The Cancer Genomics Lab has a history of transformative discoveries in cancer research. These include the first genome-wide sequence analysis in human cancers, identifying key genes and pathways dysregulated in tumorigenesis. Members of the group developed methods for global gene expression analyses and coined the word “transcriptome” to describe the expression patterns that could now be studied in cancer and other cells. This research revealed the genomic landscape of human cancers, including in breast, colorectal, brain, pancreatic, ovarian, and lung cancers. These analyses identified a variety of genes not previously known to be involved in neoplasia, including PIK3CA as one of the most highly mutated genes in human cancer. These discoveries have led to new FDA-approved therapies against PI3K and IDH1, and FDA-approved diagnostic tests for comprehensive tumor profiling. More recently, the group has created non-invasive machine learning liquid biopsy approaches for early detection and monitoring of cancer patients. This work has provided new paradigms for understanding human cancer that have paved the way for precision medicine, benefitting patients worldwide.
Examples of recent research efforts include:
Detecting cancer early using machine learning analyses of non-invasive liquid biopsies
Our group has identified unique properties of cell-free DNA (cfDNA) from cancer patients and developed artificial intelligence technologies for early cancer detection. Over a decade ago, we performed the first genome-wide analyses of changes in cfDNA in cancer patients as well as the first direct detection of tumor-specific changes in cfDNA using next generation sequencing. Our DNA Evaluation of Fragments (DELFI) approach uses machine learning to detect cancer-specific changes in the genome-wide cfDNA fragmentation profiles in the circulation. The observed “fragmentome” represents a compendium of changes in characteristics of cfDNA that can be used to detect patients with cancer and identify their tissue of origin. We have extended these analyses through genome-wide evaluation of mutation profiles and repeat landscapes in cfDNA. We have employed these methodologies for non-invasive early detection of patients with lung, liver, and ovarian and are extending these to other cancer types.
Disease interception through early detection of disease response or progression
Noninvasive approaches for detection of tumor-specific mutations in cell-free DNA (cfDNA) have the potential to identify appropriate therapies as well as track a patient’s response to treatment, enabling effective and timely decisions. We have used both focused and genome-wide approaches to identify mechanisms of primary and acquired resistance in lung, colorectal, and other cancers. We have also developed noninvasive measures of molecular response and resistance and residual disease detection to targeted therapies, chemotherapy and immune checkpoint blockade approaches in colorectal, lung and other cancer types.
Cancer genomics and the early changes in tumorigenesis
Our team provided some of the earliest insights into the complex genomic landscape of human cancer, identifying new genes and pathways not previously implicated in tumorigenesis, including changes in PIK3CA, IDH and chromatin modifying genes, and using these analyses as a foundation for precision medicine for cancer patients. More recently, the group has focused on early events in tumorigenesis, providing genomic evidence and timing of lesions in the fallopian tube as the site of origin of high-grade serous ovarian cancers, the most common type of ovarian cancer. New methodologies examining the “dark matter” of the genome have revealed the repeat landscapes of the cancer genome, identifying hundreds of changes in repeat elements in a variety of cancer types that provide valuable insights into mechanisms of disease. We are analyzing the combination of this approach together with other changes in the genome for early cancer detection.
Keywords: cancer genomics, early cancer detection, liquid biopsy, artificial intelligence, cell-free DNA, monitoring, minimal residual disease