Effective Detection of MRD May Support Earlier Implementation of ctDNA Testing in CRC

Despite an initial increase in the amount of cell-free DNA (cfDNA) 2 weeks after surgery, high cfDNA levels do not negatively affect the accuracy of minimal residual disease (MRD) detection via circulating tumor DNA (ctDNA) in patients with colorectal cancer (CRC). This finding supports the earlier application of ctDNA testing, and its value as a predictive biomarker, according to Stacey A. Cohen MD.

Postoperative, plasma cfDNA levels and ctDNA MRD-positivity has been assessed in patients with stage I-III CRC in multi-institutional, retrospective analyses.^1,2

In a presentation at the 2022 ESMO Congress, Cohen and colleagues presented an real-world analysis of ctDNA use in patients with stage I-III CRC, which showed that the marker was predictive of recurrence-free survival.1 Among patients with MRD detected via ctDNA within 2 to 8 weeks after surgery, 52.9% (n = 18 of 34) had disease recurrence during follow-up compared with 5.8% (n = 7 of 120) of patients who were MRD negative via ctDNA.¹

Findings from a second real-world data analysis using the same database were presented at the 2023 Gastrointestinal Cancers Symposium.² All patients experienced a decline in cfDNA levels after their initial increase at 0 to 2 weeks. In this timeframe 30.6% of patients were ctDNA positive. ctDNA detection rates were comparable within the standard MRD testing windows of 2 to 6 weeks and 2 to 8 weeks (20.8% and 20.9% respectively).²

Notably, patients in an annotated MRD cohort who displayed ctDNA-positivity 2 to 8 weeks after surgery and subsequent surveillance were more likely to experience poor recurrence-free survival than those who had ctDNA clearance.²

“In both the full cohort and in the annotated cohort, we were able to reliably detect ctDNA after 2 weeks,” said Cohen, who is an associate professor in the Clinical Research Division at Fred Hutchinson Cancer Center and an assistant professor in the Division of Oncology at the University of Washington. “We hope that this will be practice changing, because it’ll allow us to start testing for ctDNA earlier than we might have thought.”

In an interview with OncLive^®, Cohen discussed the advantages and limitations of ctDNA as a biomarker for detecting MRD, real-world data on the relationship between fluctuating cfDNA levels and ctDNA detection, and how implementing ctDNA testing earlier to identify MRD could serve as a prognostic tool in clinical practice.

OncLive^®: What is the importance of detecting MRD and what is the role of ctDNA in its detection in patients with CRC?

Cohen: [Using] MRD means looking for micrometastatic disease that is not detected radiographically, but that we are able to pick it up on a blood-based biomarker sample. [When] we look at MRD [in CRC, we are] extrapolating from work done in hematologic malignancies. Leveraging [MRD gives us] one more tool to figure out if patients have recurring [disease and allows us] to risk stratify, and ultimately give therapy [based] on patient benefit.

The concept of ctDNA is looking for DNA that is unique to the tumor that is being shed and is different from baseline cfDNA. [We can use it to identify] tumor-specific mutations [that indicate] micrometastatic disease circulating in the bloodstream. So far, retrospective series and some observational studies [have] noted that patients who are ctDNA positive tend to do worse, [and] if left untreated patients have a [high] recurrence risk. On the other hand, if we find [ctDNA] very early on after surgery and, for example, we give adjuvant therapy, some patients may clear their ctDNA and have better outcomes. [ctDNA] is a dynamic marker with a short half-life, so it has a lot of benefit [in] both single time point [analysis] and longitudinal testing.

What inspired this retrospective study on the effect of cfDNA levels on MRD detection rates in patients with resected stage I-III colorectal cancer?

The limitation of ctDNA is being able to distinguish what is tumor related vs our baseline cfDNA. After surgery or during chemotherapy, we know that there’s [an] increase in the amount of cfDNA that’s shed in our body. [We] thought that we needed to wait at least 1 month after surgery to reliably detect ctDNA [and avoid] a potential false negative…. We sought to [identify the] time point at which we can reliably start detecting ctDNA, because all our planning and logistics for ctDNA use [requires knowledge of] when we can start testing for it.

Please expand on the study’s design and eligibility criteria.

