of less than or equal to 0.025. And, in fact, it was statistically significant at 0.047. But because of the way the stats were done, it didn’t get to that level.

In retrospect, we may have missed an opportunity just to study it in the high expression levels. I’ll point out, the overall response rate was considerably higher in mirv.

And this will be my final slide, that this is the future. And is, again, what Dave has been leading, and that is using mirv in combination with other active agents: bev, pembro and even carbo. And I’ll only make 2 comments. I think the bev arm is fascinating. It has the same response rate, if not better, in mirv alone. And I think the bev is adding increased duration of response.

And then the carbo is interesting because of what I said before. I think if we can show safety and activity, and again, I do think we need an indication here, but could we move mirv up front and fundamentally change the way we treat ovarian cancer patients for those tumors that are expressing folate receptor alpha? Could we combine mirv with carbo and maybe add bev in there? That would be the initial treatment.

Other Novel Agents and Strategies Under Investigation for Cervical Cancer, Endometrial Cancer and Ovarian Cancer

DR O’MALLEY: So with this, really looking, again, at these antibody-drug conjugates, they all are very similar, where you have uptake of the — linking of the antibody and then delivery of the drug, where you have, then, direct effects of the drug as well as those in close proximity.

So this is MMAE, which is also a microtubule target. So not the same as mirvetuximab in its active agent, but a similar target against microtubules, so a paclitaxel-like mechanism.

Tissue factor seems to be expressed in almost all cells, especially in cervical cancer, and a very high rate also in ovarian cancer. It also seems to be a poor prognostic. I’m not sure if it has a known mechanism of action, and I don’t know that question. So we do know that it’s a poor predictor with higher metastatic disease and a worse prognosis.

So this is a multitumor trial, and this was their first in human. You see here that they really emphasize cervical, bladder, endometrial, head and neck, non-small cell lung, squamous head and neck and then ovary and prostate. So in these patients, they did have to have an expressed tissue factor. And it’s interesting, was not needed in the current trial, and we’ll talk about that.

And the trial was an interesting design, where they had it for 4 cycles, and if they were getting clinical benefit, they could continue it up to 8. So 12 cycles total. If they continued not to have progression or toxicities, then they were put on to another extension trial.

And you see here the numbers that all the other disease sites had 15, but they expanded into cervical into 55 patients, and that was really the concentration of what was reported.

Obviously it’s a Phase I. Its primary endpoint was safety and tolerability. But as you heard, we had both a bigger investigator initiative. We had RECIST response. But the secondary endpoint really was investigator initiative, overall response.

So we have some updated information with regards to this, but as you look at the median duration, quite modest with the median number of cycles of 4, with a duration of approximately 2 and a half. But you did see some toxicities, and I’d like to draw your attention to the top 2, epistaxis, which happens at any grade, at 51%. Now, most of this seems to be self limiting and not needed to be hospitalized. But this is not the bev “I have a little drippy nose.” This can be really nosebleeds that need intervention.

So when we look at this, this is something to make sure you alert your patients about. In addition, you do see those in cervix have a fair amount of risk of vaginal bleeding, probably from the cervical tumor. And again, you did see some Grade 3 toxicities, not included in these slides.

I think, again where you have — I don’t think we have a clear rationale, but with the right expression of the TNF, most likely you have irritation of those areas and those vascular membranes. Which is really interesting with the other noted eye toxicities with mirv. It’s corneal, and you have visual changes. This is true inflammation of the conjunctiva. They can get very beefy red eyes. And so when you look at all comers, you have about almost half the patients — we’ll talk a little bit more on that — even a patient with Grade 3.

Now, a very important part of this is, they do have alopecia, which is much different than what we just talked about with mirv, which is almost a 0% occurrence of the alopecia, and probably most of those that are reported came into the trial with some alopecia.

So you’re looking at about 40%. And a fair amount of these will be Grade 2. Most of them will be Grade 1, but a fair amount will be Grade 2. And you see some GI. You do see some neuropathy. Again, with the target of the MMR and the antimicrotubule, you do see some neuropathy. Again, most of this is going to be Grade 1 and 2, and a lot of these patients will have come into the trial with some preexisting. Do see some nausea and vomiting. And also, which has been interesting, there’s some abdominal pain, which I think is separate and could potentially be on the diarrhea/constipation-type mechanism that we think of, and then cramping. So once again, something to keep an eye on.

