How ctDNA Is Changing Adjuvant Decision-Making in Colorectal Cancer

Adjuvant decision-making in colorectal cancer has always been a compromise between population-level evidence and patient-level uncertainty. Stage, nodal status, T stage, lymphovascular invasion, obstruction, perforation, tumor budding, mismatch repair status, age, performance status, and neuropathy risk all matter.

However, none of them directly answer the question every physician is actually trying to solve after curative-intent surgery: Does this patient still have biologically meaningful residual disease? If so, is it enough to justify chemotherapy? 

That is why circulating tumor DNA (ctDNA) has become such a consequential development in colorectal cancer care. It offers a molecular readout of residual disease that sits closer to tumor biology than conventional clinicopathologic risk factors do. In practical terms, ctDNA is changing adjuvant conversations by moving the discussion away from broad estimates based on stage alone and toward more individualized risk assessment. 

For physicians, the important question is not whether ctDNA is prognostic. That point is already well-established. The more clinically useful question is where ctDNA is mature enough to influence adjuvant chemotherapy decisions now, and where it still remains better at refining prognosis than directing therapy. The answer is not uniform across colorectal cancer — the strongest randomized evidence is in stage II colon cancer. Stage III colon cancer is now supported by powerful prognostic data and a completed randomized de-escalation/escalation study, but those results are more cautionary than permissive. Rectal cancer remains more complicated because neoadjuvant therapy, organ-preservation strategies, and lower-shedding disease biology make postoperative ctDNA interpretation harder. So, the real story is not that ctDNA has solved adjuvant decision-making in colorectal cancer. Rather, ctDNA has started to reshape it — selectively, unevenly, and in ways that physicians need to understand with precision. 

Why adjuvant decisions have been hard in colorectal cancer

The central problem in adjuvant colorectal cancer care is overtreatment on one side and undertreatment on the other. Current postoperative management still depends heavily on clinicopathologic features. In resected high-risk stage II and stage III disease, fluoropyrimidine-based therapy with or without oxaliplatin remains standard (typically for three to six months, depending on the setting and regimen). 

Yet this approach is blunt. 

Some patients receive oxaliplatin and sustain neuropathy, even though surgery alone likely cured them. Others have apparently lower-risk disease by routine criteria but still recur because micrometastatic disease was present from the start. The TRACC protocol (short for Tracking mutations in cell free DNA to predict Relapse in eArly Colorectal Cancer) states this problem clearly. Standard management is still largely driven by stage and histopathology, even though clinicians are likely still overtreating many patients with adjuvant chemotherapy. 

This problem is especially visible in stage II colon cancer, where the benefit of adjuvant chemotherapy has always been modest and inconsistently distributed. It is also visible in stage III disease, where adjuvant chemotherapy is standard, but the absolute benefit varies greatly across subgroups. In the GALAXY report, the authors noted that recurrence risk and likely benefit from adjuvant therapy vary substantially even within stage III disease, which helps explain why a more biologically direct marker, such as ctDNA, has generated so much interest. To be clear, ctDNA does not replace staging. Instead, it addresses the gap staging leaves behind. 

A second reason adjuvant decision-making has been difficult is that physicians often blend prognostic and predictive concepts together. A prognostic biomarker identifies recurrence risk. A predictive biomarker helps determine whether a specific therapy will improve outcome. ctDNA is unquestionably strong on the first question. The second question is where the evidence is advancing but still setting-specific. That distinction matters because many clinical decisions are still being made in a gray zone, where a test meaningfully changes confidence without fully dictating the treatment plan. In colorectal cancer, ctDNA is already changing recommendations because it modifies post-test probability in a clinically important way, even before every use case has been fully proven as a treatment-selection tool. 

What ctDNA adds to adjuvant decision-making

At its most useful, ctDNA acts as a molecular residual disease assay. In this setting, the physician is not primarily asking which targetable alteration is present. The physician is asking whether there is still detectable tumor-derived DNA in circulation after a supposedly curative intervention. The NCI Colon and Rectal–Anal Task Forces white paper identified minimal residual disease as one of the four major colorectal cancer domains where ctDNA could change clinical practice, and it specifically described detectable ctDNA after surgery or after adjuvant therapy as a robust marker of residual disease and high recurrence risk. 

