Circulating tumor DNA (ctDNA) and mismatch repair (MMR) test results are crucial to inform treatment decision-making for patients with colorectal cancer (CRC), and continued use of these tests may further elucidate the roles of adjuvant chemotherapy and neoadjuvant immunotherapy, according to Benjamin Adam Weinberg, MD.

Updated findings from the CIRCULATE-Japan GALAXY trial (UMIN000039205) presented at the 

showed that post-surgical ctDNA positivity in the minimal residual disease (MRD) surveillance window, which translated to MRD-positive disease (n = 336), correlated with a higher risk of disease recurrence after curative-intent surgery in patients with clinical stage II to IV CRC vs those with MRD-negative disease (n = 1294; HR, 20.50; 95% CI, 11.16-37.66; 

 Furthermore, the 24-month overall survival (OS) rate was 83.70% (95% CI, 77.80%-88.10%) among patients with MRD-positive disease vs 99.50% (95% CI, 98.80%-99.80%) among those with MRD-negative disease. Disparities in OS and disease-free survival (DFS) between the MRD-positive and -negative populations became increasingly pronounced in the months after surgery.

Moreover, findings from the phase 2 NICHE-2 trial (NCT03026140) demonstrated that among 111 efficacy-evaluable patients with nonmetastatic, locally advanced, treatment-naive, MMR-deficient (dMMR) CRC who received neoadjuvant nivolumab (Opdivo) plus ipilimumab (Yervoy), 98% (95% CI, 94%-100%) achieved a pathological response, including major pathological responses (95%) and pathologic complete responses (pCRs; 68%).

“dMMR testing is here to stay, and ctDNA testing is powerful,” Weinberg said in an interview with 

. “Knowledge is power, and [ctDNA testing is] another tool in our toolbox and another data point that needs to be considered.”

In the interview, Weinberg, an associate professor of medicine at the Georgetown Lombardi Cancer Center in Washington, DC, highlighted how findings from CIRCULATE-Japan GALAXY show the prognostic value of adjuvant ctDNA testing in CRC, how long-term data from NICHE-2 support the continued use of neoadjuvant immune checkpoint inhibition in dMMR CRC, and unanswered questions that these trial findings have brought to the surface in the CRC research field.

: Updated findings from the CIRCULATE-Japan GALAXY trial were presented at the 2024 ESMO Congress. What are your thoughts on the data and the ability of ctDNA-based MRD detection to predict adjuvant chemotherapy benefit in patients with resected CRC?

: ctDNA MRD testing is prognostic for patients with resected CRC. With longer-term follow-up, we’re getting a more enriched dataset where we can make more conclusions as we await the results of prospective randomized studies assessing whether we can use ctDNA in the adjuvant setting to make clinical decisions regarding which patients should receive adjuvant chemotherapy. The Japanese [investigational team used] the Signatera [assay], a tumor-informed ctDNA platform. [This large, prospective, observational trial] enrolled 5996 patients with different stages of resected CRC. Stage for stage, this study showed that ctDNA is prognostic.

The focus [of this trial was] on the window period of MRD positivity or negativity at the 4-week post-operative time point, which is typically prior to [the start] of adjuvant chemotherapy. [If a patient has] positive [ctDNA results] at that time point, [their prognosis is worse], but that group seems to derive the most benefit from adjuvant chemotherapy. With longer follow-up, we’re now seeing a third group of [patients with] transient ctDNA clearance. Approximately one-third of patients with ctDNA-positive results clear [their ctDNA test] and stay sustained. They do well, and virtually none recur.

Slightly less than one-third of patients have transient clearance, whereby on chemotherapy, their ctDNA levels clear, but then recur or become positive early, usually by 6 to 12 months. By 2 years, [most patients] in that population have recurred. Then, there’s a group [of patients with ctDNA levels] that do not clear at all, approximately one-third of patients. They have by far the worst DFS and OS outcomes.

