 titled, “Updates for Treatment of Advanced Gastric and Gastro-Esophageal Cancers.”

I am Dr Johanna Bendell from the Sarah Cannon Research Institute in Nashville, Tennessee.

Joining me today in this discussion are my colleagues, Dr Daniel Catenacci from the University of Chicago, Dr Ian Chau from the Royal Marsden Hospital in London, Dr Yelena Janjigian from Memorial Sloan Kettering Cancer Center in New York, and Dr Ken Kato from the National Cancer Center Hospital in Tokyo, Japan.

Today we are going to discuss a number of topics pertaining to the use of systemic therapy for gastric and gastroesophageal malignancy. We will discuss the latest research in the field including some very exciting abstracts that are being presented during the Presidential Session at the European Society for Medical Oncology 2020 Virtual meeting and their impact on clinical practice.

Let’s set the stage for how we approach gastroesophageal cancers in our practice. We know that there are different trends in epidemiology, and histology, like how patients are treated in western countries versus eastern countries. Ian, can you give us some background on incidence, prognosis of gastroesophageal cancers, and how this might differ worldwide?

: Yes. Certainly, there are geographical variations in terms of gastric and esophageal cancers. I think perhaps it is important to look at it according to histology. We normally think about it as adenocarcinoma or squamous cell cancer. For adenocarcinoma, in the Western world certainly the incident of gastric adenocarcinoma has been declining. Instead the adenocarcinoma from the junction of the esophageal gastric junction has been increasing, and that has been due to several potential risk factors in the western world, mainly to do with the reflux disease, the body weights, obesity, and incidence of obesity in the western world. That all leads to a rise in adenocarcinoma of the esophageal gastric junction.

Now for squamous cell cancer the incidence is generally much lower in the western world compared with East Asia. I’m sure my colleague, Ken, will talk about that a lot. The vast majority of squamous cell cancer does happen in China, Japan, Korea, and also in some parts of Africa, but it is still a very important disease in the Western world where it is more related to some of the environmental factors as smoking, drinking, and perhaps with the use of very hot tea. Certainly there is geographical variations in terms of these diseases.

Prognosis is still poor generally. I mean we are hoping to diagnose patients who are still at a stage that radical treatment can be used. For example, in the UK for gastric cancer, only about 30% to 35% of gastric cancer are diagnosed in a stage where radical surgery and therefore perioperative or preoperative strategy can be used. Another 20% to 30% are probably locally advanced. Still, sadly, most patients are still diagnosed in a metastatic disease setting. That’s still a very real, high unmet needs in the Western world. Although no gastric cancer is diagnosed in the early stage in East Asia.

Similarly, for squamous cell cancer of the esophagus, I would say that the majority of our patients are still diagnosed in a stage which are beyond radical treatment, whether the radical treatment is chemoradiation or chemoradiation followed by surgery. I think all in all, although there is a geographical variation, the prognosis of this disease still remains poor, and I think there is still a high level unmet needs throughout the world.

: Ken, being from Tokyo, how did you approach gastric, gastroesophageal cancers, and cancers of the esophagus in Japan with that higher incidence rate?

: In Japan the proportion of gastric cancer is higher than the Western countries because we have a medical checking-up system for gastric cancer around the country. More than half of the patients are diagnosed at an earlier stage, and 20% to 30% of the patients received endoscopic treatment therapies.

Due to the COVID-19 coronavirus disease 2019 pandemic, the medical checking-up system has stopped for months. The recent number of patients with earlier upper gastrointestinal cancer has decreased, so I wonder if these patients will come to our hospital with more advanced cases.

: And what is the difference in risk factors that you’re seeing in Japan versus what we see in the United States—Ian was talking about obesity and reflux disease. For Japan, certainly different epidemiology results potentially in a different histology.

: Yes. Of course, in Japan, so obesity is not common, but the 

 infection influences the incident of gastric cancer. Recently, however, the 

 incidence was lower along with the gastric cancer that has been continuing to decrease recently.

Global Trends in Gastric Cancers

: In Japan, there are screening protocols that catch these cancers early so they can be cured before they get to an advanced setting. Here in the United States, we don’t have that so much. How are we trying to work to look at earlier diagnosis or education around symptoms for these patients?

: That’s a good question. Just to reiterate what’s been said so far, the way I think about it, there’s 3 diseases: esophageal squam, gastroesophageal [GE] junction adenocarcinoma, and gastric. For the most part in Western countries, in my clinic, two-thirds to 70% of the patients are GE junction adenocarcinomas. As we mentioned, we’re not screening routinely for these patients, so they’re for the most part presenting locally advanced or overt metastatic disease by the time they present.

Increasing awareness in terms of the symptoms that we’re talking about, chronic reflux, and symptoms from a Western diet, Caucasian, diabetes, chronic reflux, and obesity, are types of associated factors. Primary care physicians know how to act on these things earlier rather than waiting for several months to years before it’s too late.

The only other comment I’d make is that the different studies that we’re going to be talking about, they really have heterogeneity in terms of the eligibility for accrual to the study. Some of them overlap in squam and adeno, some of them overlap in GE junction and gastric. We know from the previous studies that they have different sensitivities to immunotherapy, so we need to look at their contributions and the evidence.

: Very good. We know that in the west, most patients are diagnosed with advanced metastatic disease. Yelena, what is the general approach when we see a patient here in the United States in the west? How are we going to look at that patient who presents with advanced or metastatic disease and treatment options in the course?

: In the United States, and the west in general, the biggest differences in approach to our Japanese, North Korean, and colleagues is the use of perioperative versus adjuvant therapy in localized and resectable disease in the United States because the tumors are more proximal, gastroesophageal junction adenocarcinoma tumors, they’re more aggressive in nature, and if surgery is possible, they often require total gastrectomy. It is a greater challenge to administer adjuvant therapy.

Therefore, because the patients are more nutritionally compromised, they often take a month or 2 months to recover fully after the operation, and at that point they lost 10% to 15% of their total body weight. It’s a big challenge to administer the adjuvant therapy in that setting. Therefore, unlike our east and Asian colleagues, we tend to administer more therapy in the neoadjuvant or preoperative setting. It gives an advantage of understanding tumor biology, whether or not the therapy is effective in vivo assessment with imaging and so on.

Also, this precludes patients from having unnecessary surgery. If the disease is still aggressive, then within a few cycles of therapy occult metastatic disease is detected, often that was subacute before. Again, you’re not wasting opportunity to cure your patient, you’re preventing them from having an unnecessary surgery.

The second point is, and increasingly you see, looking at first-line studies conducted in stage IV disease in Asia versus the west and the rest of the world, is the use of second- and third-line therapy.

In the world, outside of Japan and Korea, probably 30% to 40% of patients receive second- and third-line therapy. In fact, if you were watching European Society for Medical Oncology yesterday, it kept coming up. How is that possible? Is that really a real-world data, and are these studies somehow confounded? The truth is that it is the real-world data. Unfortunately, there are various factors that we’re going to discuss later in terms of disease burden and patient’s functional status but also somewhat in an annalistic approach. Prevalent among oncologists and patients, outside of Japan and Korea, the majority of patients do not receive second-line therapy and third-line therapy, and that’s really an important pivotal point that we need to change as we understand more about this disease.

: When we think about what we’re going to use in the first line, I know we’re going to get very much into the discussion a little bit later, but in general, what are general principles for treatment? I mean we know in colon cancer we would give chemotherapy. Then we’d go to maintenance. How do you approach that patient with metastatic disease in the first line in terms of how you’re going to set the stage for treatment down the line?

: Functional status is number 1. The earlier you test for biomarkers the better; and I always tell my patients, “Know your mutation status; ask your oncologist.” Knowing the silver lining of this whole global pandemic is that telehealth and telemedicine has really taken off in oncology in particular. I often see patients from all over the world virtually or from over the United States. Biomarker analysis, and also knowing the disease biology, is critical. Obviously 1 drug is better than no therapy. Baseline 5-FU–based therapy is the cornerstone of all gastrointestinal tumors, gastric cancer, and esophagus cancer.

Two drugs are better than 1 drug. Additionally, a third-drug therapy, although there’s still some data to support it, increasingly now the emerging data suggests that immune checkpoint inhibitors and other targeted therapies should be used in first-line setting. It’s becoming a challenge to build on third-drug baseline chemotherapies.

The majority of our patients, and, again, even though younger and younger patients are presenting to us, as specialists, most of our patients are still in their 60s or 70s. They can be frail with multiple comorbidities. Two-drug combination with FOLFOX [folinic acid, fluorouracil, oxaliplatin] or 5-FU [5-fluorouracil], oxaliplatin; or occasionally, like you said, to be oxaliplatin, is what most patients get.

So, Ken, how do you treat patients? How’s your treatment approach in Japan different from what we do here?

