ctDNA-guided adjuvant chemo in resected stage II colon cancer

When ctDNA-positivity justifies escalation and ctDNA-negativity justifies de-escalation of adjuvant chemotherapy after resection of stage II colon cancer, traceable to the trial spine (CIRCULATE-Japan, DYNAMIC, GALAXY, BESPOKE-CRC).

(colon cancer) AND (circulating tumor DNA OR ctDNA OR minimal residual disease) AND adjuvant AND stage II

Read the manuscript projection →

Active nodes

• anchor nod_ca77eb54-7919-407b-b3b1-4cb0f7d13c0d Tie J, Cohen JD, Lahouel K, et al. "Circulating Tumor DNA Analysis Guiding Adjuvant Therapy in Stage II Colon Cancer" (the DYNAMIC trial). N Engl J Med. 2022;386(24):2261-2272. A randomized controlled trial testing a ctDNA-guided approach to adjuvant chemotherapy decisions in stage II colon cancer: post-surgical ctDNA status (assayed at weeks 4 and 7) was used to direct whether adjuvant chemotherapy was given, versus standard clinicopathologic management. The landmark prospective evidence for the sub-topic's central question of whether MRD-by-ctDNA can safely de-escalate adjuvant therapy in stage II disease.
• anchor nod_6055e3f3-efa0-4781-b965-58d9c0d3e28e Nakamura Y, Watanabe J, Akazawa N, et al. "ctDNA-based molecular residual disease and survival in resectable colorectal cancer" (the CIRCULATE-Japan GALAXY observational study). Nat Med. 2024;30(11):3272-3283. A large prospective multicenter observational registry (n=2,240; stage II-III colon cancer and stage IV colorectal cancer undergoing curative-intent resection) measuring tumor-informed ctDNA across the molecular-residual-disease (MRD) window and during/after adjuvant chemotherapy. ctDNA positivity in the MRD window was associated with markedly inferior disease-free and overall survival, and sustained versus transient ctDNA clearance under adjuvant chemotherapy stratified outcomes. The principal observational evidence in the sub-topic's trial spine complementing the DYNAMIC randomized trial: it characterizes the prognostic and predictive behavior of ctDNA/MRD that motivates ctDNA-guided escalation (on positivity) and de-escalation (on sustained clearance/negativity) of adjuvant therapy. Note the population is broader than resected stage II colon cancer (it includes stage III and stage IV); its bearing on the sub-topic's stage II question is via the MRD biology it establishes, not a stage II-specific randomized comparison.
• excerpt nod_3eff4ffc-dcba-437b-aaa0-0bb7d999e2f8 From GALAXY (Nakamura et al., Nat Med 2024;30:3272-3283; PMID 39284954), the updated 23-month-follow-up analysis (n=2,240; stage II-III colon and stage IV CRC) reports that sustained ctDNA clearance under adjuvant chemotherapy (ACT), versus transient clearance, is associated with a dramatic separation in 24-month outcomes: 24-month DFS 89.0% vs 3.3% and 24-month OS 100.0% vs 82.3%. This is the central empirical observation in GALAXY that bears on the sub-topic's question: clearance dynamics under ACT — not just baseline MRD status — stratify outcomes, supplying the observational rationale for ctDNA-guided escalation (when clearance fails to sustain) and de-escalation (when sustained negativity is achieved/maintained) of adjuvant therapy. Caveat for the stage II colon focus: this clearance contrast is reported pooled across the GALAXY cohort (stage II-III colon + stage IV CRC), so its applicability to the sub-topic's specific population is via MRD biology rather than a stage-II-restricted estimate.
