ctDNA-Guided Adjuvant Chemotherapy Significantly Improves DFS, TTR in Stage II Colon Cancer

Administering adjuvant chemotherapy to patients with detectable postoperative circulating tumor DNA (ctDNA) produced significant, clinically meaningful reductions in disease recurrence and improvements in disease-free survival (DFS) for patients with mismatch repair–proficient (pMMR), microsatellite stable (MSS) stage II colon cancer from the per-protocol analysis of the phase 3 CIRCULATE trial (NCT04089631), according to results shared at the 2026 ASCO Annual Meeting.¹

In the per-protocol analysis, which compared 21 ctDNA-positive patients who received chemotherapy with 15 ctDNA-positive patients randomly assigned to observation, chemotherapy reduced the 3-year recurrence rate from 62% to 19% (HR, 0.23; 95% CI, 0.06-0.87; P = .009) and improved the 3-year DFS rate from 38% to 77% (HR, 0.31; 95% CI, 0.09-1.03; P = .021). For context, the 3-year DFS in the ctDNA-negative population was 87%.

In the preplanned primary intention-to-treat (ITT) DFS analysis, which included all 26 patients randomly assigned to chemotherapy vs 15 randomly assigned to observation, the 3-year DFS rates were 61% vs 38%, respectively (HR, 0.55; 95% CI, 0.21-1.48; P = .12). The 3-year recurrence rate in the ITT group was 35% with chemotherapy vs 62% with observation (HR, 0.48; 95% CI, 0.17-1.33; P = .075). Although this primary end point was formally negative, the observed HR remained consistent with the prespecified clinically relevant effect of 0.62, with early closure limiting statistical power rather than the magnitude of the observed effect, investigators noted.

"Together with the existing literature, this data justifies a ctDNA-guided treatment-escalation strategy in pMMR stage II colon cancer. If you imagine this would be a new drug with a HR of less than 0.5 in a subset of 3% of patients at a drug cost of zero, I think we can also regard it as a success."
—Gunnar Folprecht, MD, lead presenter and head of the outpatient unit for medical oncology at the University Hospital Carl Gustav Carus, Dresden University of Technology, Dresden, in Germany.

What prior evidence supports the prognostic value of postoperative ctDNA in this setting?

A landmark prospective study published in 2016 first demonstrated that ctDNA detection after stage II colon cancer resection provides direct evidence of residual disease and identifies patients at very high risk of recurrence, with recurrence occurring in only 9.8% of 164 patients with negative ctDNA (HR, 18; 95% CI, 7.9-40; P < 0.001).² The observational GALAXY study (UMIN000039205), conducted as part of the CIRCULATE-Japan platform, subsequently confirmed that ctDNA-based detection of molecular residual disease was predictive of survival outcomes and benefit with adjuvant chemotherapy in a large cohort of patients with resectable colorectal cancer.³

Observational data from GALAXY and the prospective, observational BESPOKE CRC registry (NCT04264702) have also suggested a clinical benefit with adjuvant chemotherapy among ctDNA-positive patients.^3,4 Moreover, data from the primary analysis of the DYNAMIC study (ACTRN12615000381583) demonstrated that ctDNA-guided management was noninferior to standard clinicopathologic risk–based management while reducing chemotherapy utilization from 28% in the standard management group (n = 153) to 15% in the ctDNA-guided group (n = 302; relative risk, 1.82; 95% CI, 1.25-2.65).⁵ However, prior to CIRCULATE, no prospective randomized data evaluating ctDNA-guided treatment escalation in stage II disease were available.¹

How was CIRCULATE designed?

CIRCULATE was a prospective, randomized trial enrolling patients with stage II colon cancer whose tumors were not mismatch repair–deficient or microsatellite instability–high; T4 tumors and emergency resections were permitted if not already considered an indication for adjuvant therapy. ctDNA testing was performed using a tumor-informed academic assay tracking up to 71 genes, with a 71- or 78-gene FFPE panel requiring at least 2 of 3 replicates to detect a variant allele frequency (VAF) of 0.1% or greater to confirm positivity.

Following central ctDNA analysis, patients were assigned to 1 of 3 pathways:

• ctDNA-positive patients (n = 41): randomly assigned 2:1 to adjuvant chemotherapy (capecitabine 1250 mg/m², 6 cycles, or investigator's choice of oxaliplatin combination) vs observation. Patients in the observation arm remained blinded to their ctDNA result.
• ctDNA-negative patients randomized within the trial (n = 287): assigned 1:4 to observation within the trial.
• ctDNA-negative patients off-study (n = 1083): observed outside the trial.

The primary end point was DFS with chemotherapy vs observation in ctDNA-positive patients, tested within a 1-sided alpha of 0.021.

The trial was designed to include 231 ctDNA-positive patients to detect an HR of 0.62 with 154 planned events. Patients were enrolled across approximately 140 centers in Germany and Austria. Recruitment began in June 2020 and closed early in July 2025 when funding was not prolonged; the last follow-up was January 2026.

What were the baseline patient characteristics?

Of 2,126 patients screened, 1,722 underwent ctDNA analysis, and 1,396 underwent random assignment. The ctDNA positivity rate among these patients was 2.9%. Of note, an additional 4.3% of ctDNA-positive patients were identified among those not randomly assigned. Plasma samples were taken at a median of 21 days (IQR, 13-28) after resection.

