OnCo
key papersKey paper

GALAXY: tumour DNA in blood four weeks after bowel cancer surgery predicts relapse tenfold

In 1,039 Japanese patients with resected colorectal cancer, a positive Signatera test at week 4 carried a tenfold higher recurrence risk, and only ctDNA-positive patients appeared to benefit from adjuvant chemotherapy.

GALAXY is the prospective observational arm of the Japanese CIRCULATE platform. Patients with stage II-IV resectable colorectal cancer had serial tumour-informed ctDNA testing (Signatera) after surgery, with treatment at physician discretion.

ctDNA positivity at 4 weeks post-surgery was the strongest predictor of recurrence (HR about 10). Among ctDNA-positive patients, adjuvant chemotherapy was associated with markedly better disease-free survival (HR about 6.6 in favour of treatment), whereas ctDNA-negative patients showed no apparent benefit. Clearance of ctDNA during adjuvant therapy predicted good outcomes.

The study is the largest prospective MRD dataset in colorectal cancer and the basis for the randomised VEGA and ALTAIR trials.

Observational studyHas not changed practice yet1,039 participants
Authors
Kotani D, Oki E, Nakamura Y, et al.
Published
What it found
  • ctDNA positivity at week 4 associated with recurrence: HR 10.0 (95% CI 7.7-14.0)
  • 18-month DFS 38.4% for ctDNA-positive vs 90.5% for ctDNA-negative patients
  • Adjuvant chemotherapy associated with improved DFS in ctDNA-positive patients (HR 6.59, 95% CI 3.53-12.3) but not in ctDNA-negative patients
  • ctDNA clearance by week 24 predicted substantially better DFS than persistent positivity
What it means

The blood test stratifies risk far better than stage or pathology. It supports treating ctDNA-positive patients and suggests ctDNA-negative patients gain little from chemotherapy, but because treatment was not randomised the de-escalation claim needs the randomised trials that are now under way.

Be careful
  • Observational; treatment decisions were not randomised, so the chemotherapy benefit estimate is confounded
  • Short follow-up at initial publication (median about 17 months)
  • Japanese population and Japanese adjuvant practice; generalisability to other health systems is untested
  • Commercial sponsor (Natera) involvement in assay and analysis

Connected

16top