For Meeting Recording, please click here. Passcode: 27097062

Talk Title:

Methods for tumor phenotype classification from circulating tumor DNA

Abstract:

An accurate diagnosis of the tumor histology is essential for clinical care. Aggressive tumor phenotypes can emerge through subtype changes and lineage plasticity as major mechanisms of treatment resistance in cancer. Therefore, distinguishing tumor phenotypes has clinical relevance in view of therapeutic response, but the need for a biopsy to diagnose tumor histology can be challenging. Circulating tumor DNA (ctDNA) released from tumor cells into the blood is a non-invasive “liquid biopsy” solution for addressing challenges in tissue accessibility. Current research and clinical efforts have focused on the detection of genetic mutations from ctDNA sequencing as potential biomarkers. However, genomic alterations do not always define tumor phenotypes nor fully explain treatment resistance. Recent advances now demonstrate that epigenetic information can be ascertained from profiling nucleosome positioning and accessibility in ctDNA. In this talk, I will describe how we applied this concept to develop computational methods for predicting transcriptional activity and classifying tumor phenotypes in breast, prostate, and lung cancers. We established a unique resource of mouse plasma from patient-derived xenografts to dissect the transcriptional activity of tumor phenotypes directly from ctDNA and to inform a suite of prediction models that can quantify phenotype heterogeneity in patients. These tools provide a framework that extends the utility of ctDNA beyond genomic assessments with important implications for molecular classification and precision oncology.