Today, I am going to talk about where we are with first-line treatment for advanced upper GI cancer and our current needs to improve outcomes for these patients.

[00:19:09]

ESMO Guidelines for Advanced Gastric Cancer

These are the ESMO guidelines, which are international guidelines for patients with advanced gastric cancer. Of course, this also refers to gastroesophageal adenocarcinoma but not squamous cancer. Broadly speaking, we can divide our patients into 2 groups:

• Those who are HER2-positive; and
• Those who are HER2-negative.

For patients who are HER2-positive, we further subdivide those into those who are PD-L1 positive and those who are PD-L1 negative. For patients with HER2-positive cancers that are HER2-positive and PD-L1 positive, we recommend chemotherapy plus trastuzumab plus a PD-1 and anti-PD-1 antibody.

For patients who are HER2-positive and PD-L1 negative, they can be treated with chemotherapy and trastuzumab, but we do not recommend immunotherapy.

Then moving on to the patients who are HER2-negative, we can subdivide those again according to claudin status, PD-L1 status and MMR status.

For claudin-positive patients, that means those who express claudin at an IHC 2 or 3 in more than 75% of cells, we can use chemotherapy and zolbetuximab. For patients who are MMR-deficient, we should add an immune checkpoint inhibitor to chemotherapy.

For patients who are PD-L1 positive, the strength of evidence depends on the PD-L1 CPS score. For patients with a CPS score of 1 to 4, we might consider adding immune checkpoint inhibitor, but for patients with a higher score, we definitely want to add an immune checkpoint inhibitor to chemotherapy. A higher score could be CPS 5 or 10. And we can talk a little bit more about that later.

Then for the small group of patients who are biomarker negative, we use chemotherapy alone. Whether we use a doublet or triplet is something we can spend a little bit more time on.

[00:20:58]

Evidence-Based Statements on First-line Chemotherapy for Advanced Gastric Cancer

Why we use chemotherapy and immune checkpoint inhibitors and all of those other treatments for patients with gastric cancer, is even for these frail patients, we know that adding chemotherapy improves quality of life and improves symptom control. We can either use oxaliplatin or cisplatin with 5-FU and capecitabine or infusional 5-FU. These are all equivalent in terms of efficacy with differing toxicity profiles.

Of course, if there are any Asian listeners, S-1 is a very reasonable treatment choice for patients in Asia. But there are differences in pharmacogenomics globally for that, which makes it a slightly more toxic for non-Asian patients.

[00:21:41]

Advanced Gastric Cancer Chemotherapy: Who Is Suitable for Triplet Chemotherapy―TFLOT or Not?

Whether we use a triplet or not, for patients who are biomarker negative, has been a matter of debate for a number of years. Initially, a study called TAX 325 evaluated docetaxel in addition to cisplatin and 5-FU. It did show marginal benefit in terms of efficacy. But the original DCF regimen was very toxic. So this really was not used in the community.

A recent French study evaluated TFOX, which is really an alteration on our FOLFOX regimen, adding a lower dose of docetaxel on a 2-weekly basis. In the GASTFOX study, you can see that there was an improvement in overall survival for patients who were treated with triplet, with less of the toxicity that was seen with the previous DCF regimen.

However, that was limited to patients with a PS of zero and those under the age of 65. And in the era of biomarker selective therapy, it is not a required option for most patients.

[00:22:41]

First-line Anti–PD-1 Therapy for Advanced Gastric Adenocarcinoma

If we look at the studies that evaluated the efficacy of anti-PD-1 antibodies for patients with advanced untreated gastric cancer, you can see that the outcomes are pretty much equivalent for all of these medicines. For nivolumab, for tislelizumab, for pembrolizumab, when they are added to chemotherapy in patients with a moderately high PD-L1 CPS score either CPS 5 or greater for nivolumab, CPS 10 or greater for pembrolizumab, or TAP 5 or greater for tislelizumab, we see on average about a 3-month improvement in overall survival, and that is equivalent between each of these anti-PD-1 antibodies.

However, where the real benefit is, is that the tail of the curve, and we probably see 25% or 30% of patients with a longer term survival.

[00:23:35]

Advanced MMR-Deficient Gastric Cancer

Of course, MMR-deficient gastric cancer is more rare, making only 3% to 5% of the population. However, it is absolutely critical that we test this biomarker above all biomarkers. That is because if we detect a patient with MMR-deficient gastric cancer, they have an extremely good chance of an excellent long-term outcome when treated with immune checkpoint inhibitors.

These are the results for patients with MSI or MMRd gastric cancer treated with pembrolizumab and nivolumab in first-line trials. We do see really strong response rates and probably about a 50% good overall survival benefit for these patients. So please remember if you are testing 1 biomarker above all, remember to look for MMR deficiency. Not just in gastric cancer, also the patients with junctional and esophageal adenocarcinoma.

[00:24:28]

Targeting HER2 in Gastric Cancer: ToGA

Let us talk about HER2. HER2 was the first biomarker that was tested in gastric cancer, the first that was targeted successfully. About 20% to 25% of patients are HER2-positive, expressing HER2 at IHC 3+ or 2+ FISH positive, more common in junctional cancers.

In the ToGA trial, which is now 15 years old, we saw a benefit for adding trastuzumab to chemotherapy for patients who overexpressed HER2.

[00:24:58]

Targeting HER2 and PD-1 in Advanced GEA: KEYNOTE-811

We have moved on, and as I alluded to in the first slide, we are now differentiating between patients who are HER2-positive with PD-L1 positivity and HER2-positive and PD-L1 negative. The majority of patients with HER2-positive gastroesophageal cancer do also co-express PD-L1 in the tumor, about 85%, and about 15% are HER2-positive PD-L1 negative.