That is the point at which ctDNA begins to influence adjuvant decisions. A positive postoperative result shifts the conversation toward biologic concern that occult systemic disease remains. A negative result lowers concern, although it does not eliminate it. Serial testing adds another layer because dynamics often matter more than a single landmark result. The same NCI white paper noted that next-generation sequencing-based multigene approaches using serial and/or larger plasma volumes currently provide the best sensitivity for minimal residual disease (MRD) detection, which is one reason clinicians should be careful about overinterpreting a single negative assay in a patient with otherwise concerning features. 

This is also why ctDNA is influencing adjuvant care differently from radiographic surveillance. Imaging is still the standard tool for localizing recurrence, but ctDNA addresses a different biologic timescale. It can signal molecular relapse or persistent MRD before conventional imaging can show a lesion. In the adjuvant setting, that earlier biologic signal is most useful when it helps physicians decide whether to intensify, de-escalate, or in some cases withhold chemotherapy. That is where the trial data matter most. 

How ctDNA is changing adjuvant decisions in stage II colon cancer

Stage II colon cancer is where ctDNA is already most actionable

The clearest example of ctDNA changing adjuvant decision-making comes from the randomized DYNAMIC trial in stage II colon cancer. In that study, patients were randomized to ctDNA-guided vs. standard management, and in the ctDNA-guided arm, a positive result at four or seven weeks after surgery prompted adjuvant chemotherapy, while ctDNA-negative patients were not treated. That design matters because it tested not merely whether ctDNA was prognostic, but whether a management strategy built around ctDNA could safely reduce chemotherapy use. 

The mature five-year data remain the most important reason many oncologists now view ctDNA as clinically actionable in selected stage II colon cancer discussions. The long-term follow-up showed that the earlier ctDNA-guided strategy reduced adjuvant chemotherapy use without compromising recurrence outcomes. Furthermore, at a median follow-up of 59.7 months, the five-year recurrence-free survival rate was 88% with ctDNA-guided management and 87% with standard management. Meanwhile, the overall survival rate was 93.8% and 93.3%, respectively. The authors concluded that mature DYNAMIC data support a ctDNA-guided approach to adjuvant chemotherapy in stage II colon cancer. 

Of course, not every stage II colon cancer case now has an automatic ctDNA algorithm. What has changed is the quality of the recommendation. Before ctDNA, the conversation in many stage II cases was a probabilistic discussion built largely around T stage and conventional high-risk features. After DYNAMIC, a ctDNA-negative postoperative result can materially strengthen the case for observation in a patient who might otherwise have received chemotherapy mainly because of unease. Conversely, a ctDNA-positive postoperative result makes the case for adjuvant treatment much stronger because it identifies molecular evidence that surgery may not have eradicated the disease. That is a major shift in how physicians frame postoperative risk. 

There is another clinically important message in the five-year DYNAMIC update: ctDNA may eventually do more than separate positive and negative patients. Among treated ctDNA-positive patients, ctDNA clearance at the end of adjuvant therapy was observed in 87.5%, and post hoc analysis using a newer assay found a five-year recurrence-free probability of 97% in those with ctDNA clearance versus 0% in those with persistent ctDNA. That finding does not yet mean physicians should routinely change therapy mid-course based on a single persistence signal outside a protocol, but it does show why adjuvant decision-making is moving from one-time postoperative classification toward dynamic molecular monitoring. 

Why stage II has become the proof-of-concept setting

Stage II colon cancer was always the ideal setting for a ctDNA-guided strategy because conventional adjuvant decision-making is least precise there. Many patients are cured by surgery alone. The downside of overtreatment is real, and the baseline risk is low enough that a negative MRD assay meaningfully shifts post-test probability. DYNAMIC changed more than one trial result — it changed the level of confidence physicians can bring to a recommendation for observation in properly selected ctDNA-negative patients. 

That said, stage II is also where physician judgment still matters most. ctDNA should refine interpretation, not replace it. Assay timing, mismatch repair biology, T4 status, and the patient’s tolerance for recurrence risk versus chemotherapy toxicity remain part of the recommendation. The value of ctDNA is that it reduces the amount of guesswork, not that it eliminates the need for clinical synthesis.

