A DNA-mutational panel for liquid biopsy-based pan-cancer early detection.

Abstract

e13687 Background: Early detection through liquid biopsy can significantly increase the successful chances of treatment. The sensitivity and specificity of early detection are limited by lower signal-to-noise ratio. Thus, well design of liquid biopsy panel is important. Methods: We adopted comprehensive data sources for designing the liquid biopsy panel, including databases of TCGA, ICGC, COSMIC, MSK Cancer Hotspots, while the panel of HCCscreen and CancerSeek. We also considered the database of somatic mutations in Chinese cancer patients generated by us (Genecast (Beijing) Biotechnology Co., Ltd.). We first calculated pan-cancer carrier ratio of somatic mutations in each database and constructed the distribution of mutation counts on step-wise increased carrier ratio. Then, we selected the set of somatic mutations with “elbow point” carrier ratio from each database as part of the panel. Finally, we integrated collected hotspots with the panel of HCCscreen and CancerSeek. Results: We selected 408, 521, 214, 146 and 330 hotspots from databases of TCGA, ICGC, COSMIC, MSK Cancer Hotspots and somatic mutations in Chinese cancer patients, respectively. After integration with the panel of HCCscreen and CancerSeek, this designed panel contains 3,334 bases distributed on 23 chromosomes and 915 gene models. Comparison of hotspots collected from different databases showed that few of them shared the same genomic location except for ones from the database of MSK Cancer Hotspots, indicating the well complementarity between them. Especially, there are 186 unique hotspots collected from the database of somatic mutations in Chinese cancer patients, which can improve the sensitivity of early detection for Chinese and Asian population. Next, we evaluated detected sensitivity of the designed panel based on sequencing data of plasma samples from more than 12,000 Chinese cancer patients collected in Genecast. The result showed that the maximum sensitivity is 66.67% for small cell lung cancer and the overall sensitivity is 45.71% for 26 cancer types, in which hotspots uniquely collected from the database of somatic mutations in Chinese cancer patients accounted for 5.11%. Conclusions: We designed a liquid biopsy panel for pan-cancer early detection and evaluated its sensitivity based on sequencing data of plasma samples from more than 12,000 Chinese cancer patients. Satisfactory performance of our designed panel shows its potential application in cancer screening for healthy and highly risky individuals.