What we saw in the KEYNOTE-811 study was that when pembrolizumab was added to chemotherapy and trastuzumab for patients who were PD-L1 positive, this did substantially improve overall survival, as well as leading to really profound response rates.

However, for patients who were HER2-positive and PD-L1 negative, adding pembrolizumab did not benefit those patients and in fact may have been detrimental. Therefore, the recommendation is for patients who are double biomarker positive, treat with chemotherapy, trastuzumab and pembrolizumab. But for those who are HER2-positive alone, continue treatment with chemotherapy and trastuzumab only.

[00:26:06]

Targeting Claudin 18.2: SPOTLIGHT

Lastly, we have claudin. Claudin 18.2 is the new biomarker on the scene for the past couple of years in gastroesophageal cancer. This is present in about 40% of stomach and gastroesophageal junction cancers. We know that targeting claudin-positive tumors with zolbetuximab, which is the first-in-class monoclonal antibody targeting claudin 18.2 and overall survival.

This has been demonstrated in 2 global trials, SPOTLIGHT and GLOW. As you can see, the improvement in overall survival was really almost equivalent to what we see with immune checkpoint inhibitors. We also see a tail of the curve for zolbetuximab-treated patients.

[00:26:53]

Targeting Claudin 18.2: GLOW

We also see equivalent outcomes in the GLOW study. The SPOTLIGHT study tested zolbetuximab with FOLFOX and GLOW with CAPOX.

[00:27:08]

Panel Discussion

That was our rapid fire 10-minute round through gastric cancer where we are in terms of targeted therapy and immune checkpoint inhibitors. I would like to bring my copanelists on board to ask some questions in this area.

If Dr Sundar or Dr Klempner are available, let us first ask. I showed the data for immune checkpoint inhibitors in gastroesophageal cancer. We can use nivolumab, pembrolizumab, tislelizumab. Do you think there is a need to improve the first-line immunotherapy combinations for advanced gastroesophageal adenocarcinoma?

Dr Samuel Klempner (Massachusetts General Hospital Cancer Center): Yes. There is always a need to do better for our patients, for sure. Most of our therapies add, on average, somewhere in the 3-month survival improvement in absolute terms. I think no patient is satisfied with that, and none of us as oncologists are very satisfied with that.

Certainly, as you alluded to, there are some biomarker enriched among the biomarker positive patients who may do even better, and we are always happy to see that. But yes, I think there is a need for improved first-line immunotherapy and first-line targeted therapy.

Dr Smyth: I could not agree with you more. This treatment has been transformative for a small number of patients. If we look at the long-term follow up, even from CheckMate 649, 16% of patients get to 5 years. But that is only 1 in 6. So we need better treatment for the 5 and 6 who do not get to 5 years. So I could not agree with you more.

Dr Sundar?

Dr Raghav Sundar (Yale School of Medicine): Yes. I mean, I think we definitely cannot rest on our laurels of adding PD-1 inhibitor to chemotherapy in frontline. If you compare the speed at which the breast cancer and lung cancer folks are moving into improving their standards of care, we are still far behind them. So we do need to look for better treatment options for these patients beyond just chemoimmunotherapy and front-line.

Dr Smyth: Thank you very much. I think now what we will do is give us an opportunity to think about what is emerging in future. I think Dr Sundar is going to take us through some emerging treatments looking at improvements in the immune-oncology field. Thank you, Dr Sundar.

[00:29:35]

A Look to the Future: Emerging Clinical Evidence for Novel Immunotherapy Combinations

Dr Sundar: Thanks, Dr Smyth. That is great. Thank you for the opportunity to speak here. We are going to talk a little bit about what the next frontier of immunotherapy for gastroesophageal adenocarcinoma is going to look like.

[00:29:48]

Select Therapeutic Targets in UGI Cancers

Taking a step back, I think this entire space of therapeutic targets that we are going to be seeing, coming out in the next few years in gastric and esophageal cancers, is really very exciting. We have various targeted therapy options that are emerging. Not only do we have different targets, but we have different strategies in which we are approaching these different targets. We are going beyond monoclonal antibodies into antibody drug conjugates and bispecific antibodies.

In the immune space, we are looking at beyond targeting the PD-1/PD-L1, CTLA-4 axis and going beyond that into targeting other components of the tumor microenvironment. It is very likely that it is going to be a combination of both hitting better targets as well as hitting the tumor microenvironment in a better way. That is going to improve the outcomes for patients with gastroesophageal adenocarcinomas.

[00:30:54]

          Coinhibitory Immune Checkpoints Beyond PD-1/PD-L1

Let us talk a little bit about just the co-inhibition of immune checkpoints beyond PD-1/PD-L1. One of the emerging targets is TIGIT. We will talk a little bit about that.

TIGIT is 1 of several, what we could call, second-generation immune checkpoints beyond PD-1 and PD-L1 that have been discovered. There are a lot of rationale and pre-clinical data that support this idea that targeting a second immune checkpoint, along with PD-1 or PD-L1 will have increased immune activation.

When we talk about TIGIT specifically, there are at least 4 or 5 different ways in which TIGIT works in the tumor microenvironment.

TIGIT is expressed in the T cells and the NK cells and binds to CD155 on the tumor cells, which is a co-inhibitory ligand to TIGIT. The CD155, which you can think of it as being akin to the PD-L1 that is being expressed on the tumor cells or the antigen presenting dendritic cells leads to a T or NK cell intrinsic inhibition when CD155 binds to TIGIT.

The same story plays out in terms of the dendritic cells. The other way in which TIGIT leads to immune suppression is its co-stimulatory partner, which is CD226. The inhibition of CD226 as a co-stimulatory molecule, when CD115 binds to it also leads to immune suppression. Separately, there is this component of the tumor cells having Treg stability and suppression when the TIGIT binds to CD115 on the Tregs, as well as Fap2-induced T and NK cell inhibition, that is by a bacteria called fusobacterium nucleatum.

