MATTERHORN Results Yield Promise in Resectable Gastric/GEJ Cancer

Results from the phase 3 MATTERHORN trial (NCT04592913) were recently presented at the European Society for Medical Oncology (ESMO) Congress 2025.¹ This trial looked at durvalumab (Imfinzi) plus 5-fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) as treatment for patients with resectable gastric/gastroesophageal junction adenocarcinoma. Of note, on November 25, 2025, the FDA approved this regimen for this indication.²

An Around the Practice program moderated by Rutika J. Mehta, MB, BS, MD, MPH, an associate professor in medicine and attending physician at Weill Cornell Medicine and NewYork-Presbyterian Hospital, focused on the results and how they play a role in the treatment of gastrointestinal cancer. She was joined by Sharon Shiraga, MD, FACS, a general surgeon at Keck Medicine of the University of Southern California; Sreenivasa R. Chandana, MD, PhD, a medical oncologist at The Cancer & Hematology Centers; and Jacques Fontaine, MD, a surgeon at Moffitt Cancer Center and a professor at the University of South Florida.

Mehta: The MATTERHORN study was a global phase 3 study comparing durvalumab plus FLOT vs FLOT with placebo in gastric/gastroesophageal junction adenocarcinoma. We initially had the PCR [pathologic complete response] and the NPR [non-progression rate], which had improved with the addition of durvalumab to FLOT. The interim overall survival [OS] analysis shows that there was a trend in its improvement with OS as well, with the addition of durvalumab. Based on the final analysis and the association of the pathologic response to EFS [event-free survival], the data were presented at ESMO. The trial was stratified by geographical region—Asia vs non-Asia—clinical lymph node status—positive vs negative—and PD-L1 expression…. The primary end point was EFS. After completion of the post-operative durvalumab plus FLOT, durvalumab was continued up to 1 year.

Other perioperative studies that we know are the phase 2/3 FLOT4-AIO study [NCT01216644] and the MAGIC trial [ISRCTN93793971]. Dr Chandana, on the design of the MATTERHORN study, do you think it’s a clean design? If you had the opportunity to design a perioperative study using immunotherapy, would this be something that you would have thought of? Would you have kept that maintenance durvalumab continuing for 1 year, and what is your rationale for either keeping or not keeping it?

Chandana: MATTERHORN is designed very well. I agree with the study design. Not all gastric cancers are the same, and we relied on chemotherapy for a long time. Now we have the data that immunotherapy works in this cancer, and in later stages of the cancer. This is one of the efforts to involve immunotherapy in combination with chemotherapy to improve the response rates and survival. I would have designed it the same. One of the things that we still didn’t answer is the question, would you give all chemotherapy upfront? That’s probably the way MATTERHORN was designed, mimicking giving half of the chemotherapy before [surgery] and half of the chemotherapy after [surgery]. The only difference here is adding durvalumab, which is very appropriate, and maintenance therapy. I like the maintenance for a year, because this will give maybe some advantage to these patients who have “bad biology” and maybe have some minimal residual disease. One thing I would also include—I’m sure there is data collected on this part of this study—is what is happening with the ctDNA [circulating tumor DNA] and minimal residual disease. We will also learn a lot from the study when they analyze all the data. Is there a biomarker for different patients? You have different varieties of gastric cancer, with the diffuse vs non-diffuse, and the biology is a little different. I would also look at some of the responders; for example, we don’t check in this setup for HER2 and CLDN18.2. Is there any difference? Durvalumab acted differently with different biomarkers. There are a lot of studies ongoing, including the perioperative phase 2 GEMINI-Gastric trial [NCT05702229], now looking at the biomarker-based therapy CLDN18.2 vs HER2 vs other PD-L1 [biomarkers]. There’s a specific trial ongoing to improve upon the outcomes of the MATTERHORN study, but we have a lot to learn. The study design itself is very good for me.

Mehta: Is there a percentage of patients who are referred to you, maybe after surgery, who’ve not seen you prior to surgery, and therefore they didn’t get any pre-operative treatment? If so, what percentage would that be?

Chandana: It used to be all the time maybe 10 years ago, but now, everybody knows that neoadjuvant therapy is the standard of care, especially in our institution, where all the surgeons are very well trained, and they’re very specialized. Occasionally, we’ll see a community surgeon operating on a patient who came to the hospital with some bleeding or some perforation. That’s a different story. I might have seen, in the last year, 1 or 2 patients like that, but not huge numbers.

Mehta: Do you think that, if durvalumab plus FLOT becomes the standard of care, there would be any challenges in terms of getting patients initiated on durvalumab plus FLOT from day 1? How do you think the coordination of care would change between surgeons and medical oncologists? Do you incorporate [prehabilitation] in your patients, and are you concerned about the toxicities?