And, actually, we looked at the bottom left, the Grade 3 is about half, but all patients had some toxicity, which is what we see in all these trials — but I do think we need to look and continue to monitor this as we move forward with more experience. I would say in general, I tell my patients that they’re at risk for toxicity but should be better than especially a platinum doublet.

So when they changed the intervention — the intervention included continuous eye drops as well as steroids the day of your therapy and 2 days after and also vasoconstrictors at the time of the infusion if they could tolerate it and these actually quite extensive cooling masks — all of these to try to decrease the delivery of the drug to the conjunctiva.

They’re literally eye masks, just like the head, scalp, but it goes over the entire face. They’ve actually just redone an amendment to the protocol where they extended the size of this mask. It looks like kind of a sleeping cap, but bigger and blue. These form-fitting masks are actually pretty cool.

So here we go. So we see a reduction from almost 80% down to about 25%. Grade 3, again about 2.5%, but that’s really 1 patient we were looking at.

So significant improvement. And again, we did not see this pronounced of changes when we deal with mirv, but you see here, because it is a superficial process, we’re able to put those interventions in place.

And we saw the waterfall plot here. You saw about a 24% confirmed partial response, about a 31% unconfirmed. And with some patients, as you see there, who had complete responses. And again, with the waterfall plot, you can see most patients are benefiting. I always think about disease control rate of 63%. What does that mean? Does that mean they made it through 1 cycle? We really need more information. Is it 16 weeks, or what is it? But you see here on those waterfall plots that they definitely get better control.

Now, this led to the expansion here in 100 patients, which is basically designed as a large Phase II, all in cervical comers, with these interventions that we’ve described, once every 3 weeks. And really, these patients would need to be pretreated with a platinum doublet, but no more than 2 prior systemic therapies. And this has actually just completed enrollment, GOG 3023, and enrolled very, very quickly.

I think the FDA’s bar, and we’ve seen this across tumor types, about 200 patients for safety and about 100 for efficacy, it’s for a reason. They’re smart. And so I think I really need to see more information. I need to get more about the bleeding, as well as the eye toxicity — seems to be resolved. But I think this trial with 100 patients, if this shows a consistent 20% response rate where we have really no other options, I would absolutely use it.

So speaking of which, for the same reason, they’ve also designed a Phase I with expansion cohorts looking at combining this, like we heard in mirv. And the 3 arms are TV plus carbo, TV plus pembro and then expansion arms will be, including those, but also we’re looking at TV plus bev in a safety analysis. So this is going to be a little bit smaller. We're going to have 50 patients, about, in the escalation, depending on what we have to do, and then 90 in the expansion in those 3 arms, so 30, 30, 30.

So a smaller trial that is now enrolling. And has actually just undergone this pretty extensive amendment that I’ve described here, 3024.

I think we need to discuss, at least briefly, AIM2CERV, GOG 3009. Unfortunately, this really interesting Phase III trial with the Listeria drug, as most of us refer to it as, and this is a Listeria-based immunotherapy, again delivering into and against the E7 peptide of HPV-16. So once again, entering into the cell, processing, leading to cell death and a very, very interesting mechanism.

Unfortunately, the institutional trial in these patients who are placebo controlled has been put on hold twice, once for safety and then now, most recently, by a production problem. It’s once again on hold with regards to the drug. So we’re hoping that this trial reopens and that we can get these high-risk, locoregional patients better options than just observation.

The other trial that was open at the same time was with a PD-1 inhibitor, which, again, if you look at the immune-modulated aspect of this would really have made a lot of sense. Unfortunately, this trial, during its ongoing accrual, was put on hold at the same time with regards to the concern of the safety and with that was then suspended after they became active again. So we will not get the final results of this. But I am hopeful that we can see — if we can get this Phase III open again –– now we may have an option for patients, locoregional disease, that probably have a 40% to 60% risk of recurrence even with just chemoradiation.

People ask me all the time, what are you doing at Ohio State? Are you doing somatic? Are you doing germline? And really, as we look at the guidelines which changed for the NCCN, that they were really recommending both BRCA somatic and germline. We’ve learned that a lot of us are sending extended next-generation sequencing on all of our patients. So with the change in sequencing, or molecular testing, I should say, for all patients, obviously what we put in the NCCN guidelines was that should be at CLIA approved or reputable institutional, at least BRCA, looking also at microsatellite instability in HRD. Obviously, we all know about PARP and the implications of that.