[00:32:59]

          Drug Development of Coinhibitory Immune Checkpoints

There are at least 3 immune checkpoints that are currently being developed together. TIGIT is the new kid on the block. We know that LAG-3, TIM-3 and TIGIT are the 3 immune checkpoints that have had development as second immune checkpoint inhibitors that have the most clinical data.

TIGIT was kind of first came into play in the early 2000s compared to LAG-3 and TIM-3. Over the past 20 years or so, there are multiple TIGIT inhibitors that have started to enter the clinic. Today, we will see data from a few of them, and how these are playing out in the clinic.

[00:33:48]

          The Need for Coinhibition

Let us talk a little bit about the need for whether you need to co-inhibit both TIGIT as well as PD-1. This is quite important because if you look at the interactions between the PD-1/PD-L1 pathway, as well as TIGIT, we know that when PD-L1 binds to PD-1, there is a recruitment and dephosphorylation of CD226 through Shp2. That inhibits the immune activation functions of CD226.

TIGIT has a higher affinity binding to CD155 compared to CD226, and that this competitive antagonization blocks the homodimerization of CD226 and the co-stimulatory immune activation functions of CD226. This synergy between PD-1/PD-L1 inhibition and TIGIT inhibition has some biological rationale to it.

[00:34:57]

The Hype . . . Tiragolumab (Fc-Active Anti-TIGIT)

The TIGIT inhibitor that has the most clinical data is tiragolumab. In single-agent data with tiragolumab, there was no objective responses that were seen, although there were 4 stable diseases in n of 25 patients. But when tiragolumab was added to atezolizumab, we started to see some responses in the initial phase Ib studies. This then led to an expansion specifically. The few patients that responded in the early dose escalation studies were in lung cancer and esophageal cancers.

There was a dose expansion cohort where you could then see that in lung cancer as well as esophageal cancers, that the combination of atezolizumab, which is a PD-L1 inhibitor in combination with tiragolumab. Tiragolumab here is an Fc-active anti-TIGIT monoclonal antibody. We will talk about the Fc component a little bit later.

You can see that the objective response rates were around 40% in lung cancer and about 30% in esophageal cancer.

[00:36:05]

Excellent Combination Data in Phase II in NSLC and HCC

This actually led to 2 phase II trials in lung cancer and hepatocellular carcinomas. CITYSCAPE was a study that randomized patients that were chemo-naive and had a TPS score of 1 or greater to receiving either atezolizumab or atezolizumab in combination with tiragolumab.

In the PD-L1 TPS more than 50 population, we could see that there was a clear benefit for the addition of tiragolumab to atezolizumab, although this benefit was not seen as much in the population that had a lower PD-L1 TPS score, which kind of makes sense that we do not use single-agent immune checkpoint inhibitors in lung cancer for a lower PD-L1 score.

We also saw data from the MORPHEUS liver study that showed quite impressive results when patients were randomized to get tiragolumab with atezolizumab and bevacizumab, which is the current standard of care for liver cancer, and there was a very significant improvement in progression-free survival as well as overall survival, which has its ratio of 0.5 in this randomized phase II trial.

[00:37:16]

Multiple Phase III Failures

Unfortunately, after the data that we saw from the phase II, just at the ESMO annual meeting this year that read out a couple of weeks ago, we saw multiple randomized phase III failures for tiragolumab.

We saw data from SKYSCRAPER-03 where patients were randomized to receive tiragolumab as well as atezolizumab compared with durvalumab. There was no improvement in either progression-free survival or overall survival for this combination. We also saw negative data from SKYSCRAPER-09, which is a randomized phase II trial, mostly focusing on squamous cell head and neck cancers, with no prior therapy, where again, there was no improvement for the addition of tiragolumab to atezolizumab compared to atezolizumab alone, both in the TAP score of 5 or greater or TAP score of 20 or greater in the squamous cell population.

In SKYSCRAPER-14, which was in the advanced HCC patients, the phase II data that we saw with very impressive results of the addition of tiragolumab, again did not play out when it was in a randomized phase III trial with completely no improvement in either progression-free survival or overall survival for tiragolumab. This was quite disappointing to see multiple phase IIIs that read out negative for tiragolumab, leading to actually Roche shutting down the tiragolumab program completely.

[00:38:57]

…Including Esophageal Cancer

We also saw negative data from SKYSCRAPER-07 in esophageal cancers, which was a randomized phase III trial in unresectable esophageal squamous cell cancers that had not progressed after chemo rads. This is very similar to the PACIFIC data for lung cancer, and again in this space, interestingly, the addition of tiragolumab to atezolizumab did not improve progression-free survival, where if you look at just atezolizumab alone as a single-agent, it seemed to have a benefit compared to placebo and almost questioning whether the tiragolumab was doing less when it is combined with atezolizumab.

[00:39:42]

To Fc or Not to Fc: That Is the TIGIT Question

Then this brings in the question of does the fragment crystallizable component, the Fc component of the TIGIT inhibition make a difference? This is an area of active debate at this point of time. The Fc domain debate is mostly is there an functional Fc that is required for anti-TIGIT therapeutic efficacy?

The proponents for the Fc-enabled components, for example, tiragolumab show that the preclinical activity mostly occurs through the intratumoral Treg depletion. However, in an Fc-silenced TIGIT antibody, this requires anti-tumor PD-1 blockade for anti-tumor efficacy because they do not deplete the Tregs.

How important depletion of Tregs is going to be in the efficacy in the clinic is yet to be seen. We know that TIGIT-CD155 blockade effects, when the CD226 is present and partially repressed by the anti-PD-1, seems to suggest that maybe silencing of the Fc with an antibody that has an Fc-silence component may have a different clinical activity compared to an FC-active antibody.

