Optimizing blood biopsy in cancers with low mutation burden and high structural complexity

点击翻译按钮获取中文摘要

PROJECT SUMMARY Liquid biopsy is a non-invasive technique that can be used to help diagnose and monitor cancer. It is based on the principle that tumor cells release small pieces of DNA and RNA into circulation. In several human cancers, FDA-approved liquid biopsy tests are designed to look for common disease-associated mutations. These liquid biopsy tests are most successful in tumors with a well-defined mutation landscape, such as lung and breast cancer. However, looking for common mutations is less successful in structurally complex tumors with a lower incidence of mutations, as is the case with many sarcomas, such as osteosarcoma (OS) and Ewing’s sarcoma. Recent data indicate that mutation-independent liquid biopsy techniques, including assessment of circulating DNA fragment size patterns and methylation status, can increase sensitivity of the assay and identify the tissue of origin and histologic subtype of human cancers. Additionally, evidence now suggests that unique gene expression and methylation signatures measured by liquid biopsy have the potential to act as a surrogate for response to treatment and/or identify early emergence of treatment resistance. As such, there is potential for using an advance liquid biopsy tool to inform patient-specific therapies more effectively, particularly in instances where repeat imaging/tumor sampling is challenging. As such, the hypothesis underlying this proposal is that gene expression and epigenetic metastatic signatures can be identified in RNA and DNA isolated from plasma in canine OS and integrated using machine learning to improve the sensitivity of liquid biopsy. It is further predicted that this improved liquid biopsy platform will be capable of identifying treatment specific signatures reflective of response or resistance to therapy. We will use canine OS, which has a structurally chaotic tumor genome, as a large animal disease model of human sarcomas. Using patient-matched plasma samples from dogs with OS taken at multiple timepoints throughout treatment, we will evaluate cell-free DNA and RNA using a comprehensive mutation-independent liquid biopsy assay. This will incorporate evaluation multiple parameters, including cell-free DNA fragment sizes, methylation, and gene expression alterations and use machine learning to optimize parameter integration. The liquid biopsy tool will be further validated for detection of early disease progression in OS patients. Lastly, we will begin to dissect how drug exposure alters disease-specific signatures in circulation. Ultimately, the tools and techniques developed from this work will have broad applicability to both canine and human sarcomas, facilitating enhanced accuracy for cancer detection and clinical decision-making. Importantly, the work outlined in this proposal provides a unique opportunity for expansion of genomic skill sets in the context of translational medicine, thereby further supporting my development as a successful independent clinician scientist.