Management of gastroesophageal cancers — Daniel Catenacci, MD

DR CATENACCI: It’s a pleasure to be here talking about gastroesophageal cancers.

This is the overview of what I’ll be talking about, first just noting the current standards and then moving on to some newer agents and beyond progression of anti-HER2 therapy, and also in the first line in the perioperative setting with immunotherapy drugs.

So starting with current standards, this schematic here is quite complex and just as a cheat-sheet summary to review through various lines of therapy, first line to third line, and by anatomical group, esophageal squamous, GE junction adenocarcinoma, and gastric cancer, various studies and their eligibility criteria that have led to the current standards and emerging standards. And also at the bottom a number of immunotherapy studies, recently presented, in the last few years, and also ones that are likely practice changing at the recent ESMO annual meeting, and also ongoing studies. And we’ll come back to this after we’ve gone through some of the data to bring it together.

So the standard of care here is based on the best supportive care, median overall survival of about 3 months, and a number of chemotherapeutic regimens, 2-drug and 3-drug regimens, have median overall survivals, plotted here, just shy of 1 year. A preferred regimen is the FOLFOX chemotherapy, and you can see here landmark analyses at 1, 2, and 5 years in terms of what one would expect in terms of longer-term survival.

So the first pivotal study that changed this with respect to targeted therapies was the ToGA study. This was a study published now 10 years ago that led to the approval of trastuzumab along with chemotherapy. You can see clearly an improvement in median overall survival and progression-free survival in all comers. However, when you look at the subgroup analysis, prespecified subgroup analyses, you can see that the patients that were low-level expressors, IHC 0 and 1+, although they were FISH positive, derived no benefit, as you can see in the forest plot. And this made up about a quarter of the patients enrolled.

When we looked at then, in that exploratory analysis below, the patients who had at least 2+ expression plus FISH positivity or 3+ expression, that’s where the benefit was really being driven from, and you can see the median overall survival delta was much better. And indeed, if you look at the 3+ FISH positive patients, those mostly enriched for benefit, the median survival approached 1 1/2 years.

So that’s now superimposed now on this curve, where you can see the original intention to treat analysis, the first bar, the dark purple, is placebo, and the lighter purple the addition of trastuzumab. And then has you increase the level of positivity stringency, you can see the benefit is more pronounced. And now you can see the landmark analyses at 1, 2, and 5 years are not insignificant now. And remember, these plotted numbers are the medians, and there’s a range around those medians. And 2-, 3-, 4-, and 5-year survival is being seen here with trastuzumab patients.

In the second-line setting there was pivotal studies called REGARD and RAINBOW, which was looking at ramucirumab, an anti-VEGF receptor 2 antibody, either as monotherapy in regard or in combination with paclitaxel in RAINBOW. And you can see that both of these studies were positive for overall survival. This led to approval in the second-line setting of this regimen and is a standard of care.

There are other anti-angiogenesis inhibitors that were looked at, including apatinib, a small molecule inhibitor, that was positive in a randomized Phase III study in the solely Chinese population, but in a follow up global study, called ANGEL, this study was negative. Ramucirumab, notably, and also bevacizumab, another anti-VEGF receptor — antibody, were negative in the first-line setting, and therefore this is not approved in the first-line setting.

So you can see here the NCCN Guidelines that ramucirumab plus paclitaxel has a Category 1 indication, but you can also see that there is, in certain circumstances, that FOLFIRI with ramucirumab, with a Category 2B, could be considered. And this was based on retrospective analysis that showed that the outcomes were similar, and in situations like, for example, oxaliplatin-induced neuropathy from first line to proceed with another neurotoxic agent may not be the best thing for the patient, and rather changing it to a nonoverlapping toxicity profile with FOLFIRI might be preferred. And also in patients who failed perioperative chemotherapy with taxane and platinum quickly to proceed with more paclitaxel in the next-line setting may not be ideal either.