[00:41:03]

EDGE-Gastric (Arm A1): Domvanalimab + Zimberelimab + FOLFOX for Previously Untreated Advanced HER2- GC/GEJC/EAC

Herein comes the data from domvanalimab, which is an Fc-silent anti-TIGIT antibody in combination with zimberelimab. Again, we saw data from the EDGE-Gastric study that looked at locally advanced unresectable metastatic gastric and gastroesophageal adenocarcinoma patients. These patients were all treated with domvanalimab, zimberelimab as well as FOLFOX.

[00:41:33]

EDGE-Gastric (Arm A1): Response

We saw updated data from this trial presented by Dr Janjigian at ESMO again, where we saw an objective response rate regardless of the PD-L1 score of 59%, that being increased as we go with an increase in the TAP score, increasing to 62% with a TAP score of 1 or greater and 69% to the TAP score of 5 or greater.

[00:41:56]

EDGE-Gastric (Arm A1): PFS

We saw that the progression-free survival for this study was 12.9 months, with a 24-month PFS of 26%. If you break it down based on PD-L1 score, the PFS for the TAP score of 5 or greater was 14.5 months, compared to less than 5%.

If you look at it with a TAP score of 1 or greater, it was 13.2 months. Again, seeming to suggest that PD-L1 may be a biomarker in this space.

[00:42:26]

EDGE-Gastric (Arm A1): OS

In terms of overall survival, we saw median OS of 26.7 months for the overall population with the median OS not being reached in the TAP of 5% or greater.

[00:42:46]

EDGE-Gastric (Arm A1): Safety Summary

In terms of safety, we did not see a major increase in the toxicities when domvanalimab was added to zimberelimab and chemotherapy. Any immune treatment-related adverse events was 27% with no grade 3 toxicity seen. In general, the safety signal of the addition of the domvanalimab to the chemotherapy did not seem to have a new safety signal in this study.

[00:43:26]

STAR-221: 1L Domvanalimab + Zimberelimab + CT vs Nivo + CT for Advanced Gastroesophageal Cancer

We are now waiting for data from STAR-221, which is a frontline trial of domvanalimab and zimberelimab plus chemotherapy, a 5-FU platinum doublet of either FOLFOX or CAPOX. This is compared against the same chemotherapy backbone with nivolumab.

The primary endpoint in this study is overall survival, particularly in the patients who have an enriched PD-L1 expression of a TAP score of 5 or greater, with secondary endpoints of progression-free survival, overall objective response rate, as well as duration of response.

[00:44:04]

Beyond Monoclonal Antibody Inhibition of TIGIT

Beyond monoclonal antibody inhibition of TIGIT, we do have multiple bispecific antibodies either targeting both TIGIT as well as PD-1 or other immune checkpoints such as TIGIT and 4-1BB. And there are early data from these bispecifics that are emerging as well.

[00:44:25]

GEMINI-Gastric: 1L Rilvegostomig + XELOX or FOLFOX for HER2-Negative Advanced Gastric Cancer

Rilvegostomig was combined with XELOX or FOLFOX in the GEMINI-Gastric study. Here we see early data from this trial where the combination of rilvegostomig plus chemotherapy had a median PFS of 11.1 months in a PD-L1 CPS 5 or greater population in a small subgroup of 16 patients that were treated in this.

[00:44:49]

ARTEMIDE-Gastric-01: 1L Rilvegostomig + T-DXd + Fluoropyrimidine for Advanced HER2+/PD-L1 CPS ≥1 GC

We are waiting for the ARTEMIDE-Gastric-01 trial, which is a randomized trial looking at the combination of rilvegostomig plus trastuzumab deruxtecan and 5-FU in front-line treatment for HER2-positive gastric cancer.

[00:45:05]

Let's Return to a Few Earlier Questions

Based on these data that we have seen on TIGIT, let us return to some of the earlier questions. Can we have the polls up, please?

[00:45:16]

          Posttest 3

We are going to go back to the same questions, now that you have seen some of these data. Which of the following is true of data presented at ESMO on the EDGE-Gastric phase II trial for gastroesophageal cancers?

All right. Can we close the poll? Okay. This is great. Looks like folks have been paying attention. At least half of the answers were correct.

[00:46:13]

          Posttest 3: Rationale

The objective response rates for all patients exceeded 50%, with a median PFS of 12.9 months.

Let us move on to the next poll question.

[00:46:26]

          Posttest 2

Which of the following accurately describes the mechanism of action of domvanalimab, which has shown promising efficacy in patients with upper GI tumors?

Great. Again, looks like most of the group got it correct. Domvanalimab is an Fc-silence monoclonal antibody targeting TIGIT.

Next question.

I am going to pass it over to Dr Klempner for the next part of the talk.

[00:47:36]

Impact of Angiogenesis on the Immune System

Dr Klempner: Thank you, Dr Sundar and Dr Smyth. We have been hearing a little bit about ways of targeting the environment in which the tumor cells live with pressing on other mechanisms of reversing T-cell exhaustion. There are other ways to go about this.

Certainly, we have known for many years that there is activity of anti-angiogenic therapies in gastric cancer. Ramucirumab has been approved since 2014, and there are also multiple retrospective studies looking at expression of VEGF-A in genomic landscape papers in chromosomally unstable tumors. There is also measurement of serum markers for anti-neoangiogenesis as being prognostic.

This is a tumor, like many other tumors, that is characterized by VEGF in the tumor microenvironment. There are multiple angles by which VEGF and other pro-angiogenic ligands can essentially remodel the tumor microenvironment. What we mean by that is both shift the composition, so the proportions of various immune cell populations but also remodel the distribution of cell states.

Macrophages can exist on more immunosuppressive vs a more anti-tumor spectrum. Similarly, T cells can exist more on a proliferative effector function vs more exhausted. And there is a very wide spectrum of dendritic cell states that are involved in T-cell maturation, and VEGF elevations can negatively impact essentially all of these things. So you wind up generating a more tumor permissive microenvironment.

