This transcript has been edited for clarity. For more episodes, download the Medscape app or subscribe to the podcast on Apple Podcasts, Spotify, or your preferred podcast provider.
Samuel J. Klempner, MD: Hello. My name is Dr Sam Klempner. I'm a gastrointestinal medical oncologist at Massachusetts General in Boston. Welcome to the Medscape InDiscussion: Esophageal Cancer podcast series. Today we'll be discussing the gastroesophageal junction (GEJ) with one of the global leaders and a good friend, Dr Yelena Janjigian. Dr Janjigian is a full professor and chief of gastrointestinal oncology service in the Department of Medicine at Memorial Sloan Kettering.
As many of you know, her research, career, and clinical practice are focused on improving outcomes for this patient population. So, thank you, Yelena, for joining me.
Yelena Y. Janjigian, MD: Thanks for having me.
Klempner: You're probably starting to prepare for your clinic later this week and are looking over your new patients. I'm sure there are many of them. When you're looking at a new patient, let's say this is a localized patient, what are you making sure has been tested so you have a full set of information when you're walking into clinic? Biomarkers in particular.
Janjigian: In patients with early, nonmetastatic disease, it's important to assess tumor location, imaging, and endoscopic assessment of extent of the tumor into the gastric cardia and up into the GEJ and the esophagus.
It is also important to have unstained slides or tissue block. We often request that ahead of the visit to make sure that we have critical biomarkers for immune response and chemotherapy resistance, such as microsatellite instability (MSI) and DNA mismatch repair (MMR) deficiency, which is probably the most important biomarker in early-stage disease. We also get routine testing for HER2, PDL1, claudin testing, and so forth, because often the metastatic disease declares itself, and then we need to act on it quickly.
Klempner: I totally agree. Speaking along those lines, a lot of times we'll see an outside endoscopy report that says "distal esophageal mass," but then maybe you have a PET scan and it looks a little bit more GEJ or even cardia; do you routinely rescope these patients? Do you think about laparoscopic staging just to get, as you said, the full set?
Janjigian: Yes. The true esophageal adenocarcinomas, mostly in the United States and in the West, extend and cross the GEJ. So, it's critical to repeat endoscopies, and we try to do it at Memorial Sloan Kettering essentially the day after they see me and do a retroflexion to see exactly what you're describing. To extend into the gastric cardia is important for surgical consideration, for perioperative therapy consideration, understanding whether or not a surgeon will be able to achieve R0 resection — perhaps with just chemotherapy, or whether chemo radiation is needed.
And you're right on point about the laparoscopy. In the United States, not enough patients are staged aggressively with laparoscopies, and often metastatic peritoneal disease is missed. In patients with large bulky tumors extending into gastric cardia, laparoscopies are important and are done routinely at our site.
Klempner: Yes, there's nothing harder than walking into surgery and then the surgeon opens up and finds small peritoneal disease, and you never know whether it was a response or resistance or stability; it's impossible to tell without a baseline.
You've led several trials and have been involved with multiple nonmetastatic trials. Most recently, I think we're all eagerly awaiting MATTERHORN, but as we've seen with KEYNOTE-585, maybe give us a broad rationale for moving immunotherapy into the nonmetastatic setting in GEJ.
Janjigian: Sure. We know that immune checkpoint blockade helps chemotherapy and, in combination with chemotherapy, improves survival in stage IV disease. We know that in our clinics, most of these patients present in early-stage disease, even, with T3N1 tumors. And then after standard chemoradiation, we know that adjuvant nivolumab, if there's persistent and residual disease after complete resection, adjuvant PD1 blockade can improve survival in those patients.
Biologically, we also know from melanoma, lung cancer, and bladder cancer that neoadjuvant treatment while the tumor is still in place, striking the iron while it's hot, helps, and immunotherapy helps expand antitumor immune response early on, and then you can survey for micrometastasis.
So with that rationale, and encouraged by a positive adjuvant study, KEYNOTE-585 took chemotherapy and immunotherapy for a neoadjuvant approach. Of course, it was somewhat of a culture shift because it was mostly a study that accrued patients in Asia, where most patients get treated with surgery first and then get adjuvant chemotherapy. And this study also came before fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) became a globally accepted regimen. So, in retrospect, KEYNOTE-585 should have probably focused on triplet chemotherapy regimens, but it really started with a doublet chemotherapy regimen in combination with pembrolizumab vs placebo.
