ctDNA Moves from Prognostic Marker to Treatment Selector in Resected Colorectal Cancer
Resected colorectal cancer is no longer being judged only by tumor stage and pathologic margins. Circulating tumor DNA has moved the field toward a molecular view of recurrence risk, one based on the direct detection of microscopic disease after surgery. In the postoperative setting, ctDNA behaves as a real-time biomarker of molecular residual disease: it clears within hours after a macroscopically complete resection, so a persistent signal is unlikely to represent anything other than viable tumor that remains in the body. That biologic principle explains why ctDNA has become so attractive in colon and rectal cancer, where conventional staging still leaves a substantial fraction of patients overtreated or undertreated.
The clinical stakes are large. Colorectal cancer causes roughly 850,000 deaths each year and remains the second leading cause of cancer mortality worldwide. Even after curative-intent surgery, recurrence still occurs in about 15% to 20% of stage II disease and 30% to 40% of stage III disease. Pathology captures the anatomic extent of the tumor that was removed, but it cannot fully represent the burden of residual microscopic disease left behind in the bloodstream, marrow, or distant organs. ctDNA answers a different question: whether cancer remains biologically active after resection.
De-escalation Data in ctDNA-Negative Patients Are the Strongest Proof of Utility
The most persuasive clinical use of ctDNA has been treatment de-escalation. The DYNAMIC trial showed that ctDNA-guided management reduced adjuvant chemotherapy use from 27.9% to 15.3% in stage II colon cancer while preserving outcomes. Five-year recurrence-free survival was 88% in the ctDNA-guided group and 87% in the standard management group, a difference of 1.1% with a 95% confidence interval ranging from 5.8% to 8.0%. Those numbers matter because they show that withholding chemotherapy in carefully selected ctDNA-negative patients did not sacrifice disease control over long follow-up.
The biology fits the result. A patient with no detectable ctDNA after surgery is, in effect, showing no measurable molecular trace of viable residual cancer. In that setting, the marginal benefit of chemotherapy may be small, while toxicity remains very real. ctDNA therefore functions less like a general screening test and more like a postoperative filter for risk, identifying patients who can avoid treatment that may never have helped them.
This is where the field has found its clearest practical value: lowering exposure to chemotherapy without worsening recurrence-free survival in a postoperative population that includes many patients who would otherwise be treated on the basis of stage alone.
Escalation in ctDNA-Positive Stage III Disease Has Not Yet Delivered Benefit
The story changes sharply once ctDNA is already positive. In stage III colon cancer, DYNAMIC-III tested whether increasing chemotherapy intensity in ctDNA-positive patients would improve recurrence-free survival. It did not. The hazard ratio was 1.11, with a p value of 0.6, indicating no measurable advantage from escalation in that setting.
That result is important because it challenges a tempting assumption in biomarker-guided oncology: if minimal residual disease is present, then more treatment should automatically help. The current evidence does not support that logic in ctDNA-positive stage III disease. A positive ctDNA signal identifies high risk, but risk identification is not the same as treatment sensitivity. Microscopic residual disease may already be established in a form that is less responsive to standard adjuvant regimens, or the tumor clones responsible for postoperative relapse may require different drugs altogether.
The GALAXY study reinforced ctDNA’s prognostic power. In resected colorectal cancer, ctDNA carried a disease-free survival hazard ratio of 11.99, making it the strongest biomarker in that context. Prognostic strength, however, has not translated into straightforward predictive power for chemotherapy intensification. The difference is clinically decisive. A marker can be excellent at identifying who is likely to relapse and still fail to tell clinicians which therapy will prevent that relapse.
Post-adjuvant Molecular Recurrence Still Lacks a Proven Rescue Strategy
The phase III ALTAIR trial examined a different scenario: treatment at the point of molecular recurrence after adjuvant therapy. Trifluridine/tipiracil failed to significantly improve disease-free survival, with a hazard ratio of 0.79 and a p value of 0.107. The trial therefore did not establish a clear benefit for post-adjuvant intervention once molecular recurrence had already been detected.
This is a sobering result for a field that has hoped ctDNA might not only detect relapse early but also create a window for effective intervention before overt metastatic disease appears. Early detection alone is not enough. If the agent chosen for molecular recurrence does not alter the underlying biology of the residual clone, then the lead time provided by ctDNA may not translate into better survival or longer disease-free intervals.
The therapeutic implication is direct. ctDNA can identify the moment when microscopic disease reappears, but the right drug at that moment remains uncertain. The failure of trifluridine/tipiracil to deliver a significant disease-free survival benefit means the biomarker has outpaced the available treatment strategy in this setting.
Clinical Utility Is Real, But the Next Step Requires Better Matching of Biomarker and Drug
The current evidence now separates three uses of ctDNA-guided care. First, it can support de-escalation in ctDNA-negative patients, where randomized data have already shown reduced chemotherapy use without compromised recurrence-free survival. Second, it identifies ctDNA-positive patients with very high recurrence risk, although present trials have not shown that simply intensifying standard chemotherapy improves outcomes. Third, it detects molecular recurrence after adjuvant therapy, but the first major phase III test of treatment in that setting did not prove a disease-free survival advantage.
What emerges is a more disciplined view of ctDNA in resected colorectal cancer. The biomarker is no longer just prognostic; it has proven clinical utility in sparing some patients unnecessary chemotherapy. But the most intuitive extension of the concept — treating harder when ctDNA is positive, or treating earlier at molecular relapse — has not yet produced the expected survival gains.
Future progress will depend on biomarker-matched escalation strategies, adaptive trial designs, and formal validation of ctDNA clearance as a surrogate endpoint. Without those steps, ctDNA will remain a powerful detector of residual disease without a fully established therapeutic roadmap for the patients most likely to relapse.