• excerpt nod_11726864-69d3-4156-bd40-c5d641ed2083 DYNAMIC's pre-specified de-escalation finding: in stage II colon cancer randomized 2:1 to ctDNA-guided vs standard clinicopathological management (N=455; ctDNA assayed at weeks 4/7 post-resection), adjuvant chemotherapy use was substantially reduced in the ctDNA-guided arm (15% vs 28%) while 2-year recurrence-free survival met the pre-specified noninferiority criterion (lower 95% CI bound −4.1 percentage points vs noninferiority margin of −8.5 pp). This is the strongest randomized evidence directly addressing the sub-topic's central question — whether ctDNA-negativity in the MRD window justifies de-escalation of adjuvant chemotherapy after resection of stage II colon cancer. Scope match is exact (stage II colon cancer, post-resection MRD window via weeks 4/7 ctDNA, adjuvant decision as outcome of guidance), so unlike the GALAXY/BESPOKE observational anchors, no population caveat is needed. Note: the chemotherapy-use difference is the secondary endpoint that operationalizes "de-escalation"; the noninferior RFS is the primary endpoint that licenses it. Both pieces are needed to interpret the trial's bearing.
• anchor nod_0ee5b925-e844-4dd9-96cd-e18e80680db5 Kasi PM, Sawyer S, Guilford J, et al. "BESPOKE study protocol: a multicentre, prospective observational study to evaluate the impact of circulating tumour DNA guided therapy on patients with colorectal cancer." BMJ Open. 2021;11(9):e047831. The peer-reviewed design paper for BESPOKE CRC (ClinicalTrials.gov NCT04264702), a multicenter prospective observational study of stage II/III/IV colorectal cancer using a tumor-informed ctDNA assay to characterize the impact of ctDNA-guided decisions on adjuvant chemotherapy choice and downstream recurrence outcomes. Establishes the design that the sub-topic's spine names alongside CIRCULATE-Japan, DYNAMIC, and GALAXY: post-resection ctDNA assessment in the MRD window plus longitudinal surveillance, with adjuvant chemotherapy decisions left to clinicians and ctDNA dynamics treated as the prognostic/predictive readout. Notes on bearing: this is the protocol paper, not the trial readout; the interim analysis is currently in abstract form (JCO 2024;42(3 suppl):9, ASCO GI) and is expected to be superseded by a full primary publication, at which point a propose_supersedes is the natural follow-up. Like CIRCULATE-Japan/GALAXY, BESPOKE-CRC is observational rather than randomized, and its population is broader than resected stage II colon cancer (includes stage III and stage IV); its bearing on the sub-topic's stage II question is via the ctDNA-MRD biology and the observed clinician/patient adjuvant-decision behavior it documents, not a stage II-specific randomized comparison.
• excerpt nod_cc8f4dd3-6e2b-40a4-b1b5-537033bc98e7 BESPOKE-CRC protocol prespecifies adjuvant-treatment decisions as a primary endpoint: the study will observe how tumor-informed ctDNA testing changes clinicians' adjuvant chemotherapy choices (not a randomized de-escalation/escalation mandate like DYNAMIC). For the sub-topic, this anchor documents the observational spine's design intent—linking post-resection ctDNA dynamics to real-world adjuvant use and asymptomatic recurrence detection across stage I–IV CRC—while a secondary endpoint explicitly compares survival in MRD-negative patients who receive adjuvant chemotherapy versus active surveillance, which is the closest protocol-level analogue to the stage II de-escalation question (but not stage II–specific and not randomized). Population caveat remains: enrollment is stage I–IV and observational, so bearing on resected stage II colon cancer is inferential pending readout/supersede of interim abstracts.