So these situations now, we know that we can use FOLFIRI/ramucirumab. And indeed, a recent study at ASCO 2020 virtually, which was a randomized Phase II study of FOLFIRI/ram versus paclitaxel/ram showed exactly what might be expected, and that is that overall survival is quite similar. Progression-free survival is similar, if not trending towards benefit for — for FOLFIRI with ramucirumab. And particularly in patients who had prior taxane therapy, either as a triplet in the first-line setting or, as I mentioned earlier, quick progression after perioperative therapy, these patients tended to derive the most benefit with overall survival, progression-free survival, and response rate. And so again, this is a Phase II/III study, the Phase III component is still ongoing, but because this is now in the NCCN Guidelines, in the United States we can consider this in those indications.

The next big news was the unprecedented approval of pembrolizumab in MSI-high patients irrespective of tumor type, of which of course gastroesophageal cancer did contribute to this. And you can see that here in the plot that 5 of the patients did have gastroesophageal cancer and 60% of them had a response. And so, as we know, this led to second line or higher approval for any solid tumor with MSI high, and this was also partly due to fact that those who respond have a long duration of response. You can see a plateau on the curve there.

Then when we look at specific studies that we’re going to go into more detail later, but with respect to MSI, for the third-line KEYNOTE-059 study, the second-line KEYNOTE-061 study, and the first-line KEYNOTE-062 studies, you can see that the MSI-high group amongst those studies, which again is only about 3% of the patients, they did better than the — than the MSS patients than non-MSI tumors on the left. And also in the randomized studies, in the 061 and 062, you can see that — the MSI-high tumors getting immunotherapy with pembrolizumab did far better than those not getting it.

And so this is approved in the second-line setting, but there was already suggestion that in the first-line setting that this would be beneficial. And we also know now the other evidence of KEYNOTE-177 in colorectal cancer in the first-line setting, so this is the — the trend is to try to give these patients immunotherapy as soon as possible.

In the third-line setting there is trifluridine/tipiracil, and oral chemotherapeutic agent that was evaluated in a large study called TAGS, and it was a large, global, randomized Phase III study that was placebo controlled. And you can see clearly here that patients derived benefit in improved median overall survival, and they were randomized in a 2:1 fashion. So this became a standard of care and FDA approval for third-line setting or higher.

Focusing now on immunotherapy drug, checkpoint inhibitors in the third-line setting or higher, we have 3 main studies. The first is ATTRACTION-2, which was a study done solely in Asian population, and it was a study that was randomizing nivolumab to placebo in a 2:1 fashion and in all comers irrespective of PD-L1 expression. And you can see that this study showed an improvement in median overall survival, just over 1 month, and this led to approval only in Asia for nivolumab in the third-line setting.

Then in the middle we see the KEYNOTE-059 study, which was a nonrandomized Phase II study encompassing about 250 patients, and you can see the breakdown of about 150 patients who were PD-L1 positive, and the remainder were PD-L1 negative by the combined positivity score using the 22C3 antibody. And the primary endpoint of this study was response rate, and you can see there was a decent response rate and enriched for PD-L1 positive tumors. And the response rate for microsatellite stable PD-L1-positive tumors was 13.3%. And again, because of the duration of response in the third-line setting, this led to an accelerated conditional approval of pembrolizumab only in PD-L1-positive tumors, only in the United States.

And then finally there’s the third study on the right, which is the JAVELIN 300, which was a global study, and it was randomized to physician’s choice chemotherapy with either taxane or irinotecan monotherapy, and this study was negative. And in fact actually appeared to be worse with avelumab in a subgroup of patients compared to chemotherapy. So this did not lead to approval. This is the landscape of immunotherapy at this time in the third-line setting.

So to summarize the current standard of care for gastroesophageal adenocarcinoma at the moment, there are a number of cytotoxic agents that can be used in various regimens, preferred as doublets in tandem. My preference is FOLFOX, then FOLFIRI, and then taxane and trifluridine/tipiracil. And then below you can see the few targeted immuno-oncologic therapies that are incorporating the care based on their incidence of any biomarkers when they are indicated and in their therapeutic line, and when they were approved and the anticipated benefit by hazard ratio.

So now moving on, there’s been talk about the benefit of anti-HER2 therapy beyond first progression in the first-line setting based on the ToGA study, and certainly in breast cancer this is a standard of care with a number of anti-HER2 therapies.