We have known from other tumor types, colon cancers in particular, hepatocellular carcinomas, renal cell cancers that direct inhibition—so in many cases, that involves directly soaking up the ligand with bevacizumab—can essentially withdraw the stimulus from the tumor microenvironment and, to certain degrees, reverse the phenotype of some of those immunosuppressive cells.

The left panel here is essentially showing the direct effects of VEGF on some of the immune cells as well as the vessel lining. Then, of course, at the more macro level, looking at perfusion to certain areas of the tumor.

On the right is the hypothetical benefits of blocking pro-angiogenic therapies. You can see that perhaps 1 of the predominant mechanisms is allowing easier recruitment of anti-tumor immune cell populations into the tumor microenvironment. This sets up a stage or at least a plausible biology where it would be quite favorable to combine an anti-angiogenic therapy with an angle that stimulates T cells, for example.

[00:50:46]

Combining VEGFi + IO in Gastroesophageal Cancer

That is a little bit what I will be talking about. As I mentioned, there are multiple ways to go about doing this. This is a smattering of trials primarily in later line populations. You can see some of these are a little bit older. Some of them are a little bit more recent.

I think the take-home message is that beyond what you would expect for PD-1 monotherapy in a second-line population, it does look like the numerical response rate in these phase I and phase II trials is perhaps more than you would expect from single-agent PD-1 or single-agent PD-L1.

It suggests that maybe there is a biology buried in this combination and warrants further exploration in larger data sets and also combination therapies.

[00:51:37]

Early Data of Regorafenib + Nivolumab Success in Gastric Cancer: REGONIVO

Here is a little bit deeper dive into some of the data that I just showed you. Here is the REGONIVO trial. This is published now a few years ago, and generated a fair amount of enthusiasm when it was presented and then published in JCO. This is the same rationale. So Regorafenib, which is of course a multi-kinase inhibitor. Many of the kinases that it hits are involved in pro-angiogenic signaling in combination with nivolumab.

Again, this was at a time where we did not have checkpoint inhibitors approved in the frontline setting. Although you can see in the waterfall plot on the right that there were some patients who had prior PD-1 or PD-L1.

This was encouraging. You see that in the swim lane plots both in colorectal and gastric cancer, there were some durable disease control. Then in the waterfall plot, these are relatively encouraging response rates in small subsets of patients. This looks to be both, activity in some PD-L1 previously exposed patients, as well as PD-1 naive patients.

[00:52:49]

INTEGRATE IIb: Regorafenib + Nivolumab vs Chemotherapy in Previously Treated Advanced GA

That generated the randomized phase III trial. This was just recently presented at ESMO a few weeks ago. This was the logical follow on to that, which was regorafenib plus nivolumab vs investigator choice chemotherapy. You can see the available options to investigators there in the orange.

Importantly, this is patients who had 2 or more prior lines of therapy. Primary endpoint here was overall survival. That makes a lot of sense for a late line trial. It is really the only endpoint that probably matters in this later line populations.

[00:53:27]

INTEGRATE IIb: Survival

This is humbling, I think, for all of us who take care of this disease. It is a stark reminder of just the expectations we get with chemotherapy. Overall survival shown on the left and progression-free survival shown on the right.

You can see hazard ratio greater than 0.85 or greater, not statistically significant. Really the numerical numbers here is essentially the time to your first scan is your progression-free survival. That shape of the curve with an immediate dive is it is just, unfortunately, the current reality of late line gastric and GE junction patients. You see really a minimal difference numerically in the overall survival and no statistically significant difference.

At the very end of the curve, as Dr Smyth mentioned earlier, you do not see the number at risk or number censored here below the Kaplan-Meier curve, but you can see the shape of the blue line that there are probably some of those excellent nivolumab responders in here, perhaps some quite high PD-L1 expression, etc. We do not know all of the details, but certainly the orange curve crosses the axis at a little after 1 year.

[00:54:46]

Regorafenib + Nivolumab + Chemotherapy for Previously Untreated Metastatic GEA

This is another data set. This is regorafenib-nivolumab and chemotherapy. So essentially trying to move that activity from the REGONIVO trial that I showed you earlier in disease. Regorafenib, I think, as many of you know, not the easiest drug to combine with chemotherapy. So this needs to be taken in the context of a fair amount of toxicity.

This was a 40-patient single-arm phase II. You can see that the waterfall plot is relatively encouraging, with a high objective response rate and a long median progression-free survival and median overall survival. This is actually similar to the EDGE-Gastric that Dr Sundar just showed. Again, this is a selected phase II population at an expert institution in Memorial Sloan Kettering.

[00:55:44]

Lenvatinib + Pembrolizumab for Advanced Gastric Cancer

Similarly, Lenvatinib is another multi-kinase TKI with activity in several other VEGF-high cancers such as kidney cancer and hepatocellular carcinomas. This was very similar to the REGONIVO idea, lenvatinib plus pembrolizumab. You can see here again very encouraging response rates, relatively small trial, 29 patients.

Again, you can see some of these are durable. If you look at the 1 year, you can see that there are several people who continued on therapy. Notably, there were some patients treated in earlier lines on this trial. You can see again, similar to the FOLFOX, regorafenib-nivolumab. This is encouraging, shape of these OS and PFS curves.

[00:56:35]

LEAP-015: Pembrolizumab + Lenvatinib + Chemotherapy for Advanced Gastroesophageal Cancer

All of this led to attempts to validate this in a phase III setting. We saw LEAP-015 at ASCO in 2023. Then again I believe in 2024 pembrolizumab and lenvatinib with chemotherapy for first-line advanced disease. You can see again these are typical newly diagnosed patients. They had a safety run-in, because we know that lenvatinib at high doses like 20 mg can be quite toxic. So you can see the safety run-in and the dosing of lenvatinib here at 8.

