Advances in MRD Testing Help Guide Optimal Colorectal Cancer Therapy

Benjamin A. Weinberg, MD

Although early-stage colorectal cancer (CRC) is traditionally treated with curative- intent surgery and adjuvant therapy, disease recurrence is common and there remains an unmet need for better approaches to inform treatment selection and identify patients likely to experience relapsed or refractory disease.¹ Investigators are hopeful that circulating tumor DNA (ctDNA) assessment may represent a new frontier as a component of better guided therapy for patients with all stages of CRC.

“It’s important to clarify what test we’re talking about because we often use ctDNA to mean multiple things,” Benjamin A. Weinberg, MD, said. “We’re finally at the point where this technology has evolved, and we can use it in the cancers we treat. When we say things like liquid biopsy, we may be referring to broad-based next-generation sequencing of blood, such as the Guardant360 assay, where we’re doing what we would do in tissue but we’re looking in the blood—the theory being that these tumors secrete abnormal DNA that can be picked up in the bloodstream. Now we’re at the point where we can use that technology to really try to assess MRD [minimal residual disease]. We know that for patients with stage II, III, and even stage IV CRC after a definitive, potentially curative surgery, we’re probably overtreating a lot of people with things like systemic chemotherapy [that] we know have a survival benefit on a population level but may not on an individual level. The hope is that this is really the ‘secret sauce’ to help us decide who needs treatment and who doesn’t.”

During a recent OncLive Peer Exchange^®, a panel of expert clinicians in CRC discussed studies seeking to further define the role of MRD testing in CRC that were presented during the 2024 American Society of Clinical Oncology Gastrointestinal Cancers Symposium (ASCO GI 2024), held in January in San Francisco, California. The panelists also previewed the future of MRD testing in CRC. “A lot of data have emerged in this space and with that has also come a lot of confusion. [We are] balancing the excitement of this technology and what it may be telling us with confusion about how sure we are with the result,” John L. Marshall, MD, commented. “How much can we drive a decision one way or the other?”

Updates and Advancements for MRD Testing in CRC

In the observational BESPOKE CRC study (NCT04264702), the authors evaluated the utility of a tumor-informed, personalized ctDNA assay in informing the use of adjuvant chemotherapy in patients with stage II or III CRC. The multicenter study enrolled patients in an MRD cohort (n = 623) and a surveillance cohort (n = 655). The MRD window was 2 to 12 weeks before the start of adjuvant chemotherapy and the surveillance window was more than 2 weeks post adjuvant chemotherapy or more than 12 weeks post surgery if patients were on observation.²

“At ASCO GI this year, we had the opportunity of presenting the first look at the BESPOKE study, which was one of the largest prospective studies looking at the utility of ctDNA in the MRD setting,” Pashtoon Murtaza Kasi, MD, MS, explained. “One of the main things that was a benchmark that this study provided was estimates of the percentage or proportion of individuals who would be MRD positive [by ctDNA]. This was a custom BESPOKE assay that looked at 16 total aberrations and if a patient had 2 or more that were positive, then they were considered [to have] tested positive.”

Findings from an interim analysis of the study demonstrated that ctDNA positivity at the MRD time point was predictive of inferior disease-free survival (DFS). The median DFS among patients who were MRD negative (n = 530) was not reached (NR) compared with 15.98 months (95% CI, 13.77-20.22) among patients who were MRD positive (n = 93; HR, 12.1; 95% CI, 8.0-18.3; P < .0001); the 2-year DFS rates post surgery were 91.59% (95% CI, 88.38%-93.94%) vs 29.86% (95% CI, 13.26%-48.54%), respectively. Additionally, ctDNA positivity was associated with inferior DFS in patients with stage II (HR, 18.8; 95% CI, 8.9-39.5; P < .0001) and stage III (HR, 9.9; 95% CI, 5.9-16.7; P < .0001) disease.

Moreover, a benefit from adjuvant chemotherapy vs surveillance was observed among patients who were MRD positive (HR, 3.06; 95% CI, 1.43-6.56; P = .0025), but not among those who were MRD negative (HR, 1.47; 95% CI, 0.78-2.78; P = .2316). Notably, ctDNA positivity during surveillance was predictive of inferior DFS regardless of adjuvant therapy; patients who received adjuvant chemotherapy (HR, 59.98; 95% CI, 27.3-131.9; P < .0001) as well as those on observation (HR, 80.10; 95% CI, 30.0-207.0; P < .0001) were both much more likely to experience inferior DFS if they were ctDNA positive.