But in this disease, there have been a number of negative Phase III studies in the second-line setting. The first one is the TyTAN study with lapatinib, and the second one was the GATSBY study with T-DM1. And so there’s no standard recommendation for anti-HER2 therapy passed first progression.

Partly now we understand that gastroesophageal cancer’s quite heterogeneous and that a mechanism of resistance to anti-HER2 therapy is to evolve and acquire HER2-negative clones. And so the eligibility criteria for classic second or later-line studies was to allow for any HER2-positive cancer that was called HER2 positive prior to first-line therapy. And now we understand that up to half of those patients who aren’t even HER2 positive and may explain why second or later-line studies were negative.

So to go along this line, a recent Phase II study, a single-arm study evaluating a novel HER2 therapy called margetuximab evaluated this margetuximab, which is trastuzumab backbone with an enhanced FC domain to enhance for antibody dependent cell-mediated cytotoxicity, or ADCC, which recruits immune cells like natural killer cells, along with pembrolizumab anti-PD-1 therapy. And in this study, again, patients were not selected based on their HER2 status prior to second line, but rather based on their historical biomarker status prior to first line.

But you can see that of the 17 patients responding to this therapy, out of 92, all of them were higher expressing IHC 3+, not 2+, at their prior to first-line therapy. And 80%, almost, were PD-L1 positive prior to first-line therapy. And almost 90% of the responses had ctDNA positive for HER2 amplification just prior to starting therapy. So in other words, if you select for the appropriate biomarkers for these 2 targeted agents, you can enrich for benefit. And you can see that the response rate amongst double-positive patients was 50% to 60% of the patients with this chemo-free second-line option.

And on the right, we see that indeed that’s exactly what was seen for survival, where those that were double positive did the best. Those that were double negative for the biomarker did the worst. And then if you had one or the other you were in between and in moderate outcomes.

So this has led to a first-line study with this approach with margetuximab in 2 cohorts of the study. The first, Module A, is the exact same chemotherapy-free regimen except the PD-L1 inhibitor’s now the company’s PD-1 inhibitor, retifanlimab and — along with margetuximab. And again, in this really highly selected biomarker group of IHC 3+ and PD-L1-positive tumors with the primary endpoint of response rate in order to see if that hypothesis holds true, and potentially proceed with an accelerated approval if the responses are that good.

And then Module B, which is your more traditional randomized study for your classic HER2-positive patient, and then irrespective of PD-1 status, in a 4-arm study comparing just standard of care, trastuzumab and chemo, then chemo with margetuximab, or chemotherapy with margetuximab with either a PD-1 inhibitor or a bivalent PD-1 LAG-3 inhibitor as a Phase II study to choose the winner to go on to a Phase III. This is an ongoing study called the MAHOGANY study.

And then another big study and news came at ASCO this year, and recently published in The New England Journal of Medicine, is trastuzumab deruxtecan. And this is an antibody-drug conjugate of trastuzumab to deruxtecan, which is an irinotecan analog. And this study was done in solely Asian patients and randomized 2:1 to trastuzumab deruxtecan versus physician’s choice of chemotherapy with monotherapy irinotecan or taxane. And you can see here clearly that the response rate, which was the primary endpoint, was substantially better in the patients who had trastuzumab deruxtecan. And you can also see that the overall survival was better in the investigational arm, as was the progression-free survival.

So this — it needs to be balanced with the side effect profile of this agent, where you can see here that although it is a so-called targeted agent, the Grade 3 and 4 toxicity was higher, actually, than the chemotherapy arm in all comers. And specifically something to pay attention the patient is the pneumonitis that’s seen in about 10% of patients, a quarter of which were higher grade, 3 and 4. There were no Grade 5 toxicities in this study, but there have been reported Grade 5 toxicities in breast cancer and colorectal cancer studies.

So this has led to, first of all, approval in Japan recently, in the third-line setting or higher, and also the FDA has granted breakthrough therapy designation for this agent in HER2-positive disease in the third-line setting or higher. And you can see here that this was based on not only the DESTINY-Gastric01 study, which I just discussed, but also the preceding Phase I study.

