Key Points:

• Circulating tumor DNA (ctDNA) is considered to be a surrogate marker for minimal residual disease (MRD) and a predictive marker to determine the risk of colon cancer recurrence.
• Patients with colon cancer with detectable ctDNA post–curative-intent therapy (MRD-positive) are at a remarkable risk of recurrence (95% to 100%) if not offered systemic therapy, which suggests that MRD positivity is not just a high-risk marker for recurrence but also a measure of persistent disease.
• ctDNA-guided management can be considered to identify the cohort of patients with stage II colon cancer who may not benefit from adjuvant therapy, based on the DYNAMIC study.

Surgery followed by adjuvant chemotherapy based on clinical and pathologic risk stratification is the recommended approach for all patients with stage II (ie, high-risk stage II) and stage III colon cancer. Despite this approach, approximately 15% and 30% of patients with stage II and III disease, respectively, may experience recurrence, while 40% to 50% of the patients are cured by surgery alone.^1-3 Biomarker-guided follow-up and personalized treatment are vital for patients with resected colon cancer to aid in clinical decision-making and appropriate risk stratification.

Recent Data Support the Predictive and Prognostic Utility of ctDNA as a Biomarker

ctDNA is considered a surrogate marker for MRD and a predictive marker to determine the risk of recurrence. Tumor-informed and tumor-agnostic testing strategies are currently available for ctDNA assessment.³ Both have demonstrated improved specificity and sensitivity of ctDNA as a strong predictive and prognostic biomarker compared with carcinoembryonic antigen across multiple studies.^4-8 There is strong evidence to suggest that MRD status outperforms known clinicopathologic factors in predicting relapse. However, it is important to note that improved sensitivity and specificity are associated with serial testing of ctDNA, while single timepoint testing has similar sensitivity and specificity to carcinoembryonic antigen. Patients with colon cancer with detectable ctDNA post–curative-intent therapy, defined as MRD-positive, are at a remarkable risk of recurrence (95% to 100%) if not offered systemic therapy, which suggests that MRD positivity is not just a high-risk marker for recurrence but also a measure of persistent disease. Furthermore, the growth rate of ctDNA is also prognostic of outcomes, with exponential growth associated with significantly inferior survival outcomes compared with a slow rise in ctDNA.⁷ The optimal timing of ctDNA is essential for a reliable test with improved sensitivity. Historically, ctDNA testing was advised to be performed at least 4 weeks following surgery, as the increased shed of cell-free DNA in the immediate postoperative period could lead to false-negative results. However, a real-world analysis of patients who received commercial, tumor-informed ctDNA testing at 2 to 4 weeks postoperatively demonstrated similar sensitivity for MRD detection compared with 4 to 8 weeks postoperatively despite the presence of detectable cell-free DNA levels.⁹

The updated results of the GALAXY arm of the CIRCULATE-Japan study, which prospectively followed serial ctDNA in patients with stage I-III colorectal cancer (CRC), were presented at the European Society for Medical Oncology Congress 2023.¹⁰ Detectable postoperative ctDNA was the most significant prognostic factor for disease recurrence (HR 10.44). Adjuvant chemotherapy prolonged disease-free survival (DFS) in patients who were MRD-positive compared with no therapy (38.6% vs 16.1%; P < .01). Furthermore, no significant difference in DFS at 2 years was observed in patients with ctDNA-negative stage I-III CRC, either with or without adjuvant chemotherapy (88.3% vs 89.9%; P = .156).

How Can ctDNA Be Used in Routine Practice to Personalize Treatment?

Prospective clinical trials on ctDNA-guided management of colon cancer using adjuvant therapy is the perceivable next step after establishing ctDNA as a strong biomarker (Table 1).^11-12 The DYNAMIC trial led by Tie et al compared tumor-informed, ctDNA-guided adjuvant treatment to standard treatment in patients with stage II colon cancer postoperatively.¹¹ The primary endpoint of recurrence-free survival (RFS) at 2 years with ctDNA-directed management was noninferior to standard management, reminiscent of the fact that a ctDNA-guided approach can perhaps limit the use of adjuvant chemotherapy without compromising RFS in stage II colon cancer.¹¹ PEGASUS is a multicenter prospective trial that used ctDNA-guided (tumor-agnostic) adjuvant treatment for patients with high-risk stage II or stage III colon cancer.¹² Patients who were MRD-positive were treated with CAPOX for 3 months, while patients that were MRD-negative were treated with capecitabine for 6 months. Out of 135 patients, 35 patients had detectable postoperative ctDNA (26%). Patients with persistent ctDNA despite CAPOX received FOLFIRI. Seroconversion after CAPOX or FOLFIRI was observed in 14 out of 35 patients (40%) after a median follow-up of 21.2 months. Interestingly, patients with MRD-negative stage III and high-risk stage II disease demonstrated a low relapse rate of 7% despite de-escalated treatment with capecitabine. It is important to note that the false-positive test results associated with plasma-based assay can lead to overtreatment; 7 patients demonstrated persistent ctDNA positivity with no evidence of disease recurrence after the completion of the CAPOX and FOLFIRI treatment regimens.¹²

Can ctDNA Identify Patients Who May Not Derive Benefit From Adjuvant Therapy Without Compromising RFS?

Tumor-informed, ctDNA-guided treatment can be considered to help determine the cohort of patients with stage II colon cancer who may not benefit from adjuvant therapy based on existing prospective data. However, we need long-term follow-up and prospective confirmatory data from larger phase 3 trials to consolidate this evidence. Multiple prospective trials focused on ctDNA-guided management have been ongoing in the MRD space of adjuvant treatment of colon cancer (Table 2). Unfortunately, the COBRA study (NCT04068103) has been terminated because of higher-than-expected false-positive ctDNA results with tumor-agnostic assay, likely attributed to methylation and epigenomic markers that could have limited the sensitivity of the assay. An ideal assay should provide greater than 90% sensitivity and specificity, with minimal false-positive and -negative test results; this is onerous given the setting of limited ctDNA quantity across limited-stage (I, II, III) colon cancer.

[ctDNA] is on the path to revolutionize the adjuvant treatment of patients with colon cancer by minimizing the risk of both under- and overtreatment.

Ongoing Education and Clinical Trial Efforts Continue to Improve ctDNA Use

ctDNA continues to evolve as a tool to assess MRD and treatment response monitoring and is on the path to revolutionize the adjuvant treatment of patients with colon cancer by minimizing the risk of both under- and overtreatment. Ideal testing assay with excellent sensitivity and specificity; interpretation and application of testing results in the clinical context; patient anxiety associated with a positive test; availability; and health care reimbursement across the world are some of the challenges that remain toward its incorporation into clinical practice. Although both tumor-informed and -uninformed assays are available commercially, tumor-informed assay seems to have stronger data, with the availability of larger datasets to date, as well as improved sensitivity and specificity of the assay. Ongoing educational efforts through international organizations such as ASCO and the European Society for Medical Oncology, along with the involvement of patient advocate societies, continue to be immensely helpful in improving the interpretation and application of ctDNA testing. Ongoing clinical trials can help answer the question of whether ctDNA can be a surrogate marker for DFS and whether earlier initiation of treatment for patients at high risk for recurrence can lead to prolonged survival outcomes.