Shiraga: The study is well done and done safely and smartly. Durvalumab is an anti-PD-L1 drug. Instead of just selecting for the patient, they gave it to all comers, including 10% of those who were PD-L1-negative in each group, with 5% of the microsatellite instability [MSI]-negative group as well. Even with that, they showed overall improvement. This is advocating for durvalumab plus FLOT for all comers. In the US, that is coverage driven; even when a patient would benefit from immunotherapy, sometimes a current state requires authorization, and there is perhaps some delay in getting the medication…. This will change our practice. We’re starting to adapt this, and this will make a difference. As far as the adverse effects, this is interesting. We have seen immunotherapy tolerated well by patients. However, adding immunotherapy to FLOT may change it a little bit.

In the trial, the overall safety profile was good, but there was a 10% discontinuation of either durvalumab or FLOT. We have to understand more. This is a new thing. It’s new in the upper gastrointestinal [GI] arena. Overall, it is very safe, but it’s up to us to be more diligent about who we should continue to follow up with of these patients to see how well they’re doing from the immunochemotherapy combination, and how well we can best serve them. We’re going to continue to do the same as we do right now, but we will have to think about the long term. When we get more numbers, I think there will be a different side of the adverse effect profile that we have to put our detective hats on and observe, as diligent doctors.

Mehta: Do you anticipate that you’ll have to give more time for patients to recover after pre-operative chemotherapy, before you take them for surgery, if your institution decides to do durvalumab plus FLOT when it gets approved?

[Editor’s note: The filming of this discussion occurred prior to the November 25, 2025, FDA approval of durvalumab plus FLOT in resectable gastric or gastroesophageal junction adenocarcinoma.]

Fontaine: Patients who have received FLOT or durvalumab plus FLOT have a tendency, unless they’re extremely well selected, to have delays in surgery. Then, a certain percentage of them, unfortunately, don’t make it to surgery. That’s why it’s important to be able to select patients appropriately for durvalumab plus FLOT, and if the patient may not make it to surgery, then we have to either decide to give them definitive chemotherapy plus radiation or to change the dosage of the regimen of the neoadjuvant therapy. It’s a good study design. Unfortunately, a minority of patients were not PD-L1-positive, and it’s hard to do subgroup analysis in these types of patients. We do know from multiple other studies and multiple other solid organs that patients who have a PD-L1 [expression] of at least 1% will do better with neoadjuvant or adjuvant immune checkpoint inhibitors…. We should include patients who have a PD-L1 of 0% but in some places in Europe, if your PD-L1 is not 1%, you’re not going to be approved to get durvalumab plus FLOT.

The second point is that I can extrapolate from lung cancer data that preoperative checkpoint inhibitors, if they work, they work well, and you don’t need to give [patients] a year’s worth. After 2 to 4 cycles, a significant percentage of patients get a PCR. From lung cancer, we know that those patients have an excellent cancer survival. It’s only a minority of patients in MATTERHORN who received durvalumab plus FLOT who had a PCR, but do we need to give them an adjuvant checkpoint inhibitor for a long period of time? I would venture to say that the answer is going to be no. In that small percentage of patients who do have a PCR after durvalumab plus FLOT in the neoadjuvant setting, they may not need, in my opinion, immune checkpoint inhibitors for several months afterwards.

Mehta: The primary end point for this study was EFS. With EFS vs OS for a trial like this, looking at the locally advanced setting in patients who are resectable, do you think EFS is an acceptable primary end point in lieu of OS? Do you think, ultimately, it's the OS that matters?

Shiraga: Overall, in the purest of scientific evaluation, OS is what we look at. Patients may have other issues, so we have to look at the OS of patients as a cohort. However, it is also important to look at EFS as an important end point to see how a disease affects the patient’s life. Both of them are important, and this is a good paper for both. They both have good, positive outcomes.

Mehta: If a patient has had a PCR, they probably could be spared from adjuvant durvalumab for that 1 year. Do you think in the upper GI setting, as opposed to lung cancer, PCR is a good marker to follow that?

Fontaine: PCR is an excellent, early marker that can tell us that the patient’s going to have a good OS. Patients like to hear that as well, so it gives them encouragement, especially if they need adjuvant therapy or maybe don’t need adjuvant therapy. The gold standard is OS. However, I think EFS is clinically significant because patients who live with their cancer much longer have a poorer quality of life than patients who live maybe not as long, but they're cancer-free for a longer period. EFS is clinically relevant in terms of quality of life for the patients. Does PCR influence my choice? I would say yes.

Mehta: Updated MATTERHORN data was presented at ESMO, looking at the OS in the entire population of durvalumab plus FLOT vs placebo plus FLOT. There’s a Delta difference between the OS at 18 months, 24 months, 36 months, and this is going almost up to 5 years. You still see that separation of the curves and the median OS has not yet been reached, so that is always great. The follow-up period was somewhere around 42 months, and if we remember correctly, the FLOT4 study had shown a median OS of 15 months in the FLOT arm. [The durvalumab plus FLOT OS advantage] was statistically significant, and the P value threshold was less than .04, and the P value was 0.21, with an HR of .78.

