The field of cancer clinical research and treatment is at a historic turning point, galvanized by the accepted understanding that cancer is fundamentally driven by genomic alterations. This understanding is being accelerated by affordable genomic sequencing technologies that detect such alterations, and the corresponding connections of this data to disease progression and treatments. As a result, cancer treatment is shifting from a “tumor of origin” (e.g. breast, bladder, etc.) treatment protocol to protocols based on targetable and actionable genomic alterations that are identified through molecular diagnostics.
Along with the ability to understand the mechanisms related to cancer, we also have corresponding data from large cohorts of cancer patients and the computational ability to correlate attributes of a single patient to one or more relevant cohorts. Such correlation allows researchers to uncover shared disease drivers, and allows physicians to develop treatment plans based on protocols that were effective or not effective for similar patients. PrecisionProfile is focused on enabling the precision medicine community to “make sense” of genomic profiles to create personalized treatment plans for patients. The University of Colorado Cancer Center has developed a genomic analytics platform and a “patients like mine” cohort analytics approach, and PrecisionProfile has been formed as an Anschutz Medical Campus spinoff to create a product to advance the practice of cancer precision medicine.
The Problem We Solve
A major obstacle that oncologists and researchers face is that cancer is not one, but many different diseases, driven by many genomic aberrations: mutations, deletions, copy number alterations and many interactions: transcriptional, signaling, metabolic pathway networks. This inherent complexity means that the analysis and guidelines for the future treatment of cancer will be significantly more complex than they are currently. For breast cancer, where molecular analysis has yielded significant understanding, this complexity can already be seen in the current ASCObiomarker guidelines, with 36 unique recommendations on the use of 23 different biomarkers, and 8 distinct molecular diagnostic reports.
Our vision is to harness the complexity of this information to enable oncologists to comprehend and create Precision Medicine treatment plans from it, and concurrently enable researchers to discover new causes of cancers and explore new treatments. PrecisionProfile accomplishes this by:
- applying leading edge analytics technologies to
- analyze, integrate, manage and interpret the large volumes of data required
- and leverage machine learning to match patients to cohorts with similar profiles.
We see the following benefits of the PrecisionProfile platform:
- individual patient molecular profiles can be compared to other patients, leading to insights on disease development and better treatment decisions
- shared insights and predictive analytics enabled by machine learning across extensive patient cohorts will provide treatment insights for practicing oncologists
- reduction in the R&D cycle and clinical diagnostic/prognostic analysis
While we focus here on cancer, our solution can be applied to all genomic driven diseases.
- Dynamic Data Filtering: easily adjust cohort characteristics to discover new trends and similarities
- Customizable, Interactive Visualizations: Kaplan-Meier, coMut, violin, lollipop, pathways, raw data
- Session Recording, Project Document Management
- Data Search: find additional data sources to enhance analysis
- Bayesian Machine Learning: enables continuous enhancement of analytics algorithms
- Automated Data Shaping: harmonizes differing formats of similar data, enabling integration of varied data sources
- Hadoop & Spark: optimize platform responsiveness
- Share: upload projects, pose questions or requests to community
- Discover: find similar projects and potential collaborators