Then a large randomized phase III comparing chemotherapy alone vs pembrolizumab-lenvatinib, chemotherapy and then followed by lenalidomide-pembrolizumab maintenance.

One thing to note is that at the time of this chemotherapy alone here, you can argue that chemotherapy plus immune therapy would be the more modern control arm, but no fault to the study design at the time it was launched.

[00:57:43]

LEAP-015: ORR and OS

This is the data. The punch line is over on the left. This is the overall survival in the PD-L1 positive patients. You can see absolutely no difference, essentially overlapping curves, again, as with immunotherapy trials that Dr Smyth showed earlier. After about 15 months, we do see a separation here. There are probably a selected group of patients on that blue line at that time. But this was a negative phase III trial and is not entering our practice.

You can see the response rate data there. Adding drugs does improve response rate slightly. That is shown with the 44% vs 58% response rate.

[00:58:32]

LEAP-015: PFS

This is the PFS which I will not dwell on much, but you can see, again, no real numerical difference here in the median progression-free survival. Hazard ratios are shown there. Again, we see this separation as we have seen with immune therapy curves. For what it is worth, it is hard to say that you would not see the same with chemotherapy plus pembrolizumab leaving out the lenvatinib. But that is, of course, cross-trial comparisons.

[00:59:06]

Combining Novel Approaches: RC118 and RC148

Few other approaches. The multi-kinase inhibitors have not really panned out despite the biologic rationale for targeting VEGF in the immune microenvironment, but bispecific agents such as RC148 have gained a lot of steam based on, 1, this cooperative binding mechanism of action and daisy chaining of these bispecific antibodies when they are bound fully engaged to 2 PD-1 and 2 VEGF molecules. Usually, that is the design.

This was a combination of a bispecific with a claudin 18 ADC. So very interesting quite novel combination presented at ESMO. You can see really ADC plus toripalimab vs ADC plus PD-1 VEGF. So kind of nice because it does give you some degree perhaps to assess the contribution of VEGF on top or the bispecific vs the naked antibody.

[01:00:05]

RC118 and RC148: Response

This is the data. You can see that numerically a little bit higher response rates with the VEGF PD-1, particularly in the low PD-L1 groups. This has been seen in some lung cancer data as well, where these agents have certainly larger data sets. I will move on here. Keep us on time.

[01:00:33]

DKK1 and PD-1 Inhibition

This was an another agent. As I had mentioned, macrophages in the tumor microenvironment exist along a spectrum. Some of them can be immunosuppressive. Tumor cells, 1 of the mechanisms, they want to have a immune suppressive environment around them. So oftentimes they will secrete various cytokines or signaling molecules. One of these may be DKK1.

DKK1 high tumors have lesser response to chemotherapy and immunotherapy. And a neutralizing antibody, DKN-01, also known as sirexatamab was tested in this setting. You can see that in the population shown here, patients who were DKK1 high. So trying to select the patients who might be using this mechanism to generate an immune suppressive microenvironment, there was a separation suggesting you could find a biomarker in this population. Unfortunately, when this was launched into a larger trial, it did not really pan out.

[01:01:38]

DisTinGuish: 1L Sirexatamab + Tislelizumab + Chemotherapy for Advanced Gastric/GEJ Adenocarcinoma

This was the single-arm study based on what I just showed you, combining chemotherapy, DKN-01 plus tislelizumab. You see the waterfall plot, again, similar to some of the other relatively small phase II trials I showed you encouraging early activity response rate of 73%, encouraging median overall survival about 20 months. Unfortunately, this went on to be tested in a randomized fashion, as it should be, but did not pan out. There was a press release that this did not meet the endpoint, and this drug is not currently moving forward in gastric or GE junction cancers.

[01:02:19]

Panel Discussion

I am going to turn it back over to Dr Smyth to guide us through the panel discussion. Thanks for having me.

Dr Smyth: Thanks, Sam. Thanks for bringing us through that discussion. Also to Dr Sundar, especially seeing as I know you have a cold, I think you are just doing a brilliant job. Well done. Let us just think about where we are with VEGF/anti-PD-1 after integrating LEAP-015. I mean, do you see any further mileage in tyrosine kinase inhibitor studies, or are you excited about the next-generation of anti-PD-1, anti-VEGF bispecifics?

Dr Klempner: Yes. Well, we do not really have any bispecific plus chemo data in gastric or GE junction yet in terms of like PD-1, VEGF plus FOLFOX. I think it is a worthwhile exercise to explore and hopefully we can find patient selection. Again, I think we know our best biomarker for immunotherapies, but we do not really have a biomarker for anti-angiogenic therapies, which is always a little bit frustrating. I think we all would love to have a double biomarker to select the patients who might be optimal for this strategy.

Dr Smyth: I agree. I think that we have been quite excited by the data presented at ESMO in lung cancer for combination PD-1 and anti-VEGF. And we nearly got there with anti-VEGF in the past with bevacizumab. We have ramucirumab. So we know this is an active pathway. Let us see what happens.

Okay. I do not know if we can have a question for Dr Sundar[?]. It is a tough pathway to understand I think. Are we hoping if STAR-221 was a positive trial, would that change your practice? Is that is combination that you would be prescribing for your patients immediately? What type of bar would you need to see in terms of improvement in outcomes for this to be a practice changing trial?

Dr Sundar: Yes. If I am not wrong, STAR-221 has already finished with recruitment, so we are just waiting for readout of the data. I think it is 1 of the first few trials that has been randomized against a proper control arm of chemo-nivolumab. In that sense, at least the trial is well designed against the right comparator arm.

Based on statistical calculations, if it reads out positive means the clinically meaningful part will come later when we actually look at the Kaplan-Meier curves and the hazard ratios and things like that. But if it reads out as a positive trial, it will be the first trial that has beaten chemo-immunotherapy in a randomized phase III setting. So I think that is promising enough for us to give it a deeper look.

