Quantification of circulating tumor DNA (ctDNA) using a methylation-based, tissue-free colorectal cancer (CRC) test for the detection of molecular residual disease (MRD).

Abstract

217 Background: MRD detection by measuring ctDNA has been shown to predict disease recurrence. We have previously demonstrated that a tissue-free ctDNA test (Latitude, Natera, Inc.) for MRD detection in CRC patients is prognostic and predictive of outcomes in both the post-operative and surveillance windows. Here, we provide the quantitative assessment of this assay to further refine its prognostic and predictive capabilities. Methods: The methylation-based, tissue-free ctDNA assay uses next-generation sequencing to query regions of the human genome that are differentially methylated in patients with CRC compared to cancer-free individuals. Plasma samples from Signatera-tested CRC patients (N=105) and from cancer-free individuals (N=223) were analyzed in a cross-validation setup. To evaluate the quantification performance of this assay, we compared ctDNA levels measured by the methylation-based assay (measured as the differential methylation allele fraction, DMAF) with those from a tumor-informed ctDNA assay (Signatera, measured as variant allele fraction, VAF), which served as the reference. The DMAF in plasma samples from cancer-free individuals was also calculated. Results: Among the 105 CRC patients, the median age at testing was 65 years (range: 45–92 years), and 51% (N=54) were male. A total of 51% (N=54) were stage I/II and 49% (N=51) were stage III//IV. The ctDNA levels from the tissue-free methylation-based assay were strongly correlated with ctDNA levels from the tumor-informed assay (mean squared error, 0.064, Pearson’s Correlation Coefficient: 0.955). Among CRC patients, DMAF levels were not different based on sex or age, but were higher in stage IV (N=4) compared to stages I-III. Among cancer-free individuals, 95% had a predicted DMAF of ≤ 8.0 x 10-5. In both CRC cases and cancer-free individuals, DMAF levels were independent of cell-free DNA input. Conclusions: These data demonstrate that the abundance of differential methylation in CRC patients correlates with tumor-informed ctDNA levels, supporting the tissue-free methylation-based assay as a promising tool for informing CRC prognosis and patient management. Future studies will investigate the quantitative abilities of the tissue-free assay in different patients with specific clinicopathological features.