We’ve gotten a lot more data [showing that ctDNA] is a worthwhile test. However, we await the results of prospective, randomized studies investigating treating these populations differently [from each other], knowing that the ctDNA-positive group derives the most significant benefit from adjuvant chemotherapy, and the ctDNA-negative group derives little to no benefit, depending on stage. Hopefully, we can use this test in the clinic to decide which patients need adjuvant chemotherapy.

What prospective, randomized trials are ongoing to confirm the benefit of ctDNA testing in patients with CRC?

The [phase 3] CIRCULATE-NORTH AMERICA trial [NCT05174169] is a large trial being conducted throughout the United States and Canada. It’s enrolling patients with stage III or stage II/ctDNA-positive disease. Additionally, in 1 side of the trial, patients with negative or undetectable ctDNA are randomly assigned to undergo observation with serial ctDNA testing vs the standard 3 months of adjuvant doublet chemotherapy with CAPOX [capecitabine and oxaliplatin] or FOLFOX [folinic acid, fluorouracil, and oxaliplatin]. If these patients convert to ctDNA positivity, there’s a second randomization to 6 months of CAPOX or FOLFOX vs 6 months of FOLFIRINOX [folinic acid, fluorouracil, oxaliplatin, and irinotecan], adding irinotecan in an intensified chemotherapy strategy, because we know the ctDNA-positive group is at high risk of recurrence.

This is a nice study because on the one hand, [it is asking]: Can we get away with observing, especially in the [patients with] low-risk, stage III, T1, N1 [disease] who [derive little] benefit from chemotherapy? Most of those patients are probably cured by surgery, and we’re probably overtreating them with adjuvant chemotherapy. On the other hand, if a patient is ctDNA positive, regardless of stage, should we be intensifying chemotherapy?

This trial has been open, but it’s going to take years to enroll and read out. As we’ve seen with the CIRCULATE-Japan GALAXY data, with longer-term follow-up, we’re getting a lot more useful information. The hope is that CIRCULATE-NORTH AMERICA will allow us to potentially observe patients who may not derive benefit from adjuvant chemotherapy.

What is the value of serial ctDNA testing and the importance of sustained vs transient ctDNA clearance?

The real power of ctDNA testing comes in repeating the testing serially in a longitudinal fashion. One negative or undetectable test is good. However, there are patients who have residual disease that is just below the level of detection of the test. Repeating the test at various time points adds to the strength and reliability of the test. I can tell patients [that they if have] been ctDNA negative not just at 1 time point, but at 3 or 4 time points over a year, their odds of recurrence drop and are eventually close to 0.

As this technology was first being used, we do not know what we’re doing with these data, but we know that if [a patient has] persistently undetectable ctDNA, that is good. Historically, we still surveilled these patients for up to 5 years following completion of therapy. That’s a long time, and [patients developed] a lot of scan anxiety and now, a lot of ctDNA anxiety, related to those results. However, persistently undetectable [ctDNA results are] good, and fortunately, we see a lot of patients who are persistently undetectable. We’re comfortable [saying], especially years out, that their odds of recurrence are basically 0.

How did the 3-year DFS data from NICHE-2 add to the current knowledge on the benefits of checkpoint inhibitors in patients with dMMR CRC?

Patients with dMMR rectal and colon cancer benefit from immunotherapy, especially in the neoadjuvant setting. This has changed the paradigm of how we treat these patients. It’s imperative to know MMR status at diagnosis.

In rectal cancer, this is even more paramount because [findings with] single-agent PD-1 inhibition with dostarlimab-gxly [Jemperli] have shown that these patients do not need chemoradiation or surgery. They can get by with just immunotherapy. That has improved the outcomes and quality of life for those patients.

Then, the question was: does this [benefit with neoadjuvant immunotherapy] also translate into colon cancer, where typically the surgery and long-term adverse effects are better than in rectal cancer, where [patients often receive] radiation and surgery in the pelvis? [NICHE-2 showed that] by giving 2 doses of nivolumab and 1 dose of ipilimumab, almost all patients had a pCR, where there was no tumor left at the time of surgery. Some patients had residual disease, but most had some degree of tumor shrinkage, which is impressive. We see these impressive [findings] where each patient has almost a CR, if not a CR.