: As they are in remission, so we use a drug combination with fluoropyrimidine plus an agent for the first-line treatment for the gastric cancer as well. The patients who have gastric cancer have a comorbidity such as a disorder for eating and appetite loss. So, in Japan, so they have a tendency to move to the oxaliplatin from the cisplatin.

The oxaliplatin had the adverse event of the peripheral neuropathy. We frequently use the secondary chemotherapy. In Japan, 80% of the patients receive the secondary chemotherapy with paclitaxel.

We should use the oxaliplatin carefully. Usually we use the 2-drug combination with platin. And you know the S-1 is mostly used for the gastric cancer patient and the fluoropyrimidine agent in Japan.

Optimal Approach Toward Advanced Gastric Cancers in US

: Yelena mentioned earlier the concept of biomarker testing and making sure we test the biomarkers early. Dan, in your practice, you see a first-line patient coming in, what’s your absolute base biomarkers? Then, what’s your wish biomarker profile that you would want?

: Before yesterday, the absolute minimum would be human epidermal growth factor receptor 2 (HER2) testing. And so that dichotomy of HER2 positive versus negative in which about 10% to 15% of our patients are going to be HER2 positive. We know obviously from the ToGA study (NCT01041404) that that’s an improvement in survival by adding trastuzumab to the standard doublet chemotherapy.

What I normally do would be a panel next-generation sequencing [NGS] test to get the lay of the land of the genomic biomarkers that are available, including microsatellite instability [MSI]-high, including other receptor tyrosine kinase gene amplifications that have implications on study or even later on off label, if possible. Of course, TMB [tumor mutational burden], which we’re going to talk about later, and PD-L1 expression by immunohistochemistry. Having the understanding of all of these biomarkers helps me know what the best option for first-line is, and over the next line of therapy and plan their sequencing over time.

: Dr Kato, Ken had alluded to stop and go oxaliplatin and that concept of do you use maintenance? Do you give 4, 6 months and stop everything? Do you do 4, 6 months continue? I know that this is not completely data driven, but what’s your gestalt for first line, Dan?

: For first-line, I think I’d just reiterate what Yelena and Ken said. Generally speaking, we’re talking about a doublet for permitting platinum. My preference is oxaliplatin if FOLFOX [folinic acid, fluorouracil, oxaliplatin]. Most of my patients can’t swallow when they’re first presenting, and they’re not swallowing capecitabine pills, so FOLFOX is the preferred regimen. That’s preferred. Now occasionally you’ll have patients that have baseline neuropathy from diabetes or alcohol use or whatever, idiopathic. We have data that FOLFIRI [folinic acid, fluorouracil,irinotecan hydrochloride] in the first line is as effective. That’s my preferred regimen for those patients.

Then you asked about maintenance and OPTIMOX approach, which I adopt and sort of extrapolate from the colon cancer literature; that is why I push the neuropathy. Most of my patients by 3 months in, 6 cycles, max 8 cycles, their neuropathy is going to just get worse by continuing it, and most of those patients have that point where they can afford to come off the oxaliplatin and go on a maintenance 5-FU [5-fluorouracil] or even capecitabine at that point because now they can swallow; 80% or more of them will improve their ability to swallow the pills by that point.

The only other thing I would say, Yelena mentioned the triplet regimens, the FLOT [docetaxel, oxaliplatin, leucovorin, and 5-fluorouracil] or even FOLFIRINOX where there’s now data, and that is my preferred regimen because there’s nonoverlapping neuropathy toxicity. I will consider it in patients that are heavily symptomatic that have a heavy burden of disease, younger patients, to really sort of get that initial response, extrapolating from the TRIBE data from colon where we can see that, and we already have triplet versus doublet regimens in this disease with the higher risk of toxicity. It’s always balancing the potential benefit from the treatment versus the symptoms and outcome of the patient and tailoring it to the patient.

: Very good. And, Ken, I wanted to ask you, comparing to Dan’s list of biomarkers he tests. In Japan, are you doing more comprehensive profiling at baseline?

: Yeah. We profile at baseline with the HER2 status and the MSI status. In the comprehensive data, the NGS was really robust only for the patients that fail after the standard treatment, so we conducted some trials to do the NGS up front.

: Very good. Ian, when these patients present to us, Dan started to allude to not being able to swallow. How do we think about quality of life, symptomatology, use of treatment, and choice of treatment? We talked a little bit about neuropathy and affective quality of life. What’s your approach when you think about these patients with gastric cancer that tend to have very specific and serious issues?

: I think both Ken and Yelena already mentioned that a lot of these patients are highly symptomatic. When they first come to our clinic, they already lost a lot of weight. Many of them have a lot of nutritional compromise. They may have dysphagia for a while. Certainly, at current time, we are seeing the backlash of COVID-19 [coronavirus disease 2019]. Patients have had dysphagia since March and are not able to get an endoscopy because the whole service was put on hold. Now they’re all coming in with very advanced disease.

I think it’s an important thing to reoptimize their nutritional status. Early referral and involvement from our dietetic departments and thinking apart from whether their patient is able to take in enough oral intake. Do we need to start thinking about enteral feeding, like tube feeding? Occasionally, some patients do come in where you actually need to consider total parenteral nutrition, although we will always try to have enteral feeding if possible.

Now those are the patients that really need to start treatment very quickly. And as Dan already said, you probably have to start treatment and balance whether you use a more aggressive regimen with a high toxicity with trying to result or achieve a better, higher quality response as quick as possible.

Certainly, in that situation you will need to give drugs or intravenously FOLFOX, and we probably will still not really move on. We used to use a triplet regimen, and if they can’t swallow, we would use a regimen like FOLFOX. Whereas if a patient is able to swallow, we would be using regimens like CAPOX [capecitabine, oxaliplatin] or cisplatin and capecitabine.

I think the optimization of their nutrition is very important for these patients because there are multiple quality-of-life studies that have been done, and some of the most important factors to the patients actually are appetite loss and weight loss in gastric cancer patients. That seems to come up again and that also seems to have a negative prognostic impact. People who have weight loss and appetite loss always come up as an independent poor prognostic factor. Being able to optimize that to deliver effective treatment to them is probably a very important consideration.

Frontline Treatment in Metastatic Gastric Cancer

: Yesterday at European Society for Medical Oncology, they tried to figure out howwe’re going to approach treating these patients, so maybe we start with where we are before yesterday or where we were before yesterday in terms of how we would approach the treatment of an advanced metastatic gastric cancer patients, and we’ll also talk about esophageal and gastroesophageal [GE]- junction adenocarcinoma. Dan, can you set the stage for us on where we were just 48 hours ago?

: Forty-eight hours ago, from an immunotherapy standpoint, we had a number of studies that were monotherapy checkpoint inhibitor with nivolumab, pembrolizumab [PEMBRO], and others like nivolumab [NIVO].

We had only 1 study with in combination with chemotherapy. Briefly, to just set the stage of the background, we had the ATTRACTION-2 study (NCT02267343), which was a third-line study of nivolumab versus placebo in Asian patients that was positive and led to approval in all-comers, not elsewhere outside of Asia. One way I check is the KEYNOTE-059 study (NCT02335411), which was a single-arm study, third line and higher. That study showed benefit and enriched for PD-L1 status by combined positivity score of 1 or higher with pembrolizumab, and it was led to an accelerated approval for patients in the United States in the third-line setting or higher with combined positive score of 1 or higher.

We also had a pan-solid tumor approval in the United States based on microsatellite instability [MSI]-high in the second-line setting for this disease. That was the status in the United States 48 hours ago based on the gastric and GE-junction cancers. In Europe, Ian can say that none of these led to approvals based on that.

The other thing to mention is moving it into earlier lines of treatment. In the second-line setting, we had the KEYNOTE-061 (NCT02370498) study which was pembrolizumab versus paclitaxel. Overall, it was a negative study but showed some promise and certainly higher PD-L1, CPS 10 or higher, not leading to any approvals.

Then we had the KEYNOTE-062 (NCT02494583) study which was the first-line study, a 3-arm study of chemotherapy with or without pembrolizumab, placebo-controlled, and an open-label third on the PEMBRO alone. The PEMBRO-alone arm was again showing some signs and high PD-L1 scores; PD-L1 CPS 10 or higher, but there was still sort of a detriment in a subgroup of those patients. Overall, the claim was noninferiority, and there was a group of patients not benefiting. The chemotherapy arm showed a trend to improvement with PEMBRO, but with a hazard ratio of 0.85; this was not statistically significant and was considered negative.

Then we have the esophageal part of this story which the studies that are pertinent are the KEYNOTE-181 study (NCT02564263), which allowed squamous cell and adenocarcinoma and that was comparing PEMBRO versus taxane. The study was looking at a number of primary end points, but overall led to approvals in the United States only for squamous cell cancers that were CPS 10 or higher. When you looked at the forest plots of that study, that’s where the benefit was being driven from. The other ones were really not driving much benefit. That’s what led to the approval.

