ctDNA-MRD in Resected Colorectal Cancer: Surveillance and Adjuvant Guidance

The clinical question this addresses

When a colorectal cancer is surgically removed, the central uncertainty that follows is whether the operation was curative or whether microscopic disease remains somewhere in the body — so-called minimal (or molecular) residual disease (MRD). Conventional tools answer this question only indirectly and often late. Staging describes the tumor's anatomic extent at the time of surgery but cannot confirm whether residual cancer cells persist. Serial carcinoembryonic antigen (CEA) measurements and cross-sectional imaging (CT) are the standard surveillance approach, yet both detect recurrence only once disease has grown to a radiographically or serologically measurable burden.

This gap matters because adjuvant chemotherapy decisions hinge on recurrence risk, and many resected patients are either overtreated (they were already cured by surgery) or undertreated (microscopic disease was present but not detected). A blood test that directly reports the presence of residual tumor-derived DNA after surgery could, in principle, identify which patients harbor MRD and therefore face the highest recurrence risk. This article explains how tumor-informed circulating tumor DNA (ctDNA) assays attempt to answer that question, what the landmark evidence actually shows, and where this technology stands educationally — not as treatment advice.

What the assay is and how it works

Circulating tumor DNA refers to fragments of tumor-derived DNA shed into the bloodstream, forming a small subset of the larger pool of cell-free DNA (cfDNA) that includes DNA from normal cells. In the MRD setting, the dominant design is tumor-informed ctDNA testing. The workflow begins with sequencing the resected tumor tissue to identify a patient-specific catalog of somatic mutations. A personalized assay is then built to track those specific variants in subsequent blood draws. Because the panel is anchored to mutations known to be present in that individual's tumor, the approach can detect ctDNA at very low allele fractions with high specificity.

The foundational cohorts used several such strategies. Tie and colleagues employed Safe-SeqS, an error-corrected sequencing method, to detect postoperative ctDNA in stage II colon cancer [1]. Reinert and colleagues and Henriksen and colleagues used tumor-informed multiplex PCR–based approaches to track patient-specific variants serially over time [2,4]. The GALAXY study within CIRCULATE-Japan applied a commercially developed tumor-informed assay across a large prospective cohort [3].

A result's meaning should be stated precisely. A positive postoperative or surveillance ctDNA result indicates that tumor-derived DNA is detectable in the blood, which the evidence links strongly to residual disease and elevated recurrence risk. A negative result indicates that no tumor DNA was detected at the assay's limit of detection — it lowers, but does not eliminate, the probability of residual disease. ctDNA cannot localize where recurrent disease resides, nor does a single result establish cure. These are prognostic, risk-stratifying tools, not anatomic or definitive diagnostic ones.

The evidence

The strength of the ctDNA-MRD literature in resected colorectal cancer lies in the consistency of its central finding across independent cohorts: detectable ctDNA after surgery is among the strongest known predictors of recurrence.

Tie 2016 (Sci Transl Med). In 230 patients with resected stage II colon cancer, postoperative ctDNA detected by Safe-SeqS was strongly associated with recurrence. Patients who were ctDNA-positive after surgery had a markedly higher risk of relapse, with a hazard ratio of approximately 18 [1]. This early study established that postoperative ctDNA could stratify a population — stage II disease — where the absolute benefit of adjuvant chemotherapy is debated and many patients are cured by surgery alone.

Reinert 2019 (JAMA Oncol). This study followed 125 patients with stage I–III colorectal cancer using serial tumor-informed ctDNA. Postoperative ctDNA positivity carried a hazard ratio of about 7.2 for recurrence, while ctDNA detection during longitudinal surveillance was associated with a substantially higher hazard ratio of roughly 43.5 [2]. The serial sampling detected relapse with a sensitivity of 87.5%, and — importantly for the surveillance question — ctDNA positivity frequently preceded radiographic recurrence, illustrating the lead-time advantage over imaging [2].

