- Design
- Systematic review and meta-analysis separating prognostic and decision evidence
- Population
- 38 reports; 7 prospective cohorts in resected colorectal cancer
- Primary outcome
- Recurrence risk with postoperative ctDNA positivity
- Effect
- HR 7.60 (4.70 to 12.31), I² 80.6%; prediction interval 1.60 to 36.25
A systematic review of 38 reports separated two questions about circulating tumour DNA (ctDNA) after curative colorectal cancer surgery: does it predict recurrence, and does acting on it improve outcomes? Seven prospective cohorts were pooled for prognosis.
Postoperative ctDNA positivity was associated with a much higher recurrence risk (HR 7.60, 4.70 to 12.31), with high heterogeneity and a prediction interval from 1.6 to 36. For treatment decisions, DYNAMIC supports considering ctDNA-guided de-escalation in stage II colon cancer, but DYNAMIC-III did not show non-inferiority for de-escalation in stage III. No evidence supports routinely escalating therapy for ctDNA-positive patients.
The distinction matters because a positive ctDNA is alarming and tempting to act on. In practice, it is a strong risk marker; using it to withhold or intensify chemotherapy is supported only in stage II.
- Use ctDNA as a risk marker, not a treatment switch, in stage III colon cancer
- Consider ctDNA-guided omission of adjuvant chemotherapy only in stage II, as in DYNAMIC
- Do not escalate therapy for a positive ctDNA outside a trial
- Offer trials for ctDNA-positive patients where available
Why it matters
A test that predicts recurrence well has not been shown to tell clinicians what to do in most patients.
Don't overread it
A large hazard ratio for prognosis does not show that treatment decisions based on ctDNA improve survival.
The statistics, in plain English
A hazard ratio of 7.6 is a very strong association. But I² of 81% means studies disagreed a lot, and the prediction interval, from 1.6 to 36, is the range a new study would likely show. Prognostic accuracy and decision utility are different questions, answered by cohort studies and randomised trials respectively.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free