• synthesis nod_d3956d56-a89d-4a34-8b1c-bf32ea235890 Across the trial spine, two complementary evidence types address ctDNA-guided adjuvant modulation after curative-intent resection. DYNAMIC (stage II colon cancer RCT) directly licenses de-escalation when ctDNA is negative in the post-resection MRD window: fewer patients received adjuvant chemotherapy (15% vs 28%) without compromising 2-year RFS under a prespecified noninferiority design. GALAXY (observational CIRCULATE-Japan registry; stage II–III colon and stage IV CRC) does not randomize treatment but shows that ctDNA clearance dynamics under adjuvant chemotherapy stratify outcomes sharply—sustained versus transient clearance separates 24-month DFS/OS—supporting the biological premise that ctDNA negativity/clearance marks a lower-risk subgroup where intensified or prolonged therapy may be unnecessary, while failure to clear marks escalation need. Joint reading for the sub-topic: DYNAMIC supplies the stage II–specific causal policy test for withholding chemotherapy when MRD-negative; GALAXY supplies mechanistic/prognostic reinforcement and escalation logic when MRD persists or clears only transiently, with explicit population caveats because GALAXY is not stage II–restricted. Neither node alone answers both de-escalation licensure and escalation rationale; together they frame ctDNA-guided adjuvant decisions as bidirectional (omit when negative/sustainedly clear; treat when positive or poorly clearing) pending further stage II–restricted observational/randomized readouts (e.g., BESPOKE-CRC, ongoing RCTs).
• anchor nod_e898b202-5300-4388-bbd6-7bd389992b74 "A randomized, double-blind, phase III study comparing trifluridine/tipiracil (FTD/TPI) versus placebo in patients with molecular residual disease following curative resection of colorectal cancer (CRC): The ALTAIR study." J Clin Oncol. 2025;43(4 suppl):LBA22 (CIRCULATE-Japan / EPOC1905; late-breaking abstract, ASCO GI 2025). The escalation arm of the CIRCULATE-Japan platform: patients ctDNA/MRD-positive after curative resection (plus standard adjuvant where applicable) with no radiologic recurrence were randomized to 6 months of FTD/TPI vs placebo. FTD/TPI produced a numerical but NOT statistically significant disease-free survival improvement (median DFS 9.30 vs 5.55 months; HR 0.79, 95% CI 0.60-1.05; P=.107). It bears on the escalation side of the sub-topic's question — what to do once ctDNA is positive — by testing a specific escalation agent in the MRD-positive state; the not-significant result tempers the assumption that escalation in MRD-positive disease yields durable benefit. Caveats: the population is resected CRC broadly (includes resected stage IV, not stage II-restricted), this is a late-breaking abstract pending full publication, and FTD/TPI is not a standard adjuvant regimen — so the bearing on resected stage II colon is via the general MRD-positive escalation principle, not a stage II-specific result.
• anchor nod_39527581-3bc2-4702-b068-4cece6c67f3d Baxter NN, Kennedy EB, Bergsland E, et al. "Adjuvant Therapy for Stage II Colon Cancer: ASCO Guideline Update." J Clin Oncol. 2022;40(8):892-910. doi:10.1200/JCO.21.02538. The current ASCO clinical practice guideline governing the exact decision this sub-topic interrogates: whether/how to give adjuvant chemotherapy after resection of stage II colon cancer. Its directly load-bearing statement for the sub-topic is that circulating tumor DNA was identified by the Expert Panel as an emerging potential PREDICTIVE factor, but that insufficient evidence of its chemotherapy-predictive value was available to warrant including ctDNA in the guideline's list of high-risk features used to recommend adjuvant chemotherapy; the panel explicitly anticipated that prospective-trial data on ctDNA would be forthcoming and incorporated in a future guideline version. The surrounding recommendation framework is that routine adjuvant chemotherapy is NOT recommended for non-high-risk stage II colon cancer, with adjuvant therapy offered/discussed via shared decision-making for high-risk features (e.g., T4 disease), and that the absolute benefit of adjuvant chemotherapy in stage II is small. Bearing on the sub-topic: this node is distinct from every trial/registry/protocol node already in the graph (DYNAMIC, GALAXY/CIRCULATE-Japan, BESPOKE, COBRA, ALTAIR, DYNAMIC-III, CIRCULATE-US, PEGASUS, MEDOCC-CrEATE, Tie 2016) because it is the authoritative GUIDELINE baseline against which the trial spine's de-escalation/escalation evidence is to be judged. It supplies (a) the standard-of-care comparator that DYNAMIC's "standard clinicopathologic management" arm operationalizes, (b) documentary evidence that, as of the guideline, ctDNA was NOT yet endorsed for routine stage II adjuvant decision-making, establishing the contested-decision gap the sub-topic's evidence chain aims to close, and (c) the predictive-vs-prognostic framing that the observational (GALAXY/Kotani/BESPOKE) and randomized (DYNAMIC) nodes are marshalled to satisfy. Caveats: it is a guideline/consensus document, not primary evidence — its weight is normative/synthetic rather than a trial result; and it is a 2022 snapshot that predates the mature DYNAMIC 5-year readout and most of the spine's later readouts, so a future ASCO guideline update incorporating ctDNA is the natural propose_supersedes target.