It was a tricky study to enroll too, because in the United States, obviously most patients would get FLOT. And if they didn't get FLOT, then we wouldn't necessarily want to give them cisplatin-based regimens because we've really moved away to oxaliplatin-based regimens. So, it was a little bit of a tricky study.
We didn't participate in it; it was a big ask in terms of different interim analyses and alphas. In the end, although the Kaplan-Meier survival course for disease-free survival with placebo vs pembrolizumab in combination with chemotherapy did separate, the difference was not statistically significant, and it was a big letdown.
The other lesson learned from that study is that you cannot assume that just because you have a pathologic complete response rate improvement, which the interim analysis was positive for, you will see a survival difference. We are really getting more picky about these interim analyses because you do lose a lot of alpha with it. And we like to look only if it matters. I think that in the future — you and I chatted a lot about this — perhaps looking at node-negative status would be a more appropriate surrogate for these early looks.
Klempner: Yes, I think that data on nodal regression is impressive from the UK group and from some of the trials.
Let's take it back to the clinic a little bit. So, the patient is a 74-year-old person who got through four cycles of preoperative FLOT. They had a scan. They had some clinical response. Maybe it's by PET, maybe it's by CT, and they go to surgery. I'm going to ask you two questions. One is, what do you do with these patients who are essentially nonresponders? They have limited pathologic response, maybe they're still node-positive. Are you retesting the primary and considering outside-the-guidelines therapy? Are you rigid with the adjuvant FLOT? And then do you ever think about more than four [cycles] in the preoperative setting, based on the scan after three or four cycles?
Janjigian: Total neoadjuvant FLOT is a bit tricky. It's difficult and there's a narrow window of tolerability. Typically, I don't do it outside of a clinical trial. We had done it during COVID times if the surgery couldn't be scheduled in real time. But typically, I don't routinely do FLOT in a total neoadjuvant approach.
In terms of what I do with a patient who had almost no response in the adjuvant setting to FLOT, we know that switching completely away from chemotherapy is a bad idea. Obviously, nivolumab/ipilimumab flopped severely compared with FLOT. It was detrimental. These tumors may still be chemotherapy-sensitive. So, outside of the guidelines, I do rely on PD-L1 status in those cases. And if the patient had a total gastrectomy or even gastroesophageal surgery, R0 resection, it's very difficult to tolerate FLOT adjuvantly. So I look for any excuse to drop the docetaxel, continue with a FOLFOX, the doublet, especially in a patient in their 70s, and then often add anti–PD-1 if they're PD-L1 high.
Essentially, we know that patients with a node-positive tumor with five or six positive lymph nodes residual post-FLOT are essentially stage IV. So, the way that I justify it to myself and to the patient is that if there's high PD-L1 expression, the options are waiting for the tumor to recur and giving FOLFOX/nivolumab at that time for recurrent disease or treating the micrometastasis now. For most of the patients, we try to get them on chemoimmunotherapy before it recurs.
Klempner: That's a very fair and reasonably thought-out strategy, and I have done the same a few times. We definitely see changes in PD-L1 expression. Whether it's heterogeneity or treatment-induced upregulation remains to be determined, but we do see that. I certainly encourage listeners to retest the surgical specimens at the time of surgery.
Let me push you a little bit on future thinking. Let's say you had a regimen with a high biologic activity, something like the KEYNOTE-811 regimen in a HER2-positive, nonmetastatic patient. In these settings, do you think we'll ultimately be able to get to, say, doublet chemo with biologic plus immunotherapy — let's say, taking the entire KEYNOTE-811 regimen into the nonmetastatic space? Or do you think we'll need FLOT backbone entirely?
Janjigian: I believe that for high overall response rate regimens, we may be able to get away from FLOT. Ultimately, it depends on what our goal is, because it's a seesaw between patient tolerance and being able to get treatment in, vs perhaps even nonoperative approaches, right?
So for non-MSI tumors, in the HER2 space particularly, a game changer would be to skip surgery altogether. It would be a grid of biomarker selection of which subset of patients we can avoid surgery in and even get away with a doublet plus chemo immunotherapy. So let's say it's a MYC wild type, EGFR wild type, high-level HER2 and PD-L1 co-expression. I bet the type of P53 mutations in the future will matter. All of that. And then we can stratify patients whether or not they need a doublet or a triplet with biologic and anti–PD-1, and then close surveillance.
Klempner: Speaking of how we're going to stratify within these groups, let's say I come to you with a newly diagnosed clinical node-positive GEJ adenocarcinoma, and I tell you right from the beginning, "I want to do everything to avoid surgery." Are you checking or do you ever do ctDNA at baseline in nonmetastatic settings? And do you think we'll get to a place where the performance characteristics of these assays will ultimately allow us to think about nonoperative management, even in GEJ adenocarcinomas, non-MMR deficient and HER2 wild type?