But at just my site, just in gyn, I have 3 trials that you can use this information: WNT pathway, ATM mutations and CCNE1. One of the patients that we saw had a CCNE1 amplification.

My plea is that we need to get these patients on a clinical trial. Doing it one-off and trying to see what it does on this particular patient is not good enough. We really need to see what they are in prospective fashions.

As we’ve looked for options for our patients getting somatic testing, I’m now sending somatic on all of my recurrent patients with extended next-generation sequencing. There’s just too many opportunities for these patients to enroll onto targeted therapies.

Ongoing Trials Examining the Role of Anti-PD-1/PD-L1 Antibodies in Combination with Chemotherapy, Targeted Agents or Other Immunotherapeutics

DR WESTIN: I’ve got the most exciting talk, because I think this is what we’re most excited about for gyn cancers.

We’re seeing 10% to 15% response rates, certainly not what we’re seeing in some of the other tumors like bladder, but something. And there are a proportion of patients that will benefit with some stable disease. Of course, the exception are those MSI tumors, and specifically we saw a lot of data about MSI-high endometrial cancer, where we’re seeing much higher response rates. And certainly that’s very exciting from a single-agent standpoint, but what about the rest of the patients?

So we’ve got this huge number of patients that are not going to have a clinical response and that are not going to get clinical benefit, and how can we potentially sensitize these tumors a little bit more to checkpoint inhibition? And there are a number of different theories that are listed on the right of the slide. So can we increase the tumor cell death? Can we increase the so-called neoantigens that allow the immune system to identify the tumor and to attack the tumor? Can we increase the DNA damage, immune stimulation? And so there are a number of different opportunities that I’m going to touch on over the next few minutes, including combining with chemotherapy, anti-angiogenics, PARP inhibition and also doing dual checkpoint inhibition.

So first, a similar if not the exact same data about looking at that combination of checkpoint inhibition with chemotherapy. There are a number of different reasons why we think adding chemo in this setting can potentially sensitize these tumors, including stimulating those immune cells, again increasing that cellular death so that the immune cells can identify the tumor. And then also enhancing presentation of the tumor-specific antigens so they can be identified so that immune cells can actually work.

And we saw the combination of chemotherapy with checkpoint inhibition works better than either one alone. So I’m starting kind of negatively, I know. But I think a lot of you are familiar with these data. This was first of the studies to really read out with checkpoint inhibition and chemotherapy. So first, JAVELIN 100, now this is the up-front trial, so this was a randomized Phase III trial looking at chemotherapy, chemotherapy followed by avelumab maintenance and then chemotherapy with avelumab followed by the maintenance. It’s a very large trial. We have not seen the data, but we have seen the press release that came out December 21^st saying that it did not meet its prespecified endpoints. And so kind of took us down a peg or two, especially given all of the trials that are designed right now going on with chemotherapy and checkpoint inhibition.

And then, of course, I did not have the updated data from JAVELIN 200, but you can still get the general sense that that combination of checkpoint inhibition with pegylated liposomal doxorubicin did not meet its prespecified endpoint. So did not yield progression-free or overall survival benefit in combination.

A little bit of tantalizing data presented looking at PD-L1 expression and the potential that there may be some benefit if we do a little bit better job selecting these patients. So it’ll be interesting to see, moving forward, if that reads out.

And this is just a snapshot, more for your reference. There are too many, really, combinations of IO and chemo to spend much time on. But just know that we’re looking at this in the up-front setting. We’re looking at this in the recurrent setting, ovary, endometrial and cervical cancer is coming down the pike.

And this is just one plug for an NRG study that’s looking at up-front endometrial cancer, so randomizing these patients to chemotherapy versus chemotherapy with pembrolizumab. And there are a number of different other trials that are going to be similar to this with other agents.

Now let’s kind of switch gears a little bit. You’ve seen a little bit of data around anti-VEGF and immunotherapy, and similarly to what we see with the theory behind chemo. We know, actually, VEGF has a role in immune activity. It inhibits T-cell function, both directly and indirectly, and also can inhibit dendritic cell function. So if you can inhibit VEGF, you’re going to allow that immune system to work better and to potentially target tumor. And very similar to the graph I showed you with chemotherapy and IO, the combination of anti-VEGF and ant-PD-1 seems to work better than either one alone.

There are some things about hypoxia, too — if you can create hypoxia in the tumor, that makes it more sensitive to immune activity. So I think there’s a bunch of different reasons why we think it’s going to work.

