Tobi Arkenau:
Hi, my name is Tobi Arkenau. I’m the Executive Medical Director for Sarah Cannon Research Institute. I’m a medical oncologist and interested in GI cancers. And I’m really excited to, first of all, introduce you today to a team of experts in GI cancers who followed with me, virtually this year, the GI ASCO program. But before I introduce my colleagues, I want to thank VJOncology to really bring this new format on to discuss take-home messages from GI ASCO.

Let me introduce my esteemed team; Nataliya Uboha, who is Assistant Professor at the University of Wisconsin. She is a GI oncologist, particularly interested in upper GI cancers and neuroendocrine cancers, and also member of the NCI GI taskforce. Hello, Nataliya. Then Dr. Lizzy Smyth. Lizzy is a consultant oncologist, gastrointestinal oncology at Addenbrooke’s Hospital at the University of Cambridge. Lizzy is SMO GI faculty member and also leads the ERTC GI taskforce. Hi, Lizzy.

Lizzy Smyth:
Hi, Tobi. Thanks for inviting me.

Tobi Arkenau:
And then Kai-Keen Shiu is a colleague, an oncologist at University College, London, with me here in London. His research interests are GI cancers and he’s also running several trials in GI immunotherapy trials. He’s also leading the CAP taskforce here in the U.K. So welcome to all of you. I’m really excited to discuss with you the latest evidence and, really, also forward-looking evidence and for, hopefully, a successful 2021 in GI cancers. Kai-Keen, can I start with you? You presented the KEYNOTE-177 data. Give us a brief flavor, but also what I want to discuss with you are some questions, I will ask you about the outcomes you presented.

Kai-Keen Shiu:
Sure. Thanks, Tobi and thanks to the whole team. So, I was honored to present the updated data on this Phase III randomized trial pembrolizumab in first line versus chemotherapy. And it kind of seemed like a bit under radar because I think no one quite realized when I presented it that we were going to look at this progression-free survival 2 end point, which is important in this trial because there’s a crossover. So this was, we’d already shown that pembrolizumab was better than chemo for PFS1, 16.5 months versus 8.2. But now, with the crossover, everyone was wondering whether, actually, from a sequencing point of view, is pembro first best, followed by chemo, or chemo first followed by pembro. And that was because of that slightly concerning initial drop in the PFS in the first three to six months. So, actually, bottom line it showed that the curves continued to separate for PFS2. This is very interesting because there was a crossover of around 60% of patients who had some form of immunotherapy either within the study, in the context of the crossover design, or off-study. And so actually if you look at those lines, the median PFS2 has not reached is not reached at all, I think it’s 60% of people were non-progressing on their second minor therapy in the pembro first arm. Whereas it was only about 24 months for the chemo first arm.

Now we couldn’t answer all the questions in the chat and I’m sure you’re going to ask me lots of questions. My general take-home message on this is that it’s very promising as an interim analysis because we will get OS survival data in the next, I think, few weeks, or next few months. Personally, I think this, for me, means that it’s still a good thing for our patients to have immunotherapy in first line, as long as, there’s all these little bits of, should we sub biomarker patients and clinically biomarker them and stratify them, and that is for discussion, but it looks good. The lines do not cross so I do think, for some reason, despite that initial drop in PFS, and we can talk about pseudo progression and things like that, pembro in first line still would remain, from a Phase III randomized trial, the standard of care treatment now for MSI-high advanced bowel cancer.

Tobi Arkenau:
Okay, very interesting and very well summarized. So, one question is of course the concept of PFS2 and we’ve used it in other colorectal studies many years ago in the stop-and-go setting, so explain PFS2 to us, what does it mean and what do you want to capture and what really, for the newcomer to this question, why PFS2?

Kai-Keen Shiu:
Okay, so I think we wouldn’t have asked this PFS2 question if there hadn’t been a crossover. I think that’s the first thing. The fact is that everyone wondered why FD allowed it, crossover ethically was it the right thing, it would completely impair overall survival. I wasn’t at the beginning for the design of the study, but I think that actually if we hadn’t had a crossover with the chemo arm, we may not have recruited many patients at all, and that would have been a real shame.

Now in terms of PFS2, everyone defines it slightly differently and the difference is even within the study, I think that the only clean data for PFS2 is in the within-trial crossover because although I’ve seen some of the patient level data, it’s actually really interesting. Once you get to off-trial crossover/on-trial crossover, it’s a bit like tiger territory, you find out the patient didn’t want it, or the oncologist felt it was better, they wanted to do secondary surgery, or primary surgery, and it was quite strict actually how you crossed over within the study. So, in terms of clean data, the only really clean data is the patients who were on chemo, were fit enough to be re-screened at crossover with pembro, and seamlessly, I would say, crossed over to pembro. Everyone else, I suspect, [inaudible], was a bit of a mashup of the patient wasn’t sure what they wanted, surgeons wanted to operate, and once you have those things, it’s tiger territory or cowboy territory.

And I still think, however, despite that, you can never do a pure crossover study. So, I think that even with the, I mean you saw the data table of the variety of off-trial treatments, and trying to integrate that with patients who, for example, had had some form of curative-intent surgery or non-curative intent surgery. The numbers don’t lie, it’s still a lot of people coming [inaudible] into this study and the fact that the tail and curve remain tie, I think that whatever’s happened with the patients means that in real life, I just think we have to make sure that MDTs work properly, not just the standard MDTs but the genomic MDTs. And we continue to study this population through lines and actually, if we look at PFS2 data from the TRIBE trials, or TRIBE2 trials, it’s still better. These PFS2 results within the study after the 23.5 months is pretty good, so whatever’s happened with the patient, they are living longer than we expect, whereas previously we thought probably they weren’t going to do very well. I hope that kind of answers your question.