Dr. Fontaine, you mentioned PD-L1-negative vs -positive [status]. If durvalumab plus FLOT does become approved, do you think testing for PD-L1 in the locally advanced setting would go into the workflow at your institution, and would you advise the multidisciplinary team to wait for the PD-L1 test results to come back before they decide whether it’s durvalumab plus FLOT vs FLOT for that patient?

Fontaine: The short answer is yes. I would tell them to wait. It’s also important to get the microsatellite stability status, the mismatch repair [MMR] status, and the HER2 status, and it doesn’t take any longer to get the PD-L1 status. In a day where we’re trying to give neoadjuvant therapy, and in a day where it’s a lot more personalized to the tumor, flying blindly without getting good tissue and good genomics is ill-advised. Waiting an extra week would be important. It all has to do with reflex testing; rather than waiting until the patient gets there, it should be a reflex test wherever that biopsy was done. That could be done by better education of referring physicians to request reflex testing in their institution. If you have a good nurse navigator to look at the pathology and say, “Hey, there’s been no PD-L1 or MMR status,” and request it and not wait for the patient to physically be seen before requesting it. Preemptive logistical steps can be made so that PD-L1 reporting is not an obstacle.

Mehta: Next, EFS by pathological response was assessed in terms of whether they had node-positive vs node-negative disease at the time of surgery. Durvalumab plus FLOT does improve EFS even when patients had major pathological responses, although one may arguably say that the curves look good for even those with flawed responses as opposed to patients who did not have a major pathologic response. Durvalumab plus FLOT does offer additional benefit in terms of EFS in both those subgroups. The question that has come up in my mind is, when patients have not had a major pathological response, what did those curves look like?

The EFS curve for the entire population showed that the 24-month EFS was 67% for durvalumab plus FLOT arm vs 59% for the placebo plus FLOT arm. Then, looking at some pathologic data, a PCR was seen in about 16.5% [of patients], and 83.5% had a non-pathologic complete response rate. Major pathologic response was seen in about 26% of patients, and any pathologic response in about 87% patients. Patients who were node negative at the time of surgery were 51.5% and those who remained positive or escalated to being positive were 48.5%, and the modified Ryan score was used for pathologic staging assessment. In terms of PCR, NPR, and pathologic response for EFS, this is the newer data that we saw during ESMO. Patients who had a PCR from durvalumab plus FLOT improved their EFS, as opposed to placebo. Patients with major pathologic response given durvalumab plus FLOT seemed to improve EFS, and patients with any pathologic response given durvalumab plus FLOT seemed to improve EFS. Pathologic staging at the time of surgery was available for all patients. As we saw in about 41.6% patients in the durvalumab plus FLOT arm, patients were still upstaged to be node positive, vs 55% patients in the placebo plus FLOT arm. In patients who are node-negative, durvalumab plus FLOT still offers some benefit.

Dr. Chandana, I’ll pose this question to you because that’s something that we routinely do as medical oncologists. What would be the structure of your conversation with these patients [making this decision]?

Chandana: That’s a real question we get every day. We have to be very objective with the patients and give our subjective thoughts based on the patient scenario. I always present the data. I always tell them that achieving the PCR is a good milestone in their journey. A major pathological response is better than no pathological response. These are meaningful words I use with patients to boost their morale…because this is what our patients want to hear: that we did achieve something. That will give them an opportunity to listen to what happened during that intervention.

Mehta: What do you think are some of the key strengths or some limitations that we learned from the MATTERHORN study that, as clinicians, we should keep in mind?

Shiraga: The key strength of the study is that you include all comers, and it is also a limitation. The strength is that it’s a no-brainer. It would work. That makes the data a little bit muddy, so we must dig through, [to determine] whether it would benefit all subgroups that are included. The study is very promising. Immunotherapy or targeted therapy is a new thing for upper GI cancer. There are more studies coming out showing that these do benefit patients one way or the other.

References

1. Tabernero J, Al-Batran SE, Wainberg ZAA, et al. LBA81 Final overall survival (OS) and the association of pathological outcomes with event-free survival (EFS) in MATTERHORN: A randomised, phase III study of durvalumab (D) plus 5-fluorouracil, leucovorin, oxaliplatin and docetaxel (FLOT) in resectable gastric / gastroesophageal junction (G / GEJ) adenocarcinoma. Annal Oncol. 2025;36(suppl 2):S1738-S1739. doi:10.1016/j.annonc.2025.09.096
2. FDA approves durvalumab for resectable gastric or gastroesophageal junction adenocarcinoma. News release. FDA. November 25, 2025. Accessed December 4, 2025. https://tinyurl.com/mrsfxta8