Highlight
– ctDNA measured 5–6 weeks after surgery was a powerful prognostic classifier in stage III colon cancer: 3‑year recurrence‑free survival (RFS) 87% for ctDNA‑negative vs 49% for ctDNA‑positive (P < 0.001).
– In ctDNA‑negative patients, a protocol that de‑escalated adjuvant chemotherapy reduced oxaliplatin exposure (34.8% vs 88.6%) and hospitalizations (8.5% vs 13.2%) but narrowly failed to meet the predefined non‑inferiority margin for RFS.
– Among ctDNA‑positive patients, ctDNA burden correlated with recurrence risk, but predefined chemotherapy intensification did not improve 2‑year RFS versus standard management.
Background and clinical need
Adjuvant chemotherapy after resection reduces recurrence and improves survival for many patients with stage III colon cancer, but the absolute benefit for an individual patient varies widely. Oxaliplatin‑containing regimens (eg, FOLFOX) drive efficacy but also cumulative neurotoxicity and treatment‑related morbidity. Tools that personalize adjuvant therapy to actual micrometastatic disease burden are therefore highly desirable.
Circulating tumor DNA (ctDNA) is cell‑free tumor-derived DNA detectable in plasma that reflects residual microscopic disease after surgery. Prior trials in stage II colon cancer (notably the DYNAMIC trial series) demonstrated that ctDNA‑guided strategies can safely reduce adjuvant chemotherapy use without compromising outcomes. The DYNAMIC‑III trial extends this paradigm to stage III disease, where the baseline recurrence risk is higher and the optimal tailoring of adjuvant therapy is an unresolved clinical challenge.
Study design
DYNAMIC‑III is a multicenter, randomized phase 2/3 trial conducted by the Australasian Gastro‑Intestinal Trials Group (AGITG) with intergroup collaboration (including the Canadian Cancer Trials Group). Patients with resected stage III colon cancer underwent tumor‑informed ctDNA testing at 5–6 weeks after surgery and were randomized 1:1 to either a ctDNA‑guided management strategy or clinician‑prespecified standard management.
In the ctDNA‑guided arm, ctDNA‑negative patients received de‑escalated adjuvant therapy (clinicians prespecified the de‑escalated regimens), while ctDNA‑positive patients received escalated therapy compared with standard management. The trial’s co‑primary endpoints were 3‑year RFS for ctDNA‑negative patients (to evaluate non‑inferiority of de‑escalation) and 2‑year RFS for ctDNA‑positive patients (to evaluate benefit of escalation). Key secondary endpoints included treatment‑related hospitalization, ctDNA clearance after therapy, safety, and correlations between ctDNA burden and outcomes.
Population and follow‑up
The analysis included 968 evaluable patients. At initial postoperative testing, 702 patients (72.5%) were ctDNA negative and 266 (27.5%) were ctDNA positive. Median follow‑up was 47 months, allowing robust assessment of 2‑ and 3‑year RFS endpoints and ctDNA dynamics over the adjuvant period.
Key findings
Prognostic performance of postoperative ctDNA. Postoperative ctDNA status strongly discriminated recurrence risk. Across the combined trial population, ctDNA‑negative patients had a 3‑year RFS of 87% compared with 49% among ctDNA‑positive patients (P < 0.001). This magnitude of separation confirms ctDNA as one of the most powerful individual prognostic biomarkers for adjuvant risk stratification in stage III colon cancer.
ctDNA‑guided de‑escalation in ctDNA‑negative patients. In the subgroup of ctDNA‑negative patients randomized to ctDNA guidance, de‑escalation substantially reduced oxaliplatin use (34.8% vs 88.6% in standard management) and decreased treatment‑related hospitalizations (8.5% vs 13.2%). Three‑year RFS was 85.3% in the de‑escalation arm versus 88.1% in standard management. Although absolute differences in RFS were small, the study reported that the predefined non‑inferiority margin was not met—therefore non‑inferiority of de‑escalation could not be formally claimed despite outcome convergence and clear reductions in toxicity risk.
Escalation for ctDNA‑positive patients. Among ctDNA‑positive patients, escalated chemotherapy did not improve outcomes compared with standard management. Two‑year RFS was 51% in the escalation arm versus 61% in standard care. Importantly, within ctDNA‑positive patients, higher quantitative ctDNA burden correlated with higher recurrence risk: 3‑year RFS ranged from 77% down to 23% across ctDNA quartiles (P < 0.001), indicating a dose‑response relationship between ctDNA level and prognosis.
ctDNA dynamics during and after adjuvant therapy. Persistent ctDNA after completion of adjuvant therapy identified a subgroup with very poor outcomes: 3‑year RFS was 14% for patients with persistent ctDNA versus 79% for those who cleared ctDNA. Clearance of ctDNA therefore appears to be a meaningful intermediate biomarker of treatment effect and prognosis.
Safety. There were no unexpected toxicities reported. The de‑escalation approach clearly reduced exposure to oxaliplatin and associated adverse events, reflected in lower hospitalization rates.
Interpretation and clinical implications
The DYNAMIC‑III results reinforce several practical messages for clinicians and researchers:
- Postoperative ctDNA is a validated, strong prognostic classifier in stage III colon cancer. Its use can identify patients at very low risk of recurrence (ctDNA negative) and those at high risk (ctDNA positive), enabling more precise risk communication and shared decision‑making.