Tobi Arkenau:
Yeah super, very much. As you know, your New England Journal publication, the subgroup analysis looked into the molecular genetics of these subtypes and interestingly the RAS-mutant cancers didn’t do so well as subgroup analysis. Can you explain this and particularly, I guess you have the RAS status, but do we have other genetic alterations like we know in lung cancer, for example, KRAS FCK11, non-responses to immunotherapy, is there similar data available, we haven’t published them of course, but along those lines, is there an early hint or sign why those patients didn’t do so well in the subgroup analysis?

Kai-Keen Shiu:
Yeah, so there’s the science bit and there’s the gut feeling. The gut feeling is “oh my god” because first of all, we already are struggling with our KRAS mutant subgroup, and we’ll talk about some other maybe interesting small molecules that are coming through. But we know in the lung cancer field that KRAS mutant lung maybe don’t respond so to immunotherapy, is that going to be contextual and similar in bowel? I think that the main thing that we have to say is that I don’t think we shot ourselves in the foot but when we planned the study, all these preplanned subgroup analyses were not statistically planned in that way. We just thought it was quite interesting to know. And the second thing is, for good or bad, we were really strict on what we say whether we knew the entire genomic analysis. And when I mean entire, I mean extended KRAS, NRAS, and BRAF. So, if you look in the New England Journal, the rule was, if you were missing any one of those three, that was un-evaluable.

Now in real life, that’s not how we run, right? So, for example, you do a rapid KRAS test, if you’re KRAS mutant, the presumption is, you’re not NRAS mutant or BRAF mutant because you hardly ever cross over. Even in real life, here, you see it, we don’t always run the whole test, if we hit the KRAS mutant first. So, I think that we will, because we’ve asked the science to give us the extra data over time, we will struggle because it’s not bio-banked per se to do extended profiling, so you’ve asked the question about TMB, SDK11 and I think that will come. But as you know, in real life, this is a Phase III trial, not a Phase II trial, not a Phase I trial and in some aspects, you want to run with the clinical endpoints which are quite solid, I think. And yet, I think it’s up to us as the academics and the clinicians on the ground, to pull that data off, and we will, we’ve been told by the sponsors that are allowed to do that. But you know they want to push very much for the primary endpoints and then it’s for us to work out what to do in the future. I said in some of my other talks, I said, actually is in our gift to sort that out for you guys but also, I think in the neoadjuvant setting, we will get those biomarkers in a much more clean and quicker way for all our patients.

Tobi Arkenau:
That’s very interesting. I assume lots of new and important outcomes of this study is coming when you present it in future meetings. I think one of the key aspects of this study as well, when you presented it, was quality of life, again our patients want to hear this and how much do they improve or has their quality of life been affected. That’s the day-to-day discussion we have with patients on the ground. So, what did the study show?

Kai-Keen Shiu:
Well, there was two things. There was the standard AEs, which were significantly better, the grade 3 AEs, 22% versus 66% but also in terms of, in fact even the immune-related side effects, pretty safe, less than 3% chance of a grade three colitis or hepatitis. The ESMO virtual data was really nice because that showed, the patients had iPads, and every time they filled in their own quality of life and what they felt. Certain things that I can say now like weight gain. In fact, quite a lot of weight gain, some patients felt grossly overweight and actually said “I’m overweight now” and particularly when we were planning for surgery. Overall, I think for our patients, it’s nothing like having chemotherapy. If patients had had chemo in the neoadjuvant setting, it’s nothing like it. Even from that point of view it’s better.

Even now we go back to the KRAS subgroup which I still think is a subgroup of the subgroup analysis, you can pretend that maybe for some of these patients you regard immunotherapy as a block of therapy, not necessarily the life changing immunotherapy for some of our patients, but it’s still a good treatment that will keep them alive, with good quality of life and fit of further treatment. That in itself, I think, is worth having immunotherapy, until we untangle all the other milieu and minutiae of biomarkers.

Tobi Arkenau:
Good, that’s very well summarized. Of course, the million-dollar question is, give us an idea about what we can expect in 2021? Overall survival data, are they coming?

Kai-Keen Shiu:
I mean, I don’t know. But I can predict that with those curves not crossing at two years and if you think about some of the second line data we’ve seen from KEYNOTE-164 and from the CheckMate studies. I think it’s unlikely, as long as whether you’ve had first line and second line treatment you have not progressed after, maybe, 8 months or a year. You won’t really progress, a few of them will progress. So already you’ve lost the window of opportunity but not having pembro for them first line and you can see that the difference is around 15% with second line therapy will probably read out, even OS. And we are about to hit OS because that data we presented is almost a year old anyway. So that interim analysis was from February of last year and we’re about to, even if we don’t get 190 deaths for OS, we’ve been asked the IDMC and the statistical plan is to do the data lock in the next few weeks.

So my prediction is that it will be clinically significantly different OS. The interesting thing will be the statistical difference may be closer than we expect, but I still expect it to be less than 0.05, for example.

Tobi Arkenau:
Don’t tell too much, that’s very exciting. Let me switch and again, also your field, a real novel analysis on CAR-T cell and colorectal cancer, just talk us through, I think, it’s innovative of course, your view on this.

Kai-Keen Shiu:
So, there was a poster abstract by the team in Belgium and it was the alloSHRINK study. Now this was interesting because it was only about 16 patients but we’re all looking to see whether CAR-T cells will work in solid cancers. So, UCH we are doing currently the CHIRON study in lung cancer with Charlie Swanton and Mariam Jamal-Hanjani, very exciting. Essentially, this study is interesting because they’re using what we call non-gene editing, so it’s almost like off the shelf CAR-T and trying to engage natural killer cells. The only thing I would say is, when I really look at the data, they were using various dose strengths for safety, they were using some kind of induction FOLFOX in the refractory setting to condition these patients. And the way I looked at it was, actually there were only two responders out of nine. The rest were stable disease, very little side effects because they were also targeting TIM, which is also a thing that stopped graft-versus-host. But I didn’t know the patient level data, so my cynical view was that some of these patients were not chemo refractory. They were responding because they were pattern sensitive.

