Detection of minimal residual disease in colorectal cancer patients after surgery through circulating tumor DNA profiling

Accurate identification of minimal residual disease (MRD) after curative-intent resection remains a critical unmet need in colorectal cancer (CRC). Given the narrow postoperative window during which current adjuvant chemotherapy (ACT) improves outcomes, rapid and reliable biomarkers are essential to guide timely treatment decisions. Here, we assess the prognostic performance of a targeted liquid-biopsy mutation panel for postoperative ctDNA detection to identify MRD and support ACT decision-making in CRC. We performed targeted sequencing using a Thermo Fisher Scientific NGS platform (Ion S5 Oncomine based panel) to assess more than 200 hotspot mutations across key CRC-associated genes. All samples were sequenced to a mean depth of ~50,000× using 10-20 ng of cfDNA, enabling detection of variants with a VAF >0.05%. ctDNA was evaluated before and after treatment in 49 patients with resectable stage I-III CRC (32 colon, 17 rectal), with blood samples collected at baseline and again 4 weeks after surgery or neoadjuvant therapy. Baseline ctDNA alterations were detected in 75% of patients (37/49), with higher detection rates in stage III (78%; 21/27) compared with stage II (69%; 11/16) and stage I (67%; 4/6). Across all baseline-positive samples, 76 mutations were identified in 10 genes, predominantly TP53 (n=26), APC (n=14), KRAS (n=10), and PIK3CA (n=6), with a median VAF of 0.97% (range 0.05-17.9%). Post-treatment ctDNA positivity was observed in 57% of patients (28/49), indicating persistent MRD. A reduction in VAF after treatment occurred in 30% (15/49), including 9 patients with >50% decreases, consistent with a partial molecular response but ongoing subclinical disease. Conversely, 27% (13/49) showed increased VAF, suggesting inadequate treatment response and a higher risk of relapse. Complete ctDNA clearance occurred in 24% of patients (12/49), 8 with colon cancer and 4 with rectal cancer, all of whom remain event-free for at least 12 months. Up to date, one rectal cancer patient remains ctDNA-negative for 18 months. Finally, 18% of patients had no detectable mutations at either time point, underscoring the need for further assay refinement to improve detection sensitivity in this subgroup. In conclusion, postoperative ctDNA positivity identified patients with persistent molecular disease, while ctDNA clearance correlated with short-term event-free outcomes. These findings support the use of ctDNA profiling as a practical tool to guide postoperative treatment decisions.