- For ctDNA‑negative patients, a de‑escalated adjuvant approach can meaningfully reduce oxaliplatin exposure and hospitalizations with only a small absolute reduction in RFS that did not reach prespecified non‑inferiority. Clinicians and patients who prioritize toxicity avoidance may view this tradeoff as acceptable, particularly given cumulative neurotoxicity from oxaliplatin; however, guideline incorporation will require careful consideration of the non‑inferiority outcome and patient values.
- For ctDNA‑positive patients, intensifying conventional cytotoxic chemotherapy as tested in this trial did not improve short‑term RFS. This suggests that simply escalating standard chemotherapy may be insufficient to overcome established molecular residual disease and that alternative strategies—targeted agents, immunotherapy, sequential ctDNA‑guided adaptive approaches, or inclusion in clinical trials testing novel combinations—are needed.
- ctDNA persistence after adjuvant therapy identifies a very poor‑prognosis subgroup that could be prioritized for early therapeutic intervention and investigational strategies and might serve as an efficient endpoint in early‑phase trials.
Strengths and limitations
Strengths of DYNAMIC‑III include the randomized design, large sample size, multicenter conduct, tumor‑informed ctDNA methodology, and prolonged follow‑up allowing reliable estimates of RFS. The prespecified co‑primary endpoints addressed both the low‑risk (ctDNA‑negative) and high‑risk (ctDNA‑positive) clinical questions important for adjuvant decision‑making.
Limitations include the following: the non‑inferiority margin for the de‑escalation comparison was not met despite close outcome curves, raising questions about the trial’s power and the clinical meaningfulness of the observed absolute RFS difference; the specific de‑escalation and escalation regimens were clinician‑prespecified rather than protocol‑mandated, which can introduce heterogeneity; and the trial did not demonstrate benefit from chemotherapy intensification, but the design cannot exclude benefit from alternative intensification strategies (eg, targeted agents or immunotherapy) not tested in the study.
Generalizability will depend on ctDNA assay availability, assay sensitivity, and timing of sampling after surgery—factors that can vary across laboratories and health systems. Implementation would require robust quality assurance for tumor‑informed assays and pathways for rapid report turnaround to inform adjuvant planning.
Mechanistic and translational insights
The quantitative relationship between ctDNA burden and recurrence risk supports a biological model in which ctDNA reflects the volume and proliferative activity of residual micrometastatic disease. Rapid ctDNA clearance during therapy likely indicates effective eradication of residual clones, while persistence signals refractory micrometastases. These dynamics provide a rationale for adaptive trial designs that escalate therapy for persistent ctDNA or de‑escalate when ctDNA clears early, and for studies that integrate molecular characterization of ctDNA (eg, mutation profile, methylation) to identify therapeutic vulnerabilities.
Directions for future research
Key unanswered questions include: what is the optimal de‑escalation regimen for ctDNA‑negative stage III patients that balances toxicity and efficacy; which novel systemic or targeted strategies can eradicate ctDNA‑positive disease; can early ctDNA kinetics (eg, clearance after 1–2 cycles) be used to adapt therapy effectively; and what are the health‑economic implications of ctDNA‑guided care in routine practice?
Randomized trials testing non‑cytotoxic intensification (immune checkpoint inhibitors, targeted agents where appropriate), ctDNA‑adaptive escalation strategies, and trials with ctDNA clearance as a validated surrogate endpoint are high priorities.
Expert commentary and guideline considerations
DYNAMIC‑III provides compelling evidence that ctDNA is a clinically meaningful biomarker in stage III colon cancer. Guidelines committees will need to weigh the trial’s careful demonstration of reduced toxicity with de‑escalation against the formal non‑inferiority outcome. Physicians should counsel patients individually: for many ctDNA‑negative patients, reduction of oxaliplatin exposure may be an attractive option when informed consent acknowledges the small absolute difference observed in RFS; for ctDNA‑positive patients, intensification with conventional chemotherapy appears insufficient and enrollment in clinical trials investigating novel therapies should be prioritized.
Conclusion
DYNAMIC‑III validates postoperative ctDNA as a powerful prognostic tool in stage III colon cancer and shows that ctDNA‑guided de‑escalation reduces chemotherapy exposure and hospitalizations with outcomes approaching standard care, although the non‑inferiority threshold was not formally met. Escalation of conventional chemotherapy for ctDNA‑positive patients did not improve short‑term RFS, highlighting an urgent need for novel therapeutic strategies and ctDNA‑driven adaptive trials for molecular residual disease.
Funding and clinical trial registration
Trial registration: Australian New Zealand Clinical Trials Registry Identifier ACTRN12617001566325. Funding and specific sponsor details are reported in the original publication (Tie J et al., Nat Med. 2025).
References
1. Tie J, Wang Y, Loree JM, et al.; AGITG DYNAMIC‑III Study Group. Circulating tumor DNA‑guided adjuvant therapy in locally advanced colon cancer: the randomized phase 2/3 DYNAMIC‑III trial. Nat Med. 2025 Oct 20. doi:10.1038/s41591-025-04030-w. Epub ahead of print. PMID: 41115959.
Thumbnail image prompt (AI‑friendly)
A clinician in a modern clinic looking at a tablet screen displaying a ctDNA graph with red and green trend lines, next to a CT scan of the abdomen; soft directional light, professional medical environment, emphasis on digital molecular diagnostics and patient‑centered decision making.