But equally, I think it’s important because they did do some really nice biomarker work looking at some gene expression assays and TCR complexes and they did show that in one or two of the responding patients there was a change in expression profile. So at least they were looking. I still think it’s way too soon, and I mean, they’re jumping already into thinking about adding pembrolizumab or some other immunotherapy to their milieu. And it makes me excited, to a degree, but slightly cautious that, we’re not there yet. Though I know that all our patients are saying, look I don’t have the solid biomarker for MSI, surely by gene editing we’ve seen it. And you know those institutions in the U.S have done one or two amazing case reports, but they don’t say all the failures, so they only talk about the successes.

So, bottom line is, very interesting study. It’s an early, kind of, Phase Ia/II. They are going to recruit more; they are going to start adding other markers. I’d like to know what Lizzy and Nataliya think in the upper-GI and biliary tract field. Because at least it’s the first time I’ve seen an off-the-shelf CAR-T therapy delivered into a common, solid cancer, like a GI cancer.

Tobi Arkenau:
Nataliya, Lizzy, any comments?

Lizzy Smyth:
Well yes, that’s, I suppose, a widely expressed target and they’ve done trials also in hematologic diseases, I think they were presented in ASCO. It’s good to see the off-the-shelf or the allogeneic T-cells being developed. I agree that it’s very early days and the activity is limited. It was said that the FOLFOX was given to lymphodeplete, now I agree with you in that FOLFOX is the most lymphodepleting regime that I can imagine. So perhaps another regime would tease out whether the benefit was from 5-FU and oxaliplatin rather than the CAR-T cells.

In gastric cancer, there has been one trail that was presented as ASCO last year on Claudin 18.2 targeted CAR-T cells, it was a Chinese study. You know, it was slightly disappointing in that it did not show tremendous activity, there was some very minor responses in I think about 13 patients. But what it did show was on-target toxicity because some of the patients had GI bleeds. And I think this is the key when we’re talking about CAR-T cells, isn’t it, with our epithelial antigens. So, I’m not an expert and I’m glad to see this being developed, it’s very exciting. And, we are seeing some, also, of the autologous HLA-matched trials coming through. I’m sure Tobi you see them in Sarah Cannon as well. And that’s where we need to be, but I think we all have to up skill on lymphodepletion and management of CRS. And it’s exciting times so I think it’s early days yet.

Nataliya Uboha:
I agree, I don’t know of any data of the CAR-T cells for hepatobiliary cancers but there are data coming out for the CAR-T cell therapies for upper-GI cancers. AdaptImmune is launching a Phase II trial with a CAR-T therapy. I think in general, it’s exciting to see this technology moving to solid tumor field. I think it, hopefully, will be a breakthrough that we will see in the next few years as well.

Lizzy Smyth:
I think AdaptImmune is coming in upper-GI as well, it’s fantastic.

Nataliya Uboha:
Esophageal, exactly, that’s what they’re opening, not in pancreatic or biliary, but esophageal but they have a study ongoing in HCC already.

Lizzy Smyth:
Yep.

Tobi Arkenau:
Okay, let’s switch the gear to gastric cancer, Lizzy that’s your field. And I summarize effectively, old targets, new drugs. I mean, ten years ago we were targeting the FGFR receptor with various different small molecules. Now you see a newcomer, a new kid of the block, bemarituzumab, talk us through.

Lizzy Smyth:
Yeah, I’m very glad to see this. I was actually pleasantly surprised because when you mentioned ten years ago you mentioned we were working on FGFR in gastric cancer, that was me, so we were targeting it with small molecules TKI’s and really, these drugs worked but it was very hard to find the right patients because you needed to find really highly amplified patients for them to be effective. So that pathway of development didn’t work for these drugs, it just wasn’t feasible.

So what we have in the FIGHT trial is bemarituzumab, which is an FGFR2 beta monoclonal antibody. And they chose patients who express protein at the IHC level, and I think that was about 30% of gastro-esophageal cancer screens, so quite high level. And a smaller proportion of those were amplified, so I’m really interested to understand at a deeper level what the pathways there are in terms of the amplification being connected to the over expression. This was a small trial. And this shows how difficult it is to recruit to biomarker-selected studies in the first line for gastro-esophageal cancer because ultimately these patients have difficulty swallowing and we want to get them on treatment quickly. So sending their tissue off for biomarker selection has always posed a difficult. It started out as a Phase III trial and then came down to a Phase II. But they did one aspect, which was quite smart, which was allowing a cycle of FOLFOX before the biomarker results came back. And I think this is a good way to design trials for this patient population because it gives us that little luxury of time before biomarker result comes back. And they’re getting standard of care in any case.

So, the results were good, the bemarituzumab improved overall response rates, progression-free, overall survival by a couple of months. I will say that the control arm did well, we often see this in Phase II trials. But I was fairly convinced of the results but it’s a small trial, it’s not going to change practice. What we need to do is see results in a larger portion of the population, I think.

Tobi Arkenau:
What I found quite interesting is the level-entry was immunohistochemistry 2/3+ and then 5%, 10% expression and you clearly could see that the more than 10% expressors did much better in terms of overall survival. And do we need to, even in this selected group, like we do with PDA1 expression, have a high responder group where we can do more.

