The Case for Treatment Holidays: ctDNA in Metastatic Gastroesophageal Cancers

Posttreatment surveillance remains one of the most challenging aspects of managing gastroesophageal cancers, where recurrences are common, often aggressive, and not always readily apparent on standard imaging. Despite intensive follow-up schedules in the first 2 years after surgery, clinicians still face significant limitations in detecting disease relapse early.

Against this backdrop, interest has grown in circulating tumor DNA (ctDNA)-based minimal residual disease (MRD) testing as a potential supplement to conventional surveillance strategies. Building on emerging data in the locally advanced setting, an observational study recently published in Cancer, led by Rutika Mehta, MD, MPH, sought to better understand the burgeoning role of ctDNA in the metastatic space where patients are living longer and questions around treatment duration, toxicity, and treatment breaks or “holidays” are increasingly relevant. The study found that ctDNA clearance and deep declines (>90%) during treatment were associated with improved progression-free survival [PFS], highlighting the prognostic potential of such strategies.¹

In an interview with Targeted Oncology, Dr Mehta, associate professor of medicine and attending physician at Weill Cornell Medicine and NewYork-Presbyterian Hospital, detailed the clinical gaps and experiences that prompted the study, outlined the evidence still needed before such strategies are further integrated into routine care, and explored the feasibility of ctDNA/MRD testing in practice.

Targeted Oncology: What are some of the current challenges with posttreatment monitoring techniques in gastroesophageal cancer that prompted your study?

Rutika Mehta, MD, MPH: There was a paper published in JCO Precision Oncology a couple of years ago by Dr [Brandon] Huffman [Dana-Farber Cancer Institute] and colleagues looking at the role of ctDNA or MRD testing in the locally advanced setting.² I think, despite our best efforts when we use blood work alone, plus/minus CT imaging, there still remain challenges as to diagnosing patients earlier on with recurrences. Our surveillance periods are typically quite strict in the first 2 years post-surgery, and they become a little more relaxed year 3 onwards.

The reason we keep those strict guidelines within that first 2 years, which means routine follow-ups every 3 months and scans every 6 months for that first 2 years, is because of the increased risk for recurrence in that first 2-year period. With gastroesophageal cancers, we typically see that these recurrences with the diffuse-type gastric cancers might occur in the peritoneum, and these are much harder to detect on imaging. Patients will start to clinically show symptoms such as early satiety, maybe even abdominal distension, nausea, weight loss, but unless something more substantial shows up on imaging…it's really hard to diagnose a recurrence in these patients. Therefore, when Dr Huffman and his colleagues showed their experience—this is… a multi-institutional study—I think there is value in doing MRD testing alongside tumor markers such as CEA as well as CT testing.

By no means is MRD testing is meant to replace what we do in standard practice at this time, but it's more of an adjunct to what we already have in place. And I think they were able to beautifully show that it does behave as a great adjunct tool, and that could prompt for maybe earlier imaging techniques or maybe probing more into some of the clinical symptoms patients demonstrate that could help identify a recurrence sooner, more in the oligometastatic state, rather than when they become more floridly metastatic.

How does your study build on previous literature?

Since we had seen that it was behaving as a great adjunct tool in the locally advanced setting, we wanted to understand, could we really employ MRD testing in the metastatic space? And I say this because, as we are improving our therapeutics, patients are living longer and longer, which is great for our patients. But at the same time, we always wonder, if your scans are looking clean, there's no evidence of disease, how long should we really continue systemic therapy, such as toxic chemotherapies? And patients are constantly asking rightful questions: can they get a treatment break, a treatment holiday?

Based on some of the clinical experiences that I had with my patients, I had started utilizing MRD testing in that space just to be able to provide an additional tool for me to be absolutely sure that the treatments have worked. The treatments have actually led to quite a deep response in terms of their metastatic burden, and therefore, I became more and more comfortable offering my patients treatment holidays [or] breaks.

When I started seeing those trends with my patients, that prompted me to…ask my other colleagues across other institutions. To no surprise, a lot of my colleagues were having a very similar approach to their patients with metastatic disease as well. And therefore, we decided to have a collective platform for gathering all this data for metastatic patients [to] understand, number one, how we are really utilizing MRD testing in the metastatic space; and number two, does that really correlate with some of the responses that we see, and overall, does that really prognosticate some of the [PFS] for these patients?

What research or evidence is still needed for greater integration of ctDNA testing strategies into clinical practice?

I think there's, beyond doubt…a scope for continuing to use MRD testing in the locally advanced setting, but I think in the metastatic setting, with the current trials that we're performing, I think there would be value in understanding what the role of MRD testing is in relationship to those new therapeutic drugs.

One thing from our study I can point out [is] that…this is a retrospective study. This was by no means a prospective collection of samples during the course of treatment. These were done at different time points, and therefore, I think overall, our study population became small. But I think the key findings that we saw was yes, if you are positive and become negative in terms of your MRD—which means you clear your MRD status—then that is a better prognosis, as opposed to those that continue to remain positive over time during treatment.

And then the second point I would point out would be when we're looking at those absolute values on an MRD test, a more than 90% decline in the ctDNA values does also offer a much better prognosis, as those opposed to no change or increase in the ctDNA value.¹ I think knowing this, implementing this in a prospective trial would be much more beneficial for us to be then helping to adopt this into routine practice.

What practical considerations should clinicians keep in mind when deciding to use a ctDNA/MRD test?

Number one, I think more tests give us more information, and sometimes we wonder how to handle th[at] additional information. Again, I usually have a discussion with my patients about getting these tests done. Not everybody is savvy about interpreting these test results and to still be comfortable knowing that these results are there in place, but the treatments are not going to significantly change, just because we so far don't have any prospective data to inform our treatments based on those tests alone. So I think that's kind of a risk-benefit discussion that I need to have with every single patient of mine when I'm ordering MRD or ctDNA testing.

One important component, as I mentioned: yes, it's an added test, it does definitely impose some additional financial challenges. Again, insurance companies, payers are so far very supportive in getting these approved for patients. However, we have to understand that… there could be different payer mix with different practices, and that could pose some additional challenges in terms of approval of these tests.

Last but not least, I would say, because we do these tests, and sometimes we see certain test results that make us feel uncomfortable and concerned that the patients might be progressing, we might be doing more additional tests. For example, if the patient is not scheduled for CT another 3 months, but I see a ctDNA trend rising, I might be prompted to do a CT scan in 6 weeks. Am I going to see a progression in those 6 weeks? I'm not sure. So, the correct lead time is still an unknown factor with MRD testing, and when you really see radiographical progression on scans. I think those are some key questions that need to be answered through clinical trials, and then I think the next phase would be, once we know exactly what the lead time is on prospective trials, does intervening early on actually change outcomes for our patients? We talked about maybe diagnos[ing] them in the oligometastatic state rather than in a florid metastatic state. But does that really change outcomes for patients in the long run, would be the key question that yet remains to be answered with these tests.

I would say this is a very exciting field for ctDNA and MRD. Definitely, there is a role in continuing to do MRD testing in the locally advanced setting, as we've seen in colorectal cancers and other cancers. But I think there will be value… and [an] emerging role for MRD testing in the metastatic setting as well, especially, I would say [on a] case by case basis. [For] patients who have oligometastatic disease, lymph node metastases, those who are having great response to the systemic therapies, our challenge always is, when is the right time to give patients a treatment break? When can we stop treatment for our patients? I think ctDNA or MRD testing helps answer some of those critical questions, and thereby still maintaining the survival and potentially improving their quality of life by reducing some of those cytotoxic therapy exposures.

REFERENCE

1. Mehta R, Rivero‐Hinojosa S, Dayyani F, et al. Circulating tumor DNA informs clinical practice in patients with recurrent/metastatic gastroesophageal cancers. Cancer. 2026;132(1). doi:10.1002/cncr.70242