Shortly thereafter, the ATTRACTION-3 study (NCT02569242), which was a 95% Asian study in all-comers, showed benefit with nivolumab versus chemotherapy, and led to approvals in Asia and in the United States for NIVO in the second line for esophagus cancer.That’s where we are as of 2 days ago. Then we’ll talk about how that might change.

: If I may add and highlight 1 point about getting approvals in Asia and United States and not in Europe; it actually leads to a very important point about why some of these studies that we are about to discuss is of great importance in global studies. Europe would not accept studies which included a patient population with a vast majority being Asians. They still feel within the European Union countries that they, there are practical variations, maybe different treatment patterns. Therefore they would not approve, would not give the drug approval, on studies which are predominant based in Asia or nonrandomized studies. I think it really gives importance of the study that we are, in the past 24 hours, are really, truly global studies, and how that hopefully will change practice globally.

: It is interesting because I think we’re going to break this down, too. As Dan was alluding to, the squamous population, certainly potentially driving studies that are not only global in geography but also global in the types of tumors that they’re accepted in histology that it accepts, I think we’re going to definitely break that down. I think, Ken, I would love to hear your perspective from Japan in terms of how you are approaching the treatments of these patients in Japan and utilization of checkpoint inhibitors.

: Yeah. As Daniel mentioned, the immune checkpoint inhibitor is already approved for the second-line and the third-line patient for the gastric cancer. Of course, it had squamous cell carcinoma patient with it being very dominant in Asian countries, but recently the first-line treatment was studied in a phase 3 study for the esophageal cancer. The study included many patients, but they’re not from the European countries.

The phase 3 trial should be held globally, and of course maybe the same histology type, the same results. Global studies should be done, not only Asian trials.

: I’ll just summarize my thoughts based on these studies that were previously, before yesterday, just coming into this and how I was thinking about it. The monotherapy checkpoint inhibitor in all of these studies seemed to be optimized with a higher PD-L1 score, certainly MSI-high. Asian patients tend to do better on the forest plots; squamous cells always tend to do better than adenocarcinoma, and gastric cancers tend to do better than GE junctions and esophageal adenocarcinomas; that’s with monotherapy. That’s sort of what the principles how I would think about in later lines, and even in the first line.

: I think that’s a fantastic summary, and I think also it’s going to help us shape how we’re thinking about what happened yesterday.

Treatment Landscape in Gastric Cancers Pre-ESMO 2020

 Yelena, it’s almost like we could title it a movie: What happened yesterday? I know what you did yesterday at ESMO [European Society for Medical Oncology Congress]. What in the world happened?

 ESMO this year was really exciting for our group. As you can see, we’re all a very tightly knit group of upper gastrointestinal doctors, and it was a success. We had 4 oral presentations at the plenary session. It’s their version of the plenary session, and you can change with just that 1 day the way that we think about both esophagus and gastric cancer. There were 2 practice-changing abstracts in adenocarcinoma of esophagogastric, and that’s the CheckMate649. Certainly, adjuvant therapy has been changed as well. I agree with everything that’s been said so far about their Kaplan-Meier curves in first-line setting.

What is exciting about immunotherapy is that there’s always a plateau. What’s frustrating, at least in first-line setting, is that in the beginning, in the first 6 to 9 months, chemotherapy has always been at the top. Even though immunotherapy eventually reaches a plateau, the curves do cross in the beginning, suggesting that in some patients, there’s a detrimental effect if you avoid chemotherapy in first-line setting.

We can talk about the issues with KEYNOTE-062, but despite previous attempts, there has not been a study that’s demonstrated that immunotherapy is helpful in first-line setting until CheckMate649. We saw the data that Markus Moehler presented yesterday, and they will be in publication form very soon.

What’s refreshing about the study is that for the first time—unlike previous studies, where esophagus adenocarcinoma and squamous were grouped together—with CheckMate649, esophageal adenocarcinomas are together with GE [gastroesophageal] junction and stomach adenocarcinomas. We know biologically through TCGA, the Cancer Genome Atlas analysis, that adenocarcinomas of the esophagus are very similar to GE-junction adenocarcinomas, and they’re nothing like esophageal squamous cancer. Trials in which adenocarcinomas are grouped with squamous cell cancers should be an antiquity, and we should not be designing those studies anymore.

For CheckMate649, another refreshing approach is the real-world chemotherapy backbone that was used, FOLFOX [5-fluorouracil, leucovorin calcium, oxaliplatin]. As you heard from Dan Catenacci, myself, and Ken Kato, and I’m sure Ian Chau could chime-in as well, no one is using cisplatin, and certainly not at 60 or 80 mg/m2. Once again, it’s a refreshing approach, and a real-world backbone therapy. Our pharma partners certainly listened while we mandated this design.

For these reasons, it was a large, well-designed study—and we can talk about the subgroup and biomarker analysis—designed to look for a CPS [combined positive score] PD-L1 greater than or equal to 5. The overall survival [OS], progression-free survival [PFS], and the primary end points were met.

In second- and third-line therapy in the United States, the West, and China, where the majority of the patients were enrolled, very few people get immunotherapy in the second- and third-line setting; that’s probably 1 of the reasons why the overall survival benefit was so robust and was clearly seen. Unlike in ATTRACTION-4, which was another reassuring and practice-changing study presented. Although the PFS end point was clearly met, the overall survival end point was not met.

If you look at the data, the median overall survival for those patients was 17 months, but in CheckMate649, we’re all patting one another on the back because we cleared the 1-year mark. You know the median overall survival in the experimental arm in CPS-greater-than-5 category was 14, which is great but not as good as 17, which was what was seen in ATTRACTION-4. In my mind, that raises a question. In Japan and Korea, so many patients get I/O [immuno-oncology] in second- and third-line setting; it is approved in that. Plus, a percentage of patients get second- and third-line therapy in general. Are there more patients with gastric cancer in Japan than with colon cancer in the United States? Is OS benefit even needed? Can we design trials based on PFS? We can see clearly that ATTRACTION-4 met the PFS end point.

Nonetheless, both CheckMate649 and ATTRACTION-4 show that immunotherapy can be used in the first-line setting. It’s safe; it’s tolerable. Oxaliplatin is the new standard for design of next-generation studies. This should be a clear end of cisplatin-based backbone; the FDA set precedent. You hear about these discussions behind closed doors. You know the companies are worried that cisplatin is the standard—that’s the trial design that’s mandated. But that’s not the case. Let’s please stop using cisplatin. I’m sure Ken has comments regarding ATTRACTION-4. I’d be curious to hear what his thoughts are about PFS and OS.

First-Line I/O Successful in Gastric Adenocarcinoma

 This is a great general first discussion, and we’re lucky to have Ken here as the first author. Maybe what we could do is break this down. Let’s attack esophageal cancer first and make our way down the line. Ken, you presented some incredible data that are going to change how we treat esophageal cancers with checkpoint inhibitors. Can you tell us a little about your trial and what that’s probably going to lead to.

 According to the results of the KEYNOTE-590 study, this is a comparison of the standard of care of the 5-FU [5-fluorouracil] and cisplatin. The standard of care was not changed for esophageal squamous cell carcinoma and adenocarcinoma.

Indeed, from the results of the KEYNOTE-590 study, the effect of the pembrolizumab through the 5-FU–cisplatin is observed. Of course, the KEYNOTE-590 trial contained a mixed histologic type. Indeed, the CPS [combined positive score] is mixed. The 70% of the squamous cell carcinoma is enrolled to the KEYNOTE-590 study.

The prevalent point of this study is PFS [progression-free survival] and OS [overall survival] for the squamous cell carcinoma and with CPS 10. When the primary end point is shown, they move through the analysis to the patients. The same tendency and efficacy was seen, not only for patients with esophageal squamous cell carcinoma but for all patients, including adenocarcinoma. The standard of care of esophageal cancer is changing to the chemotherapy-nivolumab combination, according to the results of the KEYNOTE-590 trial.

 Ian, for esophageal cancer in the West, how do you break that down? Also, we know that there was a population of the patients with esophageal cancer who were included in the nivolumab first-line setting, which we saw yesterday as well. How do you put that together?

 If you think about esophageal cancer, as we heard many times, we really need to separate squamous cell cancer from adenocarcinoma. Certainly, with squamous cell cancer, with KEYNOTE-590, the majority of patients had squamous cell cancer. Chemotherapy plus pembrolizumab should be considered 1 of the new standards. There are other data coming through. Those are the first data that we heard with pembrolizumab. CheckMate 648, which is chemotherapy plus or minus nivolumab with the third-arm nivolumab-ipilimumab—that study has also finished recruitment, and we will hopefully hear the results in the next 12 months. If that is anything to go by, we hope that will also show a positive outcome.

Moving forward with squamous cell cancer, checkpoint inhibitors will be part of the treatment when it’s first line or second line. Obviously, we will need to discuss later how to sequence it. Once we use it first line, what do you do second line? We’ll have a chat about that later.