Where the DFS data add to this and are compelling, is that none of these patients went on to recur, even the ones who didn’t get a pCR. We’re now 3 years out, and none of these patients have occurred, which is impressive and shows that this is the way we should be treating these patients. There’s no benefit from adjuvant therapy, especially adjuvant chemotherapy [in this population], and most of these patients did not go on to receive that.

The traditional paradigm was surgery first followed by adjuvant chemotherapy, so we’re sparing these patients chemotherapy. The question is: do these patients even need surgery? That’s more complicated because [colon] surgery is easier than rectal surgery, but at the same time, would patients need more than 2 doses of nivolumab and 1 dose of ipilimumab?

Anecdotally, in my practice, I’ve treated a couple of these patients who did not get surgery; we’ve treated them for up to 2 years with immunotherapy and then observed them. However, the [neoadjuvant] standard nowadays is 2 doses of nivolumab and 1 dose of ipilimumab. Some patients develop toxicity even from 2 doses of nivolumab and 1 dose of ipilimumab. Some patients need to receive long-term hormone supplementation for low thyroid or adrenal function. That needs to be factored in.

However, there is a more complicated question. Should these patients go straight to surgery? Should they not get surgery at all? Those are answers we still need to sort out.

Without definitive level 1 data to guide neoadjuvant therapy and surgical decision-making, how are patients with dMMR CRC currently treated in these settings?

This is a case-by-case basis because these patients often present differently, with different symptomatology and different goals and objectives about surgery, for example. We deal with this frequently now. The default would be to start doublet immunotherapy and then see what happens.

One of the reasons I use ctDNA testing frequently in the dMMR population is because [these patients’ ctDNA levels] should decrease and often clear as early as after 1 dose of immunotherapy. [ctDNA is] a confirmatory test that we’re on the right track. If these levels remain at 0, do we need to do surgery? Some of these patients are symptomatic and receive surgery one way or another. However, for other patients, whose symptoms resolve on immunotherapy, do they need surgery, and, if not, how much immunotherapy do they need? We don’t know the answer to that question.

Extrapolating from the metastatic realm, after 2 years—some oncologists would say 1 year—we don’t think there’s a benefit from adjuvant or definitive immunotherapy. We often see how it goes and if toxicities develop, because some patients develop no toxicity at all, even from doublet immune checkpoint inhibition. I give ipilimumab at 1 mg/kg every 6 weeks, which seems to be tolerable, and then I give nivolumab every 2 weeks. With that strategy, in a small number of patients, I’ve had success in sometimes avoiding surgery and significant toxicity.

Do you believe a trial will be done to determine whether asymptomatic patients with ctDNA clearance require surgery?

We have a lot of small trials. The problem with this population, even though it can comprise 15% of earlier-stage CRCs, is it’s still not a large population. Now that these drugs are approved and used, we will see more patients treated in the community with immunotherapy and then surgery. [Academic oncologists] need to make sure these data are disseminated throughout the community, so community oncologists are aware of these data and are doing MMR testing, because [MMR testing rates are] not at 100% unfortunately. It would be difficult to do a study. There is the potential to do a study using ctDNA [and enrolling patients with] ctDNA clearance [to undergo] observation vs surgery.

The other issue is financial. If we’re continuing immunotherapy, that’s expensive. A surgery is not inexpensive, but it’s a one-off proposition. Those are the types of higher-level health policy–type arguments that need to be sorted out as well.

What is your main message for your community and academic colleagues based on this discussion?

We should all do MMR testing regardless of stage, to see which patients could benefit from immunotherapy and as a screen to see which patients may have Lynch syndrome, which should be figured out early. It’s not being used universally, for understandable reasons. We don’t know the proper way to use it in the prospective setting, especially for guiding patient treatment.

