Postoperative MRD Positivity Linked With Reduced RFS in CRC

January 26, 2021 — Patients with colorectal cancer who had molecular residual disease positivity immediately after surgery were found to have a high risk of disease recurrence and longitudinal monitoring increased the predictive power of circulating tumor DNA.

Patients with colorectal cancer (CRC) who had molecular residual disease (MRD) positivity immediately after surgery were found to have a high risk of disease recurrence and longitudinal monitoring increased the predictive power of circulating tumor DNA (ctDNA), according to data from a study presented during the virtual 2021 Gastrointestinal Cancers Symposium.

Moreover, molecular recurrence was detected using ctDNA a median of 8 months prior to radiological detection of recurrence. Longitudinal testing with ctDNA also outperformed carcinoembryonic antigen (CEA) testing in the prediction of recurrence-free survival (RFS). 

“To get ctDNA implemented in the clinic, we first have to test it in specific clinical settings using randomized trials. One such trial could be an escalation trial for the [patients with] stage I and low-risk stage II [disease],” lead study author Tenna Vesterman Henriksen, PhD, of Aarhus University, said in an oral presentation of the research. “Currently, these patients do not receive any adjuvant chemotherapy after surgery, but 10% to 15% of them may be undertreated, so they go on to have disease recurrences. These patients could potentially benefit from adjuvant chemotherapy.”

The motivation for this study was inspired by the high relapse rates reported in patients with CRC (range 20%-30%), despite curative intent treatment. Identifying MRD in patients could allow for better recurrent risk assessment, and earlier detection of recurrence could increase the proportion of patients who are treated with curative intent following recurrence, thereby improving survival in this population.

“We believe that ctDNA is a promising marker for MRD detection, and that's the marker that we've been using in this study,” Henriksen noted.

The main objective of the study was to use ctDNA to detect MRD and arrange patients into subgroups of those with high risk and low risk of recurrence. Investigators also set out to evaluate the post-therapy risk of relapse in patients with ctDNA positivity and to identify the lead time for ctDNA detection compared with recurrence on computerized tomography (CT) scans.

To do this, investigators recruited patients from surgical centers in Denmark and Spain and collected tumor tissue samples. Blood samples were also collected before and directly after surgery. In some patients with CRC, blood samples were collected every 3 months after surgery for up to 3 years. Patients were monitored by CT scans at 12 months and 36 months after surgery. The CT scans formed the clinical basis for which the ctDNA data from blood samples were compared.

The cohort for this study was comprised of a total of 260 patients with stage 1 (n = 4), stage 2 (n = 90), and stage 3 (n = 166) CRC. Of those patients, 48 had relapsed. The median follow-up for patients who did not relapse was 29.9 months (range, 1.2-51.0). Some patients in the cohort were treated with adjuvant chemotherapy, the majority of which had stage 2 and stage 3 disease.

The detection technique used for ctDNA was the Signatera approach, a tumor-informed strategy in which the tumor and peripheral blood cells of the patient are exome sequenced to identify patient- and tumor-specific mutations. Investigators then picked out 16 clonal mutations that they used to design the Signatera ctDNA assay, which was then used against the plasma samples to monitor the level of ctDNA before and at several time points after surgery. 

The first time point at which the ctDNA was examined was directly after surgery. These blood samples were drawn within 2 months following surgery and before the start of neoadjuvant chemotherapy. Investigators were able to stratify patients into 2 groups: those who were ctDNA negative (n = 198) and those who were ctDNA positive (n = 20). When looking at recurrent rates, a very low recurrence rate was observed amongst the ctDNA-negative group, at 13%, compared with a higher recurrence rate in the ctDNA-positive group, at 80%. However, not all patients with ctDNA positivity experienced disease recurrence.

“We were wondering about this until we looked into the fact that 19 out of the 20 ctDNA-positive patients received adjuvant chemotherapy after their blood samples were drawn, Henriksen said. “In fact, the 4 patients who did not experience the disease recurrence but were ctDNA positive, so 20%, all received adjuvant chemotherapy.”

This indicated that the outcome was modified by adjuvant chemotherapy, according to Henriksen; it also suggested that those patients were likely cured of residual disease following surgery by receiving adjuvant chemotherapy.

“Hopefully, this could indicate that there is a benefit for ctDNA-positive patients to, in some cases, receive adjuvant chemotherapy,” Henriksen added. 

Beyond stratifying patients into groups with high and low risk of recurrence directly after surgery, investigators also examined that data after end of treatment. When looking at a blood sample drawn after treatment with adjuvant chemotherapy, investigators found a low recurrence rate in the ctDNA-negative group, at 12.5%, and an even higher recurrence rate than they had seen before in the ctDNA-positive group, at 83.3%.

With the inclusion of longitudinal sampling, drawn every 3 months after surgery, the recurrence rate in the ctDNA-negative group dropped to 3.4%, which indicates that serial sampling has a benefit over just single time point sampling. Using serial sampling, investigators evaluated the lead time achieved by using ctDNA (n = 29) compared with clinical CT scans (n = 29).

The result was a positive lead time in most cases, and ctDNA had a median lead time of 8.1 months compared with the CT scans.

Lastly, investigators compared ctDNA with CEA, an already established biomarker. To do this, CEA measurements were done at the same time points as the ctDNA, and the single time point measurements of CEA post surgery (n = 175; hazard ratio [HR], 1.3; 95% CI, 0.56-0.32; P = .524) and post adjuvant chemotherapy (n = 99; HR, 1.4; 95% CI, 0.44-4.2; P = .596) did not have any predictive power for RFS, although longitudinal samples did show some predictive power (n = 197; HR, 4.9; 95% CI, 3.2-15; P <.0001). However, when compared with ctDNA (n = 197; HR, 95.7; 95% CI, 28-322; P <.0001), HRs were much lower. 

When compared in a multivariable analysis, the predictive power of the ctDNA samples was much greater than CEA, with hazard ratios of 80.55 (95% CI, 23.1-281) versus 1.8 (95% CI, 0.77-4.0), respectively.

“This indicates that ctDNA is a stronger biomarker compared with CEA in terms of RFS,” concluded Henriksen.


  1. Henriksen TV, Tarazona N, Reinert T, et al. Circulating tumor DNA analysis for assessment of recurrence risk, benefit of adjuvant therapy, and early relapse detection after treatment in colorectal cancer patients. J Clin Oncol. 2021;39(suppl 3):11.