Henriksen 2022 (Clin Cancer Res). In 168 patients with stage III colorectal cancer assessed with tumor-informed ctDNA, postoperative positivity was associated with a hazard ratio of about 7.0 for recurrence. After completion of adjuvant chemotherapy, the prognostic signal strengthened further: patients who remained or became ctDNA-positive after adjuvant therapy had a hazard ratio of approximately 50 [4]. This observation is clinically meaningful because it suggests ctDNA status measured after adjuvant treatment reflects whether that treatment cleared residual disease.

Kotani 2023 / GALAXY, CIRCULATE-Japan (Nat Med). The largest of the cited cohorts, GALAXY enrolled 1,039 patients and measured ctDNA at the 4-week postsurgical window. ctDNA positivity at that timepoint was associated with a hazard ratio of 10.0 for recurrence [3]. The study also examined the interaction between ctDNA status and adjuvant chemotherapy: among ctDNA-positive patients, those who received adjuvant chemotherapy appeared to benefit more than those who did not, with a hazard ratio of 6.59 favoring treatment in that subgroup [3]. This finding is frequently cited as early evidence that ctDNA may help identify which patients derive the most benefit from adjuvant chemotherapy as a class — though it reports an observational association, not a treatment prescription for any individual.

Across these four cohorts, two themes recur: postoperative ctDNA positivity predicts recurrence with hazard ratios roughly in the 7–18 range, and the prognostic signal intensifies when ctDNA is measured during surveillance or after adjuvant therapy, with hazard ratios reaching the tens [1,2,3,4]. The lead-time over imaging — ctDNA turning positive months before radiographic recurrence — is a consistent, if study-dependent, observation [2].

Where it stands

The evidence that postoperative and serial ctDNA is a powerful prognostic marker in resected colorectal cancer is strong and reproducible across independent cohorts totaling well over a thousand patients [1,2,3,4]. For risk stratification, ctDNA-MRD is among the most robust predictors of recurrence described in this disease.

Several limitations and open questions remain, and these should be stated plainly. First, prognostic power is not the same as proven clinical utility: demonstrating that acting on a ctDNA result (escalating or de-escalating therapy) improves patient outcomes requires interventional, randomized trials, and the GALAXY adjuvant-benefit finding is an observational association rather than evidence from a treatment-randomized design [3]. Second, a negative result does not guarantee cure, because ctDNA can be present below an assay's detection limit; sensitivity for relapse, while high in serial testing (87.5% in Reinert), is not perfect [2]. Third, tumor-informed assays require adequate tumor tissue to build the personalized panel, and assay platforms, sampling timepoints, and detection thresholds differ across the cited studies, which complicates direct comparison and standardization [1,2,3,4]. Fourth, the optimal sampling schedule for surveillance and the actions that should follow a positive result remain areas of active investigation.

For trainees and patients, the key educational takeaway is this: ctDNA-MRD testing after colorectal cancer surgery is a validated-in-research, strongly prognostic tool that reports residual-disease risk earlier and more directly than imaging or CEA — but how that information should change management for an individual is a question for clinical trials and the treating team, not something a test result alone can decide.

References

1. Tie J, et al. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer. Sci Transl Med. 2016. PMID 27384348; DOI 10.1126/scitranslmed.aaf6219.
2. Reinert T, et al. Analysis of plasma cell-free DNA by ultradeep sequencing in patients with stages I to III colorectal cancer. JAMA Oncol. 2019. PMID 31070691; DOI 10.1001/jamaoncol.2019.0528.
3. Kotani D, et al. Molecular residual disease and efficacy of adjuvant chemotherapy in patients with colorectal cancer (GALAXY, CIRCULATE-Japan). Nat Med. 2023. PMID 36646802; DOI 10.1038/s41591-022-02115-4.
4. Henriksen TV, et al. Circulating tumor DNA in stage III colorectal cancer, beyond minimal residual disease detection, toward assessment of adjuvant therapy efficacy and clinical behavior of recurrences. Clin Cancer Res. 2022. PMID 34625408; DOI 10.1158/1078-0432.CCR-21-2404.