Circulating tumor DNA (ctDNA) testing can identify molecular signs of colorectal cancer recurrence after surgery, but a randomized phase 3 trial published in Nature Medicine in June 2026 found that treating patients solely because their blood test became positive did not significantly improve disease-free survival. In the ALTAIR trial, 243 patients with detectable tumor DNA but no cancer visible on imaging received trifluridine/tipiracil or placebo. The result exposes an important distinction in precision oncology: identifying residual disease and proving which intervention improves outcomes are different clinical achievements.
Fragments of tumor DNA can remain detectable when imaging shows no cancer
When tumor cells die or release their contents, fragments of DNA enter the bloodstream. Most circulating cell-free DNA originates from normal cells. A fraction may carry alterations associated with a patient’s cancer. A liquid biopsy using ctDNA searches for these tumor-derived fragments in a blood sample.
For molecular residual disease (MRD) testing after surgery, some assays first sequence the removed tumor to select patient-specific mutations. Subsequent blood samples are then analyzed for those mutations. Such tumor-informed tests are different from broad blood tests marketed for screening healthy people for multiple cancers. The objective here is to detect molecular evidence of an already diagnosed cancer after an apparently curative treatment.
That distinction matters because a standard scan has a detection threshold: very small deposits may be too small to see. Finding tumor DNA in blood can indicate residual malignant cells even when computed tomography shows no measurable disease. But a negative blood test cannot rule out every microscopic focus. Tumors vary in how much DNA they shed, and assay sensitivity depends on the amount of tumor-derived material in a sample.
GALAXY found that postoperative ctDNA positivity strongly predicted recurrence
An observational analysis from the CIRCULATE-Japan GALAXY program, published in Nature Medicine in September 2024, evaluated 2,240 patients with resectable colorectal cancer, including stage II–III colon cancer and resected stage IV disease. With a median follow-up of 23 months, ctDNA positivity during the postoperative MRD window was associated with substantially worse disease-free survival (hazard ratio 11.99) and overall survival (hazard ratio 9.68). These are relative associations between groups, not the percentage of ctDNA-positive patients who recurred.
The results support ctDNA’s powerful prognostic value: a positive molecular result identifies patients with higher risk after surgery. They do not, on their own, show that changing treatment based on the test improves survival. Treatment choices made in an observational cohort may reflect differences between patients that statistical adjustment cannot fully remove.
DYNAMIC showed that ctDNA guidance can reduce chemotherapy use in selected patients
A separate randomized study, DYNAMIC, addressed an earlier decision: whether to give adjuvant chemotherapy after surgery for stage II colon cancer. Its five-year follow-up, published in Nature Medicine in March 2025, reported recurrence-free survival of 88% with a ctDNA-guided strategy and 87% with standard management. Five-year overall survival was also similar at 93.8% versus 93.3%.
The trial’s earlier primary analysis showed that a ctDNA-guided strategy reduced the proportion of stage II colon cancer patients receiving adjuvant chemotherapy without compromising two-year recurrence-free survival. The later results provide reassurance about longer follow-up, although the study was not designed to prove that ctDNA-guided care works equally well for every colorectal cancer stage or every chemotherapy decision.
These findings help explain why ctDNA matters clinically. The test may inform decisions about withholding treatment as well as decisions to intensify care. Avoiding unnecessary chemotherapy is valuable because even established cancer treatments impose toxicity and quality-of-life costs. The question is whether a particular test-informed strategy has been validated in the right patient population.
ALTAIR tested whether acting on molecular recurrence improves outcomes
In ALTAIR, investigators asked a more demanding question. Patients had undergone curative-intent resection for stage 0–IV colorectal cancer and completed standard-of-care treatment. During subsequent blood testing, tumor-informed ctDNA became detectable despite no radiological evidence of disease. Investigators randomly assigned 122 patients to oral trifluridine/tipiracil and 121 to placebo for six months. The trial was double-blind, and its primary endpoint was investigator-assessed disease-free survival.
Median disease-free survival was 9.30 months in the treatment group and 5.55 months in the placebo group. The hazard ratio was 0.79, with a 95% confidence interval of 0.60–1.05 and P = 0.107. The prespecified primary endpoint was not met. The difference between median survival times cannot be interpreted as proof that the drug improved outcomes; statistical uncertainty includes the possibility of no meaningful effect.