With adenocarcinoma, I agree with both Yelena Janjigian and Dan. You know, from a molecular point of view, adenocarcinoma of the esophagus and the junction are really the same disease. To be fair, it’s difficult for oncologists to tell the difference: Where does the esophagus stop and the OG [oesophago-gastro] junction start? Gastroenterologists, we say this, but we just need to treat it the same.

To me the CheckMate649 study is the defining trial for what we do in the future. An oxaliplatin-based chemotherapy plus nivolumab to me would be how we should be looking forward, and we still need to discuss the different benefits according to the PD-L1 expression. Certainly, for those who are CPS 5 or more, adenocarcinoma of esophagus junction and gastric stomach origin, we really should use an oxaliplatin-based chemotherapy plus nivolumab. That’s how I would view the data given so far. I still find it very difficult, and I’m very interested to hear my colleagues’ view about the junctional adenocarcinoma, with the really conflicted evidence between KEYNOTE-590 and KEYNOTE-062. There are obviously both of those 2 studies that do include a proportion of esophagogastric junction adenocarcinoma. One you would say is a negative study; the other is a positive study. To me, that is still really a gray area that is quite confusing. That’s why I said before, if in doubt, CheckMate 649 is clearer to me, so I’ll probably use the data to guide my future patient management.

 Ian, your vote is we’ve changed the standard of care for squamous, so squamous in the first line will be fluorouracil plus platinum with pembrolizumab. For the adeno esophageal, that’s a little more controversial, probably CPS greater than or equal to 5.

 I’ll comment on some of this stuff as it pertains to KEYNOTE-590. To me, when you look at the forest plot of that presentation, it looks almost identical to the KEYNOTE-181 breakdown. When you look at it, really what’s driving the outcome is CPS more than 10 and squamous. When you look at CPS less than 10, including squamous and adenocarcinoma, it’s right on the line. I would reiterate what Ian was saying. For me, for squamous that are above 10, chemotherapy, probably folinic acid, fluorouracil, oxaliplatin plus immuno-oncology. For those less than 10, I would like to see the actual data before I bite at that. For adenocarcinoma, I would really like to see the data broken down in the Kaplan-Meier curves to really see the benefits there, if there are any. I suspect that those above 10 are going to be at least trending toward benefit, and I might give them the benefit of the doubt, and that would be very consistent with what CheckMate649 is showing. We’ll talk about the gastric, but the CheckMate649 forest plot also shows that the benefit derived in the GE [gastroesophageal] junction esophagus is much less, and they’re overlapping. When we see the breakdown by expression of PD-L1 and by histology, it’s not surprising what we’re seeing in these studies. It’s exactly as would be expected. The differences are that CheckMate649 is a much larger study, so it has the power to find the differences that are smaller, that are “statistically significant.”

In general, for me, it’s squamous for sure, GE junction esophagus, probably in the higher PD-L1 levels, whether it’s CPS 5 or 10. There are different biomarker assays that we should point out. There are different antibodies. The incidence, for example, of PD-L1 CPS 1 or higher with the 22C3 antibodies is about 60%, but the incidence of CPS 5 or higher by the other antibody that we just saw in CheckMate649 is 60% at 5. They’re clearly different biomarkers giving us different incidences of who’s actually being called positive, and that’s going to add some confusion in how we’re testing our patients and which assay we are going to use. It’s all good news about these studies, but we need to be critical, and look at the subsets. From the perspective of academics, we need to be an advocate for our patients to ensure they’re getting therapy that’s actually helping them.

ESMO 2020: Practice-Changing Data in Esophageal Cancer 

 Absolutely. Moving from esophagus, we’re going to move through the gastroesophageal [GE] junction. What do you think, Ian? What’s your interpretation?

 For gastric, the data are going to be mainly driven by the data from CheckMate649 because that really is, at this moment, the only positive study of using chemotherapy plus nivolumab. Certainly, for a CPS [combined positive score] of 5 or more, I’m very convinced about the data. For a patient with gastric adenocarcinoma, what you call type 2, type 3 junction of adenocarcinoma, I think that would be my preference.

Between CPS 1 and 5 is going to be the harder part. Certainly, CPS 1 or more, you can still see benefit. In fact, Yelena already mentioned that for all-comers, it was statistically significant. That was with a big sample size and a hazard ratio of 0.8, so you may not think that benefit is enough for all-comers.

To me, when I look at CPS 1 or more—in CheckMate649 that is 80% of the patient population, which is a large proportion of patients using that particular immunohistochemistry antibody to test for PD-L1 expression—the hazard ratio is still quite impressive. If you look at the survival curves, they do clearly separate as well. I would have that discussion with the patient, to say that there is some benefit and that it has become more difficult, whether that is going to benefit to them because clearly there is a proportion of patients who don’t benefit in that kind of lower PD-L1 expression.

My thought at the moment is that a CPS 5 or more is clearly something that I would discuss and recommend to patients. Anything lower than 5 needs to be quite a careful discussion. Especially in many countries outside the Untied States, maybe southern hemisphere countries, reimbursement is a very important factor. Taking these different subgroups to the reimbursement agency and having that discussion will be the challenge in the next 12 to 24 months about who would be reimbursed.

 I would bring up what Ian brought up. I was talking about the CheckMate649, that CPS 5 or higher was the amended primary end point. Midtrial, it was amended to make the primary end point, and the sample size was increased. That was the primary end point. It was clearly positive, and we will be using FOLFOX [folinic acid, fluorouracil, oxaliplatin] and nivolumab. The question of whether less than 5 is very important.

What we’ve seen in the public data set here at the European Society for Medical Oncology Congress that was presented is the 1-or-higher group lumped together with the 5-or-higher group. That’s about 80% of all patients, as Ian mentioned, and then we have all-comers. We see the outcomes are “better and statistically significant,” but they look like a crescendo plot. I dubbed the yin-yang plot of monotherapy, where you see some detriment and then crossing of the curves. What we see here with all-comers especially is this crescendo curve, meaning there’s a bunch of patients at the beginning. They’re not deriving much benefit, and as you go on, you see this increased benefit over time.

What I need to see before I’m going to be giving my patients with less than 1 CPS or the 1 to 5 CPS are the actual data of those patients, not with the 5-or-higher pulling the curve over. I want to see that. That’s not been presented. Hopefully it’s in the publication, and if not, then I will still be asking that question. Then again, Ian’s point is about reimbursement. It will be very interesting to see how the different regulatory bodies—FDA, EMA [European Medicines Agency], and the Asian countries—approach these studies. For example, KEYNOTE-181, which again looks identical to KEYNOTE-590, they looked through and approved only CPS 10 or higher in squams. How are they going to look at KEYNOTE-590, and how are they going to look at CheckMate649? Are they going to expand to the secondary end points of 1 or higher or all-comers? Are they going to hold it at 5 or higher? Are they going to demand to see those data? That’s what I’m demanding. I hope we see those data shortly.

Applying New Data in the Real-World for GEJ Cancer

 What’s also very interesting as opposed to where we were a couple of days ago, now the concept of utilization in the first line is checkpoint inhibitor plus chemotherapy. We had thought previously that even if you’re going to use it in first line, even for a very select population, it would be the checkpoint inhibitor alone. Certainly, we’ve had some education that the role is probably more for checkpoint plus chemotherapy. What about other biomarkers? Yelena, we have talked a little about MSI [microsatellite instability], we’ve talked about PD-L1 status. What about other markers, like tumor mutational burden [TMB], that might affect response to checkpoint inhibitors in this patient population?

 We stressed the point about MSI, and we really need to drive this home. In first-line settings and even early stage disease, it is critical to know the patient’s MSI status. I would not necessarily wait for NGS [next-generation sequencing] because it may take several weeks to get a result and the tissue quality may be suboptimal. Therefore, you may not have sufficient mutation calls if you have a small quantity of the tissue. Doing MMR [mismatch repair] IHC [immunohistochemistry] is likely the most practical way, at least in perioperative setting and certainly in the stage IV first-line setting as well. Why do we do this? Because there’s a huge body of literature now, in gastric but also in colon and other diseases, that MSI tumors tend to not respond as well to chemotherapy.

In the neoadjuvant approach, there’s been a meta-analysis, but also Lizzie Smyth looked at this data through the MAGIC protocol. If I know the patient is MSI high, I tend to steer away from perioperative therapy, and I recommend surgery in those patients. We can talk about other adjuvant approaches in this setting. Similarly, in metastatic disease, 1 critical piece of data that came out of KEYNOTE-062 is that in first-line pembrolizumab compared with chemotherapy, the hazard ratio for this trial was 0.2.