Patient characteristics were generally well balanced between the chemotherapy and observation/ctDNA-positive arms and were as follows:

• Median age: 66.5 years in the chemotherapy arm vs 66.0 years in the observation/ctDNA-positive arm
• High clinical risk group (pT4 or emergency resection): 3.8% vs 27% (P = .05)
• pT4 tumor stage: 3.8% vs 20% (P = .04)
• Male sex: 46% vs 47%
• Emergency resection: 0% vs 13%
• High-grade histology: 7.7% vs 13%

The overall population was predominantly low-risk, with fewer than 10% having pT4 tumors or emergency resections. Despite stratification, patients in the observation arm had slightly more pT4 tumors, placing them at a somewhat higher baseline risk relative to the chemotherapy arm.

What did ctDNA sampling and testing data show?

The tumor-informed assay tracked a median of 2 variants per patient (range, 1-3). In ctDNA-positive patients, a median of 1 variant was detected (range, 1-2), at a median VAF of 0.089% (range, 0.011-26.6%). Of the 41 ctDNA-positive patients in the randomized group, 33 were confirmed positive directly; another 8 were confirmed positive after repeat testing of equivocal results. The timing of sampling did not influence positivity rates, though cell-free DNA concentrations were elevated in the first week after resection. At the 3-year landmark, the assay showed a sensitivity of 31% (95% CI, 19%-44%) and a specificity of 98% (95% CI, 95%-99%), reflecting the assay's high-specificity design, which was intentional for an escalation trial, Folprecht explained.

What did additional analyses indicate about the prognostic value of postoperative ctDNA status?

In the observation arm, ctDNA-positive patients (n = 15) had 3-year DFS and OS rates of 38% and 84%, respectively, compared with 87% and 98% in ctDNA-negative patients (n = 272). The DFS HR was 6.32 (95% CI, 2.88-13.9; P < .001) and the OS HR was 7.14 (95% CI, 1.68-30.5; P = .001), confirming ctDNA as a highly prognostic marker in this population.

An exploratory "as-treated" analysis was conducted to provide context for comparison with existing observational studies. All randomly assigned, untreated, ctDNA-positive patients across both arms (n = 20) were pooled and compared with treated patients (n = 21). This analysis yielded an HR for treatment effect of 0.24 (P = .001) and a DFS HR of 8.10 (P < .001) for the prognostic effect of ctDNA positivity in all untreated patients vs ctDNA-negative patients. These findings were consistent with, and slightly exceeded, data from observational studies such as GALAXY and BESPOKE.

What were the treatment exposure and safety findings?

Of 26 patients randomly assigned to chemotherapy, 21 started treatment. The remaining 5 did not receive chemotherapy for the following reasons: 2 had metastatic disease detected on imaging triggered by unblinding, and 3 withdrew informed consent.

Among the 21 treated patients, 67% received capecitabine monotherapy and 33% received an oxaliplatin combination. The median number of cycles was 6 (IQR, 3-8). Any toxicity occurred in 95% of patients, and grade 3 or higher toxicity was observed in 48%. One treatment-related death occurred. No new or unexpected toxicity signals were identified.

What trial limitations were noted?

The investigators identified 3 main categories of limitations. First was the low ctDNA positivity rate of 2.9%. This was attributed to the assay's conservative, high-specificity design from a decade ago, the limited gene panel size, the single postoperative sampling time point, the predominantly low-risk population enrolled, and the loss of early progressors before randomization.

Second, 5 patients in the chemotherapy arm (n = 26) did not receive treatment because metastases were found on imaging prompted by unblinding (n = 2) and consent withdrawal (n = 3), motivating the prespecified per-protocol analysis. Third, early trial closure due to funding expiry resulted in wide confidence intervals. As Folprecht emphasized, this limitation constrains statistical power but does not diminish the observed effect size.

Disclosures: Folprecht reported serving in a consulting or advisory role for Amgen, AstraZeneca, BeiGene, Bristol-Myers Squibb, Daiichi Sankyo, GlaxoSmithKline, Incyte, Merck KGaA, Merck Sharp & Dohme, Roche/Genentech, Servier, and Takeda; serving on a speakers' bureau for AbbVie and Incyte; and travel accommodations and expenses paid by AstraZeneca, Merck Serono, Pierre Fabre, Roche, and Takeda.

References

1. Folprecht G, Sill MW, Beffa L, et al. CIRCULATE: a prospective randomized trial of ctDNA-guided adjuvant chemotherapy in mismatch repair–proficient stage II colon cancer. J Clin Oncol. 2026;44(17 suppl):LBA3500. doi:10.1200/JCO.2026.44.17_suppl.LBA3500
2. Tie J, Wang Y, Tomasetti C, et al. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer. Sci Transl Med. 2016;8(346):346ra92. doi:10.1126/scitranslmed.aaf6219
3. Nakamura Y, Watanabe J, Akazawa N, et al. ctDNA-based molecular residual disease and survival in resectable colorectal cancer. Nat Med. 2024;30(11):3272-3283. doi:10.1038/s41591-024-03254-6
4. Shah MA, Tie J, Raghav KPS, et al. ctDNA positivity post-resection as a predictor of recurrence and adjuvant chemotherapy benefit in patients with stage II colon cancer: BESPOKE CRC. J Clin Oncol. 2025;43(15 suppl):LBA3500. doi:10.1200/JCO.2025.43.15_suppl.LBA3500
5. Tie J, Cohen JD, Lahouel K, et al. Circulating tumor DNA analysis guiding adjuvant therapy in stage II colon cancer. N Engl J Med. 2022;386(24):2261-2272. doi:10.1056/NEJMoa2200075