More than a year after announcing its 10-year strategic partnership with Google, Mayo Clinic and Google announced Wednesday that the duo would focus on using artificial intelligence to better plan radiation therapy.
Radiation therapy experts from Mayo Clinic, including radiation oncologists, medical physicists, dosimetrists and service design, will collaborate with Google Health’s experts in applying AI to medical imaging.
“Within Google Health, we’ve been exploring the many areas where AI or machine-learning has the power to improve the practice of medicine in general but radiotherapy is one of those areas where we see a lot of promise and potential,” said Dr. Cían Hughes, informatics lead at Google Health. “Together Google Heath and Mayo are now collaborating to study the use of machine-learning models to help clinicians plan radiotherapy treatments for first head and neck cancer, but hopefully in time, a whole round of different cancers.”
Mayo Clinic and Google Health teams will use de-identified data to develop, train and validate machine-learning algorithms that are able to automatically segment healthy tissues and organs within the head and neck with the aim of reducing the time it takes to plan treatment for patients with cancer in need of radiation therapy.
The intersection of Mayo’s expertise in radio-oncology and Google’s expertise in image segmentation was a natural fit to make this the first innovation project the partnership would explore.
Typically done manually or in a semiautomated manner, the process can take six hours or longer for complex cases.
“The work that we do as radiation oncologists is very labor intensive,” said Dr. Nadia Laack, chairwoman of the Department of Radiation Oncology at Mayo Clinic in Rochester and one of the principal investigators on the project.
Computer models that doctors and treatment team members currently use to design radiation plans cannot see those healthy tissue and organs without their work. That manual, tedious process needs to be precise to be safe and accurate to have the best radiation plan, Laack said. The work takes many hours per patient, which also means doctors and other clinicians have less time to spend with the patients.
In the research phase of the project, the technology will not be used in the treatment planning or any other sort of clinical setting. Parallel to the work researchers are doing to understand how well algorithms can perform, regulatory experts from Google and Mayo are also working together on a larger project to understand potential regulatory approaches for a medical device that could use these algorithms.
Laack said researchers do not expect the work to replace the expertise of the people who currently do this work, but will hopefully make the planning process more efficient, help reduce variability, drive standardization, and improve plan quality as well as spend more time with patients.
Between Rochester and Arizona, Mayo sees more proton beam patients than any other medical center in the world. About 50% of cancer patients require radiation during the course of their treatment, Laack said.
“Unfortunately, cancer incidence is not going down. We see this as a huge need for the future going forward that we get better and more efficient at being able to do radiation planning, which, right now, is very, very labor intensive,” Laack said, noting that cancer cases are likely to grow in the next decade. “Hopefully this research will demonstrate that it is something that can effectively improve the needle on improving access and efficiency in our practice.”
This is the first initiative to come out of the innovation prong of the Mayo and Google partnership. It is a Mayo Clinic Institutional Review Board-approved research project funded by both organizations under its innovation research and development budget.
When Mayo and Google announced the partnership in September 2019, they said the partnership would redefine how health care was delivered and accelerate the pace of health care innovation through digital technologies.
The partnership also includes the uploading and storing of de-identified patient medical records on Mayo Cloud — a separate space that is under Mayo’s control. That work is ongoing. Google's tools and artificial intelligence will be used to analyze the anonymous data for ways to improve medical care.
Google’s Director of Global Healthcare Solutions Aashima Gupta explained in September 2019 that the patient data would be in "a lock box" and Mayo Clinic would have the only "key."