Abstract Liquid biopsies and cell-free DNA (cfDNA) analysis have emerged as novel non-invasive diagnostic tools in oncology. Analysis of cfDNA methylation status has been proposed as promising approach for several diagnostic purposes, and robust methylation-detecting technologies are of primary importance for the clinical implementation of novel diagnostic assays. Signal-C is an accurate NGS-based methyl-sensitive assay detecting colorectal cancer (CRC) currently under clinical validation. In this study we assessed the influence of different preanalytical variables influencing plasma stability, as the storage time of plasma samples and the tube type used for blood collection, on Signal-C assay results to determine its robustness. 958 samples with 3ml to 4ml plasma were used for cfDNA extraction and run through Signal-C assay following UDX protocol. In brief, 8 to 20ng cfDNA were used as an input for methylation- specific library preparation, which was run according to manufacturer’s instructions. Hybrid- capture was performed to enrich libraries for genomic regions included in Signal-C assay and sequenced on Illumina Novaseq machines. Relevant sequencing QC metrics, such as depth filtered, read yield from raw to filtered and PCR duplicate proportion were calculated, together with cancer predictions according to Signal-C test. Pearson coefficient and Pearson's Chi- squared test with Yates' continuity correction have been used to assess correlations, and Wilcoxon test has been used to measure the significance of the differences between quality metrics with respect to the tube type. We have analyzed the influence of different plasma-related preanalytical variables on sequencing metrics and cancer predictions of Signal-C assay. Plasma age, calculated as the time elapsed between blood collection and cfDNA quantification, shows a small negative correlation with YIELD_RAW_FILTERED (R = - 0.08, p=0.013) and a small positive correlation with PCR_DUP_PROP (R = 0.098, p=0.0024), while it doesn't influence DEPTH_FILTERED values (p>0.05). Use of different blood collection tubes (EDTA or Streck tubes) influenced in a small but significative manner the YIELD_RAW_FILTERED and the PCR_DUP_PROP (p<0.001), while it doesn’t affect the DEPTH_FILTERED values (p>0.05). Despite the sequencing QC metrics being subject to small but statistically significant differences, the overall cancer classification performance of Signal-C assay was not affected by the changes in plasma age (Pearson's product-moment correlation, sensitivity p=0.87, specificity p=0.091) nor in tube type used (Pearson's Chi-squared test with Yates' continuity correction, p=0.219). Here we evaluated the effect of plasma’s storage time and the type of blood collection tube on Signal-C assay performance. While signal processing features show small differences between the tested variables, this does not affect overall clinical performance of the test, supporting its robustness. Additional studies to analyze the effect of other preanalytical variables on Signal-C performance are underway to establish its robustness. Citation Format: Fernando Trincado Alonso, Pol Canal Noguer, Kristi Kruusmaa, Arianna Bertossi. Evaluating the effect of plasma-related preanalytical variables on the robustness of the liquid biopsy early CRC detection Signal-C assay [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr A024.
Read full abstract