CtDNA-guided de-escalation of adjuvant chemotherapy in stage III colon cancer: early model-based evaluation of cost-effectiveness.

Background: All stage III colon cancer (CC) patients are recommended adjuvant chemotherapy (ACT) after surgery, while over half are already cured by surgery alone. The prognostic biomarker circulating tumor DNA (ctDNA) could potentially guide the decision on whom to withhold ACT. Objective: We assessed the cost-effectiveness of ctDNA-guided ACT de-escalation in stage III CC. Design: A decision model "PATTERN-stageIII" simulates CC from diagnosis till death in surgically treated stage III patients. Methods: We evaluated administering ACT in all patients (i.e., All ACT strategy), and three ACT de-escalation strategies by omitting ACT in patients who are (Strategy 1) both pT1-3N1 and ctDNA-negative, (Strategy 2) pT1-3N1, have no vascular invasion, and are ctDNA-negative, and (Strategy 3) pT1-2N1 and ctDNA-negative. For each strategy, costs, quality-adjusted life years (QALYs), and net monetary benefit were estimated. Sensitivity analyses assessed changes in ACT effectiveness and ctDNA-related parameters. Results: In de-escalation strategies 1, 2, and 3, respectively, 52%, 61%, and 88% of patients were predicted to receive ACT, thereby losing 0.322, 0.237, and 0.034 QALYs per person. The "All ACT strategy" was preferred in terms of cost-effectiveness. Sensitivity analyses demonstrated scenarios where ctDNA-guided de-escalation strategies were cost-effective compared to "All ACT," including improved ACT treatment effect in ctDNA-positive patients, higher ctDNA positivity rates, enhanced ctDNA prognostic value, and/or reduced ctDNA testing costs. Conclusion: ctDNA-guided strategies for ACT de-escalation are currently not cost-effective compared to an "All ACT strategy." One explanation is the non-negligible recurrence risk in ctDNA-negative patients. ctDNA-guided strategies could potentially become cost-effective if more than two ctDNA-related parameters improve simultaneously.