So it’s important to note that there are a number of studies ongoing with this agent. The first is called the DESTINY-Gastric02 study, which is in the second-line setting as a single-arm study with Western patients and also a randomized Phase III study called DESTINY-Gastric04, a global randomized Phase III study comparing this agent to standard paclitaxel/ramucirumab. So it will be important to see the results of those data to see if these results hold in non-Asian groups and also earlier as opposed to third-line setting or higher, but in the second-line setting compared to standard of care and putting into context these interstitial lung disease and pneumonitis issues with the drug.

So moving on then to immunotherapy, the first-line setting of immunotherapy has been quite active recently.

We recently had the publication of the previously reported KEYNOTE-062 study that was reported at ASCO 2019, and this study was a 3-arm study looking at chemotherapy with or without pembrolizumab as a placebo control and also a third-arm open-label pembrolizumab monotherapy. So what is shown here is the pembrolizumab monotherapy versus chemo plus placebo, and you can see this similar, what I have dubbed a “ying-yang curve,” where there’s a group of patients who don’t do well and a group of patients who do well. There’s crossing of the curves.

And even when you look at the higher cutoff of CPS 10, which was an amended coprimary endpoint, this also, although it enriched for more benefit, there was still a group that were doing worse compared to chemo. And it’s also notable that this study only selected patients for CPS 1 or higher. So CPS-negative tumors were not eligible for this study.

Then we moved to the comparison of chemo with or without pembrolizumab, and this study did show that pembrolizumab trended to better over chemotherapy plus placebo, with a hazard ratio of 0.85. But this was not statistically significant, nor was it improved in the patients who were CPS 10 or higher, at a higher cutoff. And so this was disappointing to see, but the other thing to note when we compare to the other first-line studies that we’ll talk about here, are that the numbers of patients are quite low in both of these comparisons, 250 in each arm in the 1 or higher and less than 100 in each arm in the 10 or higher. So not a lot of power to detect a lower benefit in this comparison.

So then we had 3 first-line studies reported at ESMO this year, all Phase III studies. We see the first one is the ATTRACTION-4. This was a study conducted solely in Asia. This was — you can see the eligibility here was gastric and GE junction adenocarcinomas, and mostly were gastric cancer in Asia. And you can see they were randomized 1:1 to chemotherapy with or without nivolumab in a placebo-controlled fashion, and the primary endpoint was progression-free survival and overall survival. You’ll notice that this was all comers irrespective of PD-L1 score. And PD-L1 correlates were not conducted.

So the breakdown of the baseline characteristics you can see here. All patients from Asia. You can see that the arms were balanced, and that again, the TPS scoring (tumor positivity scoring) was also balanced. But again, this was not done by CPS scoring. Then the progression-free survival endpoint was met, and you can see that this improved progression-free survival by about 2 months. That was statistically significant. Unfortunately though the overall survival was not different, and some points to note here are that the — the control arm does exceedingly well here. This was far better than what one would expect from a Western population, which is seen routinely when looking at Asian patients compared to non-Asian patients.

In terms of tolerability, you can see here that there was a slight increase in toxicity in the nivolumab arm, and in particular you can see that there were a few higher-grade, Grade 5 toxicities. With respect to immune-related toxicity, again it was a slightly higher, in the Grade 3/4 toxicity rates, across all of these various indications listed here.

The next big study was CheckMate 649, which was the largest study conducted in this disease to date. You can see the original design actually was a 2-arm study, a chemo versus a chemo-free regimen of nivolumab and ipilimumab. This was — and the primary endpoint of TPS 1 or higher patients. There was an amendment to add a third arm during the conduct of the study to add a chemotherapy plus nivolumab third arm. There was also an amendment along the way to terminate the nivolumab plus ipilimumab arm.

We don’t have any official data yet that’s presented, but that’s all we know. And we also saw that the sample size increased to 800 patients per arm. And then the final study then recently reported at ESMO was only reporting on the chemo comparisons of chemo with or without nivolumab. And due to the — the history of this, there’s no placebo control. You can also see that there was a final amendment, which was to change the primary endpoint to OS and PFS and in CPS 5 or higher combined positivity scoring, which is the first time that nivolumab was analyzed in this manner. And we can see all of the other secondary endpoints, including some that were and some weren’t, reported so far to date. We have not seen the CPS 10 or higher, for example.

