SCCA and Fred Hutch
Seattle Cancer Care Alliance and the Fred Hutchinson Cancer Research Center are headquartered in Seattle’s South Lake Union neighborhood. (Fred Hutch Photo)

A collaboration to share data between top Pacific Northwest cancer research groups has announced its first three projects, focused on genetic analysis for monitoring breast cancer; interactions between gut bacteria and cancer drugs; and machine learning for identifying types of cancers.

The Cascadia Data Alliance is tackling initiatives that tease out the unique attributes of cancers and how they behave in individual patients, recognizing variability from person to person, which uniquely tailored treatments. The projects will receive more than $1.2 million in funding and credits for Microsoft’s Azure cloud computing service.

The consortium was founded in 2019 by Microsoft and the Fred Hutchinson Cancer Research Center under the name Cascadia Data Discovery Initiative. The other alliance members are BC Cancer, the University of British Columbia, the University of Washington eScience Institute, and the Knight Cancer Institute at Oregon Health & Science University (OHSU).

The effort falls within Microsoft’s AI for Health initiative, which is part of its broader AI for Good program.

These are the initial three projects:

Genetic analysis for monitoring breast cancer

During breast cancer treatment, clusters of tumor cells can mutate over time and become resistant to drugs. The project will use two cutting-edge technologies to better characterize and monitor these changes: single-cell genome sequencing that can provide insights into individual cells, and liquid biopsies, which use blood tests to detect and study cancer.

The pilot project includes researchers from Fred Hutch, BC Cancer and the UW.

Interactions between gut bacteria and cancer drugs

Researchers know that cancer drugs that harness a patient’s immune system to fight tumors are effected by the bacteria that live in a person’s gut. What they don’t know is which bacterial species have the greatest impact on a drug’s performance or cause side effects to the treatment.

Scientists from the Fred Hutch, BC Cancer and OHSU are compiling a repository of mouth tissues and stool samples from patients. They’ll use genomic techniques, data analysis and cloud computing to identify the bacteria and correlate them with patient outcomes.

Machine learning for identifying types of cancers

Not all ovarian cancers, for example, are the same, but rather there are specific forms of the cancers that are fought with targeted treatments. Pathologists visually examine the cancer cells to diagnosis which cancer is present, and that informs the treatment approach.

A team from the Fred Hutch, BC Cancer, OHSU and UBC are working to build an international network for analyzing cancer samples using machine learning to make the identification. The project will use ovarian cancer as its test case. In addition to improving the accuracy of the diagnosis, the researchers are building a platform that ensures patient privacy and data security.

The hope is to create a model for this sort of machine-learning diagnostic network that could be used by healthcare providers worldwide.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.