“We’re not ready to just not give chemotherapy to somebody who is ctDNA negative, but this provocative finding that if somebody is ctDNA negative whether they get chemotherapy or don’t get chemotherapy and they’re getting almost the same end result helps at least give rationale to ongoing studies that are asking the question of [whether] no chemotherapy or de-escalation of chemotherapy [is better] in patients who are ctDNA negative,” Kasi said. “[Additionally],the study for the first time provided some important results regarding oligometastatic [disease]. In about 40% of the times when it did occur, you could do some sort of local or regional therapy. We know in our experience that even [for] patients with stage IV CRC, we can cure [some] who have metastases as long as [they are] in one or a few places. Again, this is more of a single-arm look, but at least it provides more than just a few anecdotes of capturing metastases early in mostly patients who are cancer free. ctDNA can [potentially] lead to a higher number of individuals who might be candidates for oligometastatic-directed therapy.”

Along with the clinical findings from BESPOKE, Kasi and coauthors presented patient-reported outcomes from the study. Most patients reported that they felt ctDNA results reduced anxiety about cancer recurrence (73%), felt they were receiving the proper treatment after receiving their results (87%), would continue using the medium to monitor their cancer (92%), and that they valued the additional information provided (96%; Figure 1).

Figure 1

“The patient-reported outcome piece that we also presented at ASCO GI was something that is equally [as important as], if not more important than, the main study results,” Kasi noted. “There [are many] myths about what patients perceive regarding ctDNA testing. The general apprehension, even from centers that do not yet do ctDNA testing, is that ctDNA testing doesn’t help everybody but it’s also not true that it doesn’t help anybody at all. This for the first time looked at dimensions of well-being from a patient standpoint and was overall very positive.”

Also during ASCO GI, investigators presented findings from an updated analysis of the GALAXY study, which was an observational arm of the ongoing CIRCULATE-Japan study (UMIN000039205). GALAXY examined patients with stage II to IV radically resected CRC and used a personalized, tumor-informed assay for detection and quantification of ctDNA. Samples were collected at 1, 3, 6, 9, 12, 18, and 24 months following surgery until recurrence; the MRD window was 2 to 10 weeks post surgery and prior to the initiation of any adjuvant therapy. The study enrolled 5781 patients between May 2020 and October 2023, of which 2998 had stage I to IV disease with ctDNA data available after surgery (Figure 2).⁴

Figure 2

“We’re collecting data on thousands of [patients]; over 2000 [patients] were included with stage I to IV [disease],” Stacey Cohen, MD, noted. “The caveat was that everyone needed to have gone through surgery, so we’re really looking at this MRD question. The tricky part is there isn’t really any caveat to what kind of neoadjuvant therapy someone might have gotten and as you’re trying to interpret the long-term data that can make a big difference. Maybe a really sensitive patient who already had neoadjuvant chemotherapy is then ctDNA negative— that outcome is going to be influenced by the fact that they’ve been exposed to chemotherapy vs someone who wasn’t.”

Findings from the 24-month DFS analysis showed that among all patients, those who were ctDNA negative (n = 2491) and those who were ctDNA positive (n = 369) experienced 24-month DFS rates of 85.9% (95% CI, 83.9%-87.7%) vs 28.9% (95% CI, 23.4%- 34.8%), respectively (HR, 10.53; 95% CI, 8.74-12.69; P < .0001). Additionally, in patients with stage II or III disease, ctDNA positivity in the MRD window was again predictive of inferior DFS; patients who were ctDNA negative (n = 1783) experienced a 24-month DFS rate of 89.3% (95% CI, 87.2%-91.1%) compared with 33.5% (95% CI, 26.5%-40.7%) in the ctDNA-positive group (n = 275; HR, 12.05; 95% CI, 9.46-15.34; P < .0001). In the overall population, sustained ctDNA clearance was significantly associated with superior DFS compared with transient clearance (HR, 25.13; 95% CI, 10.57-59.73; P < .0001) and no clearance (HR, 87.08; 95% CI, 36.14-209.84; P < .0001).