Lizzy Smyth:
Well, this is what we learned from trastuzumab in gastric cancer as well. We know that these tumors are difficult to target with biomarker selected drugs first of all because of biomarker expression heterogeneity and we’re still working to optimize, I think, biomarker selection for HER2-targeted therapies in gastric cancer. So, certainly it seemed that the tumors that expressed higher levels of the protein were more addicted to FGFR2 signaling, as you might expect. But I think in order to fully refine the biomarker, you probably need to treat more patients and, again, possibly refining it too early, before going to a Phase III trial might make a Phase III trial more difficult to recruit to and to get this as a treatment for patients with esophageal cancer. So I think there’s always a balance and very difficult to make those decisions early in treatment.

Tobi Arkenau:
You know, I was quite surprised about the… I think Kai-Keen has, go ahead.

Kai-Keen Shiu:
I was just going to ask Lizzy and Nataliya, the FGFR pathway is better known in the biliary tract setting now, cholangiocarcinoma. So one, what is this difference of this inhibitor to cholangio-world and cholangio moves in kind of snail pace with cis-gem for a long time and then added in these drugs in last time and it actually now looks as good. You look at some of the FIGHT data, maybe some of these FGFR inhibitors are better than first line cis-gem and they’ve gone straight into that kind of that Phase III setting. So the [inaudible] pathway may be different, or it is just, contextually, gastric cancer is still a very difficult cancer to treat compared to maybe now intrahepatic cholangio. So, what’s your feel of the drug versus the context compared to other tumor types?

Nataliya Uboha:
I think it’s probably all of the above. I think definitely, it’s a different way, the pathway is dysregulated in cholangiocarcinoma. In cholangiocarcinoma, the initial studies were done looking at any FGFR2 alterations, were done at FGFR2 fusions, or no FGFR2 alterations at all, that was the initial design of the inside trial. And the only place where the see the signal with FGFR inhibitors is in cholangiocarcinoma, which are positive for FGFR2 fusion. And it could be any fusion, there is multiple new partners that have been identified, the most common one is BICC1, but it really doesn’t matter what FGFR2 is fused to, it’s just the fact that the fusion results in dimerization of this tyrosine kinase receptor which is essential for its constitutive activity. It’s the fusion that matters.

And so, in cholangiocarcinoma, we use small molecule inhibitors to target this pathway and trying to turn the pathway on. And it has been quite successful and you’re right, it’s been studied in second line, although now we’re trying to move these drugs in to the first line. But I think it’s more logistical issues, right. It’s a rare cancer, it’s only intrahepatic cholangiocarcinoma, primarily intrahepatic cholangiocarcinoma, that even has the fusions, then if you see it in 15% of the rare tumors, right, and it takes a while to get the fusion results. And not even every NGS test is providing reliable results on the fusions. So there’s a lot of different limitations.

Kai-Keen Shiu:
Yeah, mm-hmm.

Nataliya Uboha:
While the way it was studied in gastric cancer, you looked at protein expression, which is different. You see similar difference in benefit in this Phase III study in what we see even in ToGA study, right, it’s a two-month benefit. In ToGA study, the addition of trastuzumab resulted in two months benefit in overall survival, or roughly that, compared to FOLFOX and yet nobody is questioning the use of trastuzumab on HER2-positive gastric cancer. So it’s probably not as strong of a driver as it is in cholangiocarcinoma.

Lizzy Smyth:
I couldn’t agree more. Also, in gastric cancer because you’ve got these really chromosomally unstable cancers, you’ve got co-amplification of multiple or TKs often and so what you do, you’re playing whack-a-mole, so you treat HER2 and FGF4 or EGF4 pops up and I think that they’re really early resistance biomarkers are probably pre-existing in gastric cancer. Maybe not to the same degree in cholangiocarcinoma.

Nataliya Uboha:
Yeah, I agree. I think chemotherapy is going to be a part of the first line treatment for gastric cancer. These patients come in with a lot of symptoms, we need responses. And this is what chemo provides, right? And that’s why when we try to compare immunotherapy to chemotherapy, well people didn’t do as well because immunotherapy, when it works, it doesn’t work fast enough in these patients. And as Lizzy said, people come into our clinic and they can’t eat and you don’t have time to wait for two to three months for something to kick in, they need responses. And so, I think that’s the difference, in gastric cancer we will be adding these targeted drugs to chemotherapy and hopefully finding the patient population who can benefit more. Although in cholangiocarcinoma, we are actually trying to see if we can avoid gem-cis and just give them FGFR inhibitor because it works well enough. Hopefully, in the first line.

Tobi Arkenau:
Lizzy, just one more question, on toxicities. I mean very classical FGFR toxicities, too much tiredness, dry eyes. But not phosphate increase, again this is probably simply the biology, we’re talking antibody against small molecule. But gastric cancer population, stomatitis, it’s not the most pleasant toxicity, how do you manage this? The investigators who did this study, how did they manage those symptoms?

Lizzy Smyth:
I didn’t recruit to this study but you’re right that this is an on-target effect of FGF4 inhibitors. Actually, it was quite impressive that almost a quarter of patients came off the study due to toxicity. And I think the stomatitis with FGF4 inhibitors, when I used these drugs before, we often used a special mouthwash. Raspberry Mucilage tended to help. But it’s actually quite difficult for the patients and I think that could impact on quality of life, as does the eye toxicity. And with these drugs you really need to make friends with your local ophthalmologist when you’re treating patients on these drugs. And early intervention is important and perhaps treatment breaks will be important, and these move forward in development. But how it was managed in the study, I can’t comment on specifically because I didn’t recruit.

Nataliya Uboha:
I was surprised to see in the abstract that is was a quarter of patients who went off bema…

Lizzy Smyth:
Yeah.