Janjigian: The short answer right now is that none of these assays are good enough for non-MSI gastroesophageal cancer, where I would feel extremely confident watching a patient, especially a young patient. We're just not that good about picking up clinical minimal residual disease on endoscopies. We don't really have a very reliable way to assess them. And ctDNA signatory, even, and all these assays are just not that good for low tumor mutational burden with these aneuploid tumors.
Klempner: Yes, it's interesting. "Squams" shed at like a 7x rate of adenocarcinomas on average, so we've thought about doing this for some squamous esophageal patients. But I agree — it's hard to think about cure without surgery in the current paradigm right now, save for the MMR-deficient group that you mentioned.
So maybe let's get into that. Remind us a little bit about why MMR testing in the nonmetastatic setting is relevant. Prognosis, response to, benefit from standard chemotherapy, etc.
Janjigian: Depending on which dataset you look at, 5%-10%, or maybe a little higher, of gastroesophageal cancers, adenocarcinomas, are MMR deficient. Some of these patients may have de novo Lynch syndrome, so it's important to do germline testing. But the quickest way to do this is to do MMR deficiency testing on unstained slides of the four proteins by immunohistochemistry. It's quick.
And then if they're positive, I think as a surgeon, it's a very satisfying thought to be able to just do an operation and cure someone; 85%-90% of those patients will be cured with surgery alone. We know that the response to anti–PD-1 therapy by itself or dual combined immune checkpoint blockade ranges anywhere from 60% upwards to maybe 80%. It's not 100% complete response, right? Not the same way that we see in the rectal data, for example, with dostarlimab, but it's pretty close. It's pretty good.
So the number one lesson is to test everyone, and number two, realize that if it's a distal gastrectomy and the patient is elderly and you don't want to risk adverse events, they could probably get away with just doing surgery and this patient will not need anything ever. We know that our data show that if the patient after resection or MSI tumor is signature negative, they're not going to recur. If they're ctDNA positive, you have to act on it. So that's where ctDNA would be very helpful, to monitor those patients.
The big leap forward is to whether or not we need surgery in those patients. And then it just depends on the patient preference, the likelihood that they're going to achieve complete response and how comfortable you are with monitoring them closely. At our center, we do offer nonoperative approaches with these patients, but we review them carefully at each [disease management team meeting]. And it's tricky; endoscopies and PET scans are just not that good at predicting clinical and pathologic complete response. You may miss it. I guess that's okay if the patient is willing to get endoscopies every 2-3 months. But it's not a fun thing to do, obviously.
Klempner: We do the same and we also pursue nonoperative management after informed discussions. But let's say you're going to go down a nonoperative pathway and the patient has received, essentially, a course of immune checkpoint inhibitor. How long do you give the checkpoint inhibitor?
Janjigian: On the study we would give it for 6 months. But because it's part of the NCCN guidelines, you could do it basically however you want to do it. So the right answer is, we don't know — probably 6 months to a year. Again, it depends on adverse events, depth, and likelihood of complete response. And now we know that not all MSI tumors are the same. We know there are some subsets of mutations that are associated with resistance to immune checkpoint blockade, certainly tumor mutational burden, heterogeneity of the disease. We've seen patients with complete regression of MMR-deficient clone without growth of MSS, non-MSI clone. So this disease is tricky. I think most patients need chemotherapy plus immunotherapy. The question is, should we give it in an MSI setting? We don't know that yet.
Klempner: I agree. We tend to do 3-6 months, image, and rarely consider doing longer in people who have had a complete clinical response.
All right. Let's take our patient who got to perioperative FLOT. Unfortunately, they recur 6 months later with — let's make it interesting — let's say they have an isolated lung metastasis. Someone like that comes in to your clinic. They say, "You know, I have a little bit of neuropathy left over from the adjuvant component of FLOT. I got through it. I'm here in your clinic with a really asymptomatic metastatic recurrence." What is your go-to nonprotocol management right there?
Janjigian: I don't approach this as oligometastatic, if that's the path you're going down. I just get them back on chemo, whether or not it's a clinical trial. We'll do claudin testing and other things. The fact that the patient recurred so quickly and has some neuropathy, we could consider FOLFIRI; ramucirumab is standard practice. But I wouldn't jump into local therapy just yet. Let the disease declare itself because often on the follow-up scan, there are going to be other things in the liver.