In our study, we use real-world analysis [of] patients who had commercial, tumor-informed ctDNA testing. [Testing] could have been [performed] at any time, and some were done months after surgery. It was up to the provider when they started to test. We homed in on a group of over 14,000 patients with colon cancer stage I-III. We specifically excluded patients with stage IV [colon cancer] and rectal cancers to [avoid] influencing [the effect of] neoadjuvant therapy on their cfDNA [levels]. [In] the overall cohort, we had a very limited amount of clinical information. [However], we had a subset of 450 patients who were fully annotated. We knew about their surgery and their chemotherapy [exposure], as well as recurrence.

What correlation between cfDNA levels and ctDNA detection was observed in the study?

We looked at cfDNA levels in these patients and noted, as we might expect, that it was higher after surgery. The greatest variability [in cfDNA] was in the first week. It was still somewhat variable in week 2 but was consistent after that. With that [information], we were able to reliably detect ctDNA after 2 weeks. There were some statistically significant differences. Compared with preoperative baseline levels, we did see a statistically significant increase in cfDNA after surgery in those first 2 weeks, which decreased [after] 2 to 4 weeks, and even further [after] 4 to 8 weeks. In our multivariate analysis, [we saw] that [this] didn’t affect our ability to detect ctDNA.

Going forward, how might this research influence the use of ctDNA testing in clinical trials?

Prior studies suggested that we shouldn’t start making decisions [based on ctDNA] until 4 weeks [after surgery], which slows down research. [It also affects] the ability to randomly assign patients, and for them to receive the clinical care they need. For example, we usually try to give adjuvant chemotherapy between 4 to 8 weeks. If we don’t test for ctDNA until 4 weeks, and it takes a few weeks to [obtain the] result, we are probably going to miss that window. [By] recognizing that we can [potentially] start testing as soon as 2 weeks later, we can shave down that time. Hopefully, [we can] give chemotherapy [to] patients in a timely manner and help the patients who don’t need chemotherapy [to] avoid it.

What unanswered questions remain regarding the implementation of ctDNA testing in clinical practice?

This is just the tip of the iceberg, and there are so many things that we need to do. One of the major areas that we need to look at is how to implement ctDNA into clinical practice. We know of [ctDNA] as a prognostic biomarker, but we need to think of it as more [of] a predictive biomarker, in terms of [asking]: Should patients in different clinical scenarios be receiving different chemotherapy regimens? Can we use it to de-escalate patients who are ctDNA negative and may benefit less from chemotherapy? Can we use it as an end point so we can [conduct] clinical trials faster? There are different ways that we can [use this metric]. We hope that understanding the test better will help us to answer those questions in a more accurate fashion.

Which other presentations or data were you excited to see at the meeting?

I [was excited] to hear about some of the ongoing studies that are using integrated ctDNA. Different studies [are] looking to maybe escalate therapy because of ctDNA positivity, and [others] are looking at it for resistance mutations to see if changing therapy faster is beneficial to patients. We recognize that there [are] times when patients develop resistant clone mutations to the therapy, but we don’t know if that means we should then immediately change treatment. There’s still a lot for us to learn about these liquid biopsies, and how [they] influence clinical care. I’m excited to see that it’s no longer a rare topic, because noninvasive biomarkers are hopefully the future for [colon cancer].

Editor’s Note: Dr Cohen reported serving as a consultant or advisor for the Association of Community Cancer Centers (ACCC), Bayer, Istari Oncology, Kallyope, Pfizer, Taiho Oncology; she received research funding from Boston Biomedical (Inst), EMD Serono/Merck (Inst), Isofol Medical (Inst), Natera, Pfizer (Inst), Polaris (Inst); she provided expert testimony for Helsell Fetterman, Robins Kaplan.

References

1. Cohen SA, Kasi PM, Aushev VN, et al. Real-world monitoring of circulating tumor DNA reliably predicts cancer recurrence in patients with resected stages I-III colorectal cancer. Ann Oncol. 2022;33(suppl 7):S683-S684. doi:10.1016/j.annonc.2022.07.457
2. Cohen SA, Kasi PM, Aushev VN, et al. Kinetics of postoperative circulating cell-free DNA and impact on minimal residual disease detection rates in patients with resected stage I-III colorectal cancer. J Clin Oncol. 2023; 41(suppl 4; abstr 5). doi:10.1200/JCO.2023.41.4_suppl.5