It’s a small subgroup analysis, and this was not unlike KEYNOTE-177 in colon cancer, where the study was purely designed to answer this question. KEYNOTE-062 looked at it in a subgroup analysis, but I find these data compelling and convincing. If the patient has an MMR-deficient or MSI-high tumor, I do not recommend doing chemotherapy in the first-line setting, and we’ve been successful in getting immunotherapy for these patients if they have small-volume disease. Certainly, both KEYNOTE-062 and CheckMate649 suggest that if you’re nervous about avoiding chemotherapy in this setting, a combination of chemotherapy and immune checkpoint inhibitors could be considered.

Beyond MSI, 1 of the biomarkers that was very exciting 2 or 3 years ago but now is becoming a bit more controversial—if you follow the literature and the controversy on Twitter and other platforms—is tumor mutational burden, which has been proposed as a biomarker for immune checkpoint inhibitors. The FDA actually approved pembrolizumab in third-line setting based on a small study that was recently published in 

. Nonetheless, TMB tumors may have a higher response rate to immune checkpoint inhibitors. In gastric cancer, most TMB-high tumors—defined by greater than 10 mutations per megabase—also happen to be MSI cancers. In our disease, it may not be as big of a driver, and the data are a bit controversial and some of the initial publications in TMB-high cancers have been refuted.

Lastly, the most recent updated analysis in 

 tumors that included other cancer types, including gastric, endometrial, and colon cancer, suggests that for immunotherapy, the immune checkpoint inhibitor in this study was nivolumab at ESMO [European Society for Medical Oncology Congress] 2020. If you have a 

 mutation of known pathogenic mutation or a mutation of variant of unknown significance, although you could argue you know it’s significant if the patient is responding to therapy, those patients tend to have as high a response rate as 30% in the treatment-refractory setting.

It was a small data set, but it suggests that NGS should be done as we discussed earlier. It may not be time efficient, so you should still do 

 [human epidermal growth factor receptor 2] IHC, PD-L1, and MMR IHC in real time. NGS is helpful because it gets you a sense of 

 amplification; it’s a predictor of outcome by NGS as well in gastric cancer. It gets you the TMB status. You’ll know the POLE status. We can talk a little about fibroblast growth factor receptors, fusions, and amplifications that may be drivers in subsets of gastric cancer. Now that NGS is considered approved at least in the United States in metastatic and locally advanced setting, one should do that if tissue quality is sufficient.

Predicting Response With Biomarkers in Gastric Cancers

 Now that we have everything changing in the first-line setting, what do we do in the second-line setting? Let’s talk a little bit about how we approach these different cancers in terms of second-line treatment. What’s the current standard? I think Dan gave us a good overview as to what that looks like, but what other data do we have once patients have progressed on first-line therapy outside of immunotherapy [I/O]? Let’s talk a little bit about antiangiogenic therapy. Ian, can you talk about ramucirumab and how we approach second-line treatments of these patients with gastroesophageal cancer?

: Certainly, as we heard, the standard has been to use a first-line platinum plus fluoropyrimidine, and for those patients who progress on first-line treatment, there were 3 randomized studies that have shown the benefits of chemotherapy versus standard of care; the use of second-line chemotherapy really has been seen as the standard of care. Then we have 2 phase 3 studies that look at ramucirumab. The REGARD study [NCT00917384] was looking at ramucirumab on its own versus placebo, and showed a significant improvement over placebo, and survival is very similar to chemotherapy alone. But then the RAINBOW study [NCT01170663], which showed a significant overall survival benefit of adding ramucirumab to paclitaxel compared to paclitaxel alone, and that really has defined our standard second-line therapy with paclitaxel plus ramucirumab. Since that RAINBOW study’s original publication, there have been a number of phase 3 studies looking at either adding another drug to paclitaxel or another novel drug versus paclitaxel. There are at least 5 randomized controlled phase 3 trials, and none of them actually showed a benefit over taxane alone, whereas the combination of ramucirumab plus paclitaxel appears to show a benefit over paclitaxel. Really, that is the standard treatment option.

Then a lot of clinicians think about, well, how about HER2-positive disease because, obviously, in the first-line setting, we have chemotherapy plus trastuzumab, and we look toward our breast cancer colleagues and think whether we should continue trastuzumab in the second line. Certainly, there are registry data to say that if you continue trastuzumab beyond progression, there may be some benefits, but then I think as Yelena Janjigian, MD, showed in her paper in 

, actually a lot of things happen in the second-line setting after patients have been retreated with trastuzumab, including loss of 

 expression and other mutations that drive trastuzumab resistance. There’s a randomized phase 2 study that was done in Japan, which again showed that continuing trastuzumab beyond progression does not show real benefit compared to using the chemotherapy alone. For all these, putting it together, in the second-line setting at the moment, it remains paclitaxel plus ramucirumab, whether HER positive or negative.

Whereas we discuss a lot about all these new data on I/O in first line, I don’t think that would particularly impact on what we do second line at the moment. There is a subgroup analysis that was presented as a poster in patients who were recruited in KEYNOTE-061 [NCT02370498], so these are the second-line pembrolizumab, and then they looked at how many people got ramucirumab afterward. It seemed to have similar benefits to get ramucirumab afterward. At least as I can see, there is no biological rationale to say if you’ve been treated with I/O in the front line, then you will have less benefit from having angiogenic in the second line. To me, even the standard first line is likely to change. The second-line setting is probably going to remain as paclitaxel plus ramucirumab. That’s my viewpoint at the moment.

: Ken, in Japan, is that also the case? Is there a lot of ramucirumab usage in Japan and do you ever use it as a single agent?

: Yes, in Japan, ramucirumab is frequently used for the patient with second-line gastric cancer, mostly with paclitaxel because paclitaxel can be administered easily, even for the patient who cannot take medication orally. Before the approval of ramucirumab, we frequently used paclitaxel for patients with second-line gastric cancer. Patients achieved some response to ramucirumab plus paclitaxel, and neurotoxicity was increased. So we used ramucirumab monotherapy for such patients.

Clinical Rationale for Antiangiogenics in Gastric Cancer 

 Dan, you’ve done some work with irinotecan-based therapies and ramucirumab. What if there is a neurotoxicity that continues? What about changing out the chemotherapy? Tell us a little bit about some work on irinotecan-based regimens with ramucirumab.

: Sure. I think Ian hit it with the RAINBOW study [NCT01170663] setting the standard with paclitaxel and ramucirumab. Now, even if immunotherapies are added in first line, it’s really not going to change what we do in second line. That said, as you mentioned, there are alternatives to using the paclitaxel backbone. Why would one consider doing that in the first place? Well, we all experience with our patients having neuropathy from the oxaliplatin previously in first line. We also have platinum- and taxane-based perioperative approaches with FLOT [fluorouracil, leucovorin, oxaliplatin, docetaxel] or with carbotaxol [carboplatin/paclitaxel]. Then patients who have quick recurrence within 6 months, we often consider those patients as having that as their first line and what to do next. Are we going to give them more taxane? They just progressed on it. In addition, paclitaxel is weekly, and other regimens can be more convenient in having to come less frequently. So that said, that’s the introduction for what other backbone. FOLFIRI [folinic acid, fluorouracil, and irinotecan], I have preferred that for all of those reasons I just mentioned since the approval of ramucirumab in second line.

We actually had published a retrospective analysis on their outcomes, and it’s essentially as you would expect, very similar to what you would see with paclitaxel/ramucirumab, and actually that led to NCCN [National Comprehensive Cancer Network] guideline approval in January of this year as an option for second line.

Now, we have randomized phase 2 data that came out at ASCO [the American Society of Clinical Oncology annual meeting] this year with the RAMIRIS study [NCT03081143], which was done in Europe; it was randomized of paclitaxel/ramucirumab versus folinic acid, fluorouracil, and irinotecan/ramucirumab. As expected, the outcomes were similar between the 2 arms, and in fact, trended in favor, particularly in patients with prior taxane exposure, as would be logical. For me, especially now we have phase 3 data in second line in colon with folinic acid, fluorouracil, and irinotecan/ramucirumab to show safety, and we have these data. It’s for me very reassuring that we can use that as an option, and it’s always good to have choice for our patients.

: Absolutely. We see a lot of patients with gastroesophageal cancers having issues with bleeding tumors. Do you have any concern about the utilization of ramucirumab for patients who might have bleeding or a slow leak from their gastric cancer? Ian, any thoughts there?

: We published a meta-analysis of ramucirumab. We compared to placebo. There’s a reduced incidence of bleeding and even GI [gastrointestinal] bleeding. Obviously, some of those data are not in gastric or colorectal, but they were included. Whereas it is a cause-effect, the bleeding risk is less problematic than bevacizumab. Having said that, if someone is bleeding, we try to see whether we can control that with some palliative radiation to stop the bleeding. Once it’s controlled, then we usually feel comfortable to use ramucirumab. That, to me, is not a major issue.

