Abstract Liquid biopsies, by analyzing circulating cell-free DNA (cfDNA), have emerged as a promising tool for noninvasive cancer diagnostics and monitoring. Compared to the limited number of scattered DNA mutations in cancer patients, the genome-wide distribution of numerous densely clustered DNA methylation alterations may enable more robust cancer signal detection and higher sensitivity in cancer diagnostics. Furthermore, the cancer type specific methylation signatures potentially can be used for the identification of cancer tissue origin. To enable multivariate analysis of genome-wide methylation markers, we have developed a targeted methylation sequencing assay that can measure the methylation status of thousands of affected CpG (5'-C-phosphate-G-3') sites simultaneously. This assay allows low cfDNA input as little as 3ng while retaining good efficiency. The targeted sites were particularly selected for hypermethylated markers in more than 20 major cancer types based on The Cancer Genome Atlas (TCGA) database. Following bisulfite sequencing of cfDNA, methylation signals from targeted CpG sites were aggregated into a single score that measures deviation of a sample’s methylation profile from a reference baseline. Cancer/normal classification was performed by comparing the methylation score of a test sample to a cutoff established by a group of plasma samples from healthy individuals. Subsequently, methylation signals from cancer type specific markers were used to perform cancer type classification. In a study to evaluate the potential clinical applications of this method, we processed 25 blood samples collected from healthy individuals and 70 blood samples collected from 63 advanced cancer patients. The assay demonstrated a specificity of 100% (n=25), and a sensitivity of 94.3% in colorectal cancer (n=36), 72.7% in breast cancer (n=11), and 52.6% in lung cancer (n=19) detection, respectively. We further performed cancer type classification on the samples that were correctly classified as true positives in the cancer/normal classification. Our results showed a classification accuracy of 82.4% for colorectal cancer (n=34), 87.5% for breast cancer (n=8), and 70% for lung cancer (n=10). In summary, we have developed a comprehensive method that allows us to perform noninvasive cancer profiling through targeted methylation sequencing assay and novel analysis algorithms. This method can potentially be used in multiple validated clinical applications, such as early cancer detection, cancer diagnostics, and monitoring of minimal residual disease. Citation Format: Li Liu, Jonathan M. Toung, Raakhee Vijayaraghavan, Ruoyu Zhang, Helen J. Huang, Toshinori Hinoue, Hui Shen, Neeraj Salathia, Marina Bibikova, Richard Shen, Karen Gutekunst, Peter W. Laird, Filip Janku, Jian-Bing Fan. A highly sensitive method for noninvasive cancer profiling through targeted methylation sequencing of circulating cell-free DNA [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5381. doi:10.1158/1538-7445.AM2017-5381