“If we continue to check for ctDNA—typically in trials it’s done for 2 years—when patients decrease their levels, when they have a response to chemotherapy, and their ctDNA decreases, that is a good thing, and if a patient remains positive, that is a bad thing,” Cohen explained. “We’re seeing that those persistently positive patients have the worst outcomes, whereas patients who clear and then become positive again don’t do as well but some of them do better than the patients who are persistently positive. We’re starting to not just think of this as a binary MRD time point, yes or no, positive or negative, but now thinking about incorporating this into the surveillance process of patients and starting to think about surveillance from a number of different angles: symptoms, carcinoembryonic antigen [test], CT scans, and now maybe ctDNA.”

Then, study authors presented findings from the phase 2 AGITG DYNAMIC-Rectal study (ACTRN12617001560381), which examined the utility of using ctDNA analysis to inform the use of adjuvant chemotherapy in patients with locally advanced rectal cancer. Patients needed to have an ECOG performance status of 2 or less, have undergone an R0 total mesorectal excision, and have a provision of adequate tumor tissue within 5 weeks following surgery. The primary end point was the proportion of patients who received adjuvant chemotherapy; 3-year RFS was a key secondary end point, with other secondary end points including overall survival and ctDNA dynamics.⁵

Patients needed to have received neoadjuvant chemoradiation. Following enrollment, patients were randomly assigned 2:1 to ctDNA-guided management or standard management per investigator’s decision. A tumor-informed personalized ctDNA assay was utilized in the ctDNA group and a positive result at 4 and/or 7 weeks following surgery led to 4 months of oxaliplatin-based or fluoropyrimidine chemotherapy. Patients who were ctDNA negative received observation in the case of a ypN0 finding or investigator’s choice of treatment in the event of a ypN+ result.

Although the study was forced to end recruitment prematurely due to the COVID-19 pandemic and increasing use of total neoadjuvant therapy, 230 patients were enrolled between July 2018 and November 2021 (Figure 3).⁵ Patients who received ctDNA-informed treatment (n = 155) and those who received standard disease management (n = 75) commenced adjuvant chemotherapy at a rate of 46% vs 77%, respectively (P < .001). The median treatment duration was 15 weeks (IQR, 11-18) vs 14 weeks (IQR, 11-17), respectively, and 80% vs 71% of patients completed planned treatment, respectively.

Figure 3

At a median follow-up of 36 months, patients in the ctDNA-informed treatment group and those in the standard disease management group achieved 24-month recurrence-free survival (RFS) rates of 84% vs 84%, respectively, and 36-month RFS rates of 74% vs 82% (HR, 1.38; 95% CI, 0.76-2.50; P = .28). Additionally, patients who were ctDNA positive (n = 42) and those who were ctDNA negative (n = 108) achieved 24-month RFS rates of 61% vs 89%, respectively, and 36-month RFS rates of 53% vs 83%, respectively (HR, 0.29; 95% CI, 0.15-0.55; P <.001). All patients in the ctDNA–positive subgroup received chemotherapy compared with 23% in the ctDNA–negative subgroup.

Beyond clinical trials, investigators also presented findings from the INTERCEPT program, which was the first prospective real-world study of a ctDNA testing program. INTERCEPT enrolled patients undergoing curative surgery for stage II to IV CRC at The University of Texas MD Anderson Cancer Center in Houston from January through December 2022 with follow-up occurring through December 2023; the postoperative timepoint was defined as any test within 6.5 months of surgery, including adjuvant therapy. The program employed the Signatera MRD assay, and a false negative result was defined as a negative test followed by radiographic detection of recurrence within 4 months.⁶

When the analysis was limited to patients with a postoperative timepoint (n = 443), at a median follow-up of 13.0 months, the median DFS among patients who were ctDNA negative was NR compared with 5.7 months among patients with at least 1 ctDNA-positive test (HR, 12.0; 95% CI, 7.4-19.5; P < .0001); the 12-month DFS rates were 86.8% vs 15.9%, respectively. Patients with postoperative ctDNA were found to have inferior DFS compared with those with no detectable ctDNA post operation across disease stages; patients with stage II disease had 12-month DFS rates of 16.7% (95% CI, 0.8%-51.7%) vs 94.9% (95% CI, 84.5%-98.4%), respectively. Patients with stage III disease had 12-month DFS rates of 34.9% (95% CI, 14.2%-56.6%) vs 93.3% (95% CI, 86.2%-96.8%), respectively, and patients with stage IV disease had 12-month DFS rates of 12.4% (95% CI, 6.1%-21.0%) vs 74.1% (95% CI, 63.8%-81.9%), respectively.