Going on Dan’s point about the different backbone, the neuropathy, all these things do affect patients’ quality of life, so having a different backbone with ramucirumab will increase the exposure of our patients to more effective second-line treatment, especially in Europe. I’m sure in the United States, perioperative chemotherapy with FLOT [fluorouracil, leucovorin, oxaliplatin, docetaxel] is standard of care, and it is a big problem when they progress on fluorouracil, leucovorin, oxaliplatin, docetaxel, shortly after fluorouracil, leucovorin, oxaliplatin, docetaxel. You’ve used up all these drugs, what are we going to do? Having more data of folinic acid, fluorouracil, and irinotecan or irinotecan plus ramucirumab in that setting, I think it gets us thinking about how treating gastric cancer is all about continuum of care, like colorectal cancer. How do we maximize exposure to all the active drugs for patients while preserving their quality of life?

: One drug we haven’t mentioned yet is TAS-102, yet another instrument in our armamentarium to treat gastroesophageal cancers. Ken, can you tell us a little bit about TAS-102 for gastric cancer? Is that utilized a lot in Japan?

: In Japan, nivolumab is used for third-line treatment for patients with gastric cancer, and even after failure of nivolumab, the so-called fourth-line treatment. Even in Japan, less than 20% of patients get to the fourth-line regimen. Then, some of the patients have metastasis, and they cannot take the drug orally. TAS-102 is one of the options for the subsequent line for gastric cancer, for the patient who appears to be in good condition even after the fourth or fifth line. So it’s one option. Not too many patients can reach the TAS-102.

Toxicity Management in Second Line for Gastric Cancer 

: Let’s try to bring it on home. We now have new treatment options, we probably have new algorithms for treatments for our patients with gastroesophageal [GE] cancers. And if we break it down into esophageal, squamous cell, GE junction adenocarcinoma, and gastric cancers, how are we now going to approach these diseases? Dan, you gave us the ghost of GE cancers past. What is now the ghost of GE cancers future? 

: I think that it’s going to depend on the regulatory approvals, which we should hear within the next months, and they may be actually different across the geographical region, just as they have been with all of these other studies previously. That said, what I would anticipate, in the United States at least, would be the approvals as they did with KEYNOTE-181 [NCT02564263] in the first-line setting, at least squamous cell CPS [combined positive score] 10 or higher. We’ll see how the FDA feels that day, whether they’re going to lower that bar and allow all esophageal squamous cell and all adenocarcinomas or not. For me, again I’d really like to see those data before I would routinely be using that, certainly now before any approvals and before publication, to really see those data further for esophageal cancer. The reassuring and nice thing is that we will be able to get this drug to the patient who will benefit most immediately in the first line.

: It’s important to consider the benefit and the duration of benefit and the options for the second- and third-line setting as our first-line treatment options are becoming more effective. I strongly believe that more patients will remain well and receive second- and third-line therapies. In the first-line setting, 5-fluorouracil platinum is the standard with combination immune checkpoint inhibitors likely for the vast majority of the patients. Looking at the numbers, 60% of the patients in CheckMate-649 [NCT02872116] were CPS greater than or equal to 5, and certainly depending on what the subset analysis in overall benefit will be in CPS 1 or CPS 0 patients, we may have those options.

Sometimes folks ask me, well what about HER2? Should we consider immunotherapy in HER2-positive patients? It’s important to know that CheckMate-649 excluded HER2+ patients. If you were HER2+, you were excluded, and patients with HER2-negative or unknown were included on the study. So it’s critical to continue to test for HER2 and to enroll those patients on different studies or outside of a trial to consider trastuzumab as the standard of care.

The practice at this point has not changed for HER2+ tumors, and there are some data to suggest that combination of immune checkpoint inhibitors plus anti-HER2 therapy is helpful, and we’ll hopefully see the confirmatory data in the first-line setting. Certainly, in the second- and third-line setting, antibody-drug conjugates such as trastuzumab deruxtecan has been dominating the field. Irreversible selective HER2 inhibitors such as tucatinib are very exciting and something to watch out for. For HER2, that’s a piece of the story that we’re still developing. Please continue to test patients for HER2 and to direct them to HER2-directed therapies.

Looking Toward the Future in Gastroesophageal Cancer 

: We have these great data now bringing checkpoint inhibitors earlier into the lines of therapy, particularly for patients with esophageal cancer. Now to throw another fly in the ointment, let’s talk a little bit about the CheckMate-577 study [NCT02743494] we also saw at ESMO [the European Society for Medical Oncology meeting] yesterday. Ian, can you tell us what was the CheckMate-577 study, and how is that going to potentially impact what we do?

: CheckMate-577 is a large phase 3 randomized study, which is being done globally, so about a third in Europe, a third in North America, and then a third in Asia and the rest of the world. Again, it’s a global study. In this study, patients would have undergone preoperative chemoradiation and then underwent surgery. Those who still had a residual tumor, those who have not achieved a pathological complete response [path CR] and had an R0 resection, were randomized to adjuvant nivolumab or placebo. The standard of care after preoperative chemoradiation is just observation. Usually we don’t give any further treatment afterward. This study randomized between nivolumab and placebo, and 794 patients were randomized in a 2:1 fashion. The primary end point in this study is disease-free survival [DFS], and a secondary end point is overall survival.

What we have heard is the primary end point was met, ie, there was a significant improvement in disease-free survival in favor of nivolumab. At the median level, there was an 11-month advantage of nivolumab over placebo. The median DFS was 22.4 months in the nivolumab arm and only 11 months in the placebo arm.

Now, in this study, again because it’s esophageal, they actually have both squamous cell cancer as well as adenocarcinoma in this study, although the majority was adenocarcinoma. Only about 30% or maybe just under 30% of the patients had squamous cell cancer. I think that probably is another interesting discussion point in this study. We heard earlier on, Dan said in the advanced disease setting, we think that squamous cell really benefits from checkpoint inhibitors. Here we have a study where the majority had adenocarcinoma, but they seem to benefit.

The important thing to me from this study is knowing that even in a so-called curative setting, after preoperative chemoradiation, after surgery, there are so many patients who relapsed, even within the first 12 months. If you take away all the path CR patients, there are so many people who relapse so quickly, and I think that is an important observation. But now we have nivolumab, which improved things; that would certainly be a great contribution.

Now I suppose one of the other main criticisms, and I’m interested to hear what Ken, and Dan, and Yelena said, is all the studies in preoperative treatment for esophageal gastric cancer, use overall survival as the primary end point. Are we happy enough to use disease-free survival as a practice-changing end point? Certainly, it’s a very convincing disease-free survival advantage. I don’t think the follow-up is enough for them to analyze for overall survival yet. We may still wait and see that. That might still come, so we’ll see.

 Let’s say we’re using nivolumab now as adjuvant therapy for these patients with esophageal cancer. What happens when they recur? Would you consider bringing a checkpoint inhibitor therapy with combination chemotherapy still as standard first-line therapy, or would you make sure that there’s a certain amount of time between the completion of nivolumab and recurrence? How would you address now what you do in the metastatic setting with these patients, if these patients recur? Ken, what are your thoughts there?

: For such patients, we consider that patient’s disease is refractory to the immune checkpoint inhibitor. There are no data that say to reintroduce the immune checkpoint inhibitor, so there’s still a challenging problem. There are some data from melanoma or lung cancer for the rechallenging of immune therapy for the metastatic setting, but there are no data for gastric cancer.

: Yes, operating in a data-free zone, which as we get so many trials that keep coming so quickly, we end up getting stuck in that zone. Is there anybody here who would consider using a checkpoint inhibitor after progression in the adjuvant setting, or would you want a time break? Again, it’s a data-free zone, so everybody’s answer is right.

: Johanna, if I may comment; in short, we don’t know, and typically for cytotoxic therapy, we use the 6 months as the magic cutoff. In other words, if the patient recurred greater than 6 months after the stop of therapy, perhaps use of platinum or other agents may be considered. Immune checkpoint inhibitors, there’s still a lot to be learned. Perhaps if there was a long enough interval since the discontinuation of adjuvant therapy, a rechallenge with a combination of chemotherapy plus immune checkpoint inhibitor, as opposed to immune checkpoint inhibitor monotherapy, could be considered.

I wanted to make a separate point about the standard of care in the United States and most of the world. Again, for esophagus and GE [gastroesophageal] junction tumors, for Siewert type 1 and 2, preoperative chemoradiation and preoperative therapies really stay the mainstay. There are several important studies that are underway looking at neoadjuvant chemotherapy or combination with immunotherapy. Yes, it was very satisfying to see the adjuvant data with nivolumab, and this is the first of its kind in the GI [gastrointestinal] and upper GI setting. It will add to the body of literature in this setting. If the neoadjuvant studies, however, read out to be positive, fingers crossed, that will be the big question, when will immune checkpoint inhibitors be used? The majority of us, again because of all the challenges of getting adjuvant therapy to patients, after 3 months of chemoradiation and then surgery, they just want to be done with therapy. It’s likely that neoadjuvant therapy will replace the adjuvant therapy if the data hold up.

