Quantification of circulating tumor DNA (ctDNA) in patients using cancer-specific, methylation-based, tissue-free tests for the detection of molecular residual disease (MRD)

While MRD detection by differential methylation patterns has been shown to predict disease recurrence, clinical decision-making and disease management are increasingly dependent on the exact level of the measured biomarker. We assessed quantitative results by methylation-based MRD assays designed for colorectal (CRC), breast, lung, and muscle-invasive bladder cancer (MIBC). Each cancer-specific, methylation-based, tissue-free ctDNA assay uses next-generation sequencing to query regions of the human genome that are differentially methylated in patients with colorectal, breast, lung, or MIBC cancer compared to cancer-free individuals. Plasma samples from Signatera-tested cancer patients with stage I-IV disease (CRC N=105, breast N=112, lung N=113, MIBC N=60) and from cancer-free individuals (N=223, sex-matched for breast cancer) were analyzed in a cross-validation setup. We compared ctDNA levels measured by the methylation-based assay (estimated by the fraction of differentially methylated alleles for circulating cell-free DNA targets across regions, DMAF) with those from a bespoke, tumor-informed ctDNA assay (SignateraTM, measured as variant allele fraction, VAF), which served as the reference, by calculating the mean squared error (MSE) and Pearson's Correlation Coefficient (⍴). For each comparison of the cancer-specific, tissue-free, methylation-based assay with the personalized tumor-informed assay, DMAF and VAF were strongly correlated for patients with CRC (VAF median [range]: 1.43% [0.003-8.26%], MSE: 0.064, ⍴: 0.955), breast cancer (VAF median [range]: 1.08% [0.002-7.80%], MSE: 0.161, ⍴: 0.863), lung cancer (VAF median [range]: 0.93% [0.005-12.9%], MSE: 0.179, ⍴: 0.829), and MIBC (VAF median [range]: 2.33% [0.005-9.26%], MSE: 0.079, ⍴: 0.954). Among patients with breast cancer, the correlation between DMAF and VAF was maintained across subtypes, including HR+ (MSE: 0.132, ⍴: 0.888), HER2+ (MSE: 0.272, ⍴: 0.815), and triple-negative (MSE: 0.109, ⍴: 0.897). Similarly, the correlation between DMAF and VAF was strong across lung cancer subtypes, non-small cell lung cancer (MSE: 0.207, ⍴: 0.802) and small cell lung cancer (MSE: 0.069, ⍴: 0.941). For all cancer-specific, tissue-free, methylation-based assays, DMAF levels were independent of patient characteristics such as sex (excluding breast), age, and stage (I-III), as well as cell-free DNA input. These data demonstrate that the abundance of differential methylation in patients with CRC, breast, lung, or MIBC cancer correlates with tumor-informed ctDNA levels, supporting each tissue-free methylation-based assay as a promising tool for informing prognosis and patient management. Future studies will investigate the quantitative abilities of the tissue-free assay in different patients with specific clinicopathological features.