Safety makes that distinction consequential. Grade 3 or higher hematologic adverse events occurred in 73.0% of treated patients versus 3.3% receiving placebo. Such events reflect substantial blood-related toxicity, rather than a negligible cost of treating a molecular marker. The result does not mean ctDNA testing is useless or that no treatment can benefit patients with MRD. It means this specific treatment strategy did not demonstrate a statistically significant disease-free survival benefit in the overall randomized population.
The trial also included exploratory analyses of subgroups, including patients with resected stage IV disease. Subgroup signals are hypotheses for further study, not grounds to overturn an unsuccessful primary outcome. The broader oncology challenge is familiar from efforts to improve personalized cancer treatment: a biologically plausible strategy requires prospective clinical testing before its benefits can be claimed.
Detecting cancer earlier does not automatically establish a survival benefit
There are at least three separate questions: whether a test detects tumor-derived DNA accurately, whether its result predicts a patient’s future risk, and whether responding to that result improves a clinical outcome. ALTAIR and DYNAMIC tested different treatment decisions in different patient populations. Their results are not contradictory.
A molecular signal may become positive before conventional imaging shows recurrence, but detecting a problem earlier also creates a period of uncertainty. Patients may undergo repeated blood draws, investigations, and treatment discussions without a proven intervention for that precise situation. More sensitive surveillance could therefore increase opportunities for care while also increasing the risk of premature treatment.
What counts as meaningful benefit should be specified in advance. Disease-free survival, recurrence-free survival, overall survival, treatment exposure, toxicity, and quality of life are distinct endpoints. A trial that improves a biomarker but not patient outcomes cannot be treated as equivalent to a trial demonstrating longer survival.
What the evidence cannot yet answer
These results apply most directly to colorectal cancer populations and the specific assays and management strategies studied. Detection sensitivity can vary by tumor biology, sampling time, disease burden, assay design, and the mutations tracked. MRD-negative blood results should not be interpreted as guarantees of cure, and ctDNA results should not automatically replace imaging or established follow-up plans.
GALAXY is observational and cannot establish that chemotherapy caused the outcomes associated with ctDNA clearance. DYNAMIC enrolled stage II colon cancer patients making an adjuvant treatment decision. ALTAIR examined a different setting: ctDNA-positive patients after standard treatment with no detectable disease on imaging. Neither trial answers whether every ctDNA-positive patient should receive additional chemotherapy, immunotherapy, or a targeted drug. Other randomized strategies remain under investigation.
The ALTAIR publication also reports disclosures relevant to interpretation. Several coauthors were employed by Natera, a company involved in ctDNA testing, and multiple investigators disclosed financial relationships with pharmaceutical or diagnostics companies. These disclosures do not negate the randomized findings, but they are important context when evaluating claims about clinical adoption.
For now, ctDNA is a strong risk signal, not a universal instruction to treat
The most defensible conclusion is narrower than many liquid-biopsy headlines suggest. Blood-based tumor DNA testing can provide information conventional imaging does not capture, and randomized evidence supports using ctDNA to guide chemotherapy decisions in a defined stage II colon cancer setting. But ALTAIR illustrates how the benefit of intervening on a molecular result must be established independently for each treatment strategy.
That distinction should guide how future studies of MRD-positive disease are evaluated, just as it shapes the interpretation of early signals from experimental solid-tumor therapies. A more sensitive test expands what clinicians can observe. It does not automatically tell them what to prescribe.
References
- Bando H, Watanabe J, Takahashi Y, et al. Post-adjuvant chemotherapy in ctDNA-positive patients with resected colorectal cancer: a randomized phase 3 trial. Nature Medicine. 2026;32:2473–2480. DOI: 10.1038/s41591-026-04428-0.
- Nakamura Y, Watanabe J, Akazawa N, et al. ctDNA-based molecular residual disease and survival in resectable colorectal cancer. Nature Medicine. 2024;30:3272–3283. DOI: 10.1038/s41591-024-03254-6.
- Tie J, et al. Circulating tumor DNA analysis guiding adjuvant therapy in stage II colon cancer: 5-year outcomes of the randomized DYNAMIC trial. Nature Medicine. 2025;31:1509–1518. DOI: 10.1038/s41591-025-03579-w.