Adjuvant Treatment in Gastroesophageal Cancer

: Let’s talk about what’s coming, what’s emerging, and what’s the next thing to come out. Currently, we have several approaches with chemotherapy that are under investigation. Ken, can you tell us a little bit about the SOLAR study [NCT02322593] that’s going on in Asia?

: The SOLAR study is first-line phase 3 trial compared to the S-1 plus cisplatin to the TAS-118. TAS-118 is combination with S-1 plus leucovorin and oxaliplatin combination therapy. The primary end point of overall survival was met. The results of the study are positive, but a slightly higher proportion of the adverse event was seen in the TAS-118 plus oxaliplatin arm, especially the grade 3 and higher diarrhea. The overall response rate was achieved in 73%, amazingly, of patients in the TAS-118 plus oxaliplatin arm, but the PMDA [Pharmaceuticals and Medical Devices Agency] is not approved, and the pharmaceutical company does not submit an application to PMDA. I think the higher response rate is very good for the patient who had a higher tumor burden and ended up with progressive disease in the symptomatic patient. Finally, the TAS-118 is not approved even in the Asian countries.

: Thank you so much. Interesting, TAS-118, a new novel fluoropyrimidine combination to try to improve, but certainly toxicity is an issue. What about this concept of maintenance and switch maintenance for gastric cancer treatment? Ian, can you talk with us a little bit about the ARMANI study [NCT02934464] and the JAVELIN Gastric 100 study [NCT02625610]?

: The PLATFORM study [NCT01943903] I’d like to talk about as well. We heard that do you treat a patient first line with 6 to 8 cycles or 4 to 6 months of treatment. Do you continue, do you do an Optimox where you drop oxaliplatin but continue 5-fluorouracil? There are certainly different practices around the world. If you take others, perhaps talking about the JAVELIN 100 study … a phase 3 study that was presented earlier in the year at the American Society of Clinical Oncology Gastrointestinal Cancers Symposium where the patient had first-line chemotherapy, and then they were randomized to continue the same chemotherapy versus avelumab. Unfortunately, in that study, a patient who switched to avelumab, although some … had additional responses, that did not translate into any improvement in overall survival in that study. Maintenance avelumab is not going to be a treatment option.

We conducted an academic study, the PLATFORM study, which also presented in the same meeting in our very initial, interim analysis, where the standard of care in the UK is that you stop chemotherapy after 6 or 8 cycles, so up to 4 to 6 months of treatment, where you stop treatment completely. Patients were randomized in 1 of the arms to continue with capecitabine alone, and then a third arm to durvalumab. We got 2 other arms that we haven’t reported data yet because they’re still recruiting, with a Poly (ADP-ribose) polymerase inhibitor in a fourth arm and a fifth arm being capecitabine and ramucirumab, so those arms are still recruiting.

From what we at least know in the interim analysis is continuing, durvalumab does not have an early detriment…. We’re still waiting for the study that we recruited just about 125 patients in that arm. The whole study has got about 400 patients so far, so we’ll wait for those data to come.

We haven’t proven that maintenance therapy is of benefit, but it is used in many different institutions in many countries…. Certainly, around the world, many clinicians will use that as a strategy, but I think it … also depends on the patient preference. Some patients want to stop treatment completely; some patients want to have their treatment going.

Maintenance Therapy in Gastric Cancers

: Dan, we also have several targeted agents … in development right now for gastroesophageal cancers. Can you tell us a little bit about the FIGHT trial [NCT01800968], SPOTLIGHT trial [NCT03504397], and some of the trials we’re seeing with other vascular endothelial growth factor receptors [VEGFR] tyrosine kinase inhibitor [TKI] that are coming into play?

: Let’s start with the Claudin study. We have the SPOTLIGHT study, which is a phase 3 study, and the GLOW study [NCT02919592], which is sort of an Asian counterpart with CAPOX backbone and SPOTLIGHT with leucovorin calcium, fluorouracil, and oxaliplatin [FOLFOX]. This study is based on a randomized phase 2 study called the FAST study [NCT01630083] that was done with a different backbone of chemotherapy, the MAGIC backbone, but certainly showed benefit in patients selected by Claudin expression by immunohistochemistry. That was defined as 2+ or 3+ immunohistochemistry at least 40% of cells or higher, and a subgroup analysis of 70% or higher looked like it was even more predictive.

After transition of hands of the drug to a different company, renaming of the drug to zolbetuximab, that study is ongoing and probably about halfway accrued. Like many studies, it has suffered the problems with coronavirus disease 2019 and with accrual being on hold. That is a promising study. It’s going to be very interesting now with these new studies from yesterday and their impact on accrual to this study now, where the control arm is with FOLFOX alone, and you’re going to be having this challenge of patients, especially the ones who have high programmed death-ligand 1 scores, microsatellite instability high, they’re probably not going to be going on this study. That’s going to be a challenge, but we’ll see if it validates the randomized phase 2 FAST study.

Then we have the FIGHT study … a fibroblast growth factor receptor 2 [FGFR2]-based study, based on an earlier trial, the phase 1 and expansion study in FGFR2 gene amplified gastroesophageal cancer, which is about 5% of the disease. In a late-line monotherapy phase 1 expansion study showed a response rate of approximately 18%, 20%, which is promising and shows activity of a single agent drug.

That led to moving that into the first-line study in combination with FOLFOX. The selection criteria for that assay for that study … was not gene amplification, which is 5%, but rather immunohistochemistry [IHC] positive, which was turning out to be like 30% to 35% of patients. The study … accrued a lot quicker than expected because it was thought that the IHC 3+ would be only around 10% by IHC. … That study was downgraded from a phase 3 to a phase 2 in size to just sort of look at the outcomes before moving ahead with a larger sample set, and we expect those data to read out in the fall to the winter; we’ll see if that merits further progression in all of those patients or only in sort of the genomically driven patients, which probably would have a higher magnitude of benefit.

Then finally, you mentioned the TKIs, the antiangiogenesis TKIs, of which there are several studies. There’s the INTEGRATE II study [NCT02773524] with regorafenib in later line compared to placebo. There’s the anlotinib study [NCT02388919], another sort of multikinase with inhibition of the vascular endothelial growth factor receptors. There’s the FRUTIGA study [NCT03223376] of fruquintinib. All those studies are ongoing and in later lines. Then there was the negative study of the ANGEL study [NCT03042611], which was with apatinib compared to placebo in third-line setting or higher, despite an earlier study that was solely in China, which was a randomized phase 3 study that suggested benefit by overall survival in third-line setting or higher.

My take on these antiangiogenesis drugs in late line is that there’s probably sort of a marginal benefit hazard ratio of 0.85 and that many of these studies may just hit the mark, may just miss the mark, and may add another option in the toolbox for later line. I will say the data of combining these TKIs with an immune-oncology agent like regorafenib and nivolumab or lenvatinib and pembrolizumab, as has been shown in colon cancer, hepatocellular carcinoma, and now gastroesophageal cancer seems to have some sort of synergy at least in these smaller phase 2 studies, mostly from Asia. If that can be validated in larger sample sizes, some of which are already embarking on first-line studies, that will be interesting and exciting to see those results.

VEGFR TKIs in Gastroesophageal Cancer

: Yelena, you have done so much work with the HER2-positive population, and particularly now there are such exciting data in combination with I/O [immunotherapy]. Can you tell us a little bit about where that is and where you’re going with it?

: Sure. We recently published our phase 2 experience, single-center, 37 patients, looking at combination of trastuzumab/pembrolizumab and chemotherapy in the first-line setting. These results were quite promising. The overall response rate was 91%, with promising overall survival and progression-free survival. What’s interesting about the study is that all the correlative work points to the fact that HER2 is the important driver in this subpopulation. It’s not about checkpoint inhibitor effect as much as HER2-positive augmenting the anti-HER2 response. In other words, even PD-L1–negative tumors had a deep response as long as the driver HER2 was there at a high level. It’s very important to characterize these tumors, and we understand that similar as we see in colon cancer, RAS activation really dampens the response to HER2 inhibitors in gastric cancer. Only if you have a very high level of 

amplification will then, luckily in gastric, 

 amplifications will take the back seat and the patients may still benefit.

We’re now looking at these data in the phase 3 setting. Hopefully the results of KEYNOTE-811 [NCT03615326] will be available in a year or so, and having CheckMate-649 [NCT02872116] data now showing benefit for I/O therapy in HER2-negative disease, really is reassuring in that setting. We know it’s safe, we know it’s tolerable; combination therapies may be the answer, but the degree of benefit is to be determined.

: Very good. Dan, tell us a little bit about MAHOGANY [NCT04082364].