• anchor nod_30c9bfa8-8353-449a-a4f3-4a5c8e359393 Schraa SJ, van Rooijen KL, van der Kruijssen DEW, et al. "Circulating tumor DNA guided adjuvant chemotherapy in stage II colon cancer (MEDOCC-CrEATE): study protocol for a trial within a cohort study." BMC Cancer. 2020;20(1):790. The peer-reviewed protocol for a Dutch randomized 'trial within cohort' (TwiCs) nested in the Prospective Dutch ColoRectal Cancer (PLCRC) cohort (NCT06434896), planning 1,320 stage II colon cancer patients who have NO guideline indication for adjuvant chemotherapy (i.e., the lower-risk stage II population in whom 15-20% still recur). Patients are randomized 1:1; in the experimental arm, post-resection plasma is tested for tumor-informed ctDNA (PGDx elio platform), and ctDNA-DETECTABLE patients are OFFERED adjuvant CAPOX (8 cycles), while ctDNA-negative experimental-arm patients and all control-arm patients follow standard guideline surveillance. The primary endpoint is the proportion of ctDNA-positive patients who accept adjuvant chemotherapy; secondary endpoints include 2- and 5-year recurrence, DFS, OS, time to recurrence, quality of life, and cost-effectiveness. Bearing on the sub-topic: MEDOCC-CrEATE is distinctive within the trial-spine landscape because it is a RANDOMIZED, exactly-in-scope (resected stage II colon cancer) test of the ESCALATION direction — treat-when-ctDNA-positive — which the existing graph's randomized, stage-II-exact evidence (DYNAMIC, nod_ca77eb54) does NOT test, since DYNAMIC randomizes only the omit-when-negative (de-escalation) policy. It therefore speaks directly to the gap the cause's staged open_question (prp_ac67c7bd) names: that the escalation arm of the sub-topic's bidirectional logic rests almost entirely on observational, broader-population sources (GALAXY, BESPOKE, Kotani 2023) where confounding-by-indication cannot be excluded, with no stage-II-restricted randomized escalation result. MEDOCC-CrEATE is the prospective, randomized, stage-II-colon-specific attempt to supply exactly that missing escalation evidence in a guideline-low-risk population, using a different (PGDx, non-Signatera) tumor-informed assay and a different health system (Netherlands) than the Japanese/Australian/US spine members. Notes on bearing/limits: (a) this is the design/protocol paper, not a results readout — it supplies the prespecified randomized framework and hypotheses (primary endpoint is ACT uptake, a behavioral feasibility measure, not RFS/OS), so the AACR 2024 feasibility abstract (Cancer Res 2024;84(6 suppl):5022) and AACR 2025 update (Cancer Res 2025;85(8 suppl 2):CT221) and any eventual recurrence/survival readout are the natural propose_anchor + propose_supersedes follow-ups; (b) scope match on population is exact (resected stage II colon cancer) but it specifically targets the guideline-low-risk subset without a standard ACT indication, so it does not address high-risk stage II in the way PEGASUS/CIRCULATE-US (IIB/IIC) do; (c) the TwiCs design means the control arm is never informed of randomization and the experimental-arm intervention is an OFFER of chemotherapy, so the primary readout measures real-world acceptance under randomized invitation rather than a per-protocol treatment-effect estimate. It complements rather than duplicates the DYNAMIC (de-escalation), GALAXY/BESPOKE/Kotani (observational MRD biology and escalation rationale), and CIRCULATE-Japan/CIRCULATE-US/PEGASUS (high-risk-stage-II/stage-III platform) nodes by being the stage-II-low-risk, randomized, escalation-direction trial that none of those supply.