At least from EDGE-Gastric, we know that there is no increased toxicities. So even incremental improvements in survival may be something that patients would be willing to go for if their toxicities are not more. So I think that is my current sense of it. But all of us are waiting to see the data from this study.

Dr Smyth: I am getting a sense of urgency[?] because we do not know the outcome yet. So we just need to see what comes along and maybe next year when we see the results. A last question for, Sam, perhaps. Sam, you talked eloquently about DKK1 and some disappointments in the first-line study that you talked about. Do you think that there is any more mileage in anti-DKK1 therapy in gastroesophageal adenocarcinoma?

Dr Klempner: One of the things to note is that in the randomized trial, there was no biomarker selection. Even though in the prior data sets the DKK1-high patients were the really the only ones that benefited significantly. The prevalence of DKK1-high in the samples for the earlier trials was about 45%, 50%. They were hoping that without selecting patients, they would have enough enriched patients that it would show up as an effect in the randomized trial because they were not selected on DKK1 expression.

Unfortunately, what wound up happening is that only about 20%-something of patients in the randomized phase II were actually DKK1-high. I actually do think there is a biology to be targeted in the myeloid populations of the tumor microenvironment. I think that DKK1 by RNA-ish expression actually did seem to perform relatively well as a selection biomarker. But unfortunately that is not the way the study was done.

I would love to see like window of opportunities, neoadjuvant trials. It is a very safe drug. I think studying the biology in patient samples to see if we are really modulating the pathway would actually be quite cool.

Dr Smyth: I think this happens a lot in gastric cancer. We go too broad whether by accident or design. Then we know that there is a group whom we could help, but the study has unfortunately not selected for those patients. So lessons that we perennially do not learn unfortunately, I think in our trial design.

Okay. Thank you both for those questions. We are running slightly behind time. I am sure we have a couple of slides that where we are going to discuss some equity in terms of clinical trial access, I believe. Maybe we could move on to those.

[01:08:43]

Supporting Equitable Access to Novel Immunotherapy Combinations and Clinical Trials

Fantastic. Thanks so much. We would like to spend a little bit of time discussing how we support access to not just, I suppose, novel immunotherapy combinations but also to clinical trials for patients with cancer.

Could I have the next slide, please?

[01:09:08]

Why Equity in Clinical Trials Matters

I think what is really important is in terms of clinical trial, and this is a safety as well as a scientific issue, is that often we do clinical trials to test new drugs, and they do not reflect the patients that we see in the clinic. This compromises the external validity of the results, especially when really important subgroups are omitted.

Also we know that in omitting many patients from clinical trials in terms of poor access, we fail to recruit to many trials. That is a reason why many trials close prematurely and we do not get to ask those important scientific questions.

[01:09:48]

Older Adults: The Biggest Gap

There are a number of groups that we could reflect on and how we can help them to access trials. In particular older patients. So most patients, the average age of patients in gastroesophageal adenocarcinoma trials is about 60. The average age of patients in the clinic is probably more like 70s or late 70s. Those patients are definitely underrepresented in trials in GEA and in other cancers. So only 15% of patients are over 65.

There are reasons for this. Sometimes those patients are less well, sometimes they take a number of different medications that might exclude them. But I think it is important that we do not consider [inaudible] permit, when we are designing trials and discussing these with our industry colleagues, that we should broaden the eligibility, such that those patients might be included.

[01:10:47]

Sex and Gender: Representation and Analysis

If we are concerned about patients, we should be performing geriatric assessment in order to facilitate them going on trials, because we may be able to optimize them thus that they could participate in studies.

Women are also important. Women also perennially less represented in gastroesophageal cancer trials, also because it is less common in females. But often we find that women are less than the proportion of patients who have the cancer. And it would be reasonable because there may be pharmacogenomic, there may be pharmacologic reasons, pharmacokinetic reasons why women have different outcomes with specific medicines, that we need to understand the efficacy in both genders.

There are very reasonable requests to have sex based analysis in studies. We should, in sex neutral tumors, I think try to recruit to parity. Not an issue in esophageal and gastric cancer at this moment in time.

[01:11:42]

Race/Ethnicity: Global and Local Gaps

Of course, I think this has been particularly important in the FDA recently, also in terms of other regulatory authorities that we have representation of different ethnicities. This is also important from a pharmacogenomic perspective. As I mentioned in my talk, for example, S-1 is differentially metabolized in East Asian vs non-East Asian populations. And there are other important pharmacogenomic differences in other medicines.

Also DPYD, for example, we know the characteristics of DPYD deficiency in Caucasian populations, but less so in non-Caucasian populations. Just are 2 important examples. And how we reach out to different populations is by expanding trials beyond academic centers and also working specifically with different communities, because they may have less experience of clinical trials and therefore less understanding. And so it is about education and maybe sometimes working with community leaders to dispel myths that surround clinical trials.

[01:12:45]

Adolescents and Young Adults (AYA): Should We Include?

Young patients are also important, maybe less so for gastroesophageal cancers. But certainly we should not have arbitrary age cutoffs like 18. Sometimes we do see young patients who are 16 with specific cancers. We need to work very closely with our AYA colleagues to deliver care which is age appropriate, whilst not excluding this increasingly important group of patients from clinical trials. We are seeing more young onset cancers in particularly in colon cancer, but also in stomach and gastroesophageal cancer globally.

[01:13:17]

Geography, Rurality, and Travel Burden

Then there are barriers that exist for patients participating in trials, perhaps who live long distances from the cancer center, maybe less of a challenge in England, where I am from, it is pretty small, but certainly in the US, I know patients have much further to travel. And we need to think about how we radically change the delivery of trials to help patients participate. Whether we can have studies opening at satellite sites, we should certainly include travel expenses for our patients. We should also be considering local phlebotomy and allowing local imaging, for example, if possible.

