ctDNA, Neoadjuvant Immunotherapy Combos Could Help Reshape Management of CRC

The use of circulating tumor DNA (ctDNA) to drive adjuvant chemotherapy decisions for patients with colorectal cancer (CRC) could ultimately affect treatment decision-making in clinical practice, according to Stacey A. Cohen, MD, who added that prospective data from larger trials remain a key hurdle to integrating this strategy into real-world use.

She also highlighted that data for neoadjuvant immunotherapy regimens in patients with mismatch repair–deficient (dMMR)/microsatellite instability–high (MSI-H) CRC underscore the need to determine MSI status at diagnosis before patients are sent to surgery.

“Coming out of all this great published data from 2024…are 2 major trends. One is that we are seeing that ctDNA is an exciting factor to help us [better] understand how to treat patients,” Cohen explained in an interview with OncLive®. “[Second,] the data [for neoadjuvant immunotherapy] are exciting. There is more room to grow, but we are making great progress.”

The potential utility of tumor-informed ctDNA as a prognostic tool for guiding chemotherapy decisions in patients with CRC has been underscored by data from the observational CIRCULATE-Japan GALAXY trial (UMIN000039205). Updated findings from the study presented at the 2024 ESMO Congress showed that ctDNA positivity was associated with significantly worse disease-free survival (DFS) and overall survival (OS), and the findings also demonstrated the potential benefit of administering adjuvant chemotherapy to patients positive for minimal residual disease (MRD).1

In the interview, Cohen discussed the updated findings from CIRCULATE-Japan Galaxy and the need to prospectively evaluate the role of ctDNA as a prognostic marker in CRC. She also provided an overview on the current landscape of neoadjuvant immunotherapy in dMMR/MSI-H, highlighting insights from both the phase 2 NICHE-2 and NICHE-3 trials (NCT03026140) in this patient population.2,3

Cohen is a physician and an associate professor in the Clinical Research Division of Fred Hutchinson Cancer Center, as well as an associate professor in the Division of Hematology and Oncology at the University of Washington School of Medicine in Seattle.

OncLive: What did findings from the subgroup analysis from the CIRCULATE-Japan GALAXY trial reveal about the use of ctDNA to inform adjuvant chemotherapy decisions in patients with resected CRC?

Cohen: It is great that we continue to see updates from the CIRCULATE-Japan GALAXY study. When [findings] were first published and presented, there were approximately 1000 patients [included in the trial]. There are [now] over 2000 patients, and [investigators] aim to get closer to 5000 patients. It is great getting updates along the way because it's a large, impressive set of observational data. It helps to get strength in bigger numbers.

We're seeing some different trends in how to interpret the data, but this remains an observational cohort. Therefore, some of the issues that have been present throughout [the study] continue to be an issue in terms of how we [apply] these data [to clinical practice]. We need some strong, prospective studies, and that is not what CIRCULATE-Japan GALAXY is aiming to do. However, it continues to offer us some great direction on where the most fruitful areas of a prospective clinical trial could come.

As we review some of the updates from GALAXY, ctDNA continues to be a very strong prognostic risk factor, meaning that the detection of ctDNA is associated with much worse survival. We're learning about how we can start to use that [information] to help make clinical decisions. Ideally, you want to use the presence of ctDNA to suggest that someone's at a higher risk for recurrence, then escalate therapy or give therapy when you wouldn't otherwise. [Conversely,] in a patient who's ctDNA negative, you want to be able to step back and perhaps avoid the toxicity of therapy or give a less-toxic therapy.

As we see updates in GALAXY, we're still noticing that [patients with] a variety of stages [of CRC] were included. All patients went to surgical resection; however, this is where some of the heterogeneity comes up. For example, some patients had chemotherapy before surgery—perhaps this was more prevalent in patients with stage IV disease. This was all done at the discretion of the clinician who was treating a given patient, but it offers some heterogeneity in the patient population.

We're seeing improvements in DFS and OS for patients who were ctDNA negative. They showed that [trend] both at the up-front time point called the MRD time point, which was after surgery before any chemotherapy was given, and then also in the surveillance time period, which is a more longitudinal dataset for patients who had completed their chemotherapy and beyond.

Some interesting updates included in the ESMO Congress presentation and the Nature Medicine publication were that we continued to see that liver metastases are higher-shedding sites of metastasis. Therefore, a patient who had a radiographic recurrence in the liver was more likely to have positive ctDNA, whereas [ctDNA] still can be detected in some patients with lung and the peritoneum [recurrences], but it may occur a little bit less often. We know that approximately 5% of patients who are ctDNA negative will still have recurrences. It may be that these [recurrences] are related to lower-volume sites of disease, or that [patients had metastatic] sites in organs that don't tend to shed as much ctDNA.

There continue to be some interesting updates. [Investigators] showed data from different molecular subgroups that mimicked patterns that we've seen before in terms of molecular markers that are prognostically better or worse.