: Building on what Yelena just said, I think there’s evidence in our disease type and also other disease types of this synergistic activity of harnessing both the innate immune system and the acquired immune system. Meaning the acquired immune system with anti–PD-1 inhibitors and the innate system with these RTK [receptor tyrosine kinase] antibodies like trastuzumab or margetuximab, which is sort of an enhanced trastuzumab molecule to enhance that ADCC [antibody-dependent cellular cytotoxicity] effect, which recruits natural killer cells and other macrophage based on the binding to the Fc [fragment crystallizable] domain of the antibody to really get a synergistic response. We also experienced the study that preceded MAHOGANY, which was a second-line study, and showed basically, as we talked about this earlier in second line, I think Ian was mentioning a lot and Yelena—how these tumors evolve to become negative over time in second line. The study with margetuximab did not require ensuring HER2 positivity prior to enrollment. What we saw would be expected, and those patients who were still HER2-positive and certainly PD-L1–positive, those were the ones who were deriving substantial benefit with this chemotherapy-free regimen of margetuximab, and at the time, pembrolizumab, with response rates of 50% to 60% in second line, and survival showing that as well, as a secondary end point.

That led to moving that regimen into the first-line setting, very akin to KEYNOTE-811, but there are 2 major cohorts of that MAHOGANY study. Cohort A is a chemotherapy-free regimen, the same as what was being used in second line. Since the response rates were 50% to 60%, the thought was if we really super-select our patients with IHC [immunohistochemistry score of] 3+ and PD-L1 CPS [combined positive score] of 1 or higher with 22C3 assay, that maybe we can afford not having chemotherapy. That is a single arm, open label, response rate primary end point in the first 40 patients to explore that possibility. Then in the cohort B, it’s more your traditional randomized study. It’s a little bit more complex because there’s a 4-arm phase 2 part of a control arm of chemotherapy plus trastuzumab, then there are 3 experimental arms. One is chemotherapy plus margetuximab; chemotherapy plus margetuximab plus the company MacroGenics, Inc’s PD-1 inhibitor, retifanlimab; and the third experimental arm is chemotherapy plus margetuximab plus a bivalent PD-1 LAG-3 inhibitor. Based on that, pick your winner for the phase 3.

The first-line setting has at least those 2 studies and probably a few others going to start relatively soon, and I think we’ll talk about the other landscape, but we will have the 4. We have the trastuzumab deruxtecan in second-line studies that are ongoing. We have tucatinib and trastuzumab. We have ZW25, and we have ZW49, which is the ADC [antibody-drug conjugate] built on ZW25. There’s a lot of excitement in the HER2 space, and how we implement them and what’s the optimal sequence of those, I think we’ll see over the next few years.

: Very good. Ken, can you tell us a little bit about DESTINY-Gastric01 [NCT03329690]? This is the trastuzumab deruxtecan study for patients with HER2-positive gastric cancer, which caused such excitement and finally hope for even improved treatments for these patients with HER2-positive gastric cancer.

: The DESTINY-Gastric01 trial was presented at the ASCO [American Society of Clinical Oncology] meeting, and they already published it in 

 by Kohei Shitara, MD. This is a study of the comparison of paclitaxel versus the T-DXd [trastuzumab deruxtecan], the ADC drug, after failure of a first-line trastuzumab-containing regimen. The survival benefit shows a hazard ratio of 0.59, and the progression-free survival shows 0.47, the superiority of T-DXd [trastuzumab deruxtecan] was shown. In Japan, it was approved this month for the second-line treatment after failure of the trastuzumab-containing regimen.

But there’s some concern about adverse events with this drug. A relatively higher proportion of ILD [interstitial lung disease] was seen during the clinical trial, including the breast cancer and gastric cancer trial. We should manage the ILD carefully, so frequently check the CT scan to monitor the symptoms of the patient. Finally, the proportion of ILD in this DESTINY-Gastric01 trial was around 10%, and most of them were grade 1 and 2. We should be careful about ILD during the course of the T-DXd [trastuzumab deruxtecan], but if we are taking care of the patient, the ILD severity will be decreased.

Emerging Approaches in HER2+ Gastric Cancers

: Thank you to all of you for a rich and informative discussion. Before we conclude, I’d like to get final thoughts from each of you. Let’s start with you, Dr Catenacci.

: It’s been a pleasure discussing this today, and it’s nice to have positive studies, let’s start with that. That said, you’ve heard my responses, and I think we have to look at the data; we have to see all the data and look at them critically. My motto is targeted therapies for targeted populations, and immunotherapy is not immune to that concept. I think we’ve seen in every study to date that there are patients who clearly benefit; there are patients who don’t, and in the studies that we’ve seen yesterday, we don’t have the data yet to really comment on whether chemotherapy plus immunotherapy checkpoint inhibitors will enhance their benefit compared to monotherapy. Before I comment on that, I’d like to see those data. If it shows it, great, because that’s what we would love to see. If it doesn’t, then we have to look at the data, and it may be sobering, but it’s the reality. That’s my final summary of all of those studies, including CheckMate-577 [NCT02743494], which I didn’t comment on specifically, but follows those exact same principals, and you can see that in the forest plots.

: I echo what Dan said. Sometimes the devil is in the details, and some of it, we want to see more detailed data, and for some of it, you can get more from the studies already done. Because we talk about exactly what’s happening between CPS [combined positive score] 1 and 5, but there should be data available for ATTRACTION-4 [NCT02746796]. They have not reported any CPS data, and that’s another 700-patient data set that we can get more out of to supplement what we hear from CheckMate-649 [NCT02872116].

: And for CheckMate-577, there are no CPS data; that is really important.

: Yes, absolutely. From all these studies that we’ve heard the top-line data, we really want to hear more details, and there may be things that they haven’t even done that we really want to hear about. I think it’s important, apart from looking at what could benefit our patients, we need to also push our pharmaceutical colleagues to say that these analyses need to be done to guide us.

: It’s a great discussion, and I think there’s a lot of enthusiasm. The research light is really shining bright in our corner. The next few months, before hopefully the onslaught of the vaccine trials, 

 hopefully will see a lot of these data published, and the devil will be in the details. I want to finish up with some positive thoughts. I think in the West, we need to really put all of our patients on clinical trials so that we can understand the difference in the biology and answer very important questions in an expeditious manner. As we heard earlier today, some of these trials are struggling, so please, if you have a patient who is interested in clinical trials, even for the advocates and the patients out there, use the tools available online and find these trials and participate on them because without these trials, we could not have these answers.

Biomarker analysis will be critical and needs to continue. The subset of patients that we’ve continued to struggle with are the genomically stable signet ring cell type of gastric cancer, and for those patients, we need to continue to study them both on the somatic alteration level but also germline analysis of gastric cancer. And esophagus cancer is the next frontier. Hopefully the enthusiasm in this disease will continue to carry us, and more patients will receive second- and third-line therapy because that’s the only way we can really change the curve on this disease, seeing it as a chronic disease, and as a marathon as opposed to a sprint, to try to minimize toxicity to our patients and focus on quality of life.

Another thread at ESMO [the European Society for Medical Oncology annual meeting] this year was nutritional support for our patients, how important it is early on and ongoing, and to use a multidisciplinary approach. Those are all the factors. We argue so passionately about difference in the deltas of a few weeks, when clearly data suggest that specialized care, excellent nursing care, and symptom management can improve survival of our patients in a more meaningful way than some of the targeted agents that we have.

: I’m happy to attend this historic time when the standard of care is changing for both gastric cancer and esophageal cancer; this is an historic era. We have a positive result moving forward in development in gastric and esophageal cancer. But there are some questions...and I hope that global development becomes more action-oriented for gastric and even for esophageal cancer, and we move forward to do more for the patient with esophageal cancer. I wish for further development with accelerated, real global study, I hope. Thank you.

: Thank you so much. And thank you again to our viewing audience, we hope you found this 

 discussion to be useful and informative. Thank you very much.

Global Updates in Advanced Gastric Cancer from ESMO 2020 | Peer Exchange

EP 1: Global Trends in Gastric Cancers

EP 2: Optimal Approach Toward Advanced Gastric Cancers in US

EP 3: Frontline Treatment in Metastatic Gastric Cancer

EP 4: Treatment Landscape in Gastric Cancers Pre-ESMO 2020

EP 5: First-Line I/O Successful in Gastric Adenocarcinoma

EP 6: ESMO 2020: Practice-Changing Data in Esophageal Cancer

EP 7: Applying New Data in the Real-World for GEJ Cancer

EP 8: Predicting Response With Biomarkers in Gastric Cancers

EP 9: Clinical Rationale for Antiangiogenics in Gastric Cancer

EP 10: Toxicity Management in Second Line for Gastric Cancer

EP 11: Looking Toward the Future in Gastroesophageal Cancer

EP 12: Adjuvant Treatment in Gastroesophageal Cancer

EP 13: Maintenance Therapy in Gastric Cancers

EP 14: VEGFR TKIs in Gastroesophageal Cancer

EP 15: Emerging Approaches in HER2+ Gastric Cancers

EP 16: Applying Data to the Real World in Gastric Cancer