Plasma-based, biopsy-free cancer detection assay
This technology is a cancer detection assay designed to non-invasively diagnose patients by analyzing a small blood sample for biomarkers and signature genes, leveraging deep neural network modeling to identify early and late stage cancers with high accuracy.
Unmet Need: Non-invasive methods for detecting and diagnosing various cancers
Early and precise detection of cancers is critical for improving therapeutic responses and increasing overall treatment success rates. There is strong evidence that multi-omic analysis including genomics and epigenomics provide unprecedented insight into pathology, allowing physicans to apply precision medicine approaches best tailored to treat a patient’s specific disease. Unfortunately, the current methods for diagnosing cancer and identifying tumor orign predominately rely on scans or biopsies, which are either unable to capture important -omic information such as specific methylation states, or highly invasive.
The Technology: Deep neural network model for non-invasively cancer diagnostics via DNA methylation and high-throughput sequencing
This diagnostic is a blood-based, biopsy-free method for identifying various cancers by analyzing cell-free DNA (cfDNA) released from tumors and circulating in the blood. By applying machine learning to ss-cfMeDIP-Seq data, users can determine methylation states, identify biomarkers, and screen for signature genes of specific cancers. Built on a deep neural network-based model, this technology has the potential to rapidly diagnose patients, characterize patient specific states, and provide physicians with critical information essential for precision therapies, all while generating a library of cancer-specific cfDNA biomarkers.
This technology has been validated in liver and brain cancer patients.
Applications:
- Diagnostic and prognostic assay for cancers and cancer patients
- Research tool for studying DNA mythlation and cancer biomarkers, especially in animal models
- Generating library of cancer specific DNA biomarkers
- Diagnostic tool for diseases associated with DNA methylation
- Framework for deep neural network models of other single-cell sequencing and methylation datasets
Advantages:
- Early-stage, minimally invasive diagnostic for multiple cancers
- Requirews only small blood samples
- Determines tumor origin
- Provides in-depth, patient-specific omics such as DNA methylation states for precise therapies
- Can be substantially cheaper than biopsy-based approaches when considering procedural costs
- Deep Neural Network model allows for unbiased diagnosis and unprecedented insight
Lead Inventor:
Related Publications:
Tech Ventures Reference:
IR CU23111
Licensing Contact: Joan Martinez