So some baseline characteristics. You can see that a quarter of patients were Asian, 70% of patients were gastric cancer, and as might be expected 3% to 4% were MSI-high tumors. The primary, or the amended primary endpoint of CPS 5 or higher was met. You can see that there was just over a 3-month survival benefit in the nivolumab plus chemotherapy arm. And then in secondary analysis you can see that in CPS 1 or higher that there’s still a benefit, but that this seems to be attenuated, with a hazard ratio now of 0.77. And then by adding all comers, including the negative CPS tumors, you can see that there’s still benefit, but now the hazard ratio’s 0.8, and also attenuated curve, something I call crescendo plots. And you can there that the hazard ratio going from 0.71 to 0.8, that one must ask the question what the contribution is of these smaller populations of patients in this study, and really the expectation is that it will be very minimal if at all.

So in terms of subgroup analyses and the forest plot, notably this was only the primary endpoint CPS 5 or higher patients. And you can see again that the Asian patients do better than the non-Asian patients. You can see that although the TPS scoring here is shown and does enrich for benefit, we’ve not seen the CPS scoring, say of 10 or higher versus less than 10. And the one thing to really note is the MSI-high tumors are off the charts to the left here in favor of nivolumab, as was expected.

Another thing to now contrast the 2 studies then is that the ATTRACTION-4 study not being significant for overall survival. One concept here is that there was more crossover in the second-line and later setting of getting just more second-line therapy and more second-line immunotherapy. That may have diluted out some benefit, but there’s still some unexplained phenomenon here. On this study, the Asian patients still do better with nivo than without nivo, and so did these patients not crossover as much. It’s just not clear.

And so in terms of tolerability, similar to what we’ve been seeing. There is a higher toxicity profile with nivolumab that we see here, a 15% higher Grade 3 to 4 toxicity of any treatment-related event and also higher treatment-related deaths. And with respect to immunotherapy-related etiology, you can also see that there’s a higher Grade 3/4 toxicity compared to the control arm.

Finally, we get to the KEYNOTE-590 study, which is a randomized study. Now this study is very similar in eligibility to the preceding second-line KEYNOTE-181 study, where this study allows also esophageal squamous cell tumors, which are notably more sensitive to immunotherapy, and not gastric adenocarcinoma. And randomized 1:1 in a placebo control. The other notable difference is that this cisplatin as a backbone chemo. The primary endpoints are overall survival and progression-free survival. You can see here that the makeup of this study is favorable towards immunotherapy, with more than half of patients having — from Asia, almost 75% squamous cell, and almost 50% of patients CPS 10 or higher. So all factors that we know enrich for benefit from immunotherapy.

And you can see here that there were a number of coprimary endpoints of overall survival. One was amongst squamous cell tumors with a CPS 10 or higher score. You can see there’s a clear benefit. Then the next coprimary endpoint was amongst all squamous cell irrespective of PD-L1 expression, where you see benefit but it’s somewhat attenuated. And then you see in all comers irrespective of histology that were CPS 10 or higher there is benefit, but slightly less than in the squams. And then in all comers, irrespective of histology and irrespective of PD-L1 score, where you see benefit but again it’s somewhat attenuated.

And you see that here in the forest plot, where looking at the CPS less than 10 the benefit’s really not there. This is very similar to what was seen in the preceding KEYNOTE-181 second-line study, and also that the benefit is more substantial in the squamous cell tumors. We have yet to see histology by PD-L1 score, which we want to see, and might expect that the adenocarcinomas might do best with CPS higher than 10, whereas the ones less than 10 do not derive benefit and potentially even for squamous cell tumors.

And again, I’ll point out that almost 50% of patients who were CPS 10 or higher amongst adenocarcinomas, which is at least 2 if not 3 times higher than what one would expect in all comers. So whether there was some enrichment of this study, and also CheckMate 649, we can discuss.