How ctDNA is changing adjuvant decisions in stage III colon cancer

Stage III data show why prognostic power and treatment direction are not the same

In stage III colon cancer, ctDNA has clearly changed risk assessment, but it has not yet simplified treatment selection to the same extent. Earlier observational work already showed how strongly postoperative ctDNA stratifies recurrence risk. In the 2019 JAMA Oncology cohort study, patients with detectable ctDNA after surgery had a three-year recurrence-free interval of 47% vs. 76% for those with undetectable ctDNA. After completion of chemotherapy, the gap widened further (30% vs. 77%, respectively). These are not small differences. They show that ctDNA can identify residual biologic risk even after standard adjuvant treatment. 

The larger 2023 GALAXY analysis reinforced the same point in a broader resectable colorectal cancer cohort. Postsurgical ctDNA positivity at four weeks after surgery was associated with a markedly higher recurrence risk (hazard ratio, 10.0 overall and 10.82 in stage II or III disease). In observational analyses, postsurgical ctDNA positivity also identified patients with stage II or III colorectal cancer who appeared more likely to benefit from adjuvant chemotherapy. This is important because it strengthens the biologic argument that ctDNA is not just a passive risk marker. Still, GALAXY was observational, not a randomized treatment-allocation trial. That distinction matters when translating the data into standard practice. 

The randomized DYNAMIC-III trial, reported in 2025, is therefore pivotal. In that study, ctDNA-negative patients received de-escalated therapy and ctDNA-positive patients received escalated therapy in the ctDNA-guided arm. With a median follow-up of 47 months, ctDNA-negative patients had much better outcomes than ctDNA-positive patients overall, confirming that ctDNA is a strong prognostic classifier. But de-escalation in ctDNA-negative patients did not meet the noninferiority margin compared with standard management, even though it reduced oxaliplatin exposure and hospitalizations. In ctDNA-positive patients, escalated therapy did not improve outcomes over standard management. 

That result is one of the most important things physicians need to understand about ctDNA colorectal cancer testing right now. In stage III colon cancer, ctDNA is already changing the adjuvant conversation, but mainly by improving recurrence estimates, sharpening discussions about benefit vs. burden, and helping identify patients for clinical trials. It is not yet a blanket justification to omit oxaliplatin because a patient is ctDNA-negative, nor is it a universal reason to intensify chemotherapy because a patient is ctDNA-positive. Stage III is where ctDNA has become clinically indispensable for prognosis, yet still incomplete as a universal treatment-selection switch.

What this means in real practice

For a ctDNA-negative stage III patient, the result can still matter. It can support a more nuanced discussion about the absolute benefit of oxaliplatin, the downside of neuropathy, and whether a lower-intensity strategy is being considered in a frailer patient. But the DYNAMIC-III results mean physicians should be careful about interpreting ctDNA negativity as permission to broadly de-escalate outside the evidence base. 

For a ctDNA-positive stage III patient, the result should heighten concern and raise the threshold for complacency, but the available randomized data do not yet prove that empiric escalation will improve outcome. This is precisely why trial referral is so important in ctDNA-positive stage III disease. 

Why rectal cancer is different

The phrase colorectal cancer can obscure an important clinical truth: Colon and rectal cancers are not interchangeable adjuvant settings. The NCI white paper makes this clear. In rectal cancer, the role of ctDNA in determining the need for adjuvant therapy following neoadjuvant therapy and/or surgery remains a major research need, and conclusive data are lacking. The same paper notes that nonmetastatic rectal cancer often carries a lower tumor burden and may shed even less ctDNA after neoadjuvant treatment, which means assay sensitivity becomes even more critical. 

For physicians, that means most of the clinically actionable adjuvant ctDNA evidence still belongs to colon cancer, especially stage II colon cancer. In rectal cancer, ctDNA may eventually help refine decisions after total neoadjuvant therapy, surgery, or watch-and-wait strategies, but it is not yet the same kind of adjuvant decision tool it has become in selected colon cases. When clinicians talk about ctDNA changing adjuvant decision-making in colorectal cancer, the strongest evidence still comes from colon cancer. 

Assay choice and logistics affect the decision as much as the biology

Tumor-informed and tumor-agnostic assays are not interchangeable

One reason ctDNA is changing clinical practice gradually rather than all at once is that assays differ in design and performance. Two primary MRD methodologies are used in colorectal cancer. Tumor-informed approaches begin with sequencing the patient’s tumor and then tracking patient-specific mutations in plasma. Tumor-agnostic approaches use predefined genomic or epigenomic panels directly from plasma without prior tumor sequencing. Those differences are operationally and clinically relevant because they influence turnaround time, tissue requirements, and sensitivity. 

