A key technical challenge in oncology is the detection of minimal residual disease (MRD, the presence of residual tumor in a subject post-treatment).  Recent studies have shown poor prognostic outcomes for research subjects with ultra-low residual tumor burdens after treatment across a variety of cancer types (1-4).  Since changes in ctDNA levels correlate with relative changes in tumor burden (2), in order to sensitively detect MRD, an assay must be able to detect circulating tumor DNA (ctDNA) 30X lower (5) than pre-treatment levels.  The difficulty inherent in achieving this limit of detection (LOD) is compounded by the small amount of cell free DNA (cfDNA) derived from a typical 10ml blood draw (2).  Detecting a single variant at sequencing depths commensurate with genotyping typically faces sensitivity limitations at low allele fractions (AFs), as the probability of observing even a single molecule with a specific variant are drastically low.  To address this issue, we present here a method using the Roche AVENIO ctDNA Surveillance Kit (Research Use Only) that takes advantage of multiple variants to improve the LOD for ctDNA detection below the LOD for a single variant.

To assess MRD sensitivity at relevant allele fractions, cfDNA spikes were made mimicking ctDNA allele fractions ranging from 0.002% to 0.02%.  Samples with different input masses (10-50 ng input) were run using the Surveillance Kit, which was designed for surveillance-centric research applications in lung cancer and colorectal cancer (CRC).  Known variants were queried and a Monte-Carlo-based approach was used to determine a p-value for presence of ctDNA.

Here the results of three experiments to assess the performance of the monitoring algorithm are described.  1. Using healthy donor cfDNA samples, specificity analyses were performed with different p-value cutoffs, and a p-value of 0.04 was shown to give 95% specificity for lung cancer specific cancer variants.  2. To determine the expected variability within a given sample, 10 technical replicates of 2 distinct samples were assessed, showing the range of performance expected due simply to random sampling.  3. MRD sensitivities across different input masses, number of variants, and sequencing depths are shown across 3 distinct samples with variants at 0.002%, 0.004%, and 0.01%.  With 30 ng of input, the Surveillance Kit achieves a MRD sensitivity of >99% with an AF of 0.002% and 34 known variants, and 90% MRD sensitivity at a 0.01% AF with 6 variants (the median expected number of variants per subject).  Using the algorithm described, a retrospective clinical research study looking for MRD post-surgery in stage II/III CRC subjects (4) were analyzed.  Subjects with ctDNA an average of 2 weeks after surgery recurred significantly sooner and ctDNA levels below 0.005% were enough to show prognostic significance.

Citation Format: Amrita Pati, Alex Lovejoy, Paul W. Shi, Abraham Munoz, Vera Rapoport, Katrina Mayol, Seng Saelee, Janet Jin, Dan Klass. Detection of minimal residual disease from less than one cell equivalent in liquid biopsy samples using the AVENIO ctDNA Surveillance Kit [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4574.