1232P Analytic analysis of PanSeer7, a targeted bisulfite sequencing assay for blood-based multi-cancer detection for cancer early detection and tissue-of-origin identification

Many cancers are symptoms free in early clinical stages, causing nearly half of patients diagnosed in advanced-stages when therapeutic options are limited. Early cancer detection is key to improve clinical outcomes. We developed PanSeer7, a multi-cancer detection assay based on targeted bisulfite sequencing of circulating cell-free DNA (cfDNA) and evaluated its technical performances of reproducibility and sensitivity. The PanSeer7 panel consists of 2,447 markers which are either differentially methylated between healthy and cancer samples, or distinctively methylated in a specific cancer. We assessed its reproducibility by sequencing 40 technical replicates of synthetic healthy cfDNA samples on 4 independent batches and with different inputs (2∼20 ng). For its limit of detection (LOD), we prepared samples mimicking cancer plasma DNA by diluting fragmented cancer cell line DNA, from 7 common cancer types, into GM12878 control at ratios of 1/10,000 to 1/100. We tested PanSeer7’s ability to identify tissue of origin (TOO) by analyzing DNA from formalin-fixed paraffin-embedded (FFPE) tissues and healthy plasma. PanSeer7 produced highly consistent methylation levels among replicates at a minimum of 10 ng input. Its technical LOD was 1/10,000 for lung cancer (LuC; H1650), liver cancer (LiC; HepG2), gastric cancer (GC; HGC27), esophageal cancer (EC; KYSE150) and colorectal cancer (CC; SW480), and slightly lower as 5/10,000 for pancreatic cancer (PC; PANC1) and breast cancer (BC; MDA-MB-231). To evaluate PanSeer7’s accuracy for predicting TOO, we sequenced 38 healthy plasma and 121 FFPE tissues (17 LiC, 13 PC, 21 GC, 18 EC, 16 CC, 14 LuC and 22 BC). They were clearly clustered according to their TOO based on methylation levels. We simulated 3,500 datasets by mixing reads of cancer tissue into those of healthy plasma at ratios of 1/10,000 to 1/100, and the TOO-predicting model achieved an accuracy over 95% at a ratio of 5/10,000. With excellent technical performances in both cancer detection and TOO identification, PanSeer7 is promising to be clinically applied for non-invasive multi-cancer detection.