Nataliya Uboha:
… Because of ocular toxicity. I have to say we don’t see as much ocular toxicity with small molecule inhibitors in cholangiocarcinoma. It is listed as a side effect, ophthalmology is involved early, during clinical trials now with the drug that’s approved. But I have not seen significant eye toxicity with a drug that required discontinuation of the drug, short of dry eyes. So I was surprised to see that type of toxicity with an antibody, I think it is different than the small molecule inhibitors.

Tobi Arkenau:
Well, I think the antibody half-lives are maybe affecting this as well. And effectively different to the small molecules is a much shorter half-life so patients can stop. But really, I must say the eye toxicity is certainly of concern. No phosphate levels, again, which is the other thing we see in cholangiocarcinoma often and that bothers us many times. Phosphate levels are pretty much normal.

Lizzy, one more. HER2 and zanidatamab and ZW25, an interesting drug, bispecific HER2 antibody. Talk us through this study and where do you see, particularly with in HER2, just being approved on the same day they presented these data, by the FDA. So where do you see both drugs running into the first line and give us a bit of context.

Lizzy Smyth:
So, I was quite interested… the first thing I should say is that after ToGA, HER2 treatment of gastric cancer has spent 10 years in the wilderness. So we’ve had many negative trials in the first line setting, we had JACOB, in the second line we had lapatinib, we had the T-DM1 studies. And we have been disappointed every time, sadly for our patients. So this year, we’re really seeing this massive explosion of active, anti-HER2 drugs in gastric cancer, which is fantastic. And this ZW25, zanidatamab being one of those, the other being trastuzumab deruxtecan, which just got a license as an orphan drug in the States, which is fantastic.

So this was a relatively early study, not too many patients. We saw activity of the drug as monotherapy, which is interesting. A 30% response rate and also, some activity. Progression-free survival was around, I think, 3 months, which is good for monotherapy antibody in this disease. Also, activity with taxanes. The field is getting quite crowded at the moment, we have ZW25, we have T-DXd, or trastuzumab deruxtecan, we also have margetuximab, which also presents very nice data in Lancet Oncology, in particular in combination with pembrolizumab in dual HER2-PDL1-positive patients.

So I think it’s really a race to the finish. I think we really need to consider perhaps the toxicities of trastuzumab deruxtecan, in particular ILD on patients on long-term treatment. May pose a challenge for bringing this into the first line or into earlier lines of treatment. That said, I’ve had a lot of experience using that in clinical trials, it’s not something that I have observed so far. So I think we really need to see larger numbers. We’re basing our comments on the Japanese trial, which was a small number of patients, also an Asian population, we know that they respond differently to Western patients. So, I’m really looking forward to seeing the data for Western patients for trastuzumab deruxtecan. What we will see is, combinations first of all with chemotherapy, also moving into the peri-operative setting, hopefully we can cure more patients. But what we’d be really interested in is the breaking down of ever smaller molecular subsets. So I think what’s interesting is in the MAHOGANY trial, which is the first line trial of margetuximab, plus or minus PDL1 or LAG3, that in the super selected HER2-3+ PDL1 population, they are going to investigate a chemotherapy-free regime first line.

And as Nataliya just said, that would be a tremendous achievement given the fact that we really need response rates first line. So I’m interested to see what we can achieve in that double positive population, but I think that that’s only ever going to be for very selective patients.

Tobi Arkenau:
And did you see anything compared to different chemotherapies, obviously mono, cape plus zanidatamab and then paclitaxel. There was quite a marked difference between the PFS and duration of response of the paclitaxel versus cape. Is this small patient numbers or do you see something, you know, better synergies or… your view please?

Lizzy Smyth:
I can’t imagine why… I can’t remember if they were first- or second-line patients but I imagine if they were previously treated that they may be a little bit 5-FU refractory because we use 5-FU in first line treatment. So, that would be my working hypothesis. But to be honest, I think that it is small numbers, you’re only looking at I think 10 or 15 in each of those lines so I wouldn’t rule out 5-FU as a partner at this point in time.

Kai-Keen Shiu:
Can I ask… Lizzy can I just.

Tobi Arkenau:
Right.

Kai-Keen Shiu:
Can I just ask you a quick question? You’re absolutely right, this biomarker selection stratified is one way to go but you’ve also said on the other side, chemo to rescue this patient is the auto way to go. And I think that you’re absolutely right, if you go in with the deruxtecan first, you don’t know what to do next and I think the natural feel is that if you jump in too fast you lack an option in the future. So my natural feeling is, and I don’t want to seem bias because I’m running the Phase III combo, herceptin plus or minus pembro, and I’m seeing some very good responses there. And I’ve sent some patients to you in second line for deruxtecan. To me, it feels that, before you jump into first line, I still think there’s a sequencing question that we haven’t answered. I mean, you’re right, if you look at the ToGA trial, how are we doing this for 10 years. And I think with the immunotherapy combination of LAGs, it makes you wonder that we’ve always thought that the HER2-positive, their immune landscape is not particularly good, it’s a cold tumor. So, what are these bivalent drugs doing, what is margetuximab doing to make that any different in, what I call, a very clonal way. So, unlike my MSI subgroup, there is no, unless you’re MMR deficient gastric, it’s a messy, messy micro-environment.

So do you really think that we should be pushing straight into a chemo-free regimen because BECON wasn’t great for bowel cancer, to be honest…

Lizzy Smyth:
What I would say is for the margetuximab anti-PD1 study, there is a run-in, they didn’t jump straight into a Phase III for the double-positive population. And margetuximab is optimized in terms of it’s FC-gamma expression to stimulate the innate immune response as well as an acquired immune response. So I think it’s possible that it could overcome some of those difficulties which are in the immune-evasive nature of the chromosomally unstable high copy number gain tumors. But you know what, I think the sequencing question is very true and I would absolutely love to be in a position in five years to have the luxury of sequencing anti-HER2 therapies for gastric cancer patients who are HER2 positive. I think that would be brilliant because we’ve learned a lot. It’s literally only been in the past two to three years that we realized that these patients, at least a third of them, do not benefit, cannot possibly benefit in a second line setting because they are no longer dependent on HER2 signaling. And this is emergent and I’m sure that we will learn about the integration of immune-checkpoint blockade too.