Future Directions for MRD Testing in CRC

Following their discussion of the key data updates presented during ASCO GI, the panelists looked forward to the future of MRD testing in CRC. They noted that a major area of interest is gaining a better understanding of the nuances of what a positive or negative ctDNA test indicates and improving test validation.

“This is not something that’s necessarily black and white—it comes out with a discrete answer, positive and negative, but the implications of what that means [remain] our biggest gap,” Cohen said. “I hope that as these tests evolve, there continues to be a very thoughtful and rigorous manner that we’re validating them, but also an acceptance of the fact that we should not repeat a retrospective series for every patient, followed by a 2000-patient observation trial, followed by prospective studies for each platform that comes out; it just doesn’t make sense. [We need to be] comfortable with the unknown because I don’t want our patients to wait 10 years each time the technology improves. But I also don’t want us to just jump ahead and assume the work was done when maybe it wasn’t.”

Additionally, Weinberg noted that ctDNA testing could be used to inform novel drug development for targeted populations of patients with CRC. He mentioned an ongoing Stand Up To Cancer study examining treatment options for patients who are persistently ctDNA positive and an ongoing clinical trial of atezolizumab (Tecentriq) plus bevacizumab (Avastin) in patients who are persistently ctDNA positive and have completed curative-intent therapy. Eventually, ctDNA clearance could be used much in the same way as molecular profiling to assign a targeted therapy to a given patient, Weinberg said.

“This is a great time for [ctDNA testing] to be incorporated into clinical practice, not just as a prognostic tool but as a predictive tool as well with the results that we’re seeing from GALAXY, BESPOKE, and some of the [other] ongoing trials. ctDNA presents a new opportunity for the pairing of drugs with novel therapeutics. The way I see the utility evolving is that this can give you a quick, early readout, as opposed to adjuvant therapy studies that took years to mature. This opens a unique opportunity of an early readout, a quicker end point. All of this will help accelerate progress for our patients with cancer,” Kasi concluded.

References

1. Hossain MS, Karuniawati H, Jairoun AA, et al. Colorectal cancer: a review of carcinogenesis, global epidemiology, current challenges, risk factors, preventive and treatment strategies. Cancers (Basel). 2022;14(7):1732. doi:10.3390/ cancers14071732
2. Kasi PM, Aushev VN, Ensor J, et al. Circulating tumor DNA (ctDNA) for informing adjuvant chemotherapy (ACT) in stage II/III colorectal cancer (CRC): interim analysis of BESPOKE CRC study. J Clin Oncol. 2024;42(suppl 3):9. doi:10.1200/ JCO.2024.42.3_suppl.9
3. Kasi PM, Rivero S, Aushev VN, et al. Patient-reported outcomes from the BESPOKE CRC study. J Clin Oncol. 2024;42(suppl 3):54. doi:10.1200/JCO.2024.42.3_suppl.54
4. Yukami H, Nakamura Y, Mishima S, et al. Circulating tumor DNA (ctDNA) dynamics in patients with colorectal cancer (CRC) with molecular residual disease: updated analysis from GALAXY study in the CIRCULATE-JAPAN. J Clin Oncol. 2024;42(suppl 3):6. doi:10.1200/JCO.2024.42.3_suppl.6
5. Tie J, Cohen JD, Wang Y, et al. Circulating tumor DNA analysis informing adjuvant chemotherapy in locally advanced rectal cancer: the randomized AGITG DYNAMIC-Rectal study. J Clin Oncol. 2024;42(suppl 3):12. doi:10.1200/JCO.2024.42.3_ suppl.12
6. Maddalena G, Pellatt AJ, Eluri M, et al. INTERCEPT program of circulating tumor DNA (ctDNA) testing for minimal residual disease (MRD) in colorectal cancer (CRC): results from a prospective clinical cohort. J Clin Oncol. 2024;42(suppl 3):27. doi:10.1200/JCO.2024.42.3_suppl.27