Klempner: I agree. Metastatic disease is, practically speaking, a palliative setting now, and management of toxicity and quality of life is arguably increasingly important.
Let's say it's the same 74-year-old; are you giving all these patients full FOLFOX right off the bat? 5FU bolus, 85 of oxaliplatin, 1200 of 5FU. Or are you saying we've seen some data with noninferiority with lesser dosing? And how do you think about frontline in someone who's pretty healthy, but maybe has a little bit of leftover toxicity?
Janjigian: I think most people agree, from the CheckMate 649 and other studies, that the majority of the grade 3/4 toxicity comes from chemotherapy. So if you get them through chemo side effects, immunotherapy, anti–PD-1 doesn't add that much. I mean, other than immune-related adverse events.
So how I do it is I drop the bolus, leucovorin, and the 5FU. I decrease the oxaliplatin to 70-75 and the 5FU infusion to 1000. Most patients do quite well, and as a result, you are able to actually get more treatment in, so the dose density overall is better.
Klempner: We follow a very similar strategy. That GO2 trial from the UK was quite important data to provide some reassurance about even initial dose reductions, right off the bat. Across the world, there's variability in screening dihydropyrimidine dehydrogenase (DPYD) or uridine diphosphate-glucuronosyltransferase isoform 1A1 (UGT1A1), if you're thinking irinotecan; are you guys considering this?
I know our institution doesn't do it preemptively for all patients, but we have considered it in some of our pharmacy discussions.
Janjigian: I know, and it comes up a little bit now because there's been a lot more pressure and press about it. The truth is that in the United States, these alterations are so rare and there are so many different genes that can encode the protein. And then there are still some issues; even if you don't have the mutation, there's a posttranscriptional modification, so the protein may be ineffective.
We give 5FU so often. And in many cases, for curative-intent therapy, it would be really detrimental to guess and preclude some patients from receiving 5FU. So as an institution right now, our position is not to test, but I know there's been a lot of pressure from external bodies, and regulatory authorities are currently reviewing this again. But right now, NCCN and ASCO are not recommending testing.
Klempner: I know we're getting to the end now, but let's take the same patient. You got them through partial response on FOLFOX and nivolumab. At month 8 they're starting to have real progression. When you start to think about second-line therapies in, let's say, someone without a clear biomarker that you have linked to a trial, I think we're starting to think a little bit more nuanced about just blanketly giving everyone everything or everyone the same thing.
When you think about progression on immunotherapy, do you look at the pattern? Do you look at the location? There's a lot of inertia from patients sometimes to think about ablative therapies in an oligoprogressive setting. Maybe talk us through your practical thinking when someone comes in with a scan that shows growth in that liver metastasis that had previously responded.
Janjigian: Close follow-up and symptom monitoring, but also imaging, are very important. One of the reasons why you and I probably get more patients on to second- and third-line therapy — as opposed to, if you look at just the general population or large phase 3 studies where only 40% or 50% of patients go on second- or third-line therapy — is because we really watch them carefully. And that will improve their chances of responding or at least doing better on and getting to subsequent therapies.
There's not a right or wrong approach to this. I typically don't ablate or radiate things that are not symptomatic, especially in a patient with a relatively short duration of response; 8 months is not great to first-line therapy, so that means their systemic disease is the battle. So I try to get them to another systemic regimen, and irinotecan actually has a pretty good track record.
The only time I consider ablative therapy or metastasectomy, or even removal of the primary tumor, is really for these outlier responses where, after 2 years on, only the primary tumor is here or even the primary is gone and then there's an adrenal metastasis that pops up. Otherwise, I typically only address symptomatic disease or radiated bone metastasis or things like that.
Klempner: Thank you. I think that's a very helpful framework. This has really been a great conversation, and in the future I think we'll be talking a lot more about newer therapies, where I know we spent a lot of our time thinking about how to develop newer drugs in select patients.
Today we've talked to Dr Yelena Janjigian. Thank you for tuning in. Please take a moment to download the Medscape app to listen and subscribe to this podcast series on esophageal cancer. My name is Dr Sam Klempner, and on behalf of the Medscape InDiscussion podcast, thank you very much for joining.
Resources
Adjuvant Nivolumab in Resected Esophageal or Gastroesophageal Junction Cancer
The KEYNOTE-811 Trial of Dual PD-1 and HER2 Blockade in HER2-Positive Gastric Cancer
Medscape © 2024 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Understanding, Detecting, and Treating Gastroesophageal Junction Adenocarcinoma - Medscape - Jun 26, 2024.
Comments