If we can decentralize trials by using telemed as was done very effectively during COVID, this will also provide greater access for patients who live distant from major cancer centers.

[01:14:07]

Language, Disability, and Consent Accessibility

Of course, language is another major barrier. Certainly in the UK we have many different languages spoken. And I do recruit many patients who are non-English speakers to trials. It is certainly a little bit more work we need to have. We can use telephone translator. We can use face-to-face translator. You know we must not use relatives for translator. And we should ideally have written information in a different language. These can always be provided by sponsor, but certainly important so that all members of the community, regardless of their language can access these new medications.

[01:14:44]

Financial Toxicity: The Silent Exclusion

Finally, financial toxicity. We know that for many patients, participating in clinical trials is costly because they need to travel to the center, they may need to stay overnight, they may need food. In England, certainly our ethics committees will request that patients are reimbursed for these issues. This is helpful and reduces the barriers to clinical trials participation.

So I think that we should certainly advocate to our patients early on that these payments are available, so that this does not become a barrier for them participating in studies.

[01:15:20]

Eligibility and Trial Design

Overall, I would say that when we are thinking about trial design, we should really think about not having overly restrictive eligibility criteria. We should not just copy paste our eligibility criteria from our last studies. We should really include most patients, unless it is unsafe to do so, and try to have a more pragmatic approach towards clinical trials, especially when we are getting towards the real world setting, for example, in phase III.

[01:15:51]

LMIC Inclusion and Global Partnerships

Finally, it is also important for us to think about how we can increase accessibility of clinical trials in the global population. We know that most trial activity has previously gone on in high income countries. And in low middle income countries, there have been barriers. However, happy to say that this is improving lately. There are efforts in clinical trial organizations and also in [inaudible] to include co-PIs also structure that is required for participation for patients from low and middle income countries.

[01:16:33]

Practical Aspects of Enrolling Underserved Patients

In terms of the barriers that we might approach and how we might approach those, it is really important that we discuss clinical trials right from the beginning, that we understand to patients. The clinical trials are usually part of our standard practice. Usually in clinical trials you get your standard of care plus something that we think might be helpful. These are not an option of last resort.

To reach out to less well-served communities, we should be using a language which is accessible. We should be checking that. It should be readable at probably a lower grade school level. We should be partnering with less well-served communities in terms of our patient engagement, beneficial to the patient and their community.

[01:17:28]

Practical Aspects of Enrolling Underserved Patients

These are some of the aspects which we have found helpful. For example, in the UK, we approach patients early on in their patient journey. I think that it is important that we have research nurses who help them provide information, can take them off after clinic and have a chat to further discuss the trial. Make it as easy as possible in terms of reimbursement.

We organize transport, if needed. There are some myths that we can address. We need to ensure that patients understand that this is a very well-regulated process, that it should be safe, and that this is often an option that will give them something in excess of the standard of care, which sometimes may not even be available, depending on your funding environment.

I think in terms of managing the expectations, all clinical trials are very transparent about our clinical equipoise for trials. Of course, what I say to my patients is if you participate in the study and it is not for you, you can always go back to the standard of care. We know that that is sometimes the best option for patients.

[01:18:35]

Panel Discussion

I think that if I wanted to speak with my panel for a couple of minutes. Sam and Raghav, have you found any approaches helpful for you in terms of engaging with your patient populations for clinical trials, which might be different from mine in the UK?

Dr Klempner: I will be interested to hear Dr Sundar’s approach or thoughts, because he is actually had experience in and outside the US, which is awesome. I think we spend a fair amount of time trying to provide information ahead of time through the EMR, so patients get a consent sent to them. They get a standardized information, what is a trial? We are starting to actually move towards like videos describing some of the trials.

So it is trying to maximally inform the patients even before they come in, if they have expressed interest in a trial or has a follow up to the visit. Because of course, as you know, once you say clinical trial, a lot of people do not hear everything else that you said, or they do not remember it. Some living information that they can refer back to we found has been helpful.

Dr Smyth: Yes, I could not agree with you more. We tend to send out the information sheet. I will see the patient. If they are interested, I send out the information sheet, then we bring them back. I know we are all busy, but extra visits so that you can go through the logistics, so you can address any concerns that they have is probably the most helpful thing. Time is probably the most helpful aspect that we can do.

I will be honest. I think making sure that transport is available and that the patients know the transport is available is very important. We draw patients from all over the UK and reimbursement for transport, especially in these days when petrol is quite expensive and public transport is also expensive. So knowing that transport is available and we will put patients up in a hotel beside the hospital if they have an early start. I think that is really helpful in terms of helping them, just making things as easy for them as possible.

Dr Sundar, what about you?

Dr Sundar: Yes. I think there are essentially 2 groups of patients broadly, 1 who really want to get on a trial and the other who are really concerned about getting onto a trial. You are right, in terms of the folks who are concerned about getting onto trials, we kind of need to meet them at where they are at and what are their concerns about. A lot of the mitigation factors that you brought up are things that we do actually implement on the ground, trying to understand are their barriers from them getting onto the trial more, both from a logistical or physical barrier to an emotional barrier of I do not want to be treated like a guinea pig or am I going to get a placebo and those sort of things.

Many times, if we do spend the time explaining and trying to address some of those barriers, then I think we do manage to get patients onto studies a lot more. But that does take time and effort, and we do need to think of ways in which we can find that time and effort for getting these patients onto these studies.

Dr Smyth: I also agree with Sam's point regarding videos. So having a video to explain the trial is very helpful that patients can look at on their telephone, that they can maybe share with their relatives and making sure that also some close family friends or loved ones do attend the clinic with the patient so that for support, and so that they can review the information and provide the confidence I think, the patients need to participate in studies.