• anchor nod_7e35660f-6232-49e8-8e9f-ae89dad432c7 Kotani D, Oki E, Nakamura Y, et al. "Molecular residual disease and efficacy of adjuvant chemotherapy in patients with colorectal cancer" (the CIRCULATE-Japan GALAXY observational study). Nat Med. 2023;29(1):127-134. The earlier, predictive-interaction GALAXY analysis (n=1,039; stage II-IV resectable CRC; UMIN000039205): beyond confirming that ctDNA positivity in the molecular-residual-disease (MRD) window marks markedly inferior disease-free survival, it reports that adjuvant chemotherapy benefit is concentrated in MRD-positive patients — among ctDNA-positive (high-risk stage II/III) patients adjuvant chemotherapy was associated with significantly better DFS, whereas ctDNA-negative patients derived no statistically significant DFS benefit from adjuvant chemotherapy. This is the predictive (treatment-effect-modifying), not merely prognostic, observation that underwrites the sub-topic's bidirectional logic: ctDNA positivity flags a subgroup that benefits from escalation/treatment, while ctDNA negativity flags a subgroup in whom adjuvant chemotherapy adds little — the observational counterpart to DYNAMIC's randomized de-escalation result. It complements rather than duplicates the already-anchored GALAXY readout (Nakamura et al., Nat Med 2024;30:3272-3283, nod_6055e3f3), which characterizes sustained-versus-transient clearance dynamics under chemotherapy at longer follow-up; this Kotani 2023 node is the ACT-efficacy-by-MRD-status analysis. Population caveat, as with the other CIRCULATE-Japan/GALAXY and BESPOKE nodes: the cohort is broader than resected stage II colon cancer (it spans stage II-IV and the predictive subgroup analysis pools high-risk stage II with stage III), and the design is observational rather than randomized, so its bearing on the sub-topic's stage II question is via the MRD biology and the treatment-by-MRD interaction it establishes, not a stage II-specific randomized comparison.
• anchor nod_434df1f7-8718-4e7a-87e5-360a96fe6691 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. The foundational prospective cohort (n=230 resected stage II colon cancer; 1,046 plasma samples) establishing that post-surgical ctDNA detection in the molecular-residual-disease (MRD) window identifies the subgroup at highest recurrence risk and could inform adjuvant decisions. This is the historical root of the sub-topic's entire evidence line: it supplied the prognostic signal — ctDNA-positive post-resection patients recur far more often than ctDNA-negative — that the DYNAMIC randomized trial (nod_ca77eb54) later converted into a prospective de-escalation policy test. Scope match is exact (stage II colon cancer, post-resection ctDNA in the MRD window, recurrence as the outcome), so no population caveat is needed; the design caveat is that this is an observational/prognostic cohort, not a treatment-randomized comparison — it motivates ctDNA-guided adjuvant decisions without itself testing them.
• anchor nod_a629b95c-23fd-4931-8a6b-df94b5b7c19a Taniguchi H, Nakamura Y, Kotani D, et al. "CIRCULATE-Japan: Circulating tumor DNA-guided adaptive platform trials to refine adjuvant therapy for colorectal cancer." Cancer Sci. 2021;112(7):2915-2920. The umbrella protocol/design paper for the CIRCULATE-Japan platform that the sub-topic's spine names directly. It defines the three linked arms built off a shared ctDNA-screening registry: GALAXY (prospective observation, anchored at nod_6055e3f3), VEGA (randomized de-escalation — postoperative surgery-alone vs CAPOX in ctDNA-negative high-risk stage II / low-risk stage III colon cancer), and ALTAIR (randomized escalation — trifluridine/tipiracil vs placebo in ctDNA/MRD-positive resected CRC). It establishes the adaptive-platform logic by which post-resection ctDNA status routes patients toward de-escalation (if negative) or escalation (if positive) — the structural backbone of the sub-topic's question. Design caveat: the platform spans stage II-IV CRC and this node is the protocol, not a readout; its bearing on the sub-topic's resected stage II colon question is structural (it frames how the GALAXY/VEGA/ALTAIR evidence is generated), with the individual arm readouts being the natural propose_anchor/propose_supersedes follow-ups as they publish.
