Background
Randomised trial (AIO-KRK-0217/ABCSG), Germany + Austria; 2126 screened, 1396 randomised. UICC stage II, pMMR/MSS colon cancer after resection; post-op ctDNA by academic tumour-informed NGS assay (2.9% ctDNA-positive). ctDNA+ randomised 2:1 to CHEMO vs observation (OBS); ctDNA− randomised 1:4 to OBS vs OFF-STUDY; ctDNA results not disclosed in OBS. Trial ended early due to funding expiry.
Results
Interventions and follow up: Arm A (ctDNA+): Adjuvant capecitabine ± oxaliplatin (n=26; 81% started therapy)
Arm B (ctDNA+): Observation
Primary endpoint: DFS in ctDNA-positive patients (one-sided log-rank)
mFollow up: NR in abstract
Results: Prognosis (ctDNA+ vs ctDNA−): 3-yr DFS 52% vs 87% (HR 4.28, 95% CI 2.32–7.93, P<.001); 3-yr OS 88% vs 98% (HR 5.48, 95% CI 1.64–18.28, P=.001)
DFS (ITT, CHEMO vs OBS): 3-yr 61% vs 38%; HR 0.55 (95% CI 0.21–1.48); P=.12 — primary endpoint not met
Recurrence (ITT): 3-yr 36% vs 62%; HR 0.48 (95% CI 0.17–1.33); P=.075
Per-protocol: 3-yr recurrence 19% vs 62% (HR 0.23, 95% CI 0.06–0.87, P=.009); 3-yr DFS 77% vs 38% (HR 0.31, 95% CI 0.09–1.03, P=.021)
Arm B (ctDNA+): Observation
Primary endpoint: DFS in ctDNA-positive patients (one-sided log-rank)
mFollow up: NR in abstract
Results: Prognosis (ctDNA+ vs ctDNA−): 3-yr DFS 52% vs 87% (HR 4.28, 95% CI 2.32–7.93, P<.001); 3-yr OS 88% vs 98% (HR 5.48, 95% CI 1.64–18.28, P=.001)
DFS (ITT, CHEMO vs OBS): 3-yr 61% vs 38%; HR 0.55 (95% CI 0.21–1.48); P=.12 — primary endpoint not met
Recurrence (ITT): 3-yr 36% vs 62%; HR 0.48 (95% CI 0.17–1.33); P=.075
Per-protocol: 3-yr recurrence 19% vs 62% (HR 0.23, 95% CI 0.06–0.87, P=.009); 3-yr DFS 77% vs 38% (HR 0.31, 95% CI 0.09–1.03, P=.021)
Adverse events
Safety: NR in abstract
Conclusions
ITT primary endpoint not met, likely from loss of power after premature closure (only 26 ctDNA+ patients assigned to chemotherapy). ctDNA positivity was strongly prognostic, and the per-protocol analysis suggests a large benefit of adjuvant chemotherapy in ctDNA-positive stage II colon cancer, supporting ctDNA-informed adjuvant decision-making.
Key Limitations
Very small randomised ctDNA+ cohort (n=26 to CHEMO); early termination for funding expiry; the benefit signal rests on per-protocol analysis (selection-bias prone); 19% of the CHEMO arm never started therapy. Academic assay developed a decade ago with 2.9% positivity — lower sensitivity than modern commercial tumour-informed assays.
Clinical Context
First randomised escalation data for ctDNA-positive stage II colon cancer; complements DYNAMIC (NEJM 2022, ctDNA-guided de-escalation). ctDNA-guided adjuvant therapy is not yet an NCCN/ESMO-mandated standard; confirmatory trials (CIRCULATE-US/NRG-GI008, BESPOKE, CIRCULATE-Japan ALTAIR) ongoing.