So in terms of tolerability, again we see that there was a slightly higher toxicity rate, but overall generally tolerated, with about a 4 % to 5% higher Grade 3 toxicity rate and about 5% higher immune-related toxicity rate in the pembrolizumab arm. A few notes about PD-L1 diagnostic antibodies now, because it’s becoming somewhat confusing, is that there are different lines of studies with different checkpoint inhibitors, using different diagnostic antibodies with different sensitivities and specificities, using different scoring systems, combined positivity score, not only the tumor positive score, but also the immune infiltrate, and then the cutoff of what’s positive, 1, 10, or whatever — 5 now. And therefore the difference in incidence of what’s called positive, and therefore the different performance in terms of what is enriching for benefit.

So what we do know is that the higher the level the more likely a patient’s going to respond. The ones who are negative, particularly with immunotherapy checkpoint monotherapy, is really not going to be helpful. And now we’re seeing evidence that that’s the case in combination with chemotherapy.

Then — to summarize then and put in contrast these agents and the studies across the 4 main first-line studies, you can see the size of CheckMate 649 really stands out. It’s a huge study. It also was not placebo controlled, unfortunately; different chemo backbones, different makeups of histology, different makeups of agents, different makeup of biomarker incidences, and therefore different outcomes.

And so to summarize now all of what we’ve discussed, you can see here, just going from left to right, that in the first-line setting for HER2 positives it’s HER2-positive tumors get trastuzumab, although esophageal adenocarcinomas were not eligible for that study. That’s where many patients that are HER2 amplified reside, and so many of us treat those patients with chemotherapy plus trastuzumab. We see that in the HER2-negative group we do expect approvals with nivolumab for patients with adenocarcinoma with CPS above 5. And also you can see the KEYNOTE-590 study, which straddles across squams and adenos, where we might expect approval in 10 or higher there. And there will be some choices in terms of what one uses in those GE junction adenos.

Then in second line, standard of care is paclitaxel with ramucirumab or FOLFIRI with ramucirumab, as I showed you. And again, although esophageal adenocarcinomas were technically ineligible for those studies, it’s very hard to distinguish these tumors clinically, and many of those patients are treated, as we do, as an adenocarcinoma. And then finally in the third-line setting you can see the trifluridine/tipiracil, and then some things that might become moot like pembrolizumab in the third-line setting for 1 or higher. If it’s being used earlier in the first line, those patients above 5, at least, will not be getting it again in the third-line setting.

So finally we’ll just talk briefly about perioperative therapy.

There is only 1 study in this space at the moment that looked at immunotherapy, and that was also reported at ESMO this year, and that’s CheckMate 577. That was a study that was including, again, adenocarcinoma and squamous cells of the esophagus and GE junction. Patients had to have had neoadjuvant chemoradiation of any type, although about three quarters had the CROSS regimen. Patients had to have had an R0 resection, complete resection, and patients could not have had a complete response to therapy, nor could they have had an R1 resection. So patients with the best outcome and the worst outcome were excluded, and only patients with residual disease in the specimen were eligible, which is a known poor prognostic factor, particularly in node-positive disease.

Patients were then randomized 2:1 in a nicely clean study to nivolumab versus placebo with a primary endpoint of disease-free survival. At the meeting we saw that the makeup was mostly non-Asian patients, that the makeup was 40% GE junction, 70% adeno, and you can see here that the TPS scoring was equal in both arms, but CPS scoring, which is what we care about, has not yet been reported.

The survival data, disease-free survival is shown here, which shows a clear benefit, almost double the median disease-free survival. Some things to note are something that I call the “tuning fork plot”, where you can actually see that the tail of that curve is really showing no benefit in about 30% of patients. So we’d like to know who those patients are to spare them from nivolumab. And again, this is in all comers irrespective of CPS score and irrespective of histology.

When we look at those subgroups in the forest plot, again as expected, squamous cell tumors do better than adenocarcinomas and GE junction cancers are nearly on the line. And again, these are irrespective of CPS score, which we would expect to even differentiate this further. What we do see at the bottom is TPS scoring, which does not predict outcome, which of course we know very well. So what we’re looking for here is breakdown by histology and by CPS score, longer-term follow up for disease-free survival and also overall survival, which we’ve not seen yet.

