Development of a cfDNA based trans-omics approach for early cancer detection

Circulating tumor DNA (ctDNA) in plasma of cancer patients provides valuable information about the cancer and also holds great promise for non-invasive early cancer detection. Nevertheless, since ctDNA is diluted by circulating cell-free DNA (cfDNA) of noncancerous origins, its detection poses significant challenge, especially during early stage of cancer. Previous studies on early cancer detection have mostly focused on single feature of the ctDNA, namely either cancer driver gene mutations or alterations in the methylome. We have developed a set of experimental and computational tools to measure both genetic and epigenetic signals from ctDNA using next-generation sequencing, aiming to improve its detection. Briefly, cfDNA extracted from cancer patients and healthy individuals were used for targeted deep sequencing as well as targeted bisulfite sequencing. More specifically, duplex-UMI libraries were prepared in order to suppress errors introduced by sequencing and PCR artifacts. Libraries were subsequently captured using a custom designed panel targeting cancer driver gene mutations. Bisulfite sequencing libraries were prepared from single strand cfDNA following bisulfite conversion and captured using a panel targeting gene promoter regions. These methods ensure higher quality of the cfDNA libraries and more sensitive ctDNA detection. A subset of samples also had matched tumor tissue samples sequenced using targeted sequencing and WGBS to validate mutations as well as changes of CpG methylation level found in plasma. All sequencing runs were conducted on the MGIseq platform developed by BGI genomics. To this end, we have developed classification models based on these data using machine learning approaches. Optimal model achieves a sensitivity of 52% in early stage lung cancer, and a sensitivity of 83.3% in ovarian cancer, while holding specificity > 99%. These results hold great potential to be further explored for early cancer detection application. We are currently expanding the study to additional cancer types, especially the ones with high incidence rate and/or poor survival rate in China, such as colorectal cancer and liver cancer. Citation Format: Yuying Wang, Heng Zhao, Kaijian Ling, Jianchao Zheng, Zhilong Li, Shuangshuang Chen, Jia Li, Chichuan Liu, Guanghui Yang, Hongpo Zhou, Jiaxi Peng, Lili Ye, Liuhong Zeng, Jianlong Sun, Ruijingfang Jiang, Li Deng, Yanzhou Wang, Kezhong Chen, Zhiqing Liang, Fan Yang, Taiping Shi, Mingzhi Ye. Development of a cfDNA based trans-omics approach for early cancer detection [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3318.