ctDNA Helps Predict Recurrence in Real-World Analysis of Patients With Resected CRC

In Partnership With:

Stacey A. Cohen, MD, discusses the results of real-world analysis of patients with resected stage I-III colorectal cancer, how circulating tumor DNA status correlated with recurrence rates, and how these real-world data compared with previous observational studies.

Stacey A. Cohen, MD

Stacey A. Cohen, MD

Circulating tumor DNA (ctDNA) was a reliable predictor of cancer recurrence in patients with resected stage I to III colorectal cancer (CRC), according to Stacey A. Cohen MD. She added that patients with ctDNA positivity experienced shorter recurrence-free survival (RFS).

Real-world data from a retrospective analysis presented during the 2022 ESMO Congress indicated that patients in the annotated minimal residual disease (MRD) cohort with stage III disease (n = 97; 27.8%) had higher MRD detection rates than those with stage I (n = 12; 8.3%) or II (n = 45; 13.3%) disease. The MRD detection rate for all patients in the annotated cohort (n = 154) was 22.1%.

Furthermore, patients who presented with ctDNA positivity following surgery were found to have a shorter RFS. In the MRD window, defined as 2 to 8 weeks following surgery, 5.8% of those with ctDNA negativity (n = 120) experienced disease recurrence vs 52.9% of those with ctDNA positivity (n = 34; HR, 14.1; 95% CI, 5.8-34.0; P <.0001).

Of the patients with ctDNA negativity any time post-surgery (n = 291), 6.5% experienced disease recurrence. Eighty-four (66.7%) patients with ctDNA positivity experienced recurrence (HR, 18.6; 95% CI, 10.9-31.5; P <.0001).

During a follow up period of 6 months following surgery and any adjuvant chemotherapy, 6.6% of patients with ctDNA negativity (n = 213) experienced disease recurrence vs 80.5% of those with ctDNA positivity (n = 41; HR, 20.0; 95% CI, 10.6=37.6; P <.0001).

“[These data] add to what we now know about ctDNA. This seems to be a good prognostic biomarker that is of a lot of interest,” Cohen said in an interview with OncLive®. “The more we learn about it, the better we will be able to help guide patient management in the future.”

Cohen discussed the results of real-world analysis of patients with resected stage I-III CRC, how ctDNA status correlated with recurrence rates, and how these real-world data compared with previous observational studies of ctDNA. She 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.

OncLive®: What is already known about the use of ctDNA in patients with CRC?

Cohen: ctDNA is a newer technology. It’s still emerging in clinical care, especially in solid tumors. A lot of our initial studies were observational, and the initial studies [were not meant] to share the results with the patients. This was when we were just starting to learn about how to use ctDNA.

As we have gained more clinical experience, we want to get that data out there in terms of outcomes, how patients are being treated [in relation to ctDNA] results, and if this is a helpful marker. This [retrospective analysis] adds to the body of literature that we currently have from observational studies.

This particular study was a real-world evidence using the standard clinician-driven monitoring of ctDNA that was being done by some providers for their patients. We are now looking at that data and trying to understand if this real-world experience mirrors what we’re seeing in the observational studies.

How did you select patient data for analysis?

We sampled [data] from across the United States of patients who [received testing via the Signatera] ctDNA assay. We reached out to clinicians and asked for them to provide the clinical data for their patients. This [allowed clinicians to] opt in. They became an investigator on the study as a part of that, and if those providers were interested, they got to contribute their clinical data to create the subset that we were able to analyze.

Could you expand on the results presented at the 2022 ESMO Congress?

We had 16,347 patients in the overall cohort. From that, we had a clinically-annotated subset of 417 patients. From that, we looked specifically at patients who had a blood draw in what we’re calling the MRD time point, which was 2 to 8 weeks after surgery. We had 154 patients who had a blood draw at the MRD time point [who comprised the annotated cohort]. This is key because we believe that at that time point, when someone is [MRD] positive or negative, that is driving some of the prognostication that we get from ctDNA.

We looked at our patient population and 22.1% of the annotated cohort had positive ctDNA. That was similar to what we saw in the overall cohort with the limited clinical information that we had, so we believe [the annotated cohort] is a representative sample. When we looked at that MRD cohort, we looked at clinical features, use of chemotherapy, and RFS.

In the ctDNA-negative population, recurrence has happened in 5.8% of patients. This is consistent with other [observational] studies. We don't believe that being ctDNA negative means that patients are cancer free or that patients are free of potential for recurrences, but [recurrence] is uncommon. Whereas in ctDNA-positive [patients], we saw more common and more frequent recurrences. That was especially true if we restrict our analysis to patients who were [ctDNA] positive at least 6 months after surgery and had completed their adjuvant chemotherapy. If we looked at that [specific ctDNA-positive] patient population, unfortunately, the recurrences were high.

Overall, [these retrospective data] support what we are seeing in observational studies. We see a higher frequency of ctDNA in later-stage populations, such as stage III, and [patients who are] ctDNA positive have a greater chance for recurrence than patients who are ctDNA negative.

What additional insights are you hoping to obtain with further research?

We were interested in looking at the use of chemotherapy and trying to decide if it was not only a prognostic, but also a predictive factor. Eventually, we want to be able to use [ctDNA] to help guide chemotherapy management.

We obtained chemotherapy information on the subset of patients, and when we looked at [patients who were] ctDNA negative, we didn’t see a clear benefit to chemotherapy. Admittedly, our numbers were quite small, so we don't have that information yet to say if a patient is ctDNA negative, then they should not get chemotherapy. We’re really interested in looking at that more, especially in the stage II population, where it’s a bit controversial whether patients should receive adjuvant chemotherapy.

What are the next steps?

We are going to continue to look at the questions about chemotherapy and look at longitudinal ctDNA monitoring and dynamics. We are going to try to grow our dataset, because as we’re able to get more information, we may be able to make more robust conclusions.

This is an exciting opportunity, especially because in the real-world setting, we see a larger variety of patients and more heterogeneity, which better reflects the true patient population, as opposed to what we’re seeing on observational studies.


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

Related Videos
View All
Related Content