But I agree that, certainly I wouldn’t throw chemotherapy out with the bathwater immediately and it certainly would need careful thought. Even if it was a licensed treatment, you would still do careful clinical patient selection, I think. And I think that’s something, as Nataliya alluded to, that we learned very strongly with the immunotherapy studies in gastric cancer.

Nataliya Uboha:
I think it’s a good problem to have, to have many different therapies and we’ll be able to sequence from. And I think the interesting thing about DS8201 and HER2 drug, the antibody drug conjugate, that in breast cancer, they see responses in those patients who have lower expression of HER2. So we may be able to identify more patients who benefit because this is not a drug that turns off necessarily the signaling, it’s an antibody drug conjugate that uses HER2 receptor, as a way, at least that’s how I explain to my patients, as a way to bring chemo into the right location, right. You might not need as much HER2 so it may be an option for patients who are not candidates for these other anti-HER2 agents as well, we need more data.

Lizzy Smyth:
That’s a really good point and I think the data for the HER2 low for T-DXd was presented at ESMO. I can’t remember but I think the response rate was something like 20% in the HER2-2+, but the HER2-1+, it was negligible. Again, in the Asian population. So, absolutely worth chasing for those patients too, that’s a really good point.

Tobi Arkenau:
So Lizzy, in the last two meetings we talked about immunotherapy, gastric cancer and PDA1, TMB, MSI, and I think the Japanese gave us some hints on what other biomarkers we can use. Particularly the DELIVER trial, the Japanese group, presented on the gut microbiome in response to nivolumab. What are your thoughts, comments, do we need to take poo samples in future for all our patients?

Lizzy Smyth:
Thanks Tobi, so first of all, the Japanese are so far ahead of us in terms of immune-checkpoint blockade for gastric cancer. They’ve had nivolumab licensed for a number of years in chemo-refractory patients based on the results of ATTRACTION-2 whereas unfortunately we did not have that in Europe. And the States have had pembrolizumab for a number of years based on KEYNOTE-059. So what they’ve done in DELIVER, we’ve seen data presented before in other immune-sensitive populations, melanoma, non-small cell lung cancer, that the gut microbiome may be a predictor of response to immune-checkpoint blockade. So what they have done, in a nice prospective manner, is looked at the data for, I think there were 200 patients in the exploratory data set and then validated that in a second data set. I find the analysis of the gut microbiome genomics extremely complex, and I won’t claim any special expertise in it. But what they did was identify two genus of bacteria that seemed to be more associated with response to nivolumab. That said, the number of responders was quite small, so it was difficult to be precise in the estimate.

So, in general, we think that the gut microbiome might relate to immune-checkpoint blockade through probably different bacteria causing expression of short-chain fatty acids which influence T-cells. And one of the challenges on interpreting that data is that most groups who do this work seem to use a different methodology and they all end up with different results. But that doesn’t mean that those results are not valid it’s just that we don’t have a standardized approach to this. And the main question is, is this something that exists that causes that patient to be a patient who is likely to respond to pembrolizumab or nivolumab, or is it something that we can manipulate and cause in other patients, get it to happen. So, we know in mouse models, actually, if we take a fecal transplant from responding nude mice, who will respond to immune-checkpoint blockade, that we can cause non-responding mice to respond. Now, I’m not suggesting that we do fecal transplants for our patients, I don’t think we’re there yet. But these are questions that could be asked prospectively, certainly, and all I can say is that we can see an association and I don’t know about you, but I tell all my patients on immunotherapy to have yogurts and kefir and whatever they can do to help their gut microbiome. Based on very little evidence, but I don’t think it’s likely to do them any harm. I’m not sure, does anyone else do that?

Kai-Keen Shiu:
Lizzy, all I’ll say is I think there is some data from MD Anderson or on the East Coast, they are doing fecal translate for melanoma patients and it seems to work. So, it’s coming. I suppose our general feeling as a GI-oncologists is that your gut is a little bit messed up already from having a bowel cancer, so we don’t know. I just had a patient who said, “If I change my diet completely, do I not need to have a different chemo and get a certain percentage of survival by not smoking and eating?” And I said I don’t know that data and I’m not sure we could say. If you feel good eating a Yakult, fine, do it. But I’m not sure that will get you to respond to an immune checkpoint.

Lizzy Smyth:
I totally agree, but I think it’s probably not likely to do any harm, so I don’t know. I should be practicing a more evidence-based medicine. What do Nataliya and Tobi think?

Kai-Keen Shiu:
Nataliya, what do you think?

Nataliya Uboha:
I’m going to start recommending yogurt and kefir for my patients after this meeting.

Tobi Arkenau:
Yeah, absolutely. I think what we learned from the melanoma groups and certainly the use of antibiotics, for example…

Lizzy Smyth:
Yes.

Nataliya Uboha:
Yes.

Tobi Arkenau:
… we as oncologists are trained so quickly to use antibiotics. Stop it, hold fire, because that really in melanoma patients had an impact. There was little intervention, it’s sometimes great, rather than paying whatever in first cover drugs which cost you a fortune. There was little interventions, let’s work towards including our patients into these national trials. But I think very important things to consider.