In terms of tolerability, again relatively tolerated single-agent with slightly higher Grade 3/4 toxicity, but overall tolerable. And in terms of the treatment — immune-related events, you can see slightly higher both in any grade and Grade 3/4, and so we have to take this in context in terms of the — the survival outcomes.

So to summarize the perioperative setting, we’ve seen here, again, there’s this discrepancy of whether to include GE junction cancers with gastric or versus esophageal squams, and you can see 2 major approaches, FLOT perioperatively or CROSS neoadjuvantly. And you can see the superimposed CheckMate 577 study, which is a positive study waiting for full data and for breakdown of the histology and CPS scores, as I mentioned, but you can clearly see that this will change the standard of care, at least in a subgroup of patients.

In the HER2 space unfortunately it’s been a little bit challenging to improve outcomes. We had the RTOG1010 study with CROSS with or without trastuzumab that was negative at ASCO this year, and so the standard of care for HER2-positive patients is to — to not treat with any HER2 therapies. The PETRARCA study was a Phase II randomized study with pertuzumab and trastuzumab versus just FLOT and showed suggestion, but it was terminated early, and so that will not be able to be acted upon.

So to summarize then, for current standards in the perioperative setting we have FLOT for gastric cancer, CROSS for squamous cell cancer, and for perioperative therapy in the GE junction adenocarcinoma either one, and there are debates as to which one is better, but both are considered options. HER2 therapies have failed. We’ve also seen VEGF therapies with bevacizumab fail, and we’re awaiting a ramucirumab study called RAMSES that is ongoing.

In the first-line metastatic setting we have chemotherapy or chemotherapy plus trastuzumab for first line HER2-positive disease, but other agents like lapatinib and pertuzumab have failed in the first-line setting. In the second-line setting we have chemotherapy plus ramucirumab, and for MSI-high tumors pembrolizumab and for squamous cell tumors of the esophagus, CPS 10 or higher, and also nivolumab for all squamous cells in all comers based on the ATTRACTION-3 study. And then HER2 therapies in the second line or higher setting have failed. In the third-line setting we have the remaining chemotherapeutic options if not yet used, as well as trifluridine/tipiracil, based on the TAGS study, as well as pembrolizumab for CPS 1 or higher tumors.

And we talked about new agents in the HER2 space beyond progression. They are listed here. We talked about margetuximab and trastuzumab deruxtecan, or DS8201a. There are others, as well, that are in development, also with the potential for combining with immunotherapy.

And then in terms of the first-line setting, where all the action was at ESMO this year, we had 4 — 3 first-line studies and a perioperative study. And the theme there, again, was that squamous cells did better than adenos. The higher the PD-L1 level the better. Asians tend to do better than non-Asians. Non-placebo controlled-study did better. And although there was a lot of talk about the chemo backbone, oxaliplatin and cisplatin, there were studies with oxaliplatin that were negative, like ATTRACTION-4, and there were studies with cisplatin that were positive, with KEYNOTE-590. So overall that doesn’t seem to be a major player as to why there’s such heterogeneity in outcome of these various studies.

And so in my prediction, what I anticipate will change for sure the standard of care is KEYNOTE-590 will lead to squamous cells 10 or higher approval, potentially the adeno CPS 10 or higher, but we need to see those subgroups, and then the less than 10s we really need to see for all-comer squams, for all-comer adeno. But I anticipate probably not. For ATTRACTION-4, an Asian-only study, but it was negative for overall survival. CheckMate 649, we do anticipate an approval for CPS 5 or higher for nivolumab. And again, we need to see the data for the lower level PD-L1 scores, but I anticipate probably not, as those were secondary endpoints and by the curves do not seem to add much in terms of benefit. And then finally for CheckMate 577, in the adjuvant setting, I think certainly for squamous cells with CPS above 5 or 10, potentially also for those that are even low level. And then we need to see the adenocarcinomas by CPS scoring to really make a decision whether or not those patients are going to derive benefit from a year of adjuvant nivolumab.

So I’ll stop there, and it was a pleasure to discuss this with you today.