In my practice, [ctDNA test results] change how I approach patient treatment, even on adjuvant therapy. [Currently], I’m not comfortable observing a patient with stage III CRC who has undetectable ctDNA levels, but [these ctDNA results] might tailor how aggressively we treat patients in the adjuvant setting. In the stage II realm, [ctDNA test results do] factor into how I treat those patients because most of them are overtreated. I hope these data, especially now with the longer-term CIRCULATE-Japan GALAXY follow-up, are available and discussed amongst community oncologists.

Nagata J, Nakamura Y, Watanabe N, et al. Association of ctDNA-based MRD detection and MRD clearance with short-term overall survival in patients with resectable colorectal cancer: updated analysis of CIRCULATE-Japan GALAXY. 

. 2024;35(suppl_2):S428-S481. doi:10.1016/annonc/annonc1588

Chalabi M, Verschoor YL, Tan PB, et al. Neoadjuvant immunotherapy in locally advanced mismatch repair-deficient colon cancer. 

. 2024;390(21):1949-1958. doi:10.1056/NEJMoa2400634

Articles

Benjamin Adam Weinberg, MD, discusses the prognostic value of adjuvant ctDNA testing in CRC and the role of immune checkpoint inhibition in dMMR CRC.

Increased ctDNA and MMR Testing May Alter Adjuvant and Neoadjuvant CRC Management 

Dr Weinberg on the Prognostic Use of ctDNA Testing in Resected CRC

Benjamin Adam Weinberg, MD, associate professor, medicine, Georgetown Lombardi Comprehensive Cancer Center, discusses the ability of circulating tumor DNA (ctDNA)–based minimal residual disease (MRD) detection to predict the benefit of adjuvant chemotherapy in patients with resected colorectal cancer (CRC).

ctDNA-based MRD testing has shown prognostic value for patients with resected CRC, Weinberg begins. With extended follow-up, an enriched dataset is emerging from the CIRCULATE-Japan GALAXY trial (UMIN000039205), aiding in more concrete conclusions regarding the role of ctDNA testing in this patient population, he says. CIRCULATE-Japan GALAXY was a large, prospective, observational trial conducted in Japan that used the Signatera ctDNA platform. This trial enrolled 5996 patients with various stages of CRC and showed that ctDNA status was prognostic across stages, according to Weinberg. 

CIRCULATE-Japan GALAXY focused on ctDNA MRD status at the 4-week post-operative window, prior to the receipt of adjuvant chemotherapy. Patients with positive ctDNA results at this point have a poor prognosis but seem to derive the most benefit from adjuvant chemotherapy, Weinberg explains. In the extended follow-up analysis, 3 distinct patient groups emerged. One group consisted of patients with ctDNA-positive results, one-third of whom cleared ctDNA and maintained clearance, experiencing excellent outcomes with minimal instances of recurrence, he reports. The second group, slightly less than one-third of patients, achieved transient ctDNA clearance with chemotherapy, but their ctDNA levels later became positive, typically within 6 to 12 months of surgery. By 2 years, most patients in this group had disease recurrence, Weinberg says. The final group comprised patients whose ctDNA levels never cleared; approximately one-third of patients fall into this category and have the least favorable disease-free survival and overall survival outcomes, he notes.

A growing body of evidence underscores the value of ctDNA testing as a prognostic tool, Weinberg reiterates. However, results from prospective randomized trials are still needed to refine treatment strategies for these distinct patient populations, he emphasizes. The ctDNA-positive population derives the most substantial benefit from adjuvant chemotherapy, whereas ctDNA-negative patients derive little to no benefit from this treatment approach depending on their disease stage, he notes. The increased clinical use of ctDNA testing may help oncologists tailor adjuvant CRC therapy more effectively, he concludes.

Videos

Benjamin Adam Weinberg, MD, discusses the ability of ctDNA-based MRD detection to predict the benefit of adjuvant chemotherapy in resected CRC.

Making the Most of MRD Assessment, Checkpoint Inhibition in Colorectal Cancer | Supplements and Featured Publications

Increased ctDNA and MMR Testing May Alter Adjuvant and Neoadjuvant CRC Management

Dr Weinberg on the Prognostic Use of ctDNA Testing in Resected CRC