Nataliya, for you, 2020 must have been a firework of success and simply for these very rare tumor types, particularly intrahepatic cholangiocarcinoma. Two drugs approved and again, the GI ASCO this year really kind of summarized some of those a bit deeper into this. Infigratinib, new kid on the block, the FGFR inhibitor. Talk us through. Particularly, I’m interested, I’ve been heavily involved in the Taiho compound, TAS-120. And I want to know what is the difference between the two drugs, which one would I give my patients? Would I treat patient before first line and then second line rescue them with another molecule? So, I think lots to learn at this GI ASCO.

Nataliya Uboha:
So cholangiocarcinoma has become a poster child for precision oncology in GI cancers, it’s our form of lung cancer where we actually sequence and get results that are actionable. So, it’s about 15%, as I alluded to earlier, of intrahepatic cholangiocarcinomas that have FGFR2 fusion and we now know that going after the FGFR2 fusion with small molecule inhibitors makes people live longer and improves patient survival. There’s a drug already approved for that indication, pemigatinib. But there are other FGFR inhibitors in development, one of them is infigratinib. The results of the study with infigratinib were presented as GI-ASCO but there’s also other ones, futibatinib, TAS-120, the drug from Taiho, as well, is another FGFR inhibitor. TAS-120, or futibatinib, is the only covalent inhibitor, irreversible inhibitor, the other inhibitors are reversible so I think that would distinguish TAS-120 from the other inhibitors. But all of the data for these inhibitors comes from Phase II studies, so they’re all too small studies and think logistically that’s what we can achieve in a reasonable time with these drugs in this rare patient population.

So, the study with infigratinib was another Phase II study of patients with advanced cholangiocarcinoma who have regressed or were intolerant to gemcitabine-cisplatin. And this was a 108-patient study, comparable in size to other Phase II studies, patients received infigratinib on a 3 weeks on, 1 week off schedule which is, again, slightly different to the way TAS-120 is given. Which is in the study with TAS-120, infigratinib is given on a continuous schedule. The response rate was promising in this Phase II study, it was 23%. Progression-free survival was 7.3 months, which is again comparable to the other inhibitors in the market. As we all know, you cannot do cross trial comparisons, however, the overall response rate of 23% was slightly lower to what we have seen with other inhibitors. And again, you can dig into the patient population, maybe this patient population was more heavily pretreated, it was at about 46% who had up to one prior line of therapy as opposed to 61% in the study with pemigatinib. But again, the numbers are too small to really make any final conclusions from that and all of the studies are still small, and you can’t compare across studies.

But progression-free survival was similar, was 7.3 months. The overall survival was lower again compared to what we’ve seen with pemigatinib where the overall survival was close to 2 years. The overall survival with this drug was 12 months. I think, Tobi, you bring up a good point, we have now all of these inhibitors coming on the market. One of them is FDA-approved and the other ones are not, yet. The question will be, which one is better, how do you select the right inhibitor for the patient, and we really don’t know. There was some data that was presented and published by Dr. Goyal from MGH that showed that TAS-120 has some activity even when the patients have been pretreated with other inhibitors. But again, these are small studies and single institution experiences. So, that’s the question, can we sequence them, is the mechanism or resistance different? And also, the toxicity, which will help us decide which inhibitors to use, these are all good questions to which we don’t have any answers yet.

Tobi Arkenau:
Absolutely, you’re absolutely right, this is early data but consistent data. I’ve seen consistent in terms of toxicity profile, hyperphosphatemia, stomatitis, many changes of these classical FGFR. That’s why I was quite intrigued that the antibody didn’t have hyperphosphatemia, so effectively hitting the target on a different way from an antibody point of view versus a small molecule point of view. It’s certainly different.

What I was wondering, and you kind of alluded to very early on, on the different re-arrangements fusion partners, do you seen any, so TAC1 being or TAC3 being the most common ones, do you see different response rates or is it again too early to say?

Nataliya Uboha:
There is no difference, and this was published from pemigatinib. There is no difference in activity with respect to the partner that FGFR2 is fused with. We see activity across different partners.

Tobi Arkenau:
Yes, okay, great.

Let’s move on to ClarIDHy, another big success story from last year. Ivosidenib in intrahepatic cholangiocarcinoma, talk us through, in particular I’m interested in the duration of response and duration on treatment and then aspects of quality of life particularly, because they were presented at GI-ASCO.

Nataliya Uboha:
ClarIDHy was a randomized Phase III trial that enrolled patients with IDH1-mutant cholangiocarcinoma who have regressed on first line chemotherapy. In this trial, patients were randomized in a 2:1 fashion to either IDH1 inhibitor ivosidenib or placebo. And placebo was an appropriate control arm when this study was launched because we did not have any data for second line therapies back then. This study, which enrolled around 180 patients, allowed crossover upon progression from the placebo arm. So there were 60 patients in the placebo arm and 70% of those patients were able to crossover. So the primary endpoint because of that crossover was progression-free survival. Progression-free survival was significantly improved with the treatment of IDH1 inhibitor although if you look at the absolute numbers, they are low, 2.7 versus 1.4 months. It’s hard to get excited when you look at these absolute numbers. However, if you look at progression-free survival rates at 6 months or 12 months, you do see, I think, clinically meaningful activity. One third of patients were free of progression at six months in the experimental arm as opposed to none in the placebo arm. And over 20% were progression-free at a year.

So I think it is an active drug, it has been used as a drug in clinic. I have seen how people have done on this drug. I think it does provide clinic and meaningful benefit to our patients in part because it is so well tolerated. It’s an oral drug, there’s limited toxicities and as you’re alluding to Tobi, they did present quality of life data. And it has shown that the pain scores were better, and the functional activity was better in patients who were on an experimental drug, as opposed to placebo. So we are providing, I think, a clinically meaningful benefit to patients with this drug. For the overall survival data, the absolute numbers were different in the experimental arm versus placebo arm, but they were not statistically significant. In part, because of that crossover. They did this fancy statistical analysis which ranked preserving structural failure time, I had to write I down for myself, I can’t claim that I understand how it’s done. This was something that was preplanned to account for the fact that the placebo arm will do better. When they used this analysis, then they did a see a statistically significant difference in OS between the experimental arm and the placebo arm. In the placebo arm the OS was 5 months…

Tobi Arkenau:
No PFS2? No PFS2 in this study, I’m just making a little joke.