This is not only a theoretical issue. 

A 2025 diagnostic accuracy meta-analysis found that in serial monitoring after colorectal cancer resection, tumor-informed assays had higher sensitivity than tumor-agnostic assays (0.88 vs. 0.59, respectively) without a significant difference in false-positive rates. That does not mean one platform is always superior in every workflow, but it does mean physicians should avoid treating all ctDNA results as equally interchangeable. When a test result will influence whether adjuvant chemotherapy is given, the details of the assay matter. 

Timing and turnaround are part of the clinical utility

The landmark trials also highlight a basic but underappreciated point: ctDNA is only clinically useful when it arrives in time to influence a decision. In DYNAMIC, testing occurred at four or seven weeks after surgery. In DYNAMIC-III, patients underwent ctDNA testing five to six weeks after surgery. The NCI white paper stresses that rapid turnaround is crucial in the MRD setting because curative-intent adjuvant decisions operate under real time constraints. A result that comes back after the practical window for adjuvant therapy is much less valuable, even if analytically excellent.

Serial testing can also matter more than a single postoperative blood draw. The NCI panel notes that serial and/or larger-volume approaches currently provide the best sensitivity for MRD detection. That is highly relevant in-clinic because a one-time negative assay is less reassuring in a known low-shedding tumor biology than serially negative assays over time. Physicians should think of ctDNA not just as a one-off test, but as a monitoring strategy whose value depends on the clinical question being asked. 

False positives and false negatives still matter

ctDNA is powerful, but not infallible. The NCI white paper warns that clonal hematopoiesis can cause false-positive results, including alterations in genes such as TP53 and KRAS that can be misread as tumor-derived in patients with colorectal cancer. It also notes that false negatives remain a biologic reality because some tumors are low shedders and some metastatic sites, including peritoneal disease, are less reliably detected in plasma. This is one reason ctDNA should modify clinical probability, not erase the rest of the case. 

What a positive ctDNA result should change

A positive postoperative ctDNA result should change the level of concern immediately. It means the probability that surgery alone has eradicated all disease is materially lower than it looked through a pathologic lens alone. In stage II colon cancer, that generally strengthens the recommendation for adjuvant chemotherapy. In stage III disease, it sharpens the case that the patient remains at biologically high risk even if standard therapy is still the default recommendation. In either setting, a positive result should also raise the likelihood that the patient should be offered a clinical trial if one is available. 

Persistent positivity is even more consequential. In the DYNAMIC 5-year analysis, ctDNA persistence at end of treatment was associated with profoundly worse long-term outcomes than ctDNA clearance. In DYNAMIC-III, persistent ctDNA after treatment was likewise associated with markedly worse prognosis, with a three-year recurrence-free survival rate of 14% vs. 79% for patients without it. For physicians, persistent ctDNA is the clearest signal that standard adjuvant therapy may not have fully addressed the underlying disease biology. That does not mean there is already a validated rescue algorithm for every such patient. However, it does mean the patient should not be considered reassuringly “finished” simply because chemotherapy has ended. 

What a negative ctDNA result should change

A negative ctDNA result should change the discussion, but not in the same way across all colorectal cancer settings. In stage II colon cancer, the strongest interpretation is that a negative postoperative result meaningfully lowers recurrence risk and supports observation in many patients who otherwise might have received adjuvant therapy based on imprecise clinicopathologic risk. That is where ctDNA is already most decision-altering. 

In stage III colon cancer, a negative result is still useful, but it does not currently have the same power to justify routine de-escalation. DYNAMIC-III showed that ctDNA-negative patients are indeed a lower-risk group, but it also showed that a de-escalated strategy did not meet noninferiority vs. standard management. So, the practical role of a negative result in stage III is to inform a more individualized conversation, not to automatically replace standard adjuvant therapy. 

The most balanced way to use a negative ctDNA result is as a biologically reassuring datapoint that modifies, but does not erase, clinical context. If the patient is stage II and otherwise borderline for treatment, that modification may be enough to support observation. If the patient is stage III with clear reasons to treat, ctDNA negativity may influence the discussion of absolute benefit and toxicity tolerance, but it should not currently be interpreted as a universal “no chemotherapy needed” signal. 