We’re certainly very excited about the data that we’ve seen with lenvatinib and pembrolizumab that Vicky Makker presented. This is KEYNOTE-146. I think that this is exciting if we can see this. Endometrial cancer has nothing, right? There’s nothing FDA approved in this space aside from progesterone. So to see these kinds of response rates, where 40% of patients had benefit, and the majority of these patients did not have MSI-high tumors. So that’s a really important point.

So it seems that the combination is sensitizing these tumors that we wouldn’t expect to be sensitive to checkpoint alone. And you can see that in the MSI-high tumors that they did have, there were certainly a higher response rate, but even in the MS stable, it was about 36%.

I think the majority of patients are going to have some kind of toxicity, and the Grade 3 toxicity is very high. But I think as we get better at taking care of these patients, we can hopefully avoid some of those things.

And based on these data, the FDA did grant a breakthrough designation for this combination and is undergoing a randomized trial comparing it to standard of care chemotherapy, which is either doxorubicin or weekly paclitaxel. So it will be very interesting to see if we have a consistent result here based on what we saw in the earlier-phase studies.

And this is just a snapshot of all of the different combinations that are being looked at. It’s important to note, this includes anti-angiogenics like bevacizumab but also TKIs in combination with chemo, sometimes without, across all of the different gynecologic malignancies.

So then what about IO/IO? So we focus so much on PD-1 and PD-L1, but there are a number of different checkpoints. This picture is kind of starting to hint at all of the different options that we have and not only the inhibitory checkpoints, but also checkpoints that stimulate the immune system. So now there’s a number of different trials where we can combine either different inhibitors or different agonists.

And this just one, the kind of most well developed in ovary cancer, so nivolumab and ipilimumab, which Bob Berger presented at IGCS last year, randomized to nivo by itself compared to nivo with ipilimumab. You can see the responses were higher in the combination arm. Progression-free survival appeared to be improved compared to single agent. Adverse events were about the same. Actually, it was pretty impressive that they were able to combine the two with reasonable adverse events. And so it’ll be very exciting to see this is likely moving more to the up-front setting.

That 30%, it does kind of make you excited about some type of activity with checkpoints, but it’s too soon to tell, I’d say. And then there’s a number of different combinations, really too many to list, but different opportunities, a lot of them in early-phase trials right now.

And then finally, I heard Anna Oaknin call this the “fashion combination.” So combining PARP with IO, and do we have any data to actually support it, because it’s certainly happening, especially in ovarian cancer. And there are some data that we see with — a lot of these tumors do have homologous recombination deficiency, and so they have a higher neoantigen load. They should be identified by the immune system. We see increased TILs, tumor infiltrating lymphocytes. And also, interestingly, PARP seems to, again, enhance DNA damage, but, importantly, it causes these DNA fragments that induce a STING response, which is — the STING pathway is quite important for our immune cells.

So I think there are a lot of different potential reasons why this combination should be exciting. And we do have data, right? So there are a couple of really nice studies that have been done in recurrent ovarian cancer, the TOPACIO study, which is combining niraparib with pembrolizumab, and the MEDIOLA study, which was olaparib and durvalumab.

To start with, the TOPACIO study, it’s a more all-comers group. So this was any recurrent platinum-resistant ovarian cancer. They had a beautiful response rate of about 25%. That did hold up in the population that was completely biomarker-negative. So if you look along the bottom there, you’ll see there were a proportion that had BRCA mutations in the tumor. There was a 25% response. There was a group that was HRD-positive, 25% response. And the HRD-negative group, biomarker-negative group, had a 27% response rate. So that’s very exciting.

The data that have been presented from MEDIOLA were just presented in the germline BRCA-mutant population. Although, that combination is also being explored in a biomarker wild-type group. And they saw a really impressive response rate, 77% in those patients that were second line and a little bit less, 67% in third line, and then 70% in fourth line. So really consistent response despite multiple prior lines of therapy. So it’ll be interesting to see if you see similar results in that biomarker-negative population.

And just like the IO/IO, IO/chemo, I mean, this is where it’s at right now for combinations. There’s a number, mostly in ovarian cancer, that are being explored. But we’re starting to see some trials come down the pike with endometrial and cervical cancer.

So I think it’s an exciting time. We’ve got so much going on. But I think for us what the bottom line is is that IO alone is probably not going to be enough for the majority of the patients with gynecologic malignancies, and now it’s about really trying to tailor what that combination is that’s going to get the best benefit.