Kai-Keen Shiu:
No.

Tobi Arkenau:
We discussed the drug zanidatamab in cholangiocarcinoma, also big story, a small, big story in a small study but I think there’s some interesting data in cholangiocarcinoma. Talk us through that zanidatamab as well, just to wrap this drug up, it has really had very different subgroups including the 21 patients in cholangiocarcinoma.

Nataliya Uboha:
Yeah, I know, I think, I mean, it’s the same drug Lizzy just discussed in upper-GI cancers, it’s a dual antibody that probably engages some immune activity of our own immune system against the cancer cells. It was a small study, with HER2-positive cholangiocarcinoma, and a single drug demonstrated some activity, I think there was a 40% response rate. Confirmed response rate was 47% in a tiny study. But again, I think we have a rare disease, a small percentage of patients even have HER2 expression, so I think it will be a challenge how do we decide that approval is warranted for these particular drugs that are geared to such a small percentage of patients. And yet we do see some activity in these small studies. Certainly, you need a larger study, but I was excited to see….

Tobi Arkenau:
It’s certainly likened with…

Nataliya Uboha:
Similar to gastric cancer but also some of these patients were pretreated with trastuzumab which is also interesting.

Kai-Keen Shiu:
Nataliya, can I ask a question?

Tobi Arkenau:
My pathway has shown with pertuzumab and herceptin, so effectively reassuring that the biology would treat, so often as you know, let’s say about BRAF in colorectal and melanoma. Different biology but here it seems to be that at least luckily, we could talk about BRAF in cholangiocarcinoma as well, the BRAF mutation, it is a very similar biology to the drugs approved in various other settings as well. Kai-Keen?

Kai-Keen Shiu:
Just a general question for Nataliya because I’ve got in an interest in cholangiocarcinoma. I work with Professor Bridgewater at UCH and we get a lot of second opinions where patients are literally just buying their own profile from somewhere and then they come to the clinic and say what’s this showing? And often we say, not the right profile for that trial or that study. In America, and I don’t know Lizzy how you find it at Addenbrooke’s with your academic pathway, what is the general recommendation now for the patient who is out there, not linked in with an academic center, desperate, looking at the data, I’ve got patients trying to login to ASCO GI and ASCO. They just want to be profiled, so I’m just interested particularly with all this difference in FGFR, with all these drugs, and you’re never going to do Phase III trials still, I think. How do you manage your patients who, outside of trial, come to you with a profile and then you have to tell them that, in fact, you have to spend another many thousands of pounds or dollars for another profile? What is the best way to manage our patients?

Nataliya Uboha:
So, I think in the U.S, most of our patients have access to platforms for next generation sequencing. In my practice I never look only for only FGFR2, or only for IDH1, we do this large platforms…

Kai-Keen Shiu:
But what platforms? Because there’s FDA-approved platforms and they come off trials and then a lot of companies are just, kind of, piggybacking a bit and saying we can also do very similar things.

Nataliya Uboha:
We use a lot of Foundation Medicine, KRAS and [inaudible] in the U.S. I use Guardant360 in my practice and patients who don’t have enough tissue, which is frequently an issue for patients with extrahepatic cholangiocarcinoma. Liquid biopsies are starting to be incorporated as eligibility criteria into clinical trials. It was an issue two to three years ago, you couldn’t really make clinical decisions based on these, or even enroll patients in trials but they are allowed now for clinical studies. I do see patients from the community for second opinion, the majority of my patients of cholangiocarcinoma come in with full profiles so we can make decisions based on that information. That information is available for review and it’s a full panel, usually it has microsatellite instability, TMB, amplifications, fusions, done early during the course of the disease usually at the time of, you know, in my practice, my patients get NGS testing done at the time of diagnosis of advanced disease.

Tobi Arkenau:
Great. It has been a very good discussion. I must say, we really covered everything in just a bit over one hour for colorectal, gastric, biliary. I think we can wrap it up here, I know we have other data, of course, of interest, pancreatic cancer. But I think we keep it for the next session and probably around ASCO and ask VJOncology to have the same panel on again and discuss the key highlights. It has been a real pleasure and I want to thank you very much and good luck for 2021 and stay healthy. Get vaccinated, another really important thing for this year.

Nataliya Uboha:
Thank you. Thank you so much for inviting me.

Lizzy Smyth:
Thank you.

Kai-Keen Shiu:
Thank you.

Lizzy Smyth:
Thank you for inviting me. Bye bye.

Tobi Arkenau:
Thank you. Bye.

Kai-Keen Shiu

Honoraria, travel funding or consulting:
Bristol-Myers Squibb, Guardant Health, Innovent Biologics, Merck, Merck KGaA, Roche, and Servier;

Research funding:
Bristol-Myers Squibb, Merck, Merck KGaA, and Roche

Lizzy Smyth

Personal financial interests (lecture honoraria, advisory boards, travel support):
Aptitude Health, Astra Zeneca, BMS, Celgene, Elsevier, Everest Clinical Research, First Word Group, Five Prime Therapeutics, Gritstone Oncology, Imedex, Merck, My Personal Therapeutics, Roche, Sai-Med, Servier, Zymeworks

Nataliya Uboha

Consulting:
Taiho Oncology, Ipsen, Incyte, AZ, QED

Research Funding:
Taiho, Eli Lilly, EMD Serono, Ipsen