A practical physician workflow for ctDNA in adjuvant colorectal cancer

A reasonable physician workflow, based on the available trials and protocols, looks like this.

  • Decide before ordering the test what action a positive, negative, or persistent result would realistically trigger. If no plausible action follows, the test may add anxiety more than value. 

  • Match the assay and timing to the question. Landmark postoperative testing in the major adjuvant trials occurred roughly four to six weeks after surgery, and rapid turnaround is essential if the result is going to shape adjuvant treatment. 

  • Interpret the result with stage, pathology, mismatch repair status, comorbidity, neuropathy risk, and patient goals, not in isolation. The ctDNA result should sharpen the recommendation, not replace clinical reasoning. 

  • Use the result differently by setting. In stage II colon cancer, ctDNA can already support real de-escalation decisions in selected patients. In stage III disease, it should usually be treated as a strong prognostic classifier and a trial-enrichment tool unless and until additional data clarify broader treatment-selection roles. 

  • Have a plan for persistence or conversion. ctDNA persistence after adjuvant therapy or later conversion to positivity should trigger renewed multidisciplinary thinking, not passive reassurance. 

This workflow is partly an inference from the evidence rather than a single published algorithm, but it reflects the way ctDNA is most defensibly used right now in colorectal cancer. The main practical change is that adjuvant therapy no longer has to be framed only as “stage-based standard care.” It can increasingly be discussed as a response to measured residual biologic risk. 

The next wave of trials will decide how far ctDNA goes

The most important unanswered questions are now being pushed into randomized trials. NRG-GI008, also called CIRCULATE-North America, is enrolling patients with stage III and high-risk stage II colon cancer to determine whether chemotherapy is needed for all or only some patients based on postoperative ctDNA. The study is specifically designed to test both ends of the problem: whether ctDNA-negative patients can avoid or reduce chemotherapy and whether ctDNA-positive patients need more intensive treatment. 

TRACC Part C is asking a similar question in the United Kingdom in patients with high-risk stage II and stage III colorectal cancer, comparing ctDNA-guided chemotherapy against standard-of-care adjuvant treatment. The protocol explicitly frames ctDNA-guided de-escalation as a way to reduce unnecessary toxicity without sacrificing disease-free survival. These ongoing studies matter because they will determine whether the field moves from risk refinement to fully biomarker-directed adjuvant care across broader colorectal populations. 

What the 2025 DYNAMIC-III results already suggest is that the final answer may not be uniform. Stage II colon cancer may end up as the clearest routine-use setting. Stage III disease may require more selective integration, perhaps tied to specific risk groups, regimens, or serial-response strategies rather than simple positive/negative branching. Rectal cancer may evolve on a different timeline entirely. That is not a failure of ctDNA. It is what one should expect when a biologically powerful biomarker enters several different clinical ecosystems at once. 

The bottom line

ctDNA is changing adjuvant decision-making in colorectal cancer by making the postoperative discussion more biologically informed and less dependent on blunt clinicopathologic estimates alone. In stage II colon cancer, randomized evidence now supports ctDNA-guided de-escalation of adjuvant therapy in selected patients without loss of long-term efficacy. In stage III colon cancer, ctDNA has become a strong prognostic classifier that changes counseling and trial selection, but current randomized data do not yet justify turning every negative result into de-escalation or every positive result into escalation. In rectal cancer, the role of ctDNA in adjuvant treatment planning remains promising, but inconclusive. 

For physicians, the most useful mindset is neither hype nor hesitation. ctDNA is not a replacement for stage, pathology, or judgment. It is a high-value molecular layer that can materially improve how adjuvant risk is estimated and discussed. The best current use is thoughtful, time-sensitive, and setting-specific. That is how ctDNA colorectal cancer testing is changing practice now, and why the next few years of randomized trial data will matter so much for where the standard ultimately lands. 

 

References

  • BMC Cancer. ctDNA guided adjuvant chemotherapy versus standard of care adjuvant chemotherapy after curative surgery in patients with high risk stage II or stage III colorectal cancer: a multi-centre, prospective, randomised control trial (TRACC Part C) 
    https://link.springer.com/article/10.1186/s12885-023-10699-4 

 

 

View Free Oncology Courses