Timing of circulating tumor DNA analysis aimed at guiding adjuvant treatment in colorectal cancer.

Abstract

Background: Multiple clinical trials investigate circulating tumor DNA (ctDNA) to guide adjuvant chemotherapy (ACT) in colorectal cancer. Timely ACT initiation necessitates early ctDNA testing, but the impact of postoperative cell-free DNA (cfDNA) and ctDNA dynamics remains unclear, particularly with cost-reducing input caps employed in some assays. This study investigates ctDNA detection at day14 versus day30, comparing whole-sample analysis with capping the cfDNA input, and evaluates single and dual timepoint assessments for ACT allocation.

Patients and methods: In 2019-2023, 611 stage I-III colorectal cancer patients were enrolled. Blood was collected preoperatively, and postoperatively ~day14 and ~day30. The cfDNA levels were assessed using digital PCR, and ctDNA using tumor-informed digital PCR or targeted sequencing analyzing all cfDNA from 8mL of plasma.

Results: Despite elevated cfDNA in 85% of day14 samples, performance was comparable between the two timepoints (sensitivity 31% vs 32%; specificity both 98%). A 50ng cfDNA input cap reduced ctDNA detection probability, impacting 78% of day14 samples and 65% of day30 samples. At both timepoints, ctDNA detection was prognostic of recurrence (day14: HR=9.0, 95%CI 5.5-14.8; day30: HR=12.5, 95%CI 7.6-20.4). In 74% of ctDNA-positive recurrence patients, both samples had ctDNA detected. An increase in ctDNA level from day14 to day30 was associated with shorter time to recurrence (Pearson R=-0.63, P=0.003). Combining the timepoints would increase sensitivity (36%) and allow earlier ACT start in 80% of patients.

Conclusion: Early ctDNA sampling is feasible and highly prognostic. Supplemental later testing may improve sensitivity while allowing early ACT initiation for most ctDNA positive patients.