• anchor nod_8bea9d72-b88b-4b3d-aa3f-b4cb18d02bf5 "Phase II results of circulating tumor DNA as a predictive biomarker in adjuvant chemotherapy in patients with stage II colon cancer: NRG-GI005 (COBRA) phase II/III study." J Clin Oncol. 2024;42(3 suppl):5 (NRG Oncology; presented at ASCO GI 2024). A randomized US cooperative-group trial in resected low-risk stage IIA colon cancer testing whether ctDNA can select patients for adjuvant chemotherapy. The prespecified phase II analysis was NEGATIVE and the trial was halted under its stopping rules: among the initial ctDNA-positive patients, adjuvant chemotherapy did not increase ctDNA clearance at 6 months relative to observation (clearance 1/9, 11% in the chemotherapy arm vs 3/7, 43% in the control arm). This is directly in scope (stage IIA colon, post-resection ctDNA, adjuvant decision) and an important counterweight to de-escalation enthusiasm: where DYNAMIC (nod_ca77eb54) showed ctDNA-guided withholding of chemotherapy was noninferior, COBRA found no evidence that the converse — giving chemotherapy to ctDNA-positive low-risk stage II patients — clears MRD, cautioning against assuming the escalation direction benefits low-risk stage II disease. Caveats: this is an interim/abstract readout with very small ctDNA-positive numbers at the analysis, the surrogate endpoint was ctDNA clearance (not RFS/OS), and it used the trial's original pre-amendment assay; a full publication is the natural supersede target.

Frontier — work to be done

• needs review prp_0ee3ada8-a500-4238-82a4-a340eab9a5b7
• needs review prp_3defdd86-4920-41d9-b625-c04c00404c61
• needs review prp_3f40d70f-9add-4cdd-80d8-689ec7cea78a
• needs review prp_45a0d93c-638b-42a5-bae0-dda15bf60a0e
• needs review prp_472befef-feb6-4d7b-b1c3-70f301c6e804
• needs review prp_47cf1393-cc3a-4f08-a454-414fec106efd
• needs review prp_55a5715e-9b3d-4f5b-93d7-8aea7490b66d
• needs review prp_564ac1c3-a998-475b-b57f-73198cdcd9b5
• needs review prp_580417af-6e4b-4875-bb20-03f6ab146fb6
• needs review prp_77dbf23b-5549-4d54-9caf-1c41be1356a8
• needs review prp_85ddbc00-d5cb-4344-a9fc-ad0a136da459
• needs review prp_9382a735-f101-479f-b0fb-622669949076
• needs review prp_93e8579a-b9e5-4587-bb3d-07aac9d4383b
• needs review prp_9bacfb64-50ca-4c24-86dc-c611b7225f08
• needs review prp_9c1621c8-0855-4998-9602-f3b628289b8a
• needs review prp_a0c5f19a-07ff-4b95-94af-0c7a9d86a342
• needs review prp_ac67c7bd-136a-4b3e-9da2-dc01a42ae3eb
• needs review prp_ca972b5a-b54a-4ad2-ae2b-caa9ab778114
• needs review prp_ccc053d7-970a-4d17-bb57-bbdd1b37f57a
• needs review prp_da3fc41b-9a49-4996-b8db-99cf84df2c8e
• needs review prp_e71be147-ffa9-4f71-8a40-13d16205c4c9
• needs review prp_ef9bbbcb-1af3-4c8d-af8c-5764143c7e32
• needs review prp_fc87886a-e4e3-4633-9bee-a2fb4256b47a