My takeaway from GALAXY is that we're continuing to learn how to treat patients [with CRC] by using ctDNA. None of these data are going to truly be practice-changing because it's not a prospective study integrating ctDNA, by which we change our management because of the ctDNA result. However, this continues to add to the breadth of data that encourages us to continue this line of research.

How could prospective data help inform the use of ctDNA for treatment decisions in clinical practice?

I'm hoping that as we have some strong prospective trials, we'll be able to start using ctDNA. The GALAXY study is great because it asks and addresses [questions about] so many different clinical situations. However, we need [prospective data for] a very specific situation, such as stage II colon cancer where we debate whether to give any chemotherapy, where we could use that test to help guide that management.

Beyond just using [results from the ctDNA] test, [we also want to identify] a specific chemotherapy regimen. For example, when we look at the [phase 2] DYNAMIC trial [ACTRN12615000381583] out of Australia, it is an amazing clinical trial, but it [used] an assay that's not commercially available, so we can't necessarily replicate that in our own practice. There was also a lot of physician choice about which chemotherapy regimen to use.

We need to [focus] on a very specific question with a [prospective] ctDNA trial [in CRC], and if that's found to be positive, we can go from there and start to use [ctDNA] in more settings. [There are] more data to come, and even within the CIRCULATE-Japan cohort of studies, there will be [prospective] trials that will come out soon. That's what we're lacking at this point.

How could disease management differ for a patient with sustained vs transient ctDNA clearance?

The hope is that we’ll need to modify our treatment strategies as well as our surveillance strategies. Clearance implies that someone has negative ctDNA. The question is, ‘Is that marker enough to say that a patient [with negative ctDNA] doesn’t need as frequent visits, lab checks, or radiographic studies?’ At this point, we have not scaled back in any way in terms of our surveillance. The idea, more hypothetically, could be that if someone was ctDNA negative for a given number of checks and you were able to show [with prospective data] that recurrences didn't happen [in these types of patients] outside of symptoms or rare instances, maybe you would say that the benefit of doing radiographic annual imaging is not there. You would then potentially not screen them or do as much surveillance. I don't think anybody's willing to go down that pathway just yet because we worry about every lab test having an error and the potential for that.

The idea of ctDNA clearance also gets to the fact that longitudinal testing matters, and it matters how patients respond over time. We think that risk groups determined right after surgery matter based on pathology and MRD testing, but we also need to incorporate their response to subsequent therapy. For example, if a patient got adjuvant treatment and cleared their ctDNA, we know that those patients do better than patients who remain ctDNA positive, and patients who briefly clear their ctDNA do better than those who never clear their ctDNA; that's the idea of transient clearance vs no clearance. However, we haven't yet [established how to use this information in practice].

These [data] make some nice survival curves, but they don’t help us totally answer the question regarding what to do next. We may find that patients who are ctDNA positive, remain positive, or convert back to being positive are the ones in whom we would recommend for closer radiographic surveillance compared with the patients who are ctDNA negative and may continue to be negative. Maybe those are the patients who [are going to require] less-intense surveillance.

Looking at findings from the NICHE-2 trial, did it surprise you that no patients with dMMR CRC treated with neoadjuvant nivolumab (Opdivo) plus ipilimumab (Yervoy) experienced disease recurrence at 3 years, irrespective of whether they achieved a pathological complete response (pCR)?

This is an exciting time to be using immunotherapy in CRC. We know that [patients with dMMR tumors represent] only a subset of patients with localized CRC or rectal cancer—we are talking about approximately 10% to 15% of patients. However, we're seeing some exciting data that's changing how we treat patients.

In rectal cancer, there have been interesting studies, notably a phase 2 study [NCT04165772] of dostarlimab-gxly [Jemperli] out of Memorial Sloan Kettering Cancer Center in New York, New York, showing that we can achieve a CR with immunotherapy alone in those patients.4 If we look at the data, they saw that [the median time to CR was 6.22 months], and only in a few patients [experienced a CR] by 3 months of treatment. We know from other studies—for example, the phase 3 KEYNOTE-177 study [NCT02563002] that compared [pembrolizumab (Keytruda) alone] vs chemotherapy in first-line metastatic dMMR/MSI-H CRC—that the time to response [with immunotherapy] was similar.

Therefore, when we think about the NICHE-2 data, these patients received only 1 month of [neoadjuvant] immunotherapy, and we saw very impressive responses. All patients went to surgery, and they saw very impressive pathologic responses. You may ask the question, ‘Are patients going to still do as well if they don't achieve a pCR?’ My takeaway from these data is we might not have waited long enough. Four weeks is probably insufficient to be able to convert all patients to a pCR if we look at the data in rectal cancer and CRC. What we're seeing [with the NICHE-2 data] is that many patients had a good response, and they were having regression of the tumor that didn't completely get to a pCR.

We also have to remember that all patients were mandated to have surgery in this trial. The next question becomes, ‘Are we ever going to get to a point where we could [use] non-operative management?’ That's much trickier because we don't have the same abilities to do surveillance [in CRC]. A colonoscopy is more complicated, and we don't have MRI.

The data from NICHE-2 are encouraging with this type of DFS, [and this study also included] patients with high-risk cT4 disease. We should be holding surgery until we know the MSI status of patients; even though [the dMMR/MSI-H population] only [accounts for] 10% to 15% [of the overall CRC population], knowing this status can make dramatic improvements in outcomes. We don't have the ability right now to give immunotherapy in the adjuvant setting, so it makes it that much more important to recognize [this status] early on.

Findings from NICHE-3 also showed that neoadjuvant nivolumab plus relatlimab-rmbw (Opdualag) generated pCRs in patients with dMMR colon cancer. How do you weigh these findings with the potential for non-operative management?

NICHE-3 evaluated nivolumab plus relatlimab, the latter of which is a LAG-3 inhibitor, as opposed to ipilimumab, which is a CTLA-4 inhibitor. We know from melanoma that ipilimumab may bring higher toxicity, although with the current doses, it may be more similar to what we see for nivolumab plus relatlimab. Therefore, we would expect that their outcomes would be pretty similar. However, [the combinations] haven't necessarily been tested head-to-head.

It's very exciting to see data coming out for nivolumab plus relatlimab. To me, we’re asking multiple questions in parallel. First, we want to know if we should be doing immunotherapy for these patients [with dMMR/MSI-H disease], and the data are suggesting that we should in the neoadjuvant setting. The next question would be, what's the optimal regimen? We don't have that answer yet because the NICHE-2 and NICHE-3 regimens have not been compared.

We [also have to consider if there are] certain toxicities to worry about. Prior studies have shown concern for grade 5 adverse effects or deaths related to therapy, such as myocarditis with the use of combination immunotherapy and even with single-agent immunotherapy. As we start to get excited about these drugs, we must remember that there are rare but very serious toxicities that can happen. Certainly, we know that patients may die of their disease or have poor outcomes; however, the deaths that have been observed have given everyone a moment of pause, although we continue to move forward.

As we're talking, I don't know that non-operative management is here yet, and we need more trial data to show [if that is a feasible option]. [One reason] is the inability to do adequate surveillance. We want to make sure that if patients don't go to surgery, we have good oversight so you can catch that patient if they start to have any evidence of progression. That's complicated because a colonoscopy is not an easy ask, and a patient’s ability to comply with those types of procedures may be less [likely] if we need to do it more often.

[These data] give us some information for planning future trials. My takeaway from NICHE-3 is that [nivolumab plus relatimab] remains an exciting combination to use, and I look forward to what the next [trial] iteration is going to be. Ideally, at some point, it'll be a head-to-head study.

Given all the ongoing research and emerging data surrounding ctDNA and immunotherapy, how do you see these ultimately affecting the management of CRC in clinical practice?

The presence of ctDNA suggests poor prognosis, which [could identify] patients who may need more aggressive therapies. However, questions remain on how to apply that to an individual patient. I encourage any studies that evaluate [ctDNA] because the [lack of] prospective data is preventing us from making [the use of ctDNA] more standard.

In the CRC world, we absolutely need to know the MSI status before we take someone to surgery. This is because we are seeing much more impressive outcomes with neoadjuvant immunotherapy, even in patients who don't have a PCR. However, a little bit of immunotherapy can go a long way, and we'd love to improve our survival outcomes for patients.

References

1. Kataoka K, Mori K, Nakamura Y, et al. Survival benefit of adjuvant chemotherapy based on molecular residual disease detection in resected colorectal liver metastases: Subgroup analysis from CIRCULATE-Japan GALAXY. Ann Oncol. 2024;35(suppl 2):S459-S460. doi:10.1016/j.annonc.2024.08.627
2. Chalabi M, van den Dungen LDW, Verschoor YL, et al. Neoadjuvant immunotherapy in locally advanced MMR-deficient colon cancer: 3-year disease-free survival from NICHE-2. Ann Oncol. 2024;35(suppl 2):S1217-S1218. doi:10.1016/j.annonc.2024.08.2263
3. De Gooyer P, Verschoor Y, van den Dungen LDW, et al. Neoadjuvant nivolumab plus relatlimab in MMR-deficient colon cancer: results of the NICHE-3 study. Ann Oncol. 2024;35(suppl 2):S428-S429. doi:10.1016/j.annonc.2024.08.572
4. Cercek A, Sinopoli JC, Shia J, et al. Durable complete responses to PD-1 blockade alone in mismatch repair deficient locally advanced rectal cancer. J Clin Oncol. 2024;42(suppl 17):LBA3512